linux-sg2042/kernel/tracepoint.c

635 lines
16 KiB
C
Raw Normal View History

tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/*
* Copyright (C) 2008 Mathieu Desnoyers
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <linux/module.h>
#include <linux/mutex.h>
#include <linux/types.h>
#include <linux/jhash.h>
#include <linux/list.h>
#include <linux/rcupdate.h>
#include <linux/tracepoint.h>
#include <linux/err.h>
#include <linux/slab.h>
#include <linux/sched.h>
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
extern struct tracepoint __start___tracepoints[];
extern struct tracepoint __stop___tracepoints[];
/* Set to 1 to enable tracepoint debug output */
static const int tracepoint_debug;
/*
* tracepoints_mutex nests inside module_mutex. Tracepoints mutex protects the
* builtin and module tracepoints and the hash table.
*/
static DEFINE_MUTEX(tracepoints_mutex);
/*
* Tracepoint hash table, containing the active tracepoints.
* Protected by tracepoints_mutex.
*/
#define TRACEPOINT_HASH_BITS 6
#define TRACEPOINT_TABLE_SIZE (1 << TRACEPOINT_HASH_BITS)
static struct hlist_head tracepoint_table[TRACEPOINT_TABLE_SIZE];
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/*
* Note about RCU :
* It is used to delay the free of multiple probes array until a quiescent
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
* state is reached.
* Tracepoint entries modifications are protected by the tracepoints_mutex.
*/
struct tracepoint_entry {
struct hlist_node hlist;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *funcs;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
int refcount; /* Number of times armed. 0 if disarmed. */
char name[0];
};
struct tp_probes {
union {
struct rcu_head rcu;
struct list_head list;
} u;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func probes[0];
};
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
static inline void *allocate_probes(int count)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tp_probes *p = kmalloc(count * sizeof(struct tracepoint_func)
+ sizeof(struct tp_probes), GFP_KERNEL);
return p == NULL ? NULL : p->probes;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
static void rcu_free_old_probes(struct rcu_head *head)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
kfree(container_of(head, struct tp_probes, u.rcu));
}
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
static inline void release_probes(struct tracepoint_func *old)
{
if (old) {
struct tp_probes *tp_probes = container_of(old,
struct tp_probes, probes[0]);
call_rcu_sched(&tp_probes->u.rcu, rcu_free_old_probes);
}
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
static void debug_print_probes(struct tracepoint_entry *entry)
{
int i;
if (!tracepoint_debug || !entry->funcs)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
return;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
for (i = 0; entry->funcs[i].func; i++)
printk(KERN_DEBUG "Probe %d : %p\n", i, entry->funcs[i].func);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
static struct tracepoint_func *
tracepoint_entry_add_probe(struct tracepoint_entry *entry,
void *probe, void *data)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
int nr_probes = 0;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old, *new;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
WARN_ON(!probe);
debug_print_probes(entry);
old = entry->funcs;
if (old) {
/* (N -> N+1), (N != 0, 1) probes */
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
for (nr_probes = 0; old[nr_probes].func; nr_probes++)
if (old[nr_probes].func == probe &&
old[nr_probes].data == data)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
return ERR_PTR(-EEXIST);
}
/* + 2 : one for new probe, one for NULL func */
new = allocate_probes(nr_probes + 2);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
if (new == NULL)
return ERR_PTR(-ENOMEM);
if (old)
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
memcpy(new, old, nr_probes * sizeof(struct tracepoint_func));
new[nr_probes].func = probe;
new[nr_probes].data = data;
new[nr_probes + 1].func = NULL;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
entry->refcount = nr_probes + 1;
entry->funcs = new;
debug_print_probes(entry);
return old;
}
static void *
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
tracepoint_entry_remove_probe(struct tracepoint_entry *entry,
void *probe, void *data)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
int nr_probes = 0, nr_del = 0, i;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old, *new;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
old = entry->funcs;
if (!old)
return ERR_PTR(-ENOENT);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
debug_print_probes(entry);
/* (N -> M), (N > 1, M >= 0) probes */
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
for (nr_probes = 0; old[nr_probes].func; nr_probes++) {
if (!probe ||
(old[nr_probes].func == probe &&
old[nr_probes].data == data))
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
nr_del++;
}
if (nr_probes - nr_del == 0) {
/* N -> 0, (N > 1) */
entry->funcs = NULL;
entry->refcount = 0;
debug_print_probes(entry);
return old;
} else {
int j = 0;
/* N -> M, (N > 1, M > 0) */
/* + 1 for NULL */
new = allocate_probes(nr_probes - nr_del + 1);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
if (new == NULL)
return ERR_PTR(-ENOMEM);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
for (i = 0; old[i].func; i++)
if (probe &&
(old[i].func != probe || old[i].data != data))
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
new[j++] = old[i];
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
new[nr_probes - nr_del].func = NULL;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
entry->refcount = nr_probes - nr_del;
entry->funcs = new;
}
debug_print_probes(entry);
return old;
}
/*
* Get tracepoint if the tracepoint is present in the tracepoint hash table.
* Must be called with tracepoints_mutex held.
* Returns NULL if not present.
*/
static struct tracepoint_entry *get_tracepoint(const char *name)
{
struct hlist_head *head;
struct hlist_node *node;
struct tracepoint_entry *e;
u32 hash = jhash(name, strlen(name), 0);
head = &tracepoint_table[hash & (TRACEPOINT_TABLE_SIZE - 1)];
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
hlist_for_each_entry(e, node, head, hlist) {
if (!strcmp(name, e->name))
return e;
}
return NULL;
}
/*
* Add the tracepoint to the tracepoint hash table. Must be called with
* tracepoints_mutex held.
*/
static struct tracepoint_entry *add_tracepoint(const char *name)
{
struct hlist_head *head;
struct hlist_node *node;
struct tracepoint_entry *e;
size_t name_len = strlen(name) + 1;
u32 hash = jhash(name, name_len-1, 0);
head = &tracepoint_table[hash & (TRACEPOINT_TABLE_SIZE - 1)];
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
hlist_for_each_entry(e, node, head, hlist) {
if (!strcmp(name, e->name)) {
printk(KERN_NOTICE
"tracepoint %s busy\n", name);
return ERR_PTR(-EEXIST); /* Already there */
}
}
/*
* Using kmalloc here to allocate a variable length element. Could
* cause some memory fragmentation if overused.
*/
e = kmalloc(sizeof(struct tracepoint_entry) + name_len, GFP_KERNEL);
if (!e)
return ERR_PTR(-ENOMEM);
memcpy(&e->name[0], name, name_len);
e->funcs = NULL;
e->refcount = 0;
hlist_add_head(&e->hlist, head);
return e;
}
/*
* Remove the tracepoint from the tracepoint hash table. Must be called with
* mutex_lock held.
*/
static inline void remove_tracepoint(struct tracepoint_entry *e)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
hlist_del(&e->hlist);
kfree(e);
}
/*
* Sets the probe callback corresponding to one tracepoint.
*/
static void set_tracepoint(struct tracepoint_entry **entry,
struct tracepoint *elem, int active)
{
WARN_ON(strcmp((*entry)->name, elem->name) != 0);
tracing: Move tracepoint callbacks from declaration to definition It's not strictly correct for the tracepoint reg/unreg callbacks to occur when a client is hooking up, because the actual tracepoint may not be present yet. This happens to be fine for syscall, since that's in the core kernel, but it would cause problems for tracepoints defined in a module that hasn't been loaded yet. It also means the reg/unreg has to be EXPORTed for any modules to use the tracepoint (as in SystemTap). This patch removes DECLARE_TRACE_WITH_CALLBACK, and instead introduces DEFINE_TRACE_FN which stores the callbacks in struct tracepoint. The callbacks are used now when the active state of the tracepoint changes in set_tracepoint & disable_tracepoint. This also introduces TRACE_EVENT_FN, so ftrace events can also provide registration callbacks if needed. Signed-off-by: Josh Stone <jistone@redhat.com> Cc: Jason Baron <jbaron@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Cc: Jiaying Zhang <jiayingz@google.com> Cc: Martin Bligh <mbligh@google.com> Cc: Lai Jiangshan <laijs@cn.fujitsu.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> LKML-Reference: <1251150194-1713-4-git-send-email-jistone@redhat.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
2009-08-25 05:43:13 +08:00
if (elem->regfunc && !elem->state && active)
elem->regfunc();
else if (elem->unregfunc && elem->state && !active)
elem->unregfunc();
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/*
* rcu_assign_pointer has a smp_wmb() which makes sure that the new
* probe callbacks array is consistent before setting a pointer to it.
* This array is referenced by __DO_TRACE from
* include/linux/tracepoints.h. A matching smp_read_barrier_depends()
* is used.
*/
rcu_assign_pointer(elem->funcs, (*entry)->funcs);
elem->state = active;
}
/*
* Disable a tracepoint and its probe callback.
* Note: only waiting an RCU period after setting elem->call to the empty
* function insures that the original callback is not used anymore. This insured
* by preempt_disable around the call site.
*/
static void disable_tracepoint(struct tracepoint *elem)
{
tracing: Move tracepoint callbacks from declaration to definition It's not strictly correct for the tracepoint reg/unreg callbacks to occur when a client is hooking up, because the actual tracepoint may not be present yet. This happens to be fine for syscall, since that's in the core kernel, but it would cause problems for tracepoints defined in a module that hasn't been loaded yet. It also means the reg/unreg has to be EXPORTed for any modules to use the tracepoint (as in SystemTap). This patch removes DECLARE_TRACE_WITH_CALLBACK, and instead introduces DEFINE_TRACE_FN which stores the callbacks in struct tracepoint. The callbacks are used now when the active state of the tracepoint changes in set_tracepoint & disable_tracepoint. This also introduces TRACE_EVENT_FN, so ftrace events can also provide registration callbacks if needed. Signed-off-by: Josh Stone <jistone@redhat.com> Cc: Jason Baron <jbaron@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Cc: Jiaying Zhang <jiayingz@google.com> Cc: Martin Bligh <mbligh@google.com> Cc: Lai Jiangshan <laijs@cn.fujitsu.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> LKML-Reference: <1251150194-1713-4-git-send-email-jistone@redhat.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
2009-08-25 05:43:13 +08:00
if (elem->unregfunc && elem->state)
elem->unregfunc();
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
elem->state = 0;
rcu_assign_pointer(elem->funcs, NULL);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
/**
* tracepoint_update_probe_range - Update a probe range
* @begin: beginning of the range
* @end: end of the range
*
* Updates the probe callback corresponding to a range of tracepoints.
*/
void
tracepoint_update_probe_range(struct tracepoint *begin, struct tracepoint *end)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
struct tracepoint *iter;
struct tracepoint_entry *mark_entry;
if (!begin)
return;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
mutex_lock(&tracepoints_mutex);
for (iter = begin; iter < end; iter++) {
mark_entry = get_tracepoint(iter->name);
if (mark_entry) {
set_tracepoint(&mark_entry, iter,
!!mark_entry->refcount);
} else {
disable_tracepoint(iter);
}
}
mutex_unlock(&tracepoints_mutex);
}
/*
* Update probes, removing the faulty probes.
*/
static void tracepoint_update_probes(void)
{
/* Core kernel tracepoints */
tracepoint_update_probe_range(__start___tracepoints,
__stop___tracepoints);
/* tracepoints in modules. */
module_update_tracepoints();
}
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
static struct tracepoint_func *
tracepoint_add_probe(const char *name, void *probe, void *data)
{
struct tracepoint_entry *entry;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
entry = get_tracepoint(name);
if (!entry) {
entry = add_tracepoint(name);
if (IS_ERR(entry))
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
return (struct tracepoint_func *)entry;
}
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_entry_add_probe(entry, probe, data);
if (IS_ERR(old) && !entry->refcount)
remove_tracepoint(entry);
return old;
}
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/**
* tracepoint_probe_register - Connect a probe to a tracepoint
* @name: tracepoint name
* @probe: probe handler
*
* Returns 0 if ok, error value on error.
* The probe address must at least be aligned on the architecture pointer size.
*/
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
int tracepoint_probe_register(const char *name, void *probe, void *data)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
mutex_lock(&tracepoints_mutex);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_add_probe(name, probe, data);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
mutex_unlock(&tracepoints_mutex);
if (IS_ERR(old))
return PTR_ERR(old);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
tracepoint_update_probes(); /* may update entry */
release_probes(old);
return 0;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
static struct tracepoint_func *
tracepoint_remove_probe(const char *name, void *probe, void *data)
{
struct tracepoint_entry *entry;
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
entry = get_tracepoint(name);
if (!entry)
return ERR_PTR(-ENOENT);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_entry_remove_probe(entry, probe, data);
if (IS_ERR(old))
return old;
if (!entry->refcount)
remove_tracepoint(entry);
return old;
}
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/**
* tracepoint_probe_unregister - Disconnect a probe from a tracepoint
* @name: tracepoint name
* @probe: probe function pointer
*
* We do not need to call a synchronize_sched to make sure the probes have
* finished running before doing a module unload, because the module unload
* itself uses stop_machine(), which insures that every preempt disabled section
* have finished.
*/
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
int tracepoint_probe_unregister(const char *name, void *probe, void *data)
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
{
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
mutex_lock(&tracepoints_mutex);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_remove_probe(name, probe, data);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
mutex_unlock(&tracepoints_mutex);
if (IS_ERR(old))
return PTR_ERR(old);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
tracepoint_update_probes(); /* may update entry */
release_probes(old);
return 0;
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
}
EXPORT_SYMBOL_GPL(tracepoint_probe_unregister);
static LIST_HEAD(old_probes);
static int need_update;
static void tracepoint_add_old_probes(void *old)
{
need_update = 1;
if (old) {
struct tp_probes *tp_probes = container_of(old,
struct tp_probes, probes[0]);
list_add(&tp_probes->u.list, &old_probes);
}
}
/**
* tracepoint_probe_register_noupdate - register a probe but not connect
* @name: tracepoint name
* @probe: probe handler
*
* caller must call tracepoint_probe_update_all()
*/
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
int tracepoint_probe_register_noupdate(const char *name, void *probe,
void *data)
{
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
mutex_lock(&tracepoints_mutex);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_add_probe(name, probe, data);
if (IS_ERR(old)) {
mutex_unlock(&tracepoints_mutex);
return PTR_ERR(old);
}
tracepoint_add_old_probes(old);
mutex_unlock(&tracepoints_mutex);
return 0;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_register_noupdate);
/**
* tracepoint_probe_unregister_noupdate - remove a probe but not disconnect
* @name: tracepoint name
* @probe: probe function pointer
*
* caller must call tracepoint_probe_update_all()
*/
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
int tracepoint_probe_unregister_noupdate(const char *name, void *probe,
void *data)
{
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
struct tracepoint_func *old;
mutex_lock(&tracepoints_mutex);
tracing: Let tracepoints have data passed to tracepoint callbacks This patch adds data to be passed to tracepoint callbacks. The created functions from DECLARE_TRACE() now need a mandatory data parameter. For example: DECLARE_TRACE(mytracepoint, int value, value) Will create the register function: int register_trace_mytracepoint((void(*)(void *data, int value))probe, void *data); As the first argument, all callbacks (probes) must take a (void *data) parameter. So a callback for the above tracepoint will look like: void myprobe(void *data, int value) { } The callback may choose to ignore the data parameter. This change allows callbacks to register a private data pointer along with the function probe. void mycallback(void *data, int value); register_trace_mytracepoint(mycallback, mydata); Then the mycallback() will receive the "mydata" as the first parameter before the args. A more detailed example: DECLARE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); /* In the C file */ DEFINE_TRACE(mytracepoint, TP_PROTO(int status), TP_ARGS(status)); [...] trace_mytracepoint(status); /* In a file registering this tracepoint */ int my_callback(void *data, int status) { struct my_struct my_data = data; [...] } [...] my_data = kmalloc(sizeof(*my_data), GFP_KERNEL); init_my_data(my_data); register_trace_mytracepoint(my_callback, my_data); The same callback can also be registered to the same tracepoint as long as the data registered is different. Note, the data must also be used to unregister the callback: unregister_trace_mytracepoint(my_callback, my_data); Because of the data parameter, tracepoints declared this way can not have no args. That is: DECLARE_TRACE(mytracepoint, TP_PROTO(void), TP_ARGS()); will cause an error. If no arguments are needed, a new macro can be used instead: DECLARE_TRACE_NOARGS(mytracepoint); Since there are no arguments, the proto and args fields are left out. This is part of a series to make the tracepoint footprint smaller: text data bss dec hex filename 4913961 1088356 861512 6863829 68bbd5 vmlinux.orig 4914025 1088868 861512 6864405 68be15 vmlinux.class 4918492 1084612 861512 6864616 68bee8 vmlinux.tracepoint Again, this patch also increases the size of the kernel, but lays the ground work for decreasing it. v5: Fixed net/core/drop_monitor.c to handle these updates. v4: Moved the DECLARE_TRACE() DECLARE_TRACE_NOARGS out of the #ifdef CONFIG_TRACE_POINTS, since the two are the same in both cases. The __DECLARE_TRACE() is what changes. Thanks to Frederic Weisbecker for pointing this out. v3: Made all register_* functions require data to be passed and all callbacks to take a void * parameter as its first argument. This makes the calling functions comply with C standards. Also added more comments to the modifications of DECLARE_TRACE(). v2: Made the DECLARE_TRACE() have the ability to pass arguments and added a new DECLARE_TRACE_NOARGS() for tracepoints that do not need any arguments. Acked-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: Frederic Weisbecker <fweisbec@gmail.com> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: David S. Miller <davem@davemloft.net> Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
2010-04-21 05:04:50 +08:00
old = tracepoint_remove_probe(name, probe, data);
if (IS_ERR(old)) {
mutex_unlock(&tracepoints_mutex);
return PTR_ERR(old);
}
tracepoint_add_old_probes(old);
mutex_unlock(&tracepoints_mutex);
return 0;
}
EXPORT_SYMBOL_GPL(tracepoint_probe_unregister_noupdate);
/**
* tracepoint_probe_update_all - update tracepoints
*/
void tracepoint_probe_update_all(void)
{
LIST_HEAD(release_probes);
struct tp_probes *pos, *next;
mutex_lock(&tracepoints_mutex);
if (!need_update) {
mutex_unlock(&tracepoints_mutex);
return;
}
if (!list_empty(&old_probes))
list_replace_init(&old_probes, &release_probes);
need_update = 0;
mutex_unlock(&tracepoints_mutex);
tracepoint_update_probes();
list_for_each_entry_safe(pos, next, &release_probes, u.list) {
list_del(&pos->u.list);
call_rcu_sched(&pos->u.rcu, rcu_free_old_probes);
}
}
EXPORT_SYMBOL_GPL(tracepoint_probe_update_all);
tracing: Kernel Tracepoints Implementation of kernel tracepoints. Inspired from the Linux Kernel Markers. Allows complete typing verification by declaring both tracing statement inline functions and probe registration/unregistration static inline functions within the same macro "DEFINE_TRACE". No format string is required. See the tracepoint Documentation and Samples patches for usage examples. Taken from the documentation patch : "A tracepoint placed in code provides a hook to call a function (probe) that you can provide at runtime. A tracepoint can be "on" (a probe is connected to it) or "off" (no probe is attached). When a tracepoint is "off" it has no effect, except for adding a tiny time penalty (checking a condition for a branch) and space penalty (adding a few bytes for the function call at the end of the instrumented function and adds a data structure in a separate section). When a tracepoint is "on", the function you provide is called each time the tracepoint is executed, in the execution context of the caller. When the function provided ends its execution, it returns to the caller (continuing from the tracepoint site). You can put tracepoints at important locations in the code. They are lightweight hooks that can pass an arbitrary number of parameters, which prototypes are described in a tracepoint declaration placed in a header file." Addition and removal of tracepoints is synchronized by RCU using the scheduler (and preempt_disable) as guarantees to find a quiescent state (this is really RCU "classic"). The update side uses rcu_barrier_sched() with call_rcu_sched() and the read/execute side uses "preempt_disable()/preempt_enable()". We make sure the previous array containing probes, which has been scheduled for deletion by the rcu callback, is indeed freed before we proceed to the next update. It therefore limits the rate of modification of a single tracepoint to one update per RCU period. The objective here is to permit fast batch add/removal of probes on _different_ tracepoints. Changelog : - Use #name ":" #proto as string to identify the tracepoint in the tracepoint table. This will make sure not type mismatch happens due to connexion of a probe with the wrong type to a tracepoint declared with the same name in a different header. - Add tracepoint_entry_free_old. - Change __TO_TRACE to get rid of the 'i' iterator. Masami Hiramatsu <mhiramat@redhat.com> : Tested on x86-64. Performance impact of a tracepoint : same as markers, except that it adds about 70 bytes of instructions in an unlikely branch of each instrumented function (the for loop, the stack setup and the function call). It currently adds a memory read, a test and a conditional branch at the instrumentation site (in the hot path). Immediate values will eventually change this into a load immediate, test and branch, which removes the memory read which will make the i-cache impact smaller (changing the memory read for a load immediate removes 3-4 bytes per site on x86_32 (depending on mov prefixes), or 7-8 bytes on x86_64, it also saves the d-cache hit). About the performance impact of tracepoints (which is comparable to markers), even without immediate values optimizations, tests done by Hideo Aoki on ia64 show no regression. His test case was using hackbench on a kernel where scheduler instrumentation (about 5 events in code scheduler code) was added. Quoting Hideo Aoki about Markers : I evaluated overhead of kernel marker using linux-2.6-sched-fixes git tree, which includes several markers for LTTng, using an ia64 server. While the immediate trace mark feature isn't implemented on ia64, there is no major performance regression. So, I think that we don't have any issues to propose merging marker point patches into Linus's tree from the viewpoint of performance impact. I prepared two kernels to evaluate. The first one was compiled without CONFIG_MARKERS. The second one was enabled CONFIG_MARKERS. I downloaded the original hackbench from the following URL: http://devresources.linux-foundation.org/craiger/hackbench/src/hackbench.c I ran hackbench 5 times in each condition and calculated the average and difference between the kernels. The parameter of hackbench: every 50 from 50 to 800 The number of CPUs of the server: 2, 4, and 8 Below is the results. As you can see, major performance regression wasn't found in any case. Even if number of processes increases, differences between marker-enabled kernel and marker- disabled kernel doesn't increase. Moreover, if number of CPUs increases, the differences doesn't increase either. Curiously, marker-enabled kernel is better than marker-disabled kernel in more than half cases, although I guess it comes from the difference of memory access pattern. * 2 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 4.811 | 4.872 | +0.061 | +1.27 | 100 | 9.854 | 10.309 | +0.454 | +4.61 | 150 | 15.602 | 15.040 | -0.562 | -3.6 | 200 | 20.489 | 20.380 | -0.109 | -0.53 | 250 | 25.798 | 25.652 | -0.146 | -0.56 | 300 | 31.260 | 30.797 | -0.463 | -1.48 | 350 | 36.121 | 35.770 | -0.351 | -0.97 | 400 | 42.288 | 42.102 | -0.186 | -0.44 | 450 | 47.778 | 47.253 | -0.526 | -1.1 | 500 | 51.953 | 52.278 | +0.325 | +0.63 | 550 | 58.401 | 57.700 | -0.701 | -1.2 | 600 | 63.334 | 63.222 | -0.112 | -0.18 | 650 | 68.816 | 68.511 | -0.306 | -0.44 | 700 | 74.667 | 74.088 | -0.579 | -0.78 | 750 | 78.612 | 79.582 | +0.970 | +1.23 | 800 | 85.431 | 85.263 | -0.168 | -0.2 | -------------------------------------------------------------- * 4 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.586 | 2.584 | -0.003 | -0.1 | 100 | 5.254 | 5.283 | +0.030 | +0.56 | 150 | 8.012 | 8.074 | +0.061 | +0.76 | 200 | 11.172 | 11.000 | -0.172 | -1.54 | 250 | 13.917 | 14.036 | +0.119 | +0.86 | 300 | 16.905 | 16.543 | -0.362 | -2.14 | 350 | 19.901 | 20.036 | +0.135 | +0.68 | 400 | 22.908 | 23.094 | +0.186 | +0.81 | 450 | 26.273 | 26.101 | -0.172 | -0.66 | 500 | 29.554 | 29.092 | -0.461 | -1.56 | 550 | 32.377 | 32.274 | -0.103 | -0.32 | 600 | 35.855 | 35.322 | -0.533 | -1.49 | 650 | 39.192 | 38.388 | -0.804 | -2.05 | 700 | 41.744 | 41.719 | -0.025 | -0.06 | 750 | 45.016 | 44.496 | -0.520 | -1.16 | 800 | 48.212 | 47.603 | -0.609 | -1.26 | -------------------------------------------------------------- * 8 CPUs Number of | without | with | diff | diff | processes | Marker [Sec] | Marker [Sec] | [Sec] | [%] | -------------------------------------------------------------- 50 | 2.094 | 2.072 | -0.022 | -1.07 | 100 | 4.162 | 4.273 | +0.111 | +2.66 | 150 | 6.485 | 6.540 | +0.055 | +0.84 | 200 | 8.556 | 8.478 | -0.078 | -0.91 | 250 | 10.458 | 10.258 | -0.200 | -1.91 | 300 | 12.425 | 12.750 | +0.325 | +2.62 | 350 | 14.807 | 14.839 | +0.032 | +0.22 | 400 | 16.801 | 16.959 | +0.158 | +0.94 | 450 | 19.478 | 19.009 | -0.470 | -2.41 | 500 | 21.296 | 21.504 | +0.208 | +0.98 | 550 | 23.842 | 23.979 | +0.137 | +0.57 | 600 | 26.309 | 26.111 | -0.198 | -0.75 | 650 | 28.705 | 28.446 | -0.259 | -0.9 | 700 | 31.233 | 31.394 | +0.161 | +0.52 | 750 | 34.064 | 33.720 | -0.344 | -1.01 | 800 | 36.320 | 36.114 | -0.206 | -0.57 | -------------------------------------------------------------- Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Acked-by: Masami Hiramatsu <mhiramat@redhat.com> Acked-by: 'Peter Zijlstra' <peterz@infradead.org> Signed-off-by: Ingo Molnar <mingo@elte.hu>
2008-07-19 00:16:16 +08:00
/**
* tracepoint_get_iter_range - Get a next tracepoint iterator given a range.
* @tracepoint: current tracepoints (in), next tracepoint (out)
* @begin: beginning of the range
* @end: end of the range
*
* Returns whether a next tracepoint has been found (1) or not (0).
* Will return the first tracepoint in the range if the input tracepoint is
* NULL.
*/
int tracepoint_get_iter_range(struct tracepoint **tracepoint,
struct tracepoint *begin, struct tracepoint *end)
{
if (!*tracepoint && begin != end) {
*tracepoint = begin;
return 1;
}
if (*tracepoint >= begin && *tracepoint < end)
return 1;
return 0;
}
EXPORT_SYMBOL_GPL(tracepoint_get_iter_range);
static void tracepoint_get_iter(struct tracepoint_iter *iter)
{
int found = 0;
/* Core kernel tracepoints */
if (!iter->module) {
found = tracepoint_get_iter_range(&iter->tracepoint,
__start___tracepoints, __stop___tracepoints);
if (found)
goto end;
}
/* tracepoints in modules. */
found = module_get_iter_tracepoints(iter);
end:
if (!found)
tracepoint_iter_reset(iter);
}
void tracepoint_iter_start(struct tracepoint_iter *iter)
{
tracepoint_get_iter(iter);
}
EXPORT_SYMBOL_GPL(tracepoint_iter_start);
void tracepoint_iter_next(struct tracepoint_iter *iter)
{
iter->tracepoint++;
/*
* iter->tracepoint may be invalid because we blindly incremented it.
* Make sure it is valid by marshalling on the tracepoints, getting the
* tracepoints from following modules if necessary.
*/
tracepoint_get_iter(iter);
}
EXPORT_SYMBOL_GPL(tracepoint_iter_next);
void tracepoint_iter_stop(struct tracepoint_iter *iter)
{
}
EXPORT_SYMBOL_GPL(tracepoint_iter_stop);
void tracepoint_iter_reset(struct tracepoint_iter *iter)
{
iter->module = NULL;
iter->tracepoint = NULL;
}
EXPORT_SYMBOL_GPL(tracepoint_iter_reset);
#ifdef CONFIG_MODULES
int tracepoint_module_notify(struct notifier_block *self,
unsigned long val, void *data)
{
struct module *mod = data;
switch (val) {
case MODULE_STATE_COMING:
case MODULE_STATE_GOING:
tracepoint_update_probe_range(mod->tracepoints,
mod->tracepoints + mod->num_tracepoints);
break;
}
return 0;
}
struct notifier_block tracepoint_module_nb = {
.notifier_call = tracepoint_module_notify,
.priority = 0,
};
static int init_tracepoints(void)
{
return register_module_notifier(&tracepoint_module_nb);
}
__initcall(init_tracepoints);
#endif /* CONFIG_MODULES */
#ifdef CONFIG_HAVE_SYSCALL_TRACEPOINTS
tracing: Move tracepoint callbacks from declaration to definition It's not strictly correct for the tracepoint reg/unreg callbacks to occur when a client is hooking up, because the actual tracepoint may not be present yet. This happens to be fine for syscall, since that's in the core kernel, but it would cause problems for tracepoints defined in a module that hasn't been loaded yet. It also means the reg/unreg has to be EXPORTed for any modules to use the tracepoint (as in SystemTap). This patch removes DECLARE_TRACE_WITH_CALLBACK, and instead introduces DEFINE_TRACE_FN which stores the callbacks in struct tracepoint. The callbacks are used now when the active state of the tracepoint changes in set_tracepoint & disable_tracepoint. This also introduces TRACE_EVENT_FN, so ftrace events can also provide registration callbacks if needed. Signed-off-by: Josh Stone <jistone@redhat.com> Cc: Jason Baron <jbaron@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Cc: Jiaying Zhang <jiayingz@google.com> Cc: Martin Bligh <mbligh@google.com> Cc: Lai Jiangshan <laijs@cn.fujitsu.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> LKML-Reference: <1251150194-1713-4-git-send-email-jistone@redhat.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
2009-08-25 05:43:13 +08:00
/* NB: reg/unreg are called while guarded with the tracepoints_mutex */
static int sys_tracepoint_refcount;
void syscall_regfunc(void)
{
unsigned long flags;
struct task_struct *g, *t;
if (!sys_tracepoint_refcount) {
read_lock_irqsave(&tasklist_lock, flags);
do_each_thread(g, t) {
tracing: Don't trace kernel thread syscalls Kernel threads don't call syscalls using the sysenter/sysexit path. Instead they directly call the sys_* or do_* functions that implement the syscalls inside the kernel. The current syscall tracepoints only bind the sysenter/sysexit path, then it has no effect to trace the kernel thread calls to syscalls in that path. Setting the TIF_SYSCALL_TRACEPOINT flag is then useless for these. Actually there is only one case when a kernel thread can reach the usual syscall exit tracing path: when we create a kernel thread, the child comes to ret_from_fork and is the fork() return is then traced. But this information alone is useless, then we don't want to set the TIF flags for these threads. Kernel threads have task_struct->mm set to NULL. (Thanks to Heiko for that hint ;-) The idea is then to check the mm field in syscall_regfunc() and set the flag accordingly. Signed-off-by: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Cc: Jason Baron <jbaron@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Lai Jiangshan <laijs@cn.fujitsu.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca> Cc: Jiaying Zhang <jiayingz@google.com> Cc: Martin Bligh <mbligh@google.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Paul Mundt <lethal@linux-sh.org> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: Hendrik Brueckner <brueckner@linux.vnet.ibm.com> LKML-Reference: <20090825160237.GG4639@cetus.boeblingen.de.ibm.com> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
2009-08-26 00:02:37 +08:00
/* Skip kernel threads. */
if (t->mm)
set_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
} while_each_thread(g, t);
read_unlock_irqrestore(&tasklist_lock, flags);
}
sys_tracepoint_refcount++;
}
void syscall_unregfunc(void)
{
unsigned long flags;
struct task_struct *g, *t;
sys_tracepoint_refcount--;
if (!sys_tracepoint_refcount) {
read_lock_irqsave(&tasklist_lock, flags);
do_each_thread(g, t) {
clear_tsk_thread_flag(t, TIF_SYSCALL_TRACEPOINT);
} while_each_thread(g, t);
read_unlock_irqrestore(&tasklist_lock, flags);
}
}
#endif