Folks,
I'd like to know the general opinion on this:
https://llvm.org/bugs/show_bug.cgi?id=23529
>From past experience, and my limited point of view, it seems that the
GCC community decided to change the ABI, plotted some runes, conjured
up a broken implementation (whose bugs people will exploit and GCC
will end up keeping them) and pushed distros to follow.
I don't know how much distros were involved in the original design.
Did they request the changes? Or was that an internal change? How do
you guys get to decide when to break the world?
I also don't know how will be your policy towards the bugs that people
started "using". Are you going to force them moving to a more sane
implementation (by issuing compiler/linker errors) or are you going to
support as legacy for the decades to come?
Is there a detailed document describing all the changes, and is the
GCC implementation following that to the letter, or are there still
bugs in the current version? I'm not talking about an email thread,
but an actual document on GCC's docs pages.
All I know is that people in the Clang community waited until GCC's
implementation was stable, but somehow Debian based distros started
moving as if there was something unholy about the previous ABI, and
we've been getting segmentation faults on Hello World!
I don't know much about the changes myself, nor I was involved in the
discussion, but this is now a major pain for release 3.8.0 and I
thought it would be better if I sent an email to Linaro folks and
members (which form most of the ARM GNU developers) instead of going
on the main lists.
That also gives me a bit more freedom to be more candid. I wouldn't
send those questions as they are to the GCC/binutils lists, or people
would kill me. I don't mean any of them as an offence, they are
genuine questions from someone that is not on top of that issue.
cheers,
--renato
Hi Linaro Toochain Group,
I have few questions on glibc+libm w.r.t aarch64.
If possible, please provide some insight, otherwise kindly redirect me to
the concerned person/forum.
1.It seems from the community patches that ARM/Linaro is optimizing glibc
functions such as memcpy/memmove, string for aarch64.
However, looks like some of these (e.g. memcpy/memmov) patches are still
not merged in glibc. Any comment on their availability in glibc?
e.g. https://www.sourceware.org/ml/libc-alpha/2015-12/msg00341.html
2. On the same note, is there any plan for optimizing/tuning libm functions
(e.g. trigonometric) for aarch64?
I could find any matching patches on review board. Please correct me if I
am wrong.
3. Looks like ARM have released an independent version of libm for certain
trigonometric functions.
https://github.com/ARM-software/optimized-routines.
Any plan of these optimization going in glibc's libm? Any comment on its
performance improvement over GNU libm ?
Thanks in advance for your time.
--
with regards,
Virendra Kumar Pathak
== Progress ==
- LTO and TCWG480 (6/10)
* Read and experimented with GCC's LTO codebase.
* Setup archlinux on chromebook and ran coremark with perf
* Read referenced publications
- PR66726 (2/10)
* Rebased the patch
* Regression tested on x86_64 and using Chritsope's setup
* getting ready to post upstream
- Misc (2/10)
* gcc/bug list
* abstract for connect presentation
* Meetings
== Plan ==
* LTO
* bugs
o Teaching activity (2/10)
== Progress ==
o Linaro GCC (4/10)
* Completed backports for 2016.01
* Merged FSF 5 branch into Linaro 5 one
* Delivered 2016.01 Snapshot
* Try to reproduce Linaro PR #1988 with 2016.01.
issue not reproducible
o GCC dev. (3/10)
* Fix armv8.1-a handling at configure time
patch committed upstream in trunk and backported in our branch.
* Investigate TCWG-128 (Thumb2 branch out of range with LTO)
o Misc (1/10)
* Various meetings
== Plan ==
o Look at diff between ABE branches in our validation context
o Continue on-going tasks
Port to microinstance - TCWG-432 [3/10]
* Fallout from attempts to fix race condition
* Various minor fixes - simplifications, better reporting
Backport benchmarking - TCWG-352 [1/10]
* Decoupled 'target triple' from 'toolchain' name
** Immediately, to let me benchmark with Juno-built native gcc
** But this has been causing minor pain for a while
Benchmarking infrastructure documentation - TCWG-496 [1/10]
* Finished drafts of LAVA wrapper and microinstance access docs
Unexpected 1/2 day off [1/10]
Misc [4/10]
* Unusual number of meetings
* Fiddling with my git workflow
* Some thinking about Benchmark-102 for Connect
* Figured out why our SPEC objects were disappearing
== This week ==
* Bugzilla 68543 - [AArch64] Implement overflow arithmetic standard
names (6/10)
- Re-implementing using sign/zero extends as this was not clear from
the documentation
* Bugzilla 69008 - gcc emits unneeded memory access when passing trivial
structs by value (1/10)
- Determined to be structure alignment issue
* Code review of patch posted for Bugzilla 68532 (1/10)
- Completed code review and successfully tested with my vaddw patches
* Bugzilla 67323 - Tested patch for TCWG-318 (non-unit stride loads) (1/10)
- Patch did not generated vld3 so the bug has been reopened
* TCWG-317 - No review feedback yet
* Misc (1/10)
== Next week ==
* Bugzilla 68543 - Complete re-implementation with zero/sign extends and
test
* Bugzilla 69008 - Determine where in compiler issue should be addressed
* USA Holiday (January 18th, MLK Day)
== Progress ==
* Support (5/10)
- Bug triage: reviewing the 300+ bugs on ARM
- closing fixed ones (+35)
- adding deps for meta bugs (kernel, android, ias, chromium) (~15)
- moving to ARM/AArch64 components, etc. (~30)
- Checking a few more kernel patches for Clang support with Arnd
* Plan (3/10)
- Still planning roadmap for 2016/2017
* Background (2/10)
- Code review, meetings, discussions, general support, etc.
# Progress #
* Estimate the effort of GDB kernel-awareness work. Give comments
on formal slides. [1/10]
* GDB inserts breakpoints on the wrong place if these files'
basename is the same. TCWG-491. At least, post a patch to avoid
GDB crash. Reproduced with a simpler case, open PR 19474. [3/10]
* Update document about input interrupt. [1/10]
Done. Patch is committed.
* TCWG-503, fix GDB test case by @progbits -> %progbits. Patch is
posted. [1/10]
* Clean up arm software single step code. [3/10]
In progress. Some patches are committed, but some are still in my
queue. The recent arm software single step change causes a
regression on stepping out of signal handler.
* Think about the GDB slides for Linaro Connect [1/10]
# Plan #
* Look into the arm software single step regression.
* Post the rest of patches for arm software single step clean up.
--
Yao
== Progress ==
* Validation
- a couple of fixes in the validation jobs
- looking at timeouts problems with buildfarm-master
* GCC
- TCWG-485/PR68620: sent patch for review
- TCWG-484/target attributes problems in the testsuite
Rebased, but need more pending updates from Christian
* Misc (conf calls, meetings, emails, ...)
== Next ==
* Validation: more debug. Hopefully start extending validation scope
* GCC:
- bug fixing
- check trunk regressions reported during the holidays
The Linaro Toolchain Working Group (TCWG) is pleased to announce the
2016.01 snapshot of the Linaro GCC 5 source package.
This monthly snapshot[1] is based on FSF GCC 5.3+svn232321 and
includes performance improvements and bug fixes backported from
mainline GCC. This snapshot contents will be part of the 2016.02
stable [1] quarterly release.
This snapshot tarball is available on:
http://snapshots.linaro.org/components/toolchain/gcc-linaro/5.3-2016.01/
Interesting changes in this GCC source package snapshot include:
* Updates to GCC 5.3+svn232321
* Fixes Linaro Bugzillas: #1982, #1980
* Backport of [BugFix] [AArch32] PR 68149 Fix ICE in unaligned_loaddi split
* Backport of [BugFix] [AArch32] PR target/68214: Delete
IP-reg-clobbering call-through-mem patterns
* Backport of [BugFix] [AArch32] PR target/68390
* Backport of [Bugfix] [AArch64] PR rtl-optimization/68796 Add
compare-of-zero_extract pattern
* Backport of [BugFix] [AArch64] PR target/68696 FAIL:
gcc.target/aarch64/vbslq_u64_1.c scan-assembler-times bif\tv 1
* Backport of [BugFix] PR rtl-optimization/68381
* Backport of [ARMv8.1] [AArch32] Add ACLE feature macro for ARMv8.1
instructions
* Backport of [ARMv8.1] [AArch32] Add ACLE intrinsics vqrdmlah and vqrdmlsh
* Backport of [ARMv8.1] [AArch32] Add ACLE intrinsics vqrdmlah_lane
and vqrdmlsh_lane
* Backport of [ARMv8.1] [AArch32] Add patterns for new instructions
* Backport of [ARMv8.1] [AArch32] Add support for ARMv8.1
* Backport of [ARMv8.1] [AArch32] Multilib support for ARMv8.1.
* Backport of [ARMv8.1] [AArch32] Support ARMv8.1 ARM tests
* Backport of [AArch32] [AArch32] Fix armv8.1 support at configure time
* Backport of [AArch32] Add attribute for compatibility with ARM pipeline models
* Backport of [AArch32] Fix vector TYPE_MODE in streaming-out
* Backport of [AArch64] 1/7 Add support for ARMv8.1 Adv.SIMD instructions
* Backport of [AArch64] 2/7 Add sqrdmah, sqrdmsh instructions
* Backport of [AArch64] 3/7 Add builtins for ARMv8.1 Adv.SIMD instructions
* Backport of [AArch64] 4/7 Add ACLE feature macro for ARMv8.1
Adv.SIMD instructions
* Backport of [AArch64] 5/7 Dejagnu support for ARMv8.1 Adv.SIMD
* Backport of [AArch64] 6/7 Add NEON intrinsics vqrdmlah and vqrdmlsh
* Backport of [AArch64] 7/7 Add NEON intrinsics vqrdmlah_lane and vqrdmlsh_lane
* Backport of [AArch64] Documentation fix for -fpic
* Backport of [AArch64] fix 3/7 Add builtins for ARMv8.1 Adv.SIMD instructions
* Backport of [AArch64] Fix a few failures with LSE enabled
* Backport of [AArch64] Improve add immediate expansion
* Backport of [AArch64] Improve comparison with complex immediates
followed by branch/cset
* Backport of [AArch64] Rework ARMv8.1 command line options
* Backport of [AArch64] Update patterns to support FP zero
* Backport of [AArch64] Use aarch64_sync_memory_operand in
atomic_store<mode> pattern
* Backport of [AArch64] Use vector wide add for mixed-mode adds
* Backport of [Misc] Only restrict pure simplification in mult-extend
subst case, allow other substitutions
* Backport of [Testsuite] [AArch64] Skip big-endian as well for
gcc.target/aarch64/got_mem_hoist_1.c
* Backport of [Testsuite] Make stackalign test LTO proof
* Backport of [Testsuite] Testcase for PR rtl-optimization/68381
Feedback and Support
Subscribe to the important Linaro mailing lists and join our IRC
channels to stay on top of Linaro development.
** Linaro Toolchain Development "mailing list":
http://lists.linaro.org/mailman/listinfo/linaro-toolchain
** Linaro Toolchain IRC channel on irc.freenode.net at @#linaro-tcwg@
* Bug reports should be filed in bugzilla against GCC product:
http://bugs.linaro.org/enter_bug.cgi?product=GCC
* Interested in commercial support? inquire at "Linaro support":
mailto:support@linaro.org
[1]. Stable source package releases are defined as releases where the
full Linaro Toolchain validation plan is executed.
[2]. Source package snapshots are defined when the compiler is only
put through unit-testing and full validation is not performed.
o Teaching activity (6/10)
== Progress ==
o Linaro GCC (4/10)
* Restart validation blocked by the last infrastructure issues
* Discussed and reviewed Jenkins jobs fixes.
* Tested potential upstream fixes for linaro BZ 1980 and 1982,
but reverting the offending backports is a better solution.
== Plan ==
o Complete backports and deliver 2016.01 GCC snapshot.
Centralized benchmark source - TCWG-354 [1/10]
* Fixed some bugs in Coremark reporting
* Experimented some more with clang builds, set this aside for now
Port to microinstance - TCWG-432 [3/10]
* Fixed some false assumptions exposed by change to uinstance
* Worked on LAVA side of coremark reporting
Backport benchmarking - TCWG-352 [4/10]
* Fixed some bugs that appeared over the holiday
* Got job dispatching benchmark run to microinstance, with pointer to
backport-containing toolchain
* But resulting job fails
Benchmarking infra documentation - TCWG-496 [1/10]
* Started documenting LAVA wrapper
Misc [1/10]
=Plan=
LAVA side of coremark reporting
Fix backport-dispatched benchmark job
Much more documentation
== Progress ==
LLDB development
-- Fixed thumb mode issues if user triggers changes to PC register
[TCWG-228] [3/10]
-- Triage of testsuite failures on armel chromebook [TCWG-228] [2/10]
-- Update xfail decorators for arm-linux targets. Marked traiged as
xfail. [TCWG-494] [3/10]
Miscellaneous [1/10]
-- Meetings, emails, discussions etc.
Half Day Leave [1/10]
-- Tooth extraction surgery.
== Plan ==
LLDB development
-- Triage more failures on armel chromebook [TCWG-228]
-- Submit an updated xfail decorator for arm-linux. [TCWG-494]
Miscellaneous
-- Connect visa follow up.
== This Week ==
* tcwg-72 (2/10)
- Following Jim's suggestions using XEXP (remainder, 0) worked to
resolve segfault
for DImode case
- Resolved ICE's due to (silly) mistakes.
* PR69133 (4/10)
- Reduced test-case following Markus's suggestions.
- I think it happens because node->lto_file_data is set to NULL in
lto_free_function_in_decl_state_for_node ().
Commenting out the following calls from get_untransformed_body():
lto_free_section_data (file_data, LTO_section_function_body, name,
data, len, decl_state->compressed);
lto_free_function_in_decl_state_for_node (this);
prevents the ICE. I suppose symbols with same name in same partition
share state (?), so setting lto_file_data = NULL
affects state of 2nd occurence of symbol and we hit the assert.
If this is true, gating on lto_file_data doesn't seem unreasonable IMO
to avoid ICE (the code anyway has undefined behavior due to violation of ODR).
I am still not sure why this "works" for partitioning enabled (one,
balanced, 1to1).
* 447.deallII bug (1/10)
- Bug not reproducible (verified with Kugan).
* TCWG-319 benchmarking (1/10)
- first job submission failed due to kernel panic
- a53: base fp run completed, with-patch in progress
* Misc (2/10)
- Meetings
- ipa
== Progress ==
* Validation
- a couple of fixes in the validation jobs
- improvements to the comparison scripts
* GCC
- investigated bugs #1980 and #1982.
Reverted the backports that introduced the problem.
- TCWG-485/PR68620: resumed
- TCWG-484: target attributes problems in the testsuite.
Sent an updated patch, but Christian has also submitted
a patch which modifies the compiler behavior. I'll have
to update mine when his patch is accepted.
* Misc (conf calls, meetings, emails, ....)
== Next ==
* Validation: more cleanup. Hopefully start extending validation scope
* GCC:
- bug fixing
- check trunk regressions reported during the holidays
== Progress ==
* Support (2/10)
- Having a go at PR25722, too hacky for a feature that can
be easily worked around.
- Reviewing some kernel issues with Arnd
* Planning (6/10)
- Drafting 2 year plans for LLVM
* Background (2/10)
- Code review, meetings, discussions, general support, etc.
- Catching up from long holidays
# Progress #
* Handle input interrupt in GDB. TCWG-424. Done. [3/10]
Need to update GDB manual to clarify the expected behaviour of GDB.
* Estimate the effort of GDB kernel-awareness work. Done. [1/10]
* TCWG-491. Ongoing. [2/10]. Understand symbol handling
in GDB.
* Various patch review upstream. [2/10].
* Clean up code on arm software single step after some changes from
Ericsson. Ongoing. [2/10].
# Plan #
* TCWG-491.
* Follow up of TCWG-424 to update GDB manual.
* Clean up code on arm software single step.
* Assess the ARM and AArch64 GDB test result, as 7.11 release is
coming soon.
--
Yao
Is it in the 2015.11-1 release ?
- rob -
-------- Original message --------
From: Jim Wilson <jim.wilson(a)linaro.org>
Date: 01/05/2016 19:45 (GMT-07:00)
To: Xiaofeng Ren <xiaofeng.ren(a)nxp.com>
Cc: linaro-toolchain(a)lists.linaro.org, Zhenhua Luo <zhenhua.luo(a)nxp.com>
Subject: Re: gcc-linaro-5.1 vs gcc-linaro-4.8
On Tue, Jan 5, 2016 at 4:19 PM, Xiaofeng Ren <xiaofeng.ren(a)nxp.com> wrote:
> Hello Jim,
> Appreciate for your comments.
> I will try to manually apply that patch on my side and try it.
> BTW, may I know which released Linaro gcc version include that patch? Maybe I can download it and try it quickly.
> https://gcc.gnu.org/ml/gcc-patches/2015-10/msg00025.html
It was backported to our gcc-5 branch on Nov 24 by Yvan. This is
after the latest release 2015-11 was made. The patch is in the
December snapshot, but I think that is a source only release.
http://snapshots.linaro.org/components/toolchain/gcc-linaro/5.3-2015.12/
You would have to build your own toolchain from that, perhaps by using abe.
Jim
_______________________________________________
linaro-toolchain mailing list
linaro-toolchain(a)lists.linaro.org
https://lists.linaro.org/mailman/listinfo/linaro-toolchain
Linaro TCWG,
In newer toolchains that are built with ABE, libc.a contains a lot of debugging information, including the paths to the source files on the build machine. I think that's because ABE builds the libraries with -g and never strips out the debug information. I verified this with both the 4.9-2015.05 and 5.2-2015.11 binary releases with the command:
arm-linux-gnueabih-objdump -g libc.a | grep '\.c'
In older toolchains that were built with crosstool-ng, libc.a did not contain the paths to the source files. I guess crosstool-ng either didn't build the libraries with -g, or it stripped out the debug information later. I verified this with the 4.9-2014.09 binary release.
I'm not sure whether this change was intentional, or just an oversight during the switchover to ABE. Regardless, it makes the libraries a lot bigger, and it potentially affects the end user during debugging.
The source files of libc, etc. are not typically included with the binary releases. So, when a user of an ABE-built binary release chooses to step into an extern function of libc, gdb will search for the source file. It likely won't be able to access the source file along the same path that worked for the build machine, so it will search its list of source directories. Ultimately, unless the user has downloaded the source files, gdb will likely display a message like "printf.c: No such file or directory".
In contrast, when a user of a crosstool-ng-built toolchain tries to step into an extern function of libc, gdb will be unaware of the name of the source file. As a result, the user will not get a message about a missing file.
So, should the toolchains' libraries really contain debug information? I think it could be useful for a theoretical multilib folder that covers a -g option. On the other hand, for the usual release builds, isn't the debug information a waste of space and a source of confusion?
Thanks,
Fred Peterson
Engineer - Developer Tools
NXP Semiconductors
Hello All,
I found one difference between gcc-linaro-5.1 vs gcc-linaro-4.8 while I'm doing lmbench benchmark test for our LS1043 (cortex-A53).
While using gcc-linaro-4.8, gcc will generate advanced SIMD instructions (like as ld1, etc), however, gcc-linaro-5.1 will not generate advance SIMD instructions. This will cause big performance gap between gcc-4.8 and gcc-5.1 for lmbench memory bandwidth "fcp" test (bw_mem program).
My compiler flags is "-O3 -mcpu=cortex-a53". I also tried several different compiler flags ("-O3 -mcpu=cortex-a53+fp+simd", "-O2 -ftree-vectorize -mcpu=cortex-a53", "-O3 -ftree-vectorize -mcpu=cortex-a53"), all of them doesn't work.
Gcc-5.1 toolchain was downloaded from following link:
https://snapshots.linaro.org/openembedded/sources/gcc-linaro-5.1-snapshot-2…
Can I have your comments on this?
Thanks
Ron
Hello toolchain gurus,
In the course of Linaro's kernel tinification project, the ability to
compile the Linux kernel using LTO is a frequent requirement. However
the kernel makes heavy usage of 'ld -r' with .o files resulting from LTO
build of .c files as well as .o files resulting from pure assembly code.
This mix of LTO and non-LTO object files is not supported by upstream
binutils unless a patch from H.J. Lu is applied. That patch has been
available since 2013 and was last refreshed in his 2.25.51.0.4 branch
last September. It is accessible here:
https://git.linaro.org/toolchain/binutils-gdb.git/commit/6da5456971
I've attached a very simple test case demonstrating the problem. With
the binutils-lto-mixed.patch applied, this test case compiles to a
working executable. Otherwise compilation fails at the 'ld -r' step.
One question and one request:
- What, if anything, has prevented this patch from being merged in the
master branch upstream?
- In the mean time, could we include this patch in the Linaro binutils
package and releases?
Having this available in our toolchain releases would greatly simplify
the LTO related work on the kernel. It was included in all binutils
releases from H.J. Lu since 2013 and therefore has obtained significant
exposure already.
Thanks.
Nicolas
