Commit Graph

542 Commits

Author SHA1 Message Date
Ellis Hoag 11d3074267 [InstrProf] Add single byte coverage mode
Use the llvm flag `-pgo-function-entry-coverage` to create single byte "counters" to track functions coverage. This mode has significantly less size overhead in both code and data because
  * We mark a function as "covered" with a store instead of an increment which generally requires fewer assembly instructions
  * We use a single byte per function rather than 8 bytes per block

The trade off of course is that this mode only tells you if a function has been covered. This is useful, for example, to detect dead code.

When combined with debug info correlation [0] we are able to create an instrumented Clang binary that is only 150M (the vanilla Clang binary is 143M). That is an overhead of 7M (4.9%) compared to the default instrumentation (without value profiling) which has an overhead of 31M (21.7%).

[0] https://groups.google.com/g/llvm-dev/c/r03Z6JoN7d4

Reviewed By: kyulee

Differential Revision: https://reviews.llvm.org/D116180
2022-01-27 17:38:55 -08:00
Ellis Hoag 88d81770f1 [InstrProf] Restore InstrProfData.inc to fix Fuchsia builds
https://reviews.llvm.org/D116179 introduced some changes to
`InstrProfData.inc` which broke some downstream builds. This commit
reverts those changes since they only changes two field names.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D117631
2022-01-19 10:10:58 -08:00
Ellis Hoag f21473752b [InstrProf][NFC] Do not assume size of counter type
Existing code tended to assume that counters had type `uint64_t` and
computed size from the number of counters. Fix this code to directly
compute the counters size in number of bytes where possible. When the
number of counters is needed, use `__llvm_profile_counter_entry_size()`
or `getCounterTypeSize()`. In a later diff these functions will depend
on the profile mode.

Change the meaning of `DataSize` and `CountersSize` to make them more clear.
* `DataSize` (`CountersSize`) - the size of the data (counter) section in bytes.
* `NumData` (`NumCounters`) - the number of data (counter) entries.

Reviewed By: kyulee

Differential Revision: https://reviews.llvm.org/D116179
2022-01-14 11:29:11 -08:00
Alexandre Ganea ae4c643bcd [compiler-rt] Silence warnings when building with MSVC
Differential Revision: https://reviews.llvm.org/D116872
2022-01-11 10:36:57 -05:00
Gulfem Savrun Yeniceri f261e258ec [profile] Fix writing binary id into profiles
This patch adds support to read all the PT_NOTE segments in the
executable to find the binary ids. Previously, it was only reading
the first PT_NOTE segment, and this was missing the cases where
binary id is in the following segments. As a result, binary-id.c
and binary-id-padding.c test were failing in the following cases:
1) sanitizer-x86_64-linux bot
   https://lab.llvm.org/staging/#/builders/97
2) OpenSuse Tumbleweed
   https://github.com/llvm/llvm-project/issues/52695

Differential Revision: https://reviews.llvm.org/D115830
2021-12-27 19:20:19 +00:00
Ellis Hoag ac719d7c9a [InstrProf] Don't profile merge by default in lightweight mode
Profile merging is not supported when using debug info profile
correlation because the data section won't be in the binary at runtime.
Change the default profile name in this mode to `default_%p.proflite` so
we don't use profile merging.

Reviewed By: kyulee

Differential Revision: https://reviews.llvm.org/D115979
2021-12-20 09:51:49 -08:00
Ellis Hoag 58d9c1aec8 [Try2][InstrProf] Attach debug info to counters
Add the llvm flag `-debug-info-correlate` to attach debug info to instrumentation counters so we can correlate raw profile data to their functions. Raw profiles are dumped as `.proflite` files. The next diff enables `llvm-profdata` to consume `.proflite` and debug info files to produce a normal `.profdata` profile.

Part of the "lightweight instrumentation" work: https://groups.google.com/g/llvm-dev/c/r03Z6JoN7d4

The original diff https://reviews.llvm.org/D114565 was reverted because of the `Instrumentation/InstrProfiling/debug-info-correlate.ll` test, which is fixed in this commit.

Reviewed By: kyulee

Differential Revision: https://reviews.llvm.org/D115693
2021-12-16 14:20:30 -08:00
Ellis Hoag c809da7d9c Revert "[InstrProf] Attach debug info to counters"
This reverts commit 800bf8ed29.

The `Instrumentation/InstrProfiling/debug-info-correlate.ll` test was
failing because I forgot the `llc` commands are architecture specific.
I'll follow up with a fix.

Differential Revision: https://reviews.llvm.org/D115689
2021-12-13 18:15:17 -08:00
Ellis Hoag 800bf8ed29 [InstrProf] Attach debug info to counters
Add the llvm flag `-debug-info-correlate` to attach debug info to instrumentation counters so we can correlate raw profile data to their functions. Raw profiles are dumped as `.proflite` files. The next diff enables `llvm-profdata` to consume `.proflite` and debug info files to produce a normal `.profdata` profile.

Part of the "lightweight instrumentation" work: https://groups.google.com/g/llvm-dev/c/r03Z6JoN7d4

Reviewed By: kyulee

Differential Revision: https://reviews.llvm.org/D114565
2021-12-13 17:51:22 -08:00
Pirama Arumuga Nainar 1df7289af6 [compiler-rt/profile] Include __llvm_profile_get_magic in module signature
The INSTR_PROF_RAW_MAGIC_* number in profraw files should match during
profile merging.  This causes an error with 32-bit and 64-bit variants
of the same code.  The module signatures for the two binaries are
identical but they use different INSTR_PROF_RAW_MAGIC_* causing a
failure when profile-merging is used.  Including it when computing the
module signature yields different signatures for the 32-bit and 64-bit
profiles.

Differential Revision: https://reviews.llvm.org/D114054
2021-11-23 10:03:43 -08:00
Pirama Arumuga Nainar 078279ff01 [compiler-rt/profile] Reland mark __llvm_profile_raw_version as hidden
Since libclang_rt.profile is added later in the command line, a
definition of __llvm_profile_raw_version is not included if it is
provided from an earlier object, e.g.  from a shared dependency.

This causes an extra dependence edge where if libA.so depends on libB.so
and both are coverage-instrumented, libA.so uses libB.so's definition of
__llvm_profile_raw_version.  This leads to a runtime link failure if the
libB.so available at runtime does not provide this symbol (but provides
the other dependent symbols).  Such a scenario can occur in Android's
mainline modules.
E.g.:
  ld -o libB.so libclang_rt.profile-x86_64.a
  ld -o libA.so -l B libclang_rt.profile-x86_64.a

libB.so has a global definition of __llvm_profile_raw_version.  libA.so
uses libB.so's definition of __llvm_profile_raw_version.  At runtime,
libB.so may not be coverage-instrumented (i.e. not export
__llvm_profile_raw_version) so runtime linking of libA.so will fail.

Marking this symbol as hidden forces each binary to use the definition
of __llvm_profile_raw_version from libclang_rt.profile.  The visiblity
is unchanged for Apple platforms where its presence is checked by the
TAPI tool.

Reviewed By: MaskRay, phosek, davidxl

Differential Revision: https://reviews.llvm.org/D111759
2021-11-16 11:37:00 -08:00
Simon Moll 496e7f330c [VE] Disable incompatible compiler-rt tests
Some compiler-rt tests are inherently incompatible with VE because..

* No consistent denormal support on VE. We skip denormal fp inputs in builtin tests.
* `madvise` unsupported on VE.
* Instruction alignment requirements.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D113093
2021-11-12 08:40:03 +01:00
Pirama Arumuga Nainar ab3d5d0533 Revert "[compiler-rt/profile] Hide __llvm_profile_raw_version"
This reverts commit 69708477be to unblock
instrprof-darwin-exports.c failure on MacOS bots.
2021-10-21 10:56:43 -07:00
Pirama Arumuga Nainar 69708477be [compiler-rt/profile] Hide __llvm_profile_raw_version
Hide __llvm_profile_raw_version so as not to resolve reference from a
dependent shared object.  Since libclang_rt.profile is added later in
the command line, a definition of __llvm_profile_raw_version is not
included if it is provided from an earlier object, e.g.  from a shared
dependency.

This causes an extra dependence edge where if libA.so depends on libB.so
and both are coverage-instrumented, libA.so uses libB.so's definition of
__llvm_profile_raw_version.  This leads to a runtime link failure if the
libB.so available at runtime does not provide this symbol (but provides
the other dependent symbols).  Such a scenario can occur in Android's
mainline modules.
E.g.:
  ld -o libB.so libclang_rt.profile-x86_64.a
  ld -o libA.so -l B libclang_rt.profile-x86_64.a

libB.so has a global definition of __llvm_profile_raw_version.  libA.so
uses libB.so's definition of __llvm_profile_raw_version.  At runtime,
libB.so may not be coverage-instrumented (i.e. not export
__llvm_profile_raw_version) so runtime linking of libA.so will fail.

Marking this symbol as hidden forces each binary to use the definition
of __llvm_profile_raw_version from libclang_rt.profile.

Differential Revision: https://reviews.llvm.org/D111759
2021-10-15 11:56:16 -07:00
Gulfem Savrun Yeniceri 9a70eb918e [profile] Remove emitting symbolizer markup
Fuchsia Clang code coverage pipeline started to use binary ids that are
embedded in profiles. This patch removes emitting symbolizer markup,
which is not necessary in the coverage pipeline anymore.

Differential Revision: https://reviews.llvm.org/D111674
2021-10-13 01:43:07 +00:00
Zequan Wu 4e8efff53e [Profile] Add missing fflush in __llvm_profile_set_file_object 2021-10-05 15:41:51 -07:00
Zequan Wu ab694cd845 [Profile] Add a warning when lock file failed in __llvm_profile_set_file_object with continuous mode 2021-10-01 14:37:09 -07:00
Leonard Chan b9f547e8e5 [llvm][profile] Add padding after binary IDs
Some tests with binary IDs would fail with error: no profile can be merged.
This is because raw profiles could have unaligned headers when emitting binary
IDs. This means padding should be emitted after binary IDs are emitted to
ensure everything else is aligned. This patch adds padding after each binary ID
to ensure the next binary ID size is 8-byte aligned. This also adds extra
checks to ensure we aren't reading corrupted data when printing binary IDs.

Differential Revision: https://reviews.llvm.org/D110365
2021-09-28 11:50:50 -07:00
Leonard Chan eb115aa6c8 Revert "[compiler-rt][profile] Add padding after binary IDs"
This reverts commit 6bc9c8dfe3.

Reverted because this broke some PPC buildbots.
2021-09-23 15:48:59 -07:00
Roland McGrath 80b92db02c [profile][fuchsia] Don't include extra NUL in log messages
Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D110361
2021-09-23 15:16:31 -07:00
Leonard Chan 6bc9c8dfe3 [compiler-rt][profile] Add padding after binary IDs
Some tests with binary IDs would fail with error: no profile can be merged.
This is because raw profiles could have unaligned headers when emitting binary
IDs. This means padding should be emitted after binary IDs are emitted to
ensure everything else is aligned. This patch accounts for that padding in
__llvm_write_binary_ids.

Differential Revision: https://reviews.llvm.org/D110188
2021-09-23 10:29:24 -07:00
David Carlier 08c3cdb8b8 [Sanitizers][PGO] missing return statement 2021-09-04 11:40:58 +01:00
Kazuaki Ishizaki a1e7e401d2 [compiler-rt] NFC: Fix trivial typo
Reviewed By: xgupta

Differential Revision: https://reviews.llvm.org/D77457
2021-09-04 14:12:58 +05:30
David Carlier b35e636e40 [Sanitizers] PGO implements enable/disable SIGKILL on callers on FreeBSD.
Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D108867
2021-08-28 22:53:16 +01:00
Zequan Wu 1b05245119 [Profile] Support __llvm_profile_set_file_object in continuous mode.
Replace D107203, because __llvm_profile_set_file_object is usually used when the
process doesn't have permission to open/create file. That patch trying to copy
from old profile to new profile contradicts with the usage.

Differential Revision: https://reviews.llvm.org/D108242
2021-08-27 13:06:46 -07:00
Petr Hosek 8111f2f7ee [profile] Update counter offset to account for binary ids
In the raw profile, binary ids immediately follow the header so when
computing counters offset we need to account for the new section.

Differential Revision: https://reviews.llvm.org/D108608
2021-08-23 21:43:30 -07:00
Brian Cain 68ab571e22 [sanitizer] Fix for CMAKE_CXX_FLAGS update
With unquoted ${CMAKE_CXX_FLAGS}, the REGEX fails when it's empty:

```CMake Error at lib/scudo/standalone/CMakeLists.txt:14 (string):
string sub-command REGEX, mode REPLACE needs at least 6 arguments total to
command.```
2021-08-19 12:05:55 -07:00
Jinsong Ji 66e2772e42 [InstrProfiling] Support relative CountersPtr for PlatformOther
D104556 change the CountersPtr to be relative, however, it did not
update the pointer initialization in  __llvm_profile_register_function,
so the platform (eg:AIX) that use __llvm_profile_register_function is now totaly
broken, any PGO code will SEGV.

This patch update the code to reflect that the Data->CountersPtr is now
relative.

Reviewed By: MaskRay, davidxl

Differential Revision: https://reviews.llvm.org/D108304
2021-08-18 17:45:39 +00:00
Gulfem Savrun Yeniceri 6c0e6f91d7 [profile] Add static keyword to binary id functions
This patch adds static keyword to internal functions that write
binary id to restrict visibility to the file that they are declared.

Differential Revision: https://reviews.llvm.org/D108154
2021-08-16 19:36:47 +00:00
Nikita Popov 4cfb047d6a [profile] Don't use pragma comment linker on mingw
At least when compiling with gcc, this is not supported and will
result in errors when linking against the profiler runtime. Only
use the pragma comment linker based code with MSVC, but not with
a mingw toolchain. This also undoes D107620, which shouldn't be
relevant anymore.

Differential Revision: https://reviews.llvm.org/D108095
2021-08-16 18:20:32 +02:00
Martin Storsjö 94c001d646 [profile] Build with -fms-extensions in MinGW mode
54902e00d1 added a use of
/alternatename via a #pragma comment(linker); in MinGW mode, this
requires building with -fms-extensions. (This flag is added to
SANITIZER_COMMON_CFLAGS in the toplevel CMakeLists.txt.)

This avoids a warning when building in MinGW mode (about an unknown
pragma being ignored), and presumably also makes the code work as
intended.

Differential Revision: https://reviews.llvm.org/D107620
2021-08-09 11:39:11 +03:00
Martin Storsjö 743f78ef7f [InstrProfiling] Fix warnings when building for Windows
The Headers.CountersDelta field is an uint64_t, not a pointer,
so just cast to uint32_t to truncate it.

Differential Revision: https://reviews.llvm.org/D107619
2021-08-09 11:39:11 +03:00
Roland McGrath 5a2a179695 [profile][Fuchsia] Add missing system header #include
The _zx_vmar_root_self function is not a system call but
a libc function declared in a separate header.

Reviewed By: gulfem

Differential Revision: https://reviews.llvm.org/D107616
2021-08-06 17:59:35 -07:00
Zequan Wu 2129c4a861 Fix Windows bots failure caused by 8c4208d5c1 2021-08-06 15:03:00 -07:00
Zequan Wu 8c4208d5c1 [Profile][NFC] Clean up initializeProfileForContinuousMode
Merge two versions of `initializeProfileForContinuousMode` function into one.

Differential Revision: https://reviews.llvm.org/D107591
2021-08-06 14:00:36 -07:00
Rainer Orth 779714f89b [profile] Only use NT_GNU_BUILD_ID if supported
The Solaris buildbots have been broken for some time by the unconditional
use of `NT_GNU_BUILD_ID`, e.g. Solaris/sparcv9
<https://lab.llvm.org/staging/#/builders/50/builds/4910> and Solaris/amd64
<https://lab.llvm.org/staging/#/builders/101/builds/3751>.  Being a GNU
extension, it is not defined in `<sys/elf.h>`.  However, providing a
fallback definition doesn't help because the code also relies on
`__ehdr_start`, another unportable GNU extension that most likely never
will be implemented in Solaris `ld`.  Besides, there's reallly no point in
supporting build ids since they aren't used on Solaris at all.

This patch fixes this by making the relevant code conditional on the
definition of `NT_GNU_BUILD_ID`.

Tested on `amd64-pc-solaris2.11` and `sparcv9-sun-solaris2.11`.

Differential Revision: https://reviews.llvm.org/D107556
2021-08-06 14:04:11 +02:00
Dimitry Andric 440d9712eb Work around non-existence of ElfW(type) macro on FreeBSD
Fixes PR51331. On FreeBSD, the elf headers don't (yet) provide the
ElfW(type) macro. However, there is a similar set of macros in the
<sys/elf-generic.h> header, of which `__ElfN(type)` exactly matches the
indended purpose.

Reviewed By: gulfem

Differential Revision: https://reviews.llvm.org/D107388
2021-08-04 20:12:30 +02:00
Vedant Kumar 3b0a9e7b39 [profile] Move assertIsZero to InstrProfilingUtil.c
... and rename it to 'warnIfNonZero' to better-reflect what it actually
does.

The goal is to minimize the amount of logic that's conditionally
compiled under '#if __APPLE__'.
2021-08-02 15:25:09 -07:00
Petr Hosek 83302c8489 [profile] Fix profile merging with binary IDs
This fixes support for merging profiles which broke as a consequence
of e50a38840d. The issue was missing
adjustment in merge logic to account for the binary IDs which are
now included in the raw profile just after header.

In addition, this change also:
* Includes the version in module signature that's used for merging
to avoid accidental attempts to merge incompatible profiles.
* Moves the binary IDs size field after version field in the header
as was suggested in the review.

Differential Revision: https://reviews.llvm.org/D107143
2021-07-30 18:54:27 -07:00
Petr Hosek d3dd07e3d0 Revert "[profile] Fix profile merging with binary IDs"
This reverts commit dcadd64986.
2021-07-30 18:53:48 -07:00
Petr Hosek dcadd64986 [profile] Fix profile merging with binary IDs
This fixes support for merging profiles which broke as a consequence
of e50a38840d. The issue was missing
adjustment in merge logic to account for the binary IDs which are
now included in the raw profile just after header.

In addition, this change also:
* Includes the version in module signature that's used for merging
to avoid accidental attempts to merge incompatible profiles.
* Moves the binary IDs size field after version field in the header
as was suggested in the review.

Differential Revision: https://reviews.llvm.org/D107143
2021-07-30 17:38:53 -07:00
Petr Hosek 6ea2f31f3d Revert "[profile] Fix profile merging with binary IDs"
This reverts commit 89d6eb6f8c, this
seemed to have break a few builders.
2021-07-30 14:32:52 -07:00
Petr Hosek 89d6eb6f8c [profile] Fix profile merging with binary IDs
This fixes support for merging profiles which broke as a consequence
of e50a38840d. The issue was missing
adjustment in merge logic to account for the binary IDs which are
now included in the raw profile just after header.

In addition, this change also:
* Includes the version in module signature that's used for merging
to avoid accidental attempts to merge incompatible profiles.
* Moves the binary IDs size field after version field in the header
as was suggested in the review.

Differential Revision: https://reviews.llvm.org/D107143
2021-07-30 13:30:30 -07:00
Fangrui Song a1532ed275 [InstrProfiling] Make CountersPtr in __profd_ relative
Change `CountersPtr` in `__profd_` to a label difference, which is a link-time
constant. On ELF, when linking a shared object, this requires that `__profc_` is
either private or linkonce/linkonce_odr hidden. On COFF, we need D104564 so that
`.quad a-b` (64-bit label difference) can lower to a 32-bit PC-relative relocation.

```
# ELF: R_X86_64_PC64 (PC-relative)
.quad .L__profc_foo-.L__profd_foo

# Mach-O: a pair of 8-byte X86_64_RELOC_UNSIGNED and X86_64_RELOC_SUBTRACTOR
.quad l___profc_foo-l___profd_foo

# COFF: we actually use IMAGE_REL_AMD64_REL32/IMAGE_REL_ARM64_REL32 so
# the high 32-bit value is zero even if .L__profc_foo < .L__profd_foo
# As compensation, we truncate CountersDelta in the header so that
# __llvm_profile_merge_from_buffer and llvm-profdata reader keep working.
.quad .L__profc_foo-.L__profd_foo
```

(Note: link.exe sorts `.lprfc` before `.lprfd` even if the object writer
has `.lprfd` before `.lprfc`, so we cannot work around by reordering
`.lprfc` and `.lprfd`.)

With this change, a stage 2 (`-DLLVM_TARGETS_TO_BUILD=X86 -DLLVM_BUILD_INSTRUMENTED=IR`)
`ld -pie` linked clang is 1.74% smaller due to fewer R_X86_64_RELATIVE relocations.
```
% readelf -r pie | awk '$3~/R.*/{s[$3]++} END {for (k in s) print k, s[k]}'
R_X86_64_JUMP_SLO 331
R_X86_64_TPOFF64 2
R_X86_64_RELATIVE 476059  # was: 607712
R_X86_64_64 2616
R_X86_64_GLOB_DAT 31
```

The absolute function address (used by llvm-profdata to collect indirect call
targets) can be converted to relative as well, but is not done in this patch.

Differential Revision: https://reviews.llvm.org/D104556
2021-07-30 11:52:18 -07:00
Arthur Eubanks ef2627e1fa [profile] Add underscore to /alternatename for Win/x86
/alternatename should use the mangled name. On x86 we need an extra
underscore.

Copied from sanitizer_win_defs.h

Fixes https://crbug.com/1233589.

Reviewed By: phosek

Differential Revision: https://reviews.llvm.org/D107000
2021-07-28 14:58:35 -07:00
Gulfem Savrun Yeniceri e50a38840d [profile] Add binary id into profiles
This patch adds binary id into profiles to easily associate binaries
with the corresponding profiles. There is an RFC that discusses
the motivation, design and implementation in more detail:
https://lists.llvm.org/pipermail/llvm-dev/2021-June/151154.html

Differential Revision: https://reviews.llvm.org/D102039
2021-07-23 00:19:12 +00:00
Gulfem Savrun Yeniceri fd895bc81b Revert "[profile] Add binary id into profiles"
Revert "[profile] Change linkage type of a compiler-rt func"
This reverts commits f984ac2715 and
467c719124 because it broke some builds.
2021-07-21 19:15:18 +00:00
Gulfem Savrun Yeniceri 467c719124 [profile] Change linkage type of a compiler-rt func
This patch changes the linkage type of a compiler-rt func
(__llvm_write_binary_ids) to fix the sanitizer-windows bot
build issue introduced in change f984ac271.

The issue is as the following:
C:\b\slave\sanitizer-windows\llvm-project\compiler-rt\lib\profile\InstrProfilingInternal.h(201):
error C2496: '__llvm_write_binary_ids': 'selectany' can only be applied
to data items with external linkage
2021-07-21 18:22:05 +00:00
Gulfem Savrun Yeniceri f984ac2715 [profile] Add binary id into profiles
This patch adds binary id into profiles to easily associate binaries
with the corresponding profiles. There is an RFC that discusses
the motivation, design and implementation in more detail:
https://lists.llvm.org/pipermail/llvm-dev/2021-June/151154.html

Differential Revision: https://reviews.llvm.org/D102039
2021-07-21 17:55:43 +00:00
Petr Hosek ff427909ca [NFC][profile] Move writeMMappedFile to ELF ifdef block
This avoids the compiler warning on Darwin where that function is unused.
2021-07-19 23:13:13 -07:00