13 Jun, 2014
1 commit
-
Pull more perf updates from Ingo Molnar:
"A second round of perf updates:- wide reaching kprobes sanitization and robustization, with the hope
of fixing all 'probe this function crashes the kernel' bugs, by
Masami Hiramatsu.- uprobes updates from Oleg Nesterov: tmpfs support, corner case
fixes and robustization work.- perf tooling updates and fixes from Jiri Olsa, Namhyung Ki, Arnaldo
et al:
* Add support to accumulate hist periods (Namhyung Kim)
* various fixes, refactorings and enhancements"* 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (101 commits)
perf: Differentiate exec() and non-exec() comm events
perf: Fix perf_event_comm() vs. exec() assumption
uprobes/x86: Rename arch_uprobe->def to ->defparam, minor comment updates
perf/documentation: Add description for conditional branch filter
perf/x86: Add conditional branch filtering support
perf/tool: Add conditional branch filter 'cond' to perf record
perf: Add new conditional branch filter 'PERF_SAMPLE_BRANCH_COND'
uprobes: Teach copy_insn() to support tmpfs
uprobes: Shift ->readpage check from __copy_insn() to uprobe_register()
perf/x86: Use common PMU interrupt disabled code
perf/ARM: Use common PMU interrupt disabled code
perf: Disable sampled events if no PMU interrupt
perf: Fix use after free in perf_remove_from_context()
perf tools: Fix 'make help' message error
perf record: Fix poll return value propagation
perf tools: Move elide bool into perf_hpp_fmt struct
perf tools: Remove elide setup for SORT_MODE__MEMORY mode
perf tools: Fix "==" into "=" in ui_browser__warning assignment
perf tools: Allow overriding sysfs and proc finding with env var
perf tools: Consider header files outside perf directory in tags target
...
05 Jun, 2014
1 commit
-
Usually, BUG_ON and friends aren't even evaluated in sparse, but recently
compiletime_assert_atomic_type() was added, and that now results in a
sparse warning every time it is used.The reason turns out to be the temporary variable, after it sparse no
longer considers the value to be a constant, and results in a warning and
an error. The error is the more annoying part of this as it suppresses
any further warnings in the same file, hiding other problems.Unfortunately the condition cannot be simply expanded out to avoid the
temporary variable since it breaks compiletime_assert on old versions of
GCC such as GCC 4.2.4 which the latest metag compiler is based on.Therefore #ifndef __CHECKER__ out the __compiletime_error_fallback which
uses the potentially negative size array to trigger a conditional compiler
error, so that sparse doesn't see it.Signed-off-by: James Hogan
Cc: Johannes Berg
Cc: Daniel Santos
Cc: Luciano Coelho
Cc: Peter Zijlstra
Cc: Paul E. McKenney
Acked-by: Johannes Berg
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
24 Apr, 2014
1 commit
-
Introduce NOKPROBE_SYMBOL() macro which builds a kprobes
blacklist at kernel build time.The usage of this macro is similar to EXPORT_SYMBOL(),
placed after the function definition:NOKPROBE_SYMBOL(function);
Since this macro will inhibit inlining of static/inline
functions, this patch also introduces a nokprobe_inline macro
for static/inline functions. In this case, we must use
NOKPROBE_SYMBOL() for the inline function caller.When CONFIG_KPROBES=y, the macro stores the given function
address in the "_kprobe_blacklist" section.Since the data structures are not fully initialized by the
macro (because there is no "size" information), those
are re-initialized at boot time by using kallsyms.Signed-off-by: Masami Hiramatsu
Link: http://lkml.kernel.org/r/20140417081705.26341.96719.stgit@ltc230.yrl.intra.hitachi.co.jp
Cc: Alok Kataria
Cc: Ananth N Mavinakayanahalli
Cc: Andrew Morton
Cc: Anil S Keshavamurthy
Cc: Arnd Bergmann
Cc: Christopher Li
Cc: Chris Wright
Cc: David S. Miller
Cc: Jan-Simon Möller
Cc: Jeremy Fitzhardinge
Cc: Linus Torvalds
Cc: Randy Dunlap
Cc: Rusty Russell
Cc: linux-arch@vger.kernel.org
Cc: linux-doc@vger.kernel.org
Cc: linux-sparse@vger.kernel.org
Cc: virtualization@lists.linux-foundation.org
Signed-off-by: Ingo Molnar
10 Apr, 2014
1 commit
-
Add a compiler-clang.h file to add specific macros needed for compiling the
kernel with clang.Initially the only override required is the macro for silencing the
compiler for a purposefully uninintialized variable.Author: Mark Charlebois
Signed-off-by: Mark Charlebois
Signed-off-by: Behan Webster
24 Jan, 2014
1 commit
-
Pull crypto update from Herbert Xu:
"Here is the crypto update for 3.14:- Improved crypto_memneq helper
- Use cyprto_memneq in arch-specific crypto code
- Replaced orphaned DCP driver with Freescale MXS DCP driver
- Added AVX/AVX2 version of AESNI-GCM encode and decode
- Added AMD Cryptographic Coprocessor (CCP) driver
- Misc fixes"* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (41 commits)
crypto: aesni - fix build on x86 (32bit)
crypto: mxs - Fix sparse non static symbol warning
crypto: ccp - CCP device enabled/disabled changes
crypto: ccp - Cleanup hash invocation calls
crypto: ccp - Change data length declarations to u64
crypto: ccp - Check for caller result area before using it
crypto: ccp - Cleanup scatterlist usage
crypto: ccp - Apply appropriate gfp_t type to memory allocations
crypto: drivers - Sort drivers/crypto/Makefile
ARM: mxs: dts: Enable DCP for MXS
crypto: mxs - Add Freescale MXS DCP driver
crypto: mxs - Remove the old DCP driver
crypto: ahash - Fully restore ahash request before completing
crypto: aesni - fix build on x86 (32bit)
crypto: talitos - Remove redundant dev_set_drvdata
crypto: ccp - Remove redundant dev_set_drvdata
crypto: crypto4xx - Remove redundant dev_set_drvdata
crypto: caam - simplify and harden key parsing
crypto: omap-sham - Fix Polling mode for larger blocks
crypto: tcrypt - Added speed tests for AEAD crypto alogrithms in tcrypt test suite
...
12 Jan, 2014
1 commit
-
A number of situations currently require the heavyweight smp_mb(),
even though there is no need to order prior stores against later
loads. Many architectures have much cheaper ways to handle these
situations, but the Linux kernel currently has no portable way
to make use of them.This commit therefore supplies smp_load_acquire() and
smp_store_release() to remedy this situation. The new
smp_load_acquire() primitive orders the specified load against
any subsequent reads or writes, while the new smp_store_release()
primitive orders the specifed store against any prior reads or
writes. These primitives allow array-based circular FIFOs to be
implemented without an smp_mb(), and also allow a theoretical
hole in rcu_assign_pointer() to be closed at no additional
expense on most architectures.In addition, the RCU experience transitioning from explicit
smp_read_barrier_depends() and smp_wmb() to rcu_dereference()
and rcu_assign_pointer(), respectively resulted in substantial
improvements in readability. It therefore seems likely that
replacing other explicit barriers with smp_load_acquire() and
smp_store_release() will provide similar benefits. It appears
that roughly half of the explicit barriers in core kernel code
might be so replaced.[Changelog by PaulMck]
Reviewed-by: "Paul E. McKenney"
Signed-off-by: Peter Zijlstra
Acked-by: Will Deacon
Cc: Benjamin Herrenschmidt
Cc: Frederic Weisbecker
Cc: Mathieu Desnoyers
Cc: Michael Ellerman
Cc: Michael Neuling
Cc: Russell King
Cc: Geert Uytterhoeven
Cc: Heiko Carstens
Cc: Linus Torvalds
Cc: Martin Schwidefsky
Cc: Victor Kaplansky
Cc: Tony Luck
Cc: Oleg Nesterov
Link: http://lkml.kernel.org/r/20131213150640.908486364@infradead.org
Signed-off-by: Ingo Molnar
05 Dec, 2013
1 commit
-
Disabling compiler optimizations can be fragile, since a new
optimization could be added to -O0 or -Os that breaks the assumptions
the code is making.Instead of disabling compiler optimizations, use a dummy inline assembly
(based on RELOC_HIDE) to block the problematic kinds of optimization,
while still allowing other optimizations to be applied to the code.The dummy inline assembly is added after every OR, and has the
accumulator variable as its input and output. The compiler is forced to
assume that the dummy inline assembly could both depend on the
accumulator variable and change the accumulator variable, so it is
forced to compute the value correctly before the inline assembly, and
cannot assume anything about its value after the inline assembly.This change should be enough to make crypto_memneq work correctly (with
data-independent timing) even if it is inlined at its call sites. That
can be done later in a followup patch.Compile-tested on x86_64.
Signed-off-by: Cesar Eduardo Barros
Acked-by: Daniel Borkmann
Signed-off-by: Herbert Xu
08 Apr, 2013
1 commit
-
Currently, __kprobes is defined in linux/kprobes.h which
is too big to be included in small or basic headers
that want to make use of this simple attribute.So move __kprobes definition into linux/compiler.h
in which other compiler attributes are defined.Signed-off-by: Masami Hiramatsu
Cc: Timo Juhani Lindfors
Cc: Ananth N Mavinakayanahalli
Cc: Pavel Emelyanov
Cc: Jiri Kosina
Cc: Nadia Yvette Chambers
Cc: yrl.pp-manager.tt@hitachi.com
Cc: David S. Miller
Cc: Linus Torvalds
Link: http://lkml.kernel.org/r/20130404104049.21071.20908.stgit@mhiramat-M0-7522
[ Improved the attribute explanation a bit. ]
Signed-off-by: Ingo Molnar
22 Feb, 2013
3 commits
-
Introduce compiletime_assert to compiler.h, which moves the details of
how to break a build and emit an error message for a specific compiler
to the headers where these details should be. Following in the
tradition of the POSIX assert macro, compiletime_assert creates a
build-time error when the supplied condition is *false*.Next, we add BUILD_BUG_ON_MSG to bug.h which simply wraps
compiletime_assert, inverting the logic, so that it fails when the
condition is *true*, consistent with the language "build bug on." This
macro allows you to specify the error message you want emitted when the
supplied condition is true.Finally, we remove all other code from bug.h that mucks with these
details (BUILD_BUG & BUILD_BUG_ON), and have them all call
BUILD_BUG_ON_MSG. This not only reduces source code bloat, but also
prevents the possibility of code being changed for one macro and not for
the other (which was previously the case for BUILD_BUG and
BUILD_BUG_ON).Since __compiletime_error_fallback is now only used in compiler.h, I'm
considering it a private macro and removing the double negation that's
now extraneous.[akpm@linux-foundation.org: checkpatch fixes]
Signed-off-by: Daniel Santos
Cc: Andi Kleen
Cc: Borislav Petkov
Cc: David Rientjes
Cc: Joe Perches
Cc: Josh Triplett
Cc: Paul Gortmaker
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Prior to the introduction of __attribute__((error("msg"))) in gcc 4.3,
creating compile-time errors required a little trickery.
BUILD_BUG{,_ON} uses this attribute when available to generate
compile-time errors, but also uses the negative-sized array trick for
older compilers, resulting in two error messages in some cases. The
reason it's "some" cases is that as of gcc 4.4, the negative-sized array
will not create an error in some situations, like inline functions.This patch replaces the negative-sized array code with the new
__compiletime_error_fallback() macro which expands to the same thing
unless the the error attribute is available, in which case it expands to
do{}while(0), resulting in exactly one compile-time error on all
versions of gcc.Note that we are not changing the negative-sized array code for the
unoptimized version of BUILD_BUG_ON, since it has the potential to catch
problems that would be disabled in later versions of gcc were
__compiletime_error_fallback used. The reason is that that an
unoptimized build can't always remove calls to an error-attributed
function call (like we are using) that should effectively become dead
code if it were optimized. However, using a negative-sized array with a
similar value will not result in an false-positive (error). The only
caveat being that it will also fail to catch valid conditions, which we
should be expecting in an unoptimized build anyway.Signed-off-by: Daniel Santos
Cc: Andi Kleen
Cc: Borislav Petkov
Cc: David Rientjes
Cc: Joe Perches
Cc: Josh Triplett
Cc: Paul Gortmaker
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
__linktime_error() does the same thing as __compiletime_error() and is
only used in bug.h. Since the macro defines a function attribute that
will cause a failure at compile-time (not link-time), it makes more sense
to keep __compiletime_error(), which is also neatly mated with
__compiletime_warning().Signed-off-by: Daniel Santos
Acked-by: David Rientjes
Acked-by: Borislav Petkov
Cc: Andi Kleen
Cc: Joe Perches
Cc: Josh Triplett
Cc: Paul Gortmaker
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
19 Dec, 2012
1 commit
-
Pull module update from Rusty Russell:
"Nothing all that exciting; a new module-from-fd syscall for those who
want to verify the source of the module (ChromeOS) and/or use standard
IMA on it or other security hooks."* tag 'modules-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux:
MODSIGN: Fix kbuild output when using default extra_certificates
MODSIGN: Avoid using .incbin in C source
modules: don't hand 0 to vmalloc.
module: Remove a extra null character at the top of module->strtab.
ASN.1: Use the ASN1_LONG_TAG and ASN1_INDEFINITE_LENGTH constants
ASN.1: Define indefinite length marker constant
moduleparam: use __UNIQUE_ID()
__UNIQUE_ID()
MODSIGN: Add modules_sign make target
powerpc: add finit_module syscall.
ima: support new kernel module syscall
add finit_module syscall to asm-generic
ARM: add finit_module syscall to ARM
security: introduce kernel_module_from_file hook
module: add flags arg to sys_finit_module()
module: add syscall to load module from fd
18 Dec, 2012
1 commit
-
linux/compiler.h has macros to denote functions that acquire or release
locks, but not to denote functions called with a lock held that return
with the lock still held. Add a __must_hold macro to cover that case.Signed-off-by: Josh Triplett
Reported-by: Ed Cashin
Tested-by: Ed Cashin
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
14 Dec, 2012
1 commit
-
Jan Beulich points out __COUNTER__ (gcc 4.3 and above), so let's use
that to create unique ids. This is better than __LINE__ which we use
today, so provide a wrapper.Stanislaw Gruszka reported that some module parameters
start with a digit, so we need to prepend when we for the unique id.Signed-off-by: Rusty Russell
Acked-by: Jan Beulich
18 Sep, 2012
1 commit
-
gcc 4.6+ has support for a externally_visible attribute that prevents the
optimizer from optimizing unused symbols away. Add a __visible macro to
use it with that compiler version or later.This is used (at least) by the "Link Time Optimization" patchset.
Signed-off-by: Andi Kleen
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
28 Feb, 2012
1 commit
-
Signed-off-by: Alexander Stein
Signed-off-by: Jiri Kosina
11 Jan, 2012
1 commit
-
We can place this in definitions that we expect the compiler to remove by
dead code elimination. If this assertion fails, we get a nice error
message at build time.The GCC function attribute error("message") was added in version 4.3, so
we define a new macro __linktime_error(message) to expand to this for
GCC-4.3 and later. This will give us an error diagnostic from the
compiler on the line that fails. For other compilers
__linktime_error(message) expands to nothing, and we have to be content
with a link time error, but at least we will still get a build error.BUILD_BUG() expands to the undefined function __build_bug_failed() and
will fail at link time if the compiler ever emits code for it. On GCC-4.3
and later, attribute((error())) is used so that the failure will be noted
at compile time instead.Signed-off-by: David Daney
Acked-by: David Rientjes
Cc: DM
Cc: Ralf Baechle
Acked-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds
20 Aug, 2010
1 commit
-
This commit provides definitions for the __rcu annotation defined earlier.
This annotation permits sparse to check for correct use of RCU-protected
pointers. If a pointer that is annotated with __rcu is accessed
directly (as opposed to via rcu_dereference(), rcu_assign_pointer(),
or one of their variants), sparse can be made to complain. To enable
such complaints, use the new default-disabled CONFIG_SPARSE_RCU_POINTER
kernel configuration option. Please note that these sparse complaints are
intended to be a debugging aid, -not- a code-style-enforcement mechanism.There are special rcu_dereference_protected() and rcu_access_pointer()
accessors for use when RCU read-side protection is not required, for
example, when no other CPU has access to the data structure in question
or while the current CPU hold the update-side lock.This patch also updates a number of docbook comments that were showing
their age.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul E. McKenney
Cc: Christopher Li
Reviewed-by: Josh Triplett
15 Jun, 2010
1 commit
-
This commit defines an __rcu API, but provides only vacuous definitions
for it. This breaks dependencies among most of the subsequent patches,
allowing them to reach mainline asynchronously via whatever trees are
appropriate.Signed-off-by: Arnd Bergmann
Signed-off-by: Paul E. McKenney
Cc: Christopher Li
Cc: Josh Triplett
05 Jan, 2010
1 commit
-
Conflicts:
arch/powerpc/platforms/pseries/hvCall.S
include/linux/percpu.h
06 Dec, 2009
3 commits
-
* 'x86-asm-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
include/linux/compiler-gcc4.h: Fix build bug - gcc-4.0.2 doesn't understand __builtin_object_size
x86/alternatives: No need for alternatives-asm.h to re-invent stuff already in asm.h
x86/alternatives: Check replacementlen t use the strict copy checks when branch profiling is in use
x86, 64-bit: Move K8 B step iret fixup to fault entry asm
x86: Generate cmpxchg build failures
x86: Add a Kconfig option to turn the copy_from_user warnings into errors
x86: Turn the copy_from_user check into an (optional) compile time warning
x86: Use __builtin_memset and __builtin_memcpy for memset/memcpy
x86: Use __builtin_object_size() to validate the buffer size for copy_from_user() -
…el/git/tip/linux-2.6-tip
* 'tracing-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (40 commits)
tracing: Separate raw syscall from syscall tracer
ring-buffer-benchmark: Add parameters to set produce/consumer priorities
tracing, function tracer: Clean up strstrip() usage
ring-buffer benchmark: Run producer/consumer threads at nice +19
tracing: Remove the stale include/trace/power.h
tracing: Only print objcopy version warning once from recordmcount
tracing: Prevent build warning: 'ftrace_graph_buf' defined but not used
ring-buffer: Move access to commit_page up into function used
tracing: do not disable interrupts for trace_clock_local
ring-buffer: Add multiple iterations between benchmark timestamps
kprobes: Sanitize struct kretprobe_instance allocations
tracing: Fix to use __always_unused attribute
compiler: Introduce __always_unused
tracing: Exit with error if a weak function is used in recordmcount.pl
tracing: Move conditional into update_funcs() in recordmcount.pl
tracing: Add regex for weak functions in recordmcount.pl
tracing: Move mcount section search to front of loop in recordmcount.pl
tracing: Fix objcopy revision check in recordmcount.pl
tracing: Check absolute path of input file in recordmcount.pl
tracing: Correct the check for number of arguments in recordmcount.pl
... -
Starting with version 4.5, GCC has a new built-in function
__builtin_unreachable() that can be used in places like the kernel's
BUG() where inline assembly is used to transfer control flow. This
eliminated the need for an endless loop in these places.The patch adds a new macro 'unreachable()' that will expand to either
__builtin_unreachable() or an endless loop depending on the compiler
version.Change from v1: Simplify unreachable() for non-GCC 4.5 case.
Signed-off-by: David Daney
Acked-by: Ralf Baechle
Signed-off-by: Linus Torvalds
02 Nov, 2009
1 commit
-
I wrote some code which is used as compile-time checker, and the
code should be elided after compile.So I need to annotate the code as "always unused", compared to
"maybe unused".Signed-off-by: Li Zefan
Cc: Steven Rostedt
Cc: Frederic Weisbecker
Cc: Linus Torvalds
LKML-Reference:
Signed-off-by: Ingo Molnar
29 Oct, 2009
1 commit
-
We have to make __kernel "__attribute__((address_space(0)))" so we can
cast to it.tj: * put_cpu_var() update.
* Annotations added to dynamic allocator interface.
Signed-off-by: Rusty Russell
Cc: Al Viro
Signed-off-by: Tejun Heo
03 Oct, 2009
1 commit
-
For automated testing it is useful to have the option to turn
the warnings on copy_from_user() etc checks into errors:In function ‘copy_from_user’,
inlined from ‘fd_copyin’ at drivers/block/floppy.c:3080,
inlined from ‘fd_ioctl’ at drivers/block/floppy.c:3503:
linux/arch/x86/include/asm/uaccess_32.h:213:
error: call to ‘copy_from_user_overflow’ declared with attribute error:
copy_from_user buffer size is not provably correctSigned-off-by: Arjan van de Ven
Cc: Linus Torvalds
Cc: Andrew Morton
LKML-Reference:
Signed-off-by: Ingo Molnar
01 Oct, 2009
1 commit
-
A previous patch added the buffer size check to copy_from_user().
One of the things learned from analyzing the result of the previous
patch is that in general, gcc is really good at proving that the
code contains sufficient security checks to not need to do a
runtime check. But that for those cases where gcc could not prove
this, there was a relatively high percentage of real security
issues.This patch turns the case of "gcc cannot prove" into a compile time
warning, as long as a sufficiently new gcc is in use that supports
this. The objective is that these warnings will trigger developers
checking new cases out before a security hole enters a linux kernel
release.Signed-off-by: Arjan van de Ven
Cc: Linus Torvalds
Cc: "David S. Miller"
Cc: James Morris
Cc: Jan Beulich
LKML-Reference:
Signed-off-by: Ingo Molnar
26 Sep, 2009
1 commit
-
gcc (4.x) supports the __builtin_object_size() builtin, which
reports the size of an object that a pointer point to, when known
at compile time. If the buffer size is not known at compile time, a
constant -1 is returned.This patch uses this feature to add a sanity check to
copy_from_user(); if the target buffer is known to be smaller than
the copy size, the copy is aborted and a WARNing is emitted in
memory debug mode.These extra checks compile away when the object size is not known,
or if both the buffer size and the copy length are constants.Signed-off-by: Arjan van de Ven
LKML-Reference:
Signed-off-by: Ingo Molnar
12 Jun, 2009
1 commit
-
Impact: new API
__builtin_types_compatible_p() is a little awkward to use: it takes two
types rather than types or variables, and it's just damn long.(typeof(type) == type, so this works on types as well as vars).
Signed-off-by: Rusty Russell
07 Apr, 2009
2 commits
-
Fix the branch tracer barfing on comma statements within if ()
statements.Signed-off-by: Ingo Molnar
-
One of the changes between kernels 2.6.28 and 2.6.29 is that a branch profiler
has been added for if() statements. Unfortunately this patch makes the sparse
output unusable with CONFIG_TRACE_BRANCH_PROFILING=y: when branch profiling is
enabled, sparse prints so much false positives that the real issues are no
longer visible. This behavior can be reproduced as follows:
* enable CONFIG_TRACE_BRANCH_PROFILING, e.g. by running make allyesconfig or
make allmodconfig.
* run make C=2Result: a huge number of the following sparse warnings.
...
include/linux/cpumask.h:547:2: warning: symbol '______r' shadows an earlier one
include/linux/cpumask.h:547:2: originally declared here
...The patch below fixes this by disabling branch profiling while analyzing the
kernel code with sparse.See also:
* http://lkml.org/lkml/2008/11/21/18
* http://bugzilla.kernel.org/show_bug.cgi?id=12925Signed-off-by: Bart Van Assche
Cc: Andrew Morton
Cc: Steven Rostedt
LKML-Reference:
Signed-off-by: Ingo Molnar
18 Mar, 2009
1 commit
-
Impact: better performance for if branch tracer
Use an array to count the hit and misses of a conditional instead
of using another conditional. This cuts down on saturation of branch
predictions and increases performance of modern pipelined architectures.Signed-off-by: Witold Baryluk
Signed-off-by: Steven Rostedt
03 Jan, 2009
1 commit
-
- include the gcc version-dependent header files from the generic gcc
header file, rather than the other way around (iow: don't make the
non-gcc header file have to know about gcc versions)- don't include compiler-gcc4.h for gcc 5 (for whenever it gets
released). That's just confusing and made us do odd things in the
gcc4 header file (testing that we really had version 4!)- generate the name from the __GNUC__ version directly, rather than
having a mess of #if conditionals.Signed-off-by: Linus Torvalds
23 Nov, 2008
3 commits
-
Impact: feature to profile if statements
This patch adds a branch profiler for all if () statements.
The results will be found in:/debugfs/tracing/profile_branch
For example:
miss hit % Function File Line
------- --------- - -------- ---- ----
0 1 100 x86_64_start_reservations head64.c 127
0 1 100 copy_bootdata head64.c 69
1 0 0 x86_64_start_kernel head64.c 111
32 0 0 set_intr_gate desc.h 319
1 0 0 reserve_ebda_region head.c 51
1 0 0 reserve_ebda_region head.c 47
0 1 100 reserve_ebda_region head.c 42
0 0 X maxcpus main.c 165Miss means the branch was not taken. Hit means the branch was taken.
The percent is the percentage the branch was taken.This adds a significant amount of overhead and should only be used
by those analyzing their system.Signed-off-by: Steven Rostedt
Signed-off-by: Ingo Molnar -
Impact: clean up to make one profiler of like and unlikely tracer
The likely and unlikely profiler prints out the file and line numbers
of the annotated branches that it is profiling. It shows the number
of times it was correct or incorrect in its guess. Having two
different files or sections for that matter to tell us if it was a
likely or unlikely is pretty pointless. We really only care if
it was correct or not.Signed-off-by: Steven Rostedt
Signed-off-by: Ingo Molnar -
Impact: clean up of branch check
The unlikely/likely profiler does an extra assign of the f.line.
This is not needed since it is already calculated at compile time.Signed-off-by: Steven Rostedt
Signed-off-by: Ingo Molnar
13 Nov, 2008
1 commit
-
Impact: name change of unlikely tracer and profiler
Ingo Molnar suggested changing the config from UNLIKELY_PROFILE
to BRANCH_PROFILING. I never did like the "unlikely" name so I
went one step farther, and renamed all the unlikely configurations
to a "BRANCH" variant.Signed-off-by: Steven Rostedt
Signed-off-by: Ingo Molnar
12 Nov, 2008
2 commits
-
Impact: fix bootup crash
the branch tracer missed arch/x86/vdso/vclock_gettime.c from
disabling tracing, which caused such bootup crashes:[ 201.840097] init[1]: segfault at 7fffed3fe7c0 ip 00007fffed3fea2e sp 000077
also clean up the ugly ifdefs in arch/x86/kernel/vsyscall_64.c by
creating DISABLE_UNLIKELY_PROFILE facility for code to turn off
instrumentation on a per file basis.Signed-off-by: Ingo Molnar
-
Impact: new unlikely/likely profiler
Andrew Morton recently suggested having an in-kernel way to profile
likely and unlikely macros. This patch achieves that goal.When configured, every(*) likely and unlikely macro gets a counter attached
to it. When the condition is hit, the hit and misses of that condition
are recorded. These numbers can later be retrieved by:/debugfs/tracing/profile_likely - All likely markers
/debugfs/tracing/profile_unlikely - All unlikely markers.# cat /debug/tracing/profile_unlikely | head
correct incorrect % Function File Line
------- --------- - -------- ---- ----
2167 0 0 do_arch_prctl process_64.c 832
0 0 0 do_arch_prctl process_64.c 804
2670 0 0 IS_ERR err.h 34
71230 5693 7 __switch_to process_64.c 673
76919 0 0 __switch_to process_64.c 639
43184 33743 43 __switch_to process_64.c 624
12740 64181 83 __switch_to process_64.c 594
12740 64174 83 __switch_to process_64.c 590# cat /debug/tracing/profile_unlikely | \
awk '{ if ($3 > 25) print $0; }' |head -20
44963 35259 43 __switch_to process_64.c 624
12762 67454 84 __switch_to process_64.c 594
12762 67447 84 __switch_to process_64.c 590
1478 595 28 syscall_get_error syscall.h 51
0 2821 100 syscall_trace_leave ptrace.c 1567
0 1 100 native_smp_prepare_cpus smpboot.c 1237
86338 265881 75 calc_delta_fair sched_fair.c 408
210410 108540 34 calc_delta_mine sched.c 1267
0 54550 100 sched_info_queued sched_stats.h 222
51899 66435 56 pick_next_task_fair sched_fair.c 1422
6 10 62 yield_task_fair sched_fair.c 982
7325 2692 26 rt_policy sched.c 144
0 1270 100 pre_schedule_rt sched_rt.c 1261
1268 48073 97 pick_next_task_rt sched_rt.c 884
0 45181 100 sched_info_dequeued sched_stats.h 177
0 15 100 sched_move_task sched.c 8700
0 15 100 sched_move_task sched.c 8690
53167 33217 38 schedule sched.c 4457
0 80208 100 sched_info_switch sched_stats.h 270
30585 49631 61 context_switch sched.c 2619# cat /debug/tracing/profile_likely | awk '{ if ($3 > 25) print $0; }'
39900 36577 47 pick_next_task sched.c 4397
20824 15233 42 switch_mm mmu_context_64.h 18
0 7 100 __cancel_work_timer workqueue.c 560
617 66484 99 clocksource_adjust timekeeping.c 456
0 346340 100 audit_syscall_exit auditsc.c 1570
38 347350 99 audit_get_context auditsc.c 732
0 345244 100 audit_syscall_entry auditsc.c 1541
38 1017 96 audit_free auditsc.c 1446
0 1090 100 audit_alloc auditsc.c 862
2618 1090 29 audit_alloc auditsc.c 858
0 6 100 move_masked_irq migration.c 9
1 198 99 probe_sched_wakeup trace_sched_switch.c 58
2 2 50 probe_wakeup trace_sched_wakeup.c 227
0 2 100 probe_wakeup_sched_switch trace_sched_wakeup.c 144
4514 2090 31 __grab_cache_page filemap.c 2149
12882 228786 94 mapping_unevictable pagemap.h 50
4 11 73 __flush_cpu_slab slub.c 1466
627757 330451 34 slab_free slub.c 1731
2959 61245 95 dentry_lru_del_init dcache.c 153
946 1217 56 load_elf_binary binfmt_elf.c 904
102 82 44 disk_put_part genhd.h 206
1 1 50 dst_gc_task dst.c 82
0 19 100 tcp_mss_split_point tcp_output.c 1126As you can see by the above, there's a bit of work to do in rethinking
the use of some unlikelys and likelys. Note: the unlikely case had 71 hits
that were more than 25%.Note: After submitting my first version of this patch, Andrew Morton
showed me a version written by Daniel Walker, where I picked up
the following ideas from:1) Using __builtin_constant_p to avoid profiling fixed values.
2) Using __FILE__ instead of instruction pointers.
3) Using the preprocessor to stop all profiling of likely
annotations from vsyscall_64.c.Thanks to Andrew Morton, Arjan van de Ven, Theodore Tso and Ingo Molnar
for their feed back on this patch.(*) Not ever unlikely is recorded, those that are used by vsyscalls
(a few of them) had to have profiling disabled.Signed-off-by: Steven Rostedt
Cc: Andrew Morton
Cc: Frederic Weisbecker
Cc: Theodore Tso
Cc: Arjan van de Ven
Cc: Steven Rostedt
Signed-off-by: Ingo Molnar
14 Oct, 2008
1 commit
-
The notrace define belongs in compiler.h so that it can be used in
init.hSigned-off-by: Steven Rostedt
Signed-off-by: Ingo Molnar