x86-64 has a problem: per-cpu variables are actually represented by
their absolute offsets within the per-cpu area, but the symbols are
not emitted as absolute. Thus kallsyms naively creates them as offsets
from _text, meaning their values change if the kernel is relocated
(especially noticeable with CONFIG_RANDOMIZE_BASE):
$ egrep ' (gdt_|_(stext|_per_cpu_))' /root/kallsyms.nokaslr
0000000000000000 D __per_cpu_start
0000000000004000 D gdt_page
0000000000014280 D __per_cpu_end
ffffffff810001c8 T _stext
ffffffff81ee53c0 D __per_cpu_offset
$ egrep ' (gdt_|_(stext|_per_cpu_))' /root/kallsyms.kaslr1
000000001f200000 D __per_cpu_start
000000001f204000 D gdt_page
000000001f214280 D __per_cpu_end
ffffffffa02001c8 T _stext
ffffffffa10e53c0 D __per_cpu_offset
Making them absolute symbols is the Right Thing, but requires fixes to
the relocs tool. So for the moment, we add a --absolute-percpu option
which makes them absolute from a kallsyms perspective:
$ egrep ' (gdt_|_(stext|_per_cpu_))' /proc/kallsyms # no KASLR
0000000000000000 A __per_cpu_start
000000000000a000 A gdt_page
0000000000013040 A __per_cpu_end
ffffffff802001c8 T _stext
ffffffff8099b180 D __per_cpu_offset
ffffffff809a3000 D __per_cpu_load
$ egrep ' (gdt_|_(stext|_per_cpu_))' /proc/kallsyms # With KASLR
0000000000000000 A __per_cpu_start
000000000000a000 A gdt_page
0000000000013040 A __per_cpu_end
ffffffff89c001c8 T _stext
ffffffff8a39d180 D __per_cpu_offset
ffffffff8a3a5000 D __per_cpu_load
Based-on-the-original-screenplay-by: Andy Honig <ahonig@google.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Acked-by: Kees Cook <keescook@chromium.org>
This refactors the address range checks to be generalized instead of
specific to text range checks, in preparation for other range checks.
Also extracts logic for "is the symbol absolute" into a function.
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Revert the recently applied 0f55159d09 ("kallsyms: fix absolute
addresses for kASLR"). Kees said
: This got NAKed, please don't apply -- this patch works for x86 and
: ARM, but may cause problems for others:
:
: https://lkml.org/lkml/2014/2/24/718
It appears that Kees will be fixing all this up for 3.15.
Cc: Andy Honig <ahonig@google.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Currently symbols that are absolute addresses are incorrectly displayed
in /proc/kallsyms if the kernel is loaded with kASLR.
The problem was that the scripts/kallsyms.c file which generates the
array of symbol names and addresses uses an relocatable value for all
symbols, even absolute symbols. This patch fixes that.
Several kallsyms output in different boot states for comparison:
$ egrep '_(stext|_per_cpu_(start|end))' /root/kallsyms.nokaslr
0000000000000000 D __per_cpu_start
0000000000014280 D __per_cpu_end
ffffffff810001c8 T _stext
$ egrep '_(stext|_per_cpu_(start|end))' /root/kallsyms.kaslr1
000000001f200000 D __per_cpu_start
000000001f214280 D __per_cpu_end
ffffffffa02001c8 T _stext
$ egrep '_(stext|_per_cpu_(start|end))' /root/kallsyms.kaslr2
000000000d400000 D __per_cpu_start
000000000d414280 D __per_cpu_end
ffffffff8e4001c8 T _stext
$ egrep '_(stext|_per_cpu_(start|end))' /root/kallsyms.kaslr-fixed
0000000000000000 D __per_cpu_start
0000000000014280 D __per_cpu_end
ffffffffadc001c8 T _stext
Signed-off-by: Andy Honig <ahonig@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: Rusty Russell <rusty@rustcorp.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull kbuild changes from Michal Marek:
- LTO fixes, but the kallsyms part had to be reverted
- Pass -Werror=implicit-int and -Werror=strict-prototypes to the
compiler by default
- snprintf fix in modpost
- remove GREP_OPTIONS from the environment to be immune against exotic
grep option settings
* 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
kallsyms: Revert back to 128 max symbol length
Kbuild: Ignore GREP_OPTIONS env variable
scripts: kallsyms: Use %zu to print 'size_t'
scripts/bloat-o-meter: use .startswith rather than fragile slicing
scripts/bloat-o-meter: ignore changes in the size of linux_banner
kbuild: replace unbounded sprintf call in modpost
kbuild, bloat-o-meter: fix static detection
Kbuild: Handle longer symbols in kallsyms.c
kbuild: Increase kallsyms max symbol length
Makefile: enable -Werror=implicit-int and -Werror=strict-prototypes by default
This reverts commits
f3462aa (Kbuild: Handle longer symbols in kallsyms.c) and
eea0e9c (kbuild: Increase kallsyms max symbol length)
except for the added overflow check. The reason is a regression caused
by increasing the buffer:
http://marc.info/?l=linux-kernel&m=138387700415675.
Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Joe Mario <jmario@redhat.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Commit f3462aa95 (Kbuild: Handle longer symbols in kallsyms.c) introduced the
following warning on ARM:
scripts/kallsyms.c:121:4: warning: format '%lu' expects argument of type 'long unsigned int', but argument 4 has type 'size_t' [-Wformat]
Use %zu to print 'size_t'.
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Also warn for too long symbols
v2: Add missing newline. Use 255 max (Joe Perches)
Signed-off-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
This patch uses CONFIG_PAGE_OFFSET to filter symbols which
are not in kernel address space because these symbols are
generally for generating code purpose and can't be run at
kernel mode, so we needn't keep them in /proc/kallsyms.
For example, on ARM there are some symbols which may be
linked in relocatable code section, then perf can't parse
symbols any more from /proc/kallsyms, this patch fixes the
problem (introduced b9b32bf70f)
Cc: Russell King <linux@arm.linux.org.uk>
Cc: linux-arm-kernel@lists.infradead.org
Cc: Michal Marek <mmarek@suse.cz>
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Cc: stable@vger.kernel.org
Description:
This bug hardly appears during real kernel compiling,
because the vmlinux symbols table is huge.
But we can still catch it under strict condition , as follows.
$ echo "c101b97b T do_fork" | ./scripts/kallsyms --all-symbols
#include <asm/types.h>
......
......
.globl kallsyms_token_table
ALGN
kallsyms_token_table:
Segmentation fault (core dumped)
$
If symbols table is small, all entries in token_profit[0x10000] may
decrease to 0 after several calls of compress_symbols() in optimize_result().
In that case, find_best_token() always return 0 and
best_table[i] is set to "\0\0" and best_table_len[i] is set to 2.
As a result, expand_symbol(best_table[0]="\0\0", best_table_len[0]=2, buf)
in write_src() will run in infinite recursion until stack overflows,
causing segfault.
This patch checks the find_best_token() return value. If all entries in
token_profit[0x10000] become 0 according to return value, it breaks the loop
in optimize_result().
And expand_symbol() works well when best_table_len[i] is 0.
Signed-off-by: Xiaochen Wang <wangxiaochen0@gmail.com>
Acked-by: Paulo Marques <pmarques@grupopie.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
As no error was handled, we wouldn't be able to know when an error does
occur. The fix preserves error messages while it doesn't let unnecessary
compiling warnings show up.
Signed-off-by: Jean Sacren <sakiwit@gmail.com>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Suppress a warn_unused_result warning.
fgets is called as a part of error handling. It is called just to drop a
line and return immediately. read_map is reading the file in a loop and
read_symbol reads line by line. So I think there is no point in using
return value for useful checking. Other checks like 3 items were returned
or !EOF have already been done.
Signed-off-by: Himanshu Chauhan <hschauhan@nulltrace.org>
Cc: WANG Cong <xiyou.wangcong@gmail.com>
Cc: Michal Marek <mmarek@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Michal Marek <mmarek@suse.cz>
Commit b478b782e1 "kallsyms, tracing: output
more proper symbol name" introduces a "bugfix" that introduces a segfault
in kallsyms in my configurations.
The cause is the introduction of prefix_underscores_count() which attempts
to count underscores, even in symbols that do not have them. As a result,
it just uselessly runs past the end of the buffer until it crashes:
CC init/version.o
LD init/built-in.o
LD .tmp_vmlinux1
KSYM .tmp_kallsyms1.S
/bin/sh: line 1: 16934 Done sh-linux-gnu-nm -n .tmp_vmlinux1
16935 Segmentation fault | scripts/kallsyms > .tmp_kallsyms1.S
make: *** [.tmp_kallsyms1.S] Error 139
This simplifies the logic and just does a straightforward count.
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
Reviewed-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Lai Jiangshan <laijs@cn.fujitsu.com>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Paulo Marques <pmarques@grupopie.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: <stable@kernel.org> [2.6.30.x, 2.6.31.x]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The previous commit (17b1f0de) introduced a slightly broken consolidation
of the memory text range checking.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
The Blackfin arch has a discontiguous .text layout due to having on-chip
instruction memory and no virtual memory support. As such, we need to
add explicit checks for these additional .text regions.
Signed-off-by: Robin Getz <robin.getz@analog.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Impact: bugfix, output more reliable symbol lookup result
Debug tools(dump_stack(), ftrace...) are like to print out symbols.
But it is always print out the first aliased symbol.(Aliased symbols
are symbols with the same address), and the first aliased symbol is
sometime not proper.
# echo function_graph > current_tracer
# cat trace
......
1) 1.923 us | select_nohz_load_balancer();
1) + 76.692 us | }
1) | default_idle() {
1) ==========> | __irqentry_text_start() {
1) 0.000 us | native_apic_mem_write();
1) | irq_enter() {
1) 0.000 us | idle_cpu();
1) | tick_check_idle() {
1) 0.000 us | tick_check_oneshot_broadcast();
1) | tick_nohz_stop_idle() {
......
It's very embarrassing, it ouputs "__irqentry_text_start()",
actually, it should output "smp_apic_timer_interrupt()".
(these two symbol are the same address, but "__irqentry_text_start"
is deemed to the first aliased symbol by scripts/kallsyms)
This patch puts symbols like "__irqentry_text_start" to the second
aliased symbols. And a more proper symbol name becomes the first.
Aliased symbols mostly come from linker script. The solution is
guessing "is this symbol defined in linker script", the symbols
defined in linker script will not become the first aliased symbol.
And if symbols are found to be equal in this "linker script provided"
criteria, symbols are sorted by the number of prefix underscores.
Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Paulo Marques <pmarques@grupopie.com>
LKML-Reference: <49BA06E2.7080807@cn.fujitsu.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
This reverts commit ad7a953c52.
And commit: ("allow stripping of generated symbols under CONFIG_KALLSYMS_ALL")
9bb482476c
These stripping patches has caused a set of issues:
1) People have reported compatibility issues with binutils due to
lack of support for `--strip-unneeded-symbols' with objcopy 2.15.92.0.2
Reported by: Wenji
2) ccache and distcc no longer works as expeced
Reported by: Ted, Roland, + others
3) The installed modules increased a lot in size
Reported by: Ted, Davej + others
Reported-by: Wenji Huang <wenji.huang@oracle.com>
Reported-by: "Theodore Ts'o" <tytso@mit.edu>
Reported-by: Dave Jones <davej@redhat.com>
Reported-by: Roland McGrath <roland@redhat.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Building upon parts of the module stripping patch, this patch
introduces similar stripping for vmlinux when CONFIG_KALLSYMS_ALL=y.
Using CONFIG_KALLSYMS_STRIP_GENERATED reduces the overhead of
CONFIG_KALLSYMS_ALL from 245k/310k to 65k/80k for the (i386/x86-64)
kernels I tested with.
The patch also does away with the need to special case the kallsyms-
internal symbols by making them available even in the first linking
stage.
While it is a generated file, the patch includes the changes to
scripts/genksyms/keywords.c_shipped, as I'm unsure what the procedure
here is.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Andi Kleen <andi@firstfloor.org>
reported that he saw a lot of symbols like this:
0000000000000b24 N DW.aio.h.903a6d92.2
0000000000000bce N DW.task_io_accounting.h.8d8de327.0
0000000000000bec N DW.hrtimer.h.c23659c6.0
in his System.map / kallsyms output.
Simple solution is to skip all debugging
symbols (they are marked 'N').
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: Paulo Marques <pmarques@grupopie.com>
Pointed out by Paulo:
"When I wrote this initially, it was a mistake to add a Changelog in
the first place, but I didn't know better at the time.
If you're going to make changes to this file, please remove all the
Changelog, instead of adding more entries to it. The 'Changelog'
should be kept by the version control system, and not the source code
itself."
Cc: Paulo Marques <pmarques@grupopie.com>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
Acked-by: Paulo Marques <pmarques@grupopie.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When resolving symbol names from addresses with aliased symbol names,
kallsyms_lookup always returns the first symbol, even if it is a weak
symbol.
This patch changes this by sorting the symbols with the weak symbols last
before feeding them to the kernel. This way the kernel runtime isn't
changed at all, only the kallsyms build system is changed.
Another side effect is that the symbols get sorted by address, too. So,
even if future binutils version have some bug in "nm" that makes it fail to
correctly sort symbols by address, the kernel won't be affected by this.
Mathieu says:
I created a module in LTTng that uses kallsyms to get the symbol
corresponding to a specific system call address. Unfortunately, all the
unimplemented syscalls were all referring to the (same) weak symbol
identifying an unrelated system call rather that sys_ni (or whatever
non-weak symbol would be expected). Kallsyms was dumbly returning the first
symbol that matched.
This patch makes sure kallsyms returns the non-weak symbol when there is
one, which seems to be the expected result.
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
Looks-great-to: Rusty Russell <rusty@rustcorp.com.au>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When passing a zero address to kallsyms_lookup(), the kernel thought it was
a valid kernel address, even if it is not. This is because is_ksym_addr()
called is_kernel_extratext() and checked against labels that don't exist on
many archs (which default as zero). Since PPC was the only kernel which
defines _extra_text, (in 2005), and no longer needs it, this patch removes
_extra_text support.
For some history (provided by Jon):
http://ozlabs.org/pipermail/linuxppc-dev/2005-September/019734.htmlhttp://ozlabs.org/pipermail/linuxppc-dev/2005-September/019736.htmlhttp://ozlabs.org/pipermail/linuxppc-dev/2005-September/019751.html
[akpm@linux-foundation.org: coding-style fixes]
Signed-off-by: Robin Getz <rgetz@blackfin.uclinux.org>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Jon Loeliger <jdl@freescale.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
* git://git.kernel.org/pub/scm/linux/kernel/git/sam/kbuild: (33 commits)
xtensa: use DATA_DATA in xtensa
powerpc: add missing DATA_DATA to powerpc
cris: use DATA_DATA in cris
kallsyms: remove usage of memmem and _GNU_SOURCE from scripts/kallsyms.c
kbuild: use -fno-optimize-sibling-calls unconditionally
kconfig: reset generated values only if Kconfig and .config agree.
kbuild: fix the warning when running make tags
kconfig: strip 'CONFIG_' automatically in kernel configuration search
kbuild: use POSIX BRE in headers install target
Whitelist references from __dbe_table to .init
modpost white list pattern adjustment
kbuild: do section mismatch check on full vmlinux
kbuild: whitelist references from variables named _timer to .init.text
kbuild: remove hardcoded _logo names from modpost
kbuild: remove hardcoded apic_es7000 from modpost
kbuild: warn about references from .init.text to .exit.text
kbuild: consolidate section checks
kbuild: refactor code in modpost to improve maintainability
kbuild: ignore section mismatch warnings originating from .note section
kbuild: .paravirtprobe section is obsolete, so modpost doesn't need to handle it
...
KSYM_NAME_LEN is peculiar in that it does not include the space for the
trailing '\0', forcing all users to use KSYM_NAME_LEN + 1 when allocating
buffer. This is nonsense and error-prone. Moreover, when the caller
forgets that it's very likely to subtly bite back by corrupting the stack
because the last position of the buffer is always cleared to zero.
This patch increments KSYM_NAME_LEN by one and updates code accordingly.
* off-by-one bug in asm-powerpc/kprobes.h::kprobe_lookup_name() macro
is fixed.
* Where MODULE_NAME_LEN and KSYM_NAME_LEN were used together,
MODULE_NAME_LEN was treated as if it didn't include space for the
trailing '\0'. Fix it.
Signed-off-by: Tejun Heo <htejun@gmail.com>
Acked-by: Paulo Marques <pmarques@grupopie.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
The only in-kernel user of "memmem" is scripts/kallsyms.c and it only
uses it to find tokens that are 2 bytes in size. It is trivial to
replace it with a simple function that finds 2-byte tokens.
This should help users from systems that don't have the memmem GNU
extension available.
Signed-off-by: Paulo Marques <pmarques@grupopie.com>
Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Kallsyms data is never written to, so it can as well benefit from
CONFIG_DEBUG_RODATA.
Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
o On some platforms like avr32, section init comes before .text and
not necessarily a symbol's relative position w.r.t _text is positive.
In such cases assembler detects the overflow and emits warning. This
patch fixes it.
Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Cc: Andi Kleen <ak@suse.de>
Cc: Haavard Skinnemoen <hskinnemoen@atmel.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Print the addresses of non-absolute symbols relative to _text
so that ld will generate relocations. Allowing a relocatable
kernel to relocate them. We can't actually use the symbol names
because kallsyms includes static symbols that are not exported
from their object files.
Add the _text symbol definitions to the architectures which don't
define it otherwise linker will fail.
Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Andi Kleen <ak@suse.de>
Local symbols generated by gcc start with a `$'; no point in including them
in the kernel.
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
This patch changes the way the compression algorithm works. The base
algorithm is similiar to the previous but we force the compressed token
size to 2.
Having a fixed size compressed token allows for a lot of optimizations, and
that in turn allows this code to run over *all* the symbols faster than it
did before over just a subset.
Having it work over all the symbols will make it behave better when symbols
change positions between passes, and the "inconsistent kallsyms" messages
should become less frequent.
In my tests the compression ratio was degraded by about 0.5%, but the
results will depend greatly on the number of symbols to compress.
Signed-off-by: Paulo Marques <pmarques@grupopie.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
The PPC32 kernel puts platform-specific functions into separate sections so
that unneeded parts of it can be freed when we've booted and actually
worked out what we're running on today.
This makes kallsyms ignore those functions, because they're not between
_[se]text or _[se]inittext. Rather than teaching kallsyms about the
various pmac/chrp/etc sections, this patch adds '_[se]extratext' markers
for kallsyms.
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
kallsyms does not consider SYMBOL_PREFIX of C. Consequently it does not
work on architectures using that prefix character (h8300, v850).
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
Initial git repository build. I'm not bothering with the full history,
even though we have it. We can create a separate "historical" git
archive of that later if we want to, and in the meantime it's about
3.2GB when imported into git - space that would just make the early
git days unnecessarily complicated, when we don't have a lot of good
infrastructure for it.
Let it rip!
Rusty Russell