2005-04-17 06:20:36 +08:00
|
|
|
/*
|
2011-12-22 06:29:42 +08:00
|
|
|
* CPU subsystem support
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
2012-01-10 10:59:49 +08:00
|
|
|
#include <linux/kernel.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/init.h>
|
2006-10-18 13:47:25 +08:00
|
|
|
#include <linux/sched.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/cpu.h>
|
|
|
|
#include <linux/topology.h>
|
|
|
|
#include <linux/device.h>
|
[PATCH] node hotplug: register cpu: remove node struct
With Goto-san's patch, we can add new pgdat/node at runtime. I'm now
considering node-hot-add with cpu + memory on ACPI.
I found acpi container, which describes node, could evaluate cpu before
memory. This means cpu-hot-add occurs before memory hot add.
In most part, cpu-hot-add doesn't depend on node hot add. But register_cpu(),
which creates symbolic link from node to cpu, requires that node should be
onlined before register_cpu(). When a node is onlined, its pgdat should be
there.
This patch-set holds off creating symbolic link from node to cpu
until node is onlined.
This removes node arguments from register_cpu().
Now, register_cpu() requires 'struct node' as its argument. But the array of
struct node is now unified in driver/base/node.c now (By Goto's node hotplug
patch). We can get struct node in generic way. So, this argument is not
necessary now.
This patch also guarantees add cpu under node only when node is onlined. It
is necessary for node-hot-add vs. cpu-hot-add patch following this.
Moreover, register_cpu calculates cpu->node_id by cpu_to_node() without regard
to its 'struct node *root' argument. This patch removes it.
Also modify callers of register_cpu()/unregister_cpu, whose args are changed
by register-cpu-remove-node-struct patch.
[Brice.Goglin@ens-lyon.org: fix it]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 17:53:41 +08:00
|
|
|
#include <linux/node.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/gfp.h>
|
2012-01-26 07:09:14 +08:00
|
|
|
#include <linux/slab.h>
|
2012-01-10 11:04:32 +08:00
|
|
|
#include <linux/percpu.h>
|
ACPI / processor: Use common hotplug infrastructure
Split the ACPI processor driver into two parts, one that is
non-modular, resides in the ACPI core and handles the enumeration
and hotplug of processors and one that implements the rest of the
existing processor driver functionality.
The non-modular part uses an ACPI scan handler object to enumerate
processors on the basis of information provided by the ACPI namespace
and to hook up with the common ACPI hotplug infrastructure. It also
populates the ACPI handle of each processor device having a
corresponding object in the ACPI namespace, which allows the driver
proper to bind to those devices, and makes the driver bind to them
if it is readily available (i.e. loaded) when the scan handler's
.attach() routine is running.
There are a few reasons to make this change.
First, switching the ACPI processor driver to using the common ACPI
hotplug infrastructure reduces code duplication and size considerably,
even though a new file is created along with a header comment etc.
Second, since the common hotplug code attempts to offline devices
before starting the (non-reversible) removal procedure, it will abort
(and possibly roll back) hot-remove operations involving processors
if cpu_down() returns an error code for one of them instead of
continuing them blindly (if /sys/firmware/acpi/hotplug/force_remove
is unset). That is a more desirable behavior than what the current
code does.
Finally, the separation of the scan/hotplug part from the driver
proper makes it possible to simplify the driver's .remove() routine,
because it doesn't need to worry about the possible cleanup related
to processor removal any more (the scan/hotplug part is responsible
for that now) and can handle device removal and driver removal
symmetricaly (i.e. as appropriate).
Some user-visible changes in sysfs are made (for example, the
'sysdev' link from the ACPI device node to the processor device's
directory is gone and a 'physical_node' link is present instead
and a corresponding 'firmware_node' is present in the processor
device's directory, the processor driver is now visible under
/sys/bus/cpu/drivers/ and bound to the processor device), but
that shouldn't affect the functionality that users care about
(frequency scaling, C-states and thermal management).
Tested on my venerable Toshiba Portege R500.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
2013-05-03 06:26:22 +08:00
|
|
|
#include <linux/acpi.h>
|
2013-06-17 19:58:45 +08:00
|
|
|
#include <linux/of.h>
|
2014-02-08 20:34:09 +08:00
|
|
|
#include <linux/cpufeature.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2005-10-14 00:54:41 +08:00
|
|
|
#include "base.h"
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static DEFINE_PER_CPU(struct device *, cpu_sys_devices);
|
2005-10-31 06:59:49 +08:00
|
|
|
|
ACPI / processor: Use common hotplug infrastructure
Split the ACPI processor driver into two parts, one that is
non-modular, resides in the ACPI core and handles the enumeration
and hotplug of processors and one that implements the rest of the
existing processor driver functionality.
The non-modular part uses an ACPI scan handler object to enumerate
processors on the basis of information provided by the ACPI namespace
and to hook up with the common ACPI hotplug infrastructure. It also
populates the ACPI handle of each processor device having a
corresponding object in the ACPI namespace, which allows the driver
proper to bind to those devices, and makes the driver bind to them
if it is readily available (i.e. loaded) when the scan handler's
.attach() routine is running.
There are a few reasons to make this change.
First, switching the ACPI processor driver to using the common ACPI
hotplug infrastructure reduces code duplication and size considerably,
even though a new file is created along with a header comment etc.
Second, since the common hotplug code attempts to offline devices
before starting the (non-reversible) removal procedure, it will abort
(and possibly roll back) hot-remove operations involving processors
if cpu_down() returns an error code for one of them instead of
continuing them blindly (if /sys/firmware/acpi/hotplug/force_remove
is unset). That is a more desirable behavior than what the current
code does.
Finally, the separation of the scan/hotplug part from the driver
proper makes it possible to simplify the driver's .remove() routine,
because it doesn't need to worry about the possible cleanup related
to processor removal any more (the scan/hotplug part is responsible
for that now) and can handle device removal and driver removal
symmetricaly (i.e. as appropriate).
Some user-visible changes in sysfs are made (for example, the
'sysdev' link from the ACPI device node to the processor device's
directory is gone and a 'physical_node' link is present instead
and a corresponding 'firmware_node' is present in the processor
device's directory, the processor driver is now visible under
/sys/bus/cpu/drivers/ and bound to the processor device), but
that shouldn't affect the functionality that users care about
(frequency scaling, C-states and thermal management).
Tested on my venerable Toshiba Portege R500.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
2013-05-03 06:26:22 +08:00
|
|
|
static int cpu_subsys_match(struct device *dev, struct device_driver *drv)
|
|
|
|
{
|
|
|
|
/* ACPI style match is the only one that may succeed. */
|
|
|
|
if (acpi_driver_match_device(dev, drv))
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
2013-04-30 06:08:50 +08:00
|
|
|
static void change_cpu_under_node(struct cpu *cpu,
|
|
|
|
unsigned int from_nid, unsigned int to_nid)
|
|
|
|
{
|
|
|
|
int cpuid = cpu->dev.id;
|
|
|
|
unregister_cpu_under_node(cpuid, from_nid);
|
|
|
|
register_cpu_under_node(cpuid, to_nid);
|
|
|
|
cpu->node_id = to_nid;
|
|
|
|
}
|
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
static int __ref cpu_subsys_online(struct device *dev)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2011-12-22 06:29:42 +08:00
|
|
|
struct cpu *cpu = container_of(dev, struct cpu, dev);
|
2013-05-03 06:25:49 +08:00
|
|
|
int cpuid = dev->id;
|
|
|
|
int from_nid, to_nid;
|
2013-09-26 05:08:27 +08:00
|
|
|
int ret;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
from_nid = cpu_to_node(cpuid);
|
2013-08-13 08:39:30 +08:00
|
|
|
if (from_nid == NUMA_NO_NODE)
|
2013-09-26 05:08:27 +08:00
|
|
|
return -ENODEV;
|
2013-08-13 08:39:30 +08:00
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
ret = cpu_up(cpuid);
|
|
|
|
/*
|
|
|
|
* When hot adding memory to memoryless node and enabling a cpu
|
|
|
|
* on the node, node number of the cpu may internally change.
|
|
|
|
*/
|
|
|
|
to_nid = cpu_to_node(cpuid);
|
|
|
|
if (from_nid != to_nid)
|
|
|
|
change_cpu_under_node(cpu, from_nid, to_nid);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
return ret;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
static int cpu_subsys_offline(struct device *dev)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2013-09-26 05:08:27 +08:00
|
|
|
return cpu_down(dev->id);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2013-05-14 22:46:07 +08:00
|
|
|
|
[PATCH] node hotplug: register cpu: remove node struct
With Goto-san's patch, we can add new pgdat/node at runtime. I'm now
considering node-hot-add with cpu + memory on ACPI.
I found acpi container, which describes node, could evaluate cpu before
memory. This means cpu-hot-add occurs before memory hot add.
In most part, cpu-hot-add doesn't depend on node hot add. But register_cpu(),
which creates symbolic link from node to cpu, requires that node should be
onlined before register_cpu(). When a node is onlined, its pgdat should be
there.
This patch-set holds off creating symbolic link from node to cpu
until node is onlined.
This removes node arguments from register_cpu().
Now, register_cpu() requires 'struct node' as its argument. But the array of
struct node is now unified in driver/base/node.c now (By Goto's node hotplug
patch). We can get struct node in generic way. So, this argument is not
necessary now.
This patch also guarantees add cpu under node only when node is onlined. It
is necessary for node-hot-add vs. cpu-hot-add patch following this.
Moreover, register_cpu calculates cpu->node_id by cpu_to_node() without regard
to its 'struct node *root' argument. This patch removes it.
Also modify callers of register_cpu()/unregister_cpu, whose args are changed
by register-cpu-remove-node-struct patch.
[Brice.Goglin@ens-lyon.org: fix it]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 17:53:41 +08:00
|
|
|
void unregister_cpu(struct cpu *cpu)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2011-12-22 06:29:42 +08:00
|
|
|
int logical_cpu = cpu->dev.id;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
[PATCH] node hotplug: register cpu: remove node struct
With Goto-san's patch, we can add new pgdat/node at runtime. I'm now
considering node-hot-add with cpu + memory on ACPI.
I found acpi container, which describes node, could evaluate cpu before
memory. This means cpu-hot-add occurs before memory hot add.
In most part, cpu-hot-add doesn't depend on node hot add. But register_cpu(),
which creates symbolic link from node to cpu, requires that node should be
onlined before register_cpu(). When a node is onlined, its pgdat should be
there.
This patch-set holds off creating symbolic link from node to cpu
until node is onlined.
This removes node arguments from register_cpu().
Now, register_cpu() requires 'struct node' as its argument. But the array of
struct node is now unified in driver/base/node.c now (By Goto's node hotplug
patch). We can get struct node in generic way. So, this argument is not
necessary now.
This patch also guarantees add cpu under node only when node is onlined. It
is necessary for node-hot-add vs. cpu-hot-add patch following this.
Moreover, register_cpu calculates cpu->node_id by cpu_to_node() without regard
to its 'struct node *root' argument. This patch removes it.
Also modify callers of register_cpu()/unregister_cpu, whose args are changed
by register-cpu-remove-node-struct patch.
[Brice.Goglin@ens-lyon.org: fix it]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 17:53:41 +08:00
|
|
|
unregister_cpu_under_node(logical_cpu, cpu_to_node(logical_cpu));
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
device_unregister(&cpu->dev);
|
2008-05-01 19:35:16 +08:00
|
|
|
per_cpu(cpu_sys_devices, logical_cpu) = NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
return;
|
|
|
|
}
|
2009-11-26 01:23:25 +08:00
|
|
|
|
|
|
|
#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t cpu_probe_store(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
2010-03-16 07:33:32 +08:00
|
|
|
const char *buf,
|
2009-11-26 01:23:25 +08:00
|
|
|
size_t count)
|
|
|
|
{
|
2013-08-30 08:22:07 +08:00
|
|
|
ssize_t cnt;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = lock_device_hotplug_sysfs();
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
cnt = arch_cpu_probe(buf, count);
|
|
|
|
|
|
|
|
unlock_device_hotplug();
|
|
|
|
return cnt;
|
2009-11-26 01:23:25 +08:00
|
|
|
}
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t cpu_release_store(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
2010-03-16 07:33:32 +08:00
|
|
|
const char *buf,
|
2009-11-26 01:23:25 +08:00
|
|
|
size_t count)
|
|
|
|
{
|
2013-08-30 08:22:07 +08:00
|
|
|
ssize_t cnt;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
ret = lock_device_hotplug_sysfs();
|
|
|
|
if (ret)
|
|
|
|
return ret;
|
|
|
|
|
|
|
|
cnt = arch_cpu_release(buf, count);
|
|
|
|
|
|
|
|
unlock_device_hotplug();
|
|
|
|
return cnt;
|
2009-11-26 01:23:25 +08:00
|
|
|
}
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static DEVICE_ATTR(probe, S_IWUSR, NULL, cpu_probe_store);
|
|
|
|
static DEVICE_ATTR(release, S_IWUSR, NULL, cpu_release_store);
|
2009-11-26 01:23:25 +08:00
|
|
|
#endif /* CONFIG_ARCH_CPU_PROBE_RELEASE */
|
2005-04-17 06:20:36 +08:00
|
|
|
#endif /* CONFIG_HOTPLUG_CPU */
|
|
|
|
|
2013-05-03 06:25:49 +08:00
|
|
|
struct bus_type cpu_subsys = {
|
|
|
|
.name = "cpu",
|
|
|
|
.dev_name = "cpu",
|
ACPI / processor: Use common hotplug infrastructure
Split the ACPI processor driver into two parts, one that is
non-modular, resides in the ACPI core and handles the enumeration
and hotplug of processors and one that implements the rest of the
existing processor driver functionality.
The non-modular part uses an ACPI scan handler object to enumerate
processors on the basis of information provided by the ACPI namespace
and to hook up with the common ACPI hotplug infrastructure. It also
populates the ACPI handle of each processor device having a
corresponding object in the ACPI namespace, which allows the driver
proper to bind to those devices, and makes the driver bind to them
if it is readily available (i.e. loaded) when the scan handler's
.attach() routine is running.
There are a few reasons to make this change.
First, switching the ACPI processor driver to using the common ACPI
hotplug infrastructure reduces code duplication and size considerably,
even though a new file is created along with a header comment etc.
Second, since the common hotplug code attempts to offline devices
before starting the (non-reversible) removal procedure, it will abort
(and possibly roll back) hot-remove operations involving processors
if cpu_down() returns an error code for one of them instead of
continuing them blindly (if /sys/firmware/acpi/hotplug/force_remove
is unset). That is a more desirable behavior than what the current
code does.
Finally, the separation of the scan/hotplug part from the driver
proper makes it possible to simplify the driver's .remove() routine,
because it doesn't need to worry about the possible cleanup related
to processor removal any more (the scan/hotplug part is responsible
for that now) and can handle device removal and driver removal
symmetricaly (i.e. as appropriate).
Some user-visible changes in sysfs are made (for example, the
'sysdev' link from the ACPI device node to the processor device's
directory is gone and a 'physical_node' link is present instead
and a corresponding 'firmware_node' is present in the processor
device's directory, the processor driver is now visible under
/sys/bus/cpu/drivers/ and bound to the processor device), but
that shouldn't affect the functionality that users care about
(frequency scaling, C-states and thermal management).
Tested on my venerable Toshiba Portege R500.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
2013-05-03 06:26:22 +08:00
|
|
|
.match = cpu_subsys_match,
|
2013-05-03 06:25:49 +08:00
|
|
|
#ifdef CONFIG_HOTPLUG_CPU
|
|
|
|
.online = cpu_subsys_online,
|
|
|
|
.offline = cpu_subsys_offline,
|
|
|
|
#endif
|
|
|
|
};
|
|
|
|
EXPORT_SYMBOL_GPL(cpu_subsys);
|
|
|
|
|
2006-01-10 12:51:42 +08:00
|
|
|
#ifdef CONFIG_KEXEC
|
|
|
|
#include <linux/kexec.h>
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t show_crash_notes(struct device *dev, struct device_attribute *attr,
|
2008-07-02 00:48:41 +08:00
|
|
|
char *buf)
|
2006-01-10 12:51:42 +08:00
|
|
|
{
|
2011-12-22 06:29:42 +08:00
|
|
|
struct cpu *cpu = container_of(dev, struct cpu, dev);
|
2006-01-10 12:51:42 +08:00
|
|
|
ssize_t rc;
|
|
|
|
unsigned long long addr;
|
|
|
|
int cpunum;
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
cpunum = cpu->dev.id;
|
2006-01-10 12:51:42 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Might be reading other cpu's data based on which cpu read thread
|
|
|
|
* has been scheduled. But cpu data (memory) is allocated once during
|
|
|
|
* boot up and this data does not change there after. Hence this
|
|
|
|
* operation should be safe. No locking required.
|
|
|
|
*/
|
2009-11-24 14:50:03 +08:00
|
|
|
addr = per_cpu_ptr_to_phys(per_cpu_ptr(crash_notes, cpunum));
|
2006-01-10 12:51:42 +08:00
|
|
|
rc = sprintf(buf, "%Lx\n", addr);
|
|
|
|
return rc;
|
|
|
|
}
|
2011-12-22 06:29:42 +08:00
|
|
|
static DEVICE_ATTR(crash_notes, 0400, show_crash_notes, NULL);
|
2013-03-28 16:15:35 +08:00
|
|
|
|
|
|
|
static ssize_t show_crash_notes_size(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
|
|
|
char *buf)
|
|
|
|
{
|
|
|
|
ssize_t rc;
|
|
|
|
|
2013-04-03 23:18:24 +08:00
|
|
|
rc = sprintf(buf, "%zu\n", sizeof(note_buf_t));
|
2013-03-28 16:15:35 +08:00
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
static DEVICE_ATTR(crash_notes_size, 0400, show_crash_notes_size, NULL);
|
2013-05-14 22:46:06 +08:00
|
|
|
|
|
|
|
static struct attribute *crash_note_cpu_attrs[] = {
|
|
|
|
&dev_attr_crash_notes.attr,
|
|
|
|
&dev_attr_crash_notes_size.attr,
|
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct attribute_group crash_note_cpu_attr_group = {
|
|
|
|
.attrs = crash_note_cpu_attrs,
|
|
|
|
};
|
|
|
|
#endif
|
|
|
|
|
|
|
|
static const struct attribute_group *common_cpu_attr_groups[] = {
|
|
|
|
#ifdef CONFIG_KEXEC
|
|
|
|
&crash_note_cpu_attr_group,
|
2006-01-10 12:51:42 +08:00
|
|
|
#endif
|
2013-05-14 22:46:06 +08:00
|
|
|
NULL
|
|
|
|
};
|
2006-01-10 12:51:42 +08:00
|
|
|
|
2013-05-14 22:46:07 +08:00
|
|
|
static const struct attribute_group *hotplugable_cpu_attr_groups[] = {
|
|
|
|
#ifdef CONFIG_KEXEC
|
|
|
|
&crash_note_cpu_attr_group,
|
|
|
|
#endif
|
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
2008-04-09 02:43:04 +08:00
|
|
|
/*
|
|
|
|
* Print cpu online, possible, present, and system maps
|
|
|
|
*/
|
2010-01-05 19:48:00 +08:00
|
|
|
|
|
|
|
struct cpu_attr {
|
2011-12-22 06:29:42 +08:00
|
|
|
struct device_attribute attr;
|
2010-01-05 19:48:00 +08:00
|
|
|
const struct cpumask *const * const map;
|
|
|
|
};
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t show_cpus_attr(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
2010-01-05 19:48:00 +08:00
|
|
|
char *buf)
|
2008-04-09 02:43:04 +08:00
|
|
|
{
|
2010-01-05 19:48:00 +08:00
|
|
|
struct cpu_attr *ca = container_of(attr, struct cpu_attr, attr);
|
2008-04-09 02:43:04 +08:00
|
|
|
|
2014-09-30 21:48:22 +08:00
|
|
|
return cpumap_print_to_pagebuf(true, buf, *ca->map);
|
2008-04-09 02:43:04 +08:00
|
|
|
}
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
#define _CPU_ATTR(name, map) \
|
|
|
|
{ __ATTR(name, 0444, show_cpus_attr, NULL), map }
|
2008-04-09 02:43:04 +08:00
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
/* Keep in sync with cpu_subsys_attrs */
|
2010-01-05 19:48:00 +08:00
|
|
|
static struct cpu_attr cpu_attrs[] = {
|
|
|
|
_CPU_ATTR(online, &cpu_online_mask),
|
|
|
|
_CPU_ATTR(possible, &cpu_possible_mask),
|
|
|
|
_CPU_ATTR(present, &cpu_present_mask),
|
|
|
|
};
|
2008-04-09 02:43:04 +08:00
|
|
|
|
2008-12-16 12:26:48 +08:00
|
|
|
/*
|
|
|
|
* Print values for NR_CPUS and offlined cpus
|
|
|
|
*/
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t print_cpus_kernel_max(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2008-12-16 12:26:48 +08:00
|
|
|
{
|
2009-01-01 10:08:48 +08:00
|
|
|
int n = snprintf(buf, PAGE_SIZE-2, "%d\n", NR_CPUS - 1);
|
2008-12-16 12:26:48 +08:00
|
|
|
return n;
|
|
|
|
}
|
2011-12-22 06:29:42 +08:00
|
|
|
static DEVICE_ATTR(kernel_max, 0444, print_cpus_kernel_max, NULL);
|
2008-12-16 12:26:48 +08:00
|
|
|
|
|
|
|
/* arch-optional setting to enable display of offline cpus >= nr_cpu_ids */
|
|
|
|
unsigned int total_cpus;
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static ssize_t print_cpus_offline(struct device *dev,
|
|
|
|
struct device_attribute *attr, char *buf)
|
2008-12-16 12:26:48 +08:00
|
|
|
{
|
|
|
|
int n = 0, len = PAGE_SIZE-2;
|
|
|
|
cpumask_var_t offline;
|
|
|
|
|
|
|
|
/* display offline cpus < nr_cpu_ids */
|
|
|
|
if (!alloc_cpumask_var(&offline, GFP_KERNEL))
|
|
|
|
return -ENOMEM;
|
2010-04-28 05:01:20 +08:00
|
|
|
cpumask_andnot(offline, cpu_possible_mask, cpu_online_mask);
|
2008-12-16 12:26:48 +08:00
|
|
|
n = cpulist_scnprintf(buf, len, offline);
|
|
|
|
free_cpumask_var(offline);
|
|
|
|
|
|
|
|
/* display offline cpus >= nr_cpu_ids */
|
|
|
|
if (total_cpus && nr_cpu_ids < total_cpus) {
|
|
|
|
if (n && n < len)
|
|
|
|
buf[n++] = ',';
|
|
|
|
|
|
|
|
if (nr_cpu_ids == total_cpus-1)
|
|
|
|
n += snprintf(&buf[n], len - n, "%d", nr_cpu_ids);
|
|
|
|
else
|
|
|
|
n += snprintf(&buf[n], len - n, "%d-%d",
|
|
|
|
nr_cpu_ids, total_cpus-1);
|
|
|
|
}
|
|
|
|
|
|
|
|
n += snprintf(&buf[n], len - n, "\n");
|
|
|
|
return n;
|
|
|
|
}
|
2011-12-22 06:29:42 +08:00
|
|
|
static DEVICE_ATTR(offline, 0444, print_cpus_offline, NULL);
|
2008-12-16 12:26:48 +08:00
|
|
|
|
2012-02-03 02:36:33 +08:00
|
|
|
static void cpu_device_release(struct device *dev)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* This is an empty function to prevent the driver core from spitting a
|
|
|
|
* warning at us. Yes, I know this is directly opposite of what the
|
|
|
|
* documentation for the driver core and kobjects say, and the author
|
|
|
|
* of this code has already been publically ridiculed for doing
|
|
|
|
* something as foolish as this. However, at this point in time, it is
|
|
|
|
* the only way to handle the issue of statically allocated cpu
|
|
|
|
* devices. The different architectures will have their cpu device
|
|
|
|
* code reworked to properly handle this in the near future, so this
|
|
|
|
* function will then be changed to correctly free up the memory held
|
|
|
|
* by the cpu device.
|
|
|
|
*
|
|
|
|
* Never copy this way of doing things, or you too will be made fun of
|
2013-01-15 22:27:46 +08:00
|
|
|
* on the linux-kernel list, you have been warned.
|
2012-02-03 02:36:33 +08:00
|
|
|
*/
|
|
|
|
}
|
|
|
|
|
2014-02-08 20:34:09 +08:00
|
|
|
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
|
|
|
|
static ssize_t print_cpu_modalias(struct device *dev,
|
|
|
|
struct device_attribute *attr,
|
|
|
|
char *buf)
|
|
|
|
{
|
|
|
|
ssize_t n;
|
|
|
|
u32 i;
|
|
|
|
|
|
|
|
n = sprintf(buf, "cpu:type:" CPU_FEATURE_TYPEFMT ":feature:",
|
|
|
|
CPU_FEATURE_TYPEVAL);
|
|
|
|
|
|
|
|
for (i = 0; i < MAX_CPU_FEATURES; i++)
|
|
|
|
if (cpu_have_feature(i)) {
|
|
|
|
if (PAGE_SIZE < n + sizeof(",XXXX\n")) {
|
|
|
|
WARN(1, "CPU features overflow page\n");
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
n += sprintf(&buf[n], ",%04X", i);
|
|
|
|
}
|
|
|
|
buf[n++] = '\n';
|
|
|
|
return n;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int cpu_uevent(struct device *dev, struct kobj_uevent_env *env)
|
|
|
|
{
|
|
|
|
char *buf = kzalloc(PAGE_SIZE, GFP_KERNEL);
|
|
|
|
if (buf) {
|
|
|
|
print_cpu_modalias(NULL, NULL, buf);
|
|
|
|
add_uevent_var(env, "MODALIAS=%s", buf);
|
|
|
|
kfree(buf);
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
2007-02-18 02:13:42 +08:00
|
|
|
* register_cpu - Setup a sysfs device for a CPU.
|
2006-12-07 09:14:10 +08:00
|
|
|
* @cpu - cpu->hotpluggable field set to 1 will generate a control file in
|
|
|
|
* sysfs for this CPU.
|
2005-04-17 06:20:36 +08:00
|
|
|
* @num - CPU number to use when creating the device.
|
|
|
|
*
|
|
|
|
* Initialize and register the CPU device.
|
|
|
|
*/
|
2013-06-20 03:22:41 +08:00
|
|
|
int register_cpu(struct cpu *cpu, int num)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
int error;
|
[PATCH] node hotplug: register cpu: remove node struct
With Goto-san's patch, we can add new pgdat/node at runtime. I'm now
considering node-hot-add with cpu + memory on ACPI.
I found acpi container, which describes node, could evaluate cpu before
memory. This means cpu-hot-add occurs before memory hot add.
In most part, cpu-hot-add doesn't depend on node hot add. But register_cpu(),
which creates symbolic link from node to cpu, requires that node should be
onlined before register_cpu(). When a node is onlined, its pgdat should be
there.
This patch-set holds off creating symbolic link from node to cpu
until node is onlined.
This removes node arguments from register_cpu().
Now, register_cpu() requires 'struct node' as its argument. But the array of
struct node is now unified in driver/base/node.c now (By Goto's node hotplug
patch). We can get struct node in generic way. So, this argument is not
necessary now.
This patch also guarantees add cpu under node only when node is onlined. It
is necessary for node-hot-add vs. cpu-hot-add patch following this.
Moreover, register_cpu calculates cpu->node_id by cpu_to_node() without regard
to its 'struct node *root' argument. This patch removes it.
Also modify callers of register_cpu()/unregister_cpu, whose args are changed
by register-cpu-remove-node-struct patch.
[Brice.Goglin@ens-lyon.org: fix it]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 17:53:41 +08:00
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
cpu->node_id = cpu_to_node(num);
|
2012-02-09 07:11:17 +08:00
|
|
|
memset(&cpu->dev, 0x00, sizeof(struct device));
|
2011-12-22 06:29:42 +08:00
|
|
|
cpu->dev.id = num;
|
|
|
|
cpu->dev.bus = &cpu_subsys;
|
2012-02-03 02:36:33 +08:00
|
|
|
cpu->dev.release = cpu_device_release;
|
2013-05-03 06:25:49 +08:00
|
|
|
cpu->dev.offline_disabled = !cpu->hotpluggable;
|
2013-05-30 06:30:05 +08:00
|
|
|
cpu->dev.offline = !cpu_online(num);
|
2013-06-17 19:58:45 +08:00
|
|
|
cpu->dev.of_node = of_get_cpu_node(num, NULL);
|
2014-02-08 20:34:10 +08:00
|
|
|
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
|
2014-02-08 20:34:09 +08:00
|
|
|
cpu->dev.bus->uevent = cpu_uevent;
|
2012-01-26 07:09:14 +08:00
|
|
|
#endif
|
2013-05-14 22:46:06 +08:00
|
|
|
cpu->dev.groups = common_cpu_attr_groups;
|
2013-05-14 22:46:07 +08:00
|
|
|
if (cpu->hotpluggable)
|
|
|
|
cpu->dev.groups = hotplugable_cpu_attr_groups;
|
2011-12-22 06:29:42 +08:00
|
|
|
error = device_register(&cpu->dev);
|
2005-10-31 06:59:49 +08:00
|
|
|
if (!error)
|
2011-12-22 06:29:42 +08:00
|
|
|
per_cpu(cpu_sys_devices, num) = &cpu->dev;
|
[PATCH] node hotplug: register cpu: remove node struct
With Goto-san's patch, we can add new pgdat/node at runtime. I'm now
considering node-hot-add with cpu + memory on ACPI.
I found acpi container, which describes node, could evaluate cpu before
memory. This means cpu-hot-add occurs before memory hot add.
In most part, cpu-hot-add doesn't depend on node hot add. But register_cpu(),
which creates symbolic link from node to cpu, requires that node should be
onlined before register_cpu(). When a node is onlined, its pgdat should be
there.
This patch-set holds off creating symbolic link from node to cpu
until node is onlined.
This removes node arguments from register_cpu().
Now, register_cpu() requires 'struct node' as its argument. But the array of
struct node is now unified in driver/base/node.c now (By Goto's node hotplug
patch). We can get struct node in generic way. So, this argument is not
necessary now.
This patch also guarantees add cpu under node only when node is onlined. It
is necessary for node-hot-add vs. cpu-hot-add patch following this.
Moreover, register_cpu calculates cpu->node_id by cpu_to_node() without regard
to its 'struct node *root' argument. This patch removes it.
Also modify callers of register_cpu()/unregister_cpu, whose args are changed
by register-cpu-remove-node-struct patch.
[Brice.Goglin@ens-lyon.org: fix it]
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Yasunori Goto <y-goto@jp.fujitsu.com>
Cc: Ashok Raj <ashok.raj@intel.com>
Cc: Dave Hansen <haveblue@us.ibm.com>
Signed-off-by: Brice Goglin <Brice.Goglin@ens-lyon.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-06-27 17:53:41 +08:00
|
|
|
if (!error)
|
|
|
|
register_cpu_under_node(num, cpu_to_node(num));
|
2006-01-10 12:51:42 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
struct device *get_cpu_device(unsigned cpu)
|
2005-10-31 06:59:49 +08:00
|
|
|
{
|
2008-05-01 19:35:16 +08:00
|
|
|
if (cpu < nr_cpu_ids && cpu_possible(cpu))
|
|
|
|
return per_cpu(cpu_sys_devices, cpu);
|
2005-10-31 06:59:49 +08:00
|
|
|
else
|
|
|
|
return NULL;
|
|
|
|
}
|
2011-12-22 06:29:42 +08:00
|
|
|
EXPORT_SYMBOL_GPL(get_cpu_device);
|
|
|
|
|
2014-09-30 21:48:24 +08:00
|
|
|
static void device_create_release(struct device *dev)
|
|
|
|
{
|
|
|
|
kfree(dev);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct device *
|
|
|
|
__cpu_device_create(struct device *parent, void *drvdata,
|
|
|
|
const struct attribute_group **groups,
|
|
|
|
const char *fmt, va_list args)
|
|
|
|
{
|
|
|
|
struct device *dev = NULL;
|
|
|
|
int retval = -ENODEV;
|
|
|
|
|
|
|
|
dev = kzalloc(sizeof(*dev), GFP_KERNEL);
|
|
|
|
if (!dev) {
|
|
|
|
retval = -ENOMEM;
|
|
|
|
goto error;
|
|
|
|
}
|
|
|
|
|
|
|
|
device_initialize(dev);
|
|
|
|
dev->parent = parent;
|
|
|
|
dev->groups = groups;
|
|
|
|
dev->release = device_create_release;
|
|
|
|
dev_set_drvdata(dev, drvdata);
|
|
|
|
|
|
|
|
retval = kobject_set_name_vargs(&dev->kobj, fmt, args);
|
|
|
|
if (retval)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
retval = device_add(dev);
|
|
|
|
if (retval)
|
|
|
|
goto error;
|
|
|
|
|
|
|
|
return dev;
|
|
|
|
|
|
|
|
error:
|
|
|
|
put_device(dev);
|
|
|
|
return ERR_PTR(retval);
|
|
|
|
}
|
|
|
|
|
|
|
|
struct device *cpu_device_create(struct device *parent, void *drvdata,
|
|
|
|
const struct attribute_group **groups,
|
|
|
|
const char *fmt, ...)
|
|
|
|
{
|
|
|
|
va_list vargs;
|
|
|
|
struct device *dev;
|
|
|
|
|
|
|
|
va_start(vargs, fmt);
|
|
|
|
dev = __cpu_device_create(parent, drvdata, groups, fmt, vargs);
|
|
|
|
va_end(vargs);
|
|
|
|
return dev;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(cpu_device_create);
|
|
|
|
|
2014-02-08 20:34:10 +08:00
|
|
|
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
|
2014-02-08 20:34:09 +08:00
|
|
|
static DEVICE_ATTR(modalias, 0444, print_cpu_modalias, NULL);
|
2012-01-26 07:09:14 +08:00
|
|
|
#endif
|
|
|
|
|
2011-12-22 06:29:42 +08:00
|
|
|
static struct attribute *cpu_root_attrs[] = {
|
|
|
|
#ifdef CONFIG_ARCH_CPU_PROBE_RELEASE
|
|
|
|
&dev_attr_probe.attr,
|
|
|
|
&dev_attr_release.attr,
|
|
|
|
#endif
|
|
|
|
&cpu_attrs[0].attr.attr,
|
|
|
|
&cpu_attrs[1].attr.attr,
|
|
|
|
&cpu_attrs[2].attr.attr,
|
|
|
|
&dev_attr_kernel_max.attr,
|
|
|
|
&dev_attr_offline.attr,
|
2014-02-08 20:34:10 +08:00
|
|
|
#ifdef CONFIG_GENERIC_CPU_AUTOPROBE
|
2012-01-26 07:09:14 +08:00
|
|
|
&dev_attr_modalias.attr,
|
|
|
|
#endif
|
2011-12-22 06:29:42 +08:00
|
|
|
NULL
|
|
|
|
};
|
|
|
|
|
|
|
|
static struct attribute_group cpu_root_attr_group = {
|
|
|
|
.attrs = cpu_root_attrs,
|
|
|
|
};
|
|
|
|
|
|
|
|
static const struct attribute_group *cpu_root_attr_groups[] = {
|
|
|
|
&cpu_root_attr_group,
|
|
|
|
NULL,
|
|
|
|
};
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2011-12-04 05:06:50 +08:00
|
|
|
bool cpu_is_hotpluggable(unsigned cpu)
|
|
|
|
{
|
2012-01-08 04:03:30 +08:00
|
|
|
struct device *dev = get_cpu_device(cpu);
|
|
|
|
return dev && container_of(dev, struct cpu, dev)->hotpluggable;
|
2011-12-04 05:06:50 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL_GPL(cpu_is_hotpluggable);
|
|
|
|
|
2012-01-10 11:04:32 +08:00
|
|
|
#ifdef CONFIG_GENERIC_CPU_DEVICES
|
|
|
|
static DEFINE_PER_CPU(struct cpu, cpu_devices);
|
|
|
|
#endif
|
|
|
|
|
|
|
|
static void __init cpu_dev_register_generic(void)
|
|
|
|
{
|
|
|
|
#ifdef CONFIG_GENERIC_CPU_DEVICES
|
|
|
|
int i;
|
|
|
|
|
|
|
|
for_each_possible_cpu(i) {
|
|
|
|
if (register_cpu(&per_cpu(cpu_devices, i), i))
|
|
|
|
panic("Failed to register CPU device");
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2012-01-10 10:59:49 +08:00
|
|
|
void __init cpu_dev_init(void)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2012-01-10 10:59:49 +08:00
|
|
|
if (subsys_system_register(&cpu_subsys, cpu_root_attr_groups))
|
|
|
|
panic("Failed to register CPU subsystem");
|
2011-12-22 06:29:42 +08:00
|
|
|
|
2012-01-10 11:04:32 +08:00
|
|
|
cpu_dev_register_generic();
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|