== Progress ==
* TCWG-413 - Release benchmarking (2/10)
- Analysed all the patches gone into the release for performance impact
* TCWG-521 - Analysed Coremark and Spec2k for uxt/sxt optimizations
TCWG-521(3/10)
- Studied Coremark and have additional patches in development for
missing cases.
* bswap pr43550 (No card yet) (3/10)
- Looked at the pr and it is not fixable with VRP
- An alternate approach based patch is in testing to fix this.
* 1 day off Sick (2/10)
== Paln ==
* Continue with TCWG-521 and bswap pr43550
== This week ==
* TCWG-515 - Neon intrinsic testing part 1 (1/10)
- Attempting testing on ARM targets; resolving build failures
* TCWG-524 - Neon intrinsic testing part 4 (2/10)
- Completed AARCH64 testing with no regressions found
* TCWG-525 - Neon intrinsic testing part 5 (2/10)
- Completed AARCH64 testing with no regressions found
* Launchpad 1312931- gcc 4.8 internal compiler error: in add_stores, at
var-tracking.c (3/10)
- Resolved, tested and requested code review on linaro 4.8 branch
- Resolved and tested on linaro 4.9 branch.
- Reviewing test results
* Linaro Connect Presentation (2/10)
- Met with Maxim and Will to prepare outline
- Began writing "Optimization Levels" portion of presentation
== Next week ==
- Continue Neon intrinsic testing including getting ARM portion to execute
- Complete "Optimization Levels" portion of Linaro Connect presentation
- Resolve Launchpad 1312931 on Linaro 4.9 branch
=Progress=
cbuild2 benchmarking - TCWG-360 [5/10]
* Proposed a plan for benchmark storage
* A bit more thought on cross-running
* Prototyped script to launch/run/release on LAVA targets (e.g. wg)
** The 'run' part is pretty much independent of the 'LAVA' part
lowlevellock.h commentary - ???? [2/10]
* In response to a patch review, added some comments to this fairly obscure file
* Took longer than expected, but the end's in sight so I guess I'll finish
* I know a bit more about futexes and threads now
Meetings/mail/etc [3/10]
=Plan=
Holiday. Back on 1st September.
== Progress ==
Arm gdb record-replay bug fix [4/10] [TCWG-314]
* Testing of all patches on x86 and arm machines.
* Patch submission upstream after sync up with latest gdb trunk.
AArch64 gdb record-replay bug fix [3/10] [TCWG-503][TCWG-498]
* Debugging of testsuite failures with apm and foundation mode.
* Patch series update after sync up with latest trunk.
GDB Tracepoints/Fast Tracepoints support on arm [2/10] [TCWG-480]
* Search for previously submitted patches, wiki info and other documentation.
* Started with initial study and code understanding.
Miscellaneous [1/10]
* Research GDB record/replay usage scenarios.
* Meetings/emails and installation/configurations etc
== Plan ==
AArch64 gdb record-replay bug fix [TCWG-503][TCWG-498]
* Fix test failures if possible otherwise submit updated patches.
GDB Tracepoints/Fast Tracepoints support on arm [TCWG-480]
* Further study and investigate a missing functionality on arm.
Verify GDB non-stop debug support on ARM. [TCWG-246]
== Progress ==
* Looked at implementation options for backlog cards and
closed/postponed cards that doss not benefit or cards that require
excessive re-architecture that will not be possible now. (5/10)
- TCWG-468 - Postponed after detailed study and discussion.
- TCWG-412 (Support literal/constant pool sharing (wont-fix): As it is,
intra procedularl literal pool fix happens in
TARGET_MACHINE_DEPENDENT_REORG with arm_reorg. It is quite complex and
somewahat hacky. it would become even more messy if we are to record
these information and reuse it for whole program as a way to share
literal poool. Additionally A-profile (which is Linaro's focus) dosent
benefit from this as we use movw/movt instead. Therefore decided to
clsoe this as wont fix.
* TCWG-413 - Release benchmarking (2/10)
- Benchmarking or release for a15 and a57
* Analysed coremark and spec2k for uxt/sxt optimizations (3/11)
- Studied coremark and have additional patches in development for
missing cases.
* Misc
- Looked at bug database and monitored patches relevant to arm/aasrch64
== Paln ==
* Look at open uxt/sxt bugs in gcc bugzilla
* Study coremark and then spec2k for uxt/sxt
== Progress ==
* Reviewed a big pile of benchmarking patches for cbuild2. (TCWG 1378
- 2/10).
* Worked on stabilizing "Backport" Jenkins job (TCWG 1378 - 6/10).
* Meetings and Misc (2/10)
== Plan ==
* Keep tracking down Jenkins issues with Backport job.
* Get my D01 board working finally, since I killed my Odroid U2.
* Do more thorough testing of SSH performance improvements.
=Progress=
cbuild2 benchmarking - TCWG-360 [4/10]
* Some disagreements with gerrit
* One of my patches broke bare metal builds (now fixed)
* More review of the way we store benchmark sources
Other - [5/10]
* Meetings, mail, etc (featuring newlib build options, defaults,
cbuild vs benchmark scripts)
* Learned to use git-send-email, bent it to my will
* Authentication pain on ARM-side
Holiday - [1/10]
=Plan=
cbuild2 benchmarking
* Finish source/data storage plan
* Figure out how benchmark cross-running will work, get started on it
Holiday next week (25th - 29th August)
== This week ==
* TCWG-515 - Neon intrinsic testing part 1 (3/10)
- Completed testing with internal compiler error discovered compiling
with -Og and -g
- GCC bugzilla report 62040 already exists for this issue
* TCWG-518 - Neon intrinsic testing part 2 (2/10)
- Completed testing with no regressions found
* TCWG-519 - Neon intrinsic testing part 3 (2/10)
- Completed testing with no regressions found
* Reviewed branch merge of Linaro 4.8 and 4.9 with FSF trees (3/10]
== Next week ==
* Continue with neon intrinsic testing
* Investigate bug 373 - [gcc-4.9][android][armv8] internal compiler error
* Develop outline for Linaro Connect presentation
== Progress ==
* TCWG-180 Debugging - LTO bootstrap failure in Aarch64 with GCC 4.9. (6/10)
Analyzed tree dumps between stage2 and stage3 compilers. Noted that
extra gimple declarations in stage2 compiler.
Also inline minimum parameters in gimple.c.048i.inline were different.
Richard Beiner found that way we hash the string literals was
different between stage2 and stage3 compilers.
Tested the patch, he posted on trunk but it does not solve the
comparison failures.
Also GCC garbage collector parameters for expansion and maximum heap
size varies between stage 2 and stage3. Setting this parameters to
same value in stage2 and stage3 builds fixes bootstrap failures in
Aarch64.
These fixes are for trunk and stage 1 5.0.
Ref: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=62077
Still need to find a work around for release branches
Testing these patches on AMD64 /Aarch64 machines.
* Others (2/10)
* Upstream patch review.
* 1-1 Meeting Christophe.
* Misc and internal meeting.
* setup LLVM.
* Friday national holiday (2/10).
== Plan ==
* Continue LTO bootstrap issue.
* Benchmark Core mark with LTO.
* Upstream patch review.
== MISC ==
None .
The Linaro Toolchain Working Group (TCWG) is pleased to announce the 2014.08
stable release of both Linaro GCC 4.9 and Linaro GCC 4.8 source package.
With the imminent release of ARMv8 hardware and the recent release of the
GCC 4.9 compiler the Linaro TCWG will be focusing on stabilization and
performance of the compiler as the FSF GCC compiler. The Linaro TCWG provides
stable[1] quarterly releases and monthly engineering[2] releases.
Linaro GCC 4.9 2014.08 is the fifth Linaro GCC source package release and first
stable one in the 4.9 series. It is based on FSF GCC 4.9.2-pre+svn213803 and
includes performance improvements and bug fixes.
Interesting changes in this GCC source package release include:
* Updates to GCC 4.9.2-pre+svn213803
* Backport of [AArch64] Drop ISB after FPCR write.
* Backport of [AArch64] Remove from arm_neon.h functions not in the spec
* Backport of [AArch32] Fix check for __FAST_MATH in arm_neon.h
* Backport of [AArch64] fix and enable non-const shuffle for bigendian using TBL
instruction
* Backport of [AArch64] Fix constraint vec_unpack_trunk
* Backport of [AArch32] Cortex-A5 rtx costs table
* Backport of [AArch32] Handle clz, rbit types in arm pipeline descriptions
* Backport of [AArch64] Fix argument types for some high_lane* intrinsics
implemented in assembly
* Backport of [AArch64] Handle fcvta[su] and frint in RTX cost function
* Backport of [AArch64] Prologue rewrite + performance.
Linaro GCC 4.8 2014.08 is the fourteenth release in the 4.8 series and is in
maintenance. Based off the latest GCC 4.8.4+svn213802 release, it includes
performance improvements and bug fixes.
Interesting changes in this GCC source package release include:
* Updates to GCC 4.8.4+svn213802
* Backport of Fix [AArch32] for PR sanitizer/58543
* Backport of Fix [AArch64] for PR target/59744
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 Launchpad against "Linaro GCC project":
http://bugs.launchpad.net/gcc-linaro/+filebug.
* Questions? "ask Linaro":
http://ask.linaro.org/.
* Interested in commercial support? inquire at "Linaro support":mailto:
support(a)linaro.org
[1] Stable source package releases are defined as releases where the full Linaro
Toolchain validation plan is executed.
[2] Engineering source package releases are defined as releases where the
compiler is only put through unit-testing and full validation is not
performed.
= Progress ==
* Upstream work (4/10, CARD-341)
- Patch review (gdb, binutils)
- glibc testsuites for ARM and AArch64
* Booting upstream kernel on Chromebook (4/10)
- Tried u-boot path but Google's shipped NV u-boot is broken in several ways
- Fall back to trying to boot a signed upstream kernel
- Gets halfway through boot then seems to turn the regulators off
* Malloc application benchmarking (2/10, TCWG-441)
- Add pdf library benchmark to benchmark suite
- Improve argument parsing and result plotting
- Seem to be hitting some new issue with mounting memory cgroup within docker
== Issues ==
* None
== Plan ==
* Iron out issues with malloc benchmark framework
* Get Chromebook booting
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
* Benchmarks (CARD-716 2/10)
- Wrapping up SPEC run, some results
- Fix missing tests, running with -ffast-math
* Toolchain (CARD-862 4/10)
- A few more assembler fixes
- http://llvm.org/PR20462
- http://llvm.org/PR20595
- Some vectorizer cleanups
- http://llvm.org/PR20655
* Background (4/10)
- Code review, meetings, discussions, etc.
- Trying D01s again as buildbots
== Plan ==
* Pick some more toolchain bugs to work on
* Release 3.5 tests and benchmarks
* Move bots to CMake, enable compiler-rt
* Change compilation to make Clang/LLVM non-NEON by default
* Enable D01s as buildbots
== Issues ==
* None.
== Progress ==
* Update "Keep constants in register when expanding" patch according
to comments (TCWG-486, 3/10).
* Refine ccmp patches since the old ones can not apply (TCWG-488, 2/10).
* Check PR62151, it was exposed by my commit r211885. But the root
cause was r208165.
* R/M toolchain related work (5/10)
== Plans ==
* Ping pending patches.
== Planed leaves ==
* Aug. 18 - 26.
Hi,
I am trying to build a working environment with Angstrom Linux, QT5 for
Toradex Apalis iMX6 module and additional QT5 SDK for my host machine.
The compilation of Angstrom with additional QT5 layer went fine.
However, there are some problems when building toolchain for the host
machine (command meta-toolchain-qt5). It seems that there is a problem
with configuration of linaro gcc runtime 4.8.
The reported error is: "checking dynamic linker characteristics...
configure: error: Link tests are not allowed after GCC_NO_EXECUTABLES".
The log file is in the attachment. Does anyone have any idea how to fix
this problem? I am using Ubuntu 14.04.
Thanks in advance,
Crt Gorup
Hello,
I'm trying to perform cross platform remote debugging from my x86 computer
to a remote ARM device. (ex: Beagle Bone Black)
I'm using Ubuntu 14.04 and Eclipse CDT.
I installed the Ubuntu package gcc-arm-linux-gnueabihf but the gdb debugger
is not in that package.
Therefore, I'm wondering if I should use the Linaro toolchain (which
include arm-linux-gnueabihf-gdb) or if I should install the Ubuntu
gdb-multiarch package.
What is the difference between gdb-multiarch and arm-linux-gnueabihf-gdb ?
Is it better to use gdb-multiarch ?
Thanks
ssinfod
== Progress ==
BeagleBoard and Pandaboard setup for gdb testing [3/10]
* Deployed debian linux on both boards
* Installation of required packages for gdb testing
* Run gdb native tests on Beagleboard, Pandaboard and Chromebook
Arm gdb record-replay bug fix [6/10] [TCWG-314]
* Applied previous asimd and vfp patches and test for failures
* Fixed exception due to a bug in previous patches with latest trunk.
* Fixed issues with solib trampoline code when stepping in reverse.
Miscellaneous [1/10]
* In pursuit of deploying linux on boards dd,ed my hard drive by
accident and spent some time re-installing my desktop.
== Plan ==
Arm record-replay bug fix [TCWG-314]
* Test bug fixes with all previous patches and resubmit.
Update and Submit Aarch64 record replay patches.
Resume work on Aarch64 prologue analysis.
=Progress=
cbuild2 benchmarking- TCWG-360 [6/10]
* Existed patches tidied up and gerritified
* A few drive-by bugfixes
* Cross-building support added and also gerritified
lowlevellock patch [1/10]
* Resubmitted bugfix, reacted to review
* Some wrestling with space-mangling evil in mail clients
Meetings/mail/etc [3/10]
=Plan=
Holiday Tuesday afternoon
cbuild2 benchmarking
* Get remote targets working
* Maybe look at results gathering
== This week ==
* Linaro bugzilla 331 - [4.9 Regression] ICE in final_scan_insn, at
final.c:2952 [4/10]
- Triaged and investigating
* Linaro bugzilla 307 - [Linaro 4.8] ICE in change_address_1, at
emit-rtl.c:2019 [1/10]
- Unable to validate. Further investigation before closing
* TCWG-515 - Neon intrinsic bug testing [5/10]
- Linaro Infrastructure headaches and setup.
- May split into several blueprints targeting two to three tests
- Require assistance debugging DejaGnu script error preventing some
tests from running
== Next week ==
- Continue with Neon intrinsic testing
- Continue investigation into resolution for Linaro bugzilla 331
== Progress ==
* Fixed aarch64_be-none-elf builds and cross tests. (2/10)
* Forked ancient RSH support in DejaGnu to be more SSH aware. SSH
is now takes 1/4 of the time it used to for each remote execution test
case. (TCWG 512 - 4/10)
* Stabilized new ZSSH protocol in DejaGnu, also much faster than the
current RSH/SSH support. (TCWG 512 - 3/10)
* Meetings and Misc. (1/10)
== Plan ==
* Get my D01 board working finally, since I killed my Odroid U2.
* Review benchmarking patches.
* Do more thorough testing of SSH performance improvements.
== Progress ==
* TCWG-180 Debugging - LTO bootstrap failure in Aarch64 with GCC 4.9. (6/10)
Tried to recompile object files that showed differences in GCC trunk
on amd64 machine.It turned out that GCC make system always builds LTO
files in “gcc” directory and uses prev-gcc/xg++ to build them.
No differences when building stage2 and stage3 objects under same
directory “gcc” in FSF trunk.
Doing the same experiment with gcc 4.9 branch shows failure both in
x86_64 and aarch64.
Communicated the issue to Richard Beiner. Got some feedback on
debugging from Honza and Richard.
Tried passing –save-temps to BOOT_CFLAGS using FSF trunk, getting
assembler errors. Need to come back to this problem after finding root
cause.
Read about LTO passes and wpa (partitioned ) vs non portioned mode.
* Others (2/10)
* Upstream patch review.
* 1-1 Meeting Christophe.
* Misc and internal meeting.
* setup LLVM.
* Friday off (2/10).
== Plan ==
* Continue LTO bootstrap issue.
* Benchmark Core mark with LTO.
* Upstream patch review.
== MISC ==
* National holiday on 15th August .
== Progress ==
TCWG-445 - AArch64 does not generate post-decrement stores.
* Issue resolved with Jiong's prologue/epilogue patch committed to trunk.
* Closed the card.
TCWG-291 - Zero/sign extensions (5/10)
* More review and posted patch based on that which was accepted
* Ran full set of validation (including s390x, aarch64 be, x86 and arm)
* Committed two outstanding patches
* Closed the card.
TCWG-413 - Release benchmarking (2/10)
* Benchmarking or release for a15 and a57
TCWG-468: Sha1 regress (2/10)
* Experimented with back-end patterns. Not much improvement for the
test-case.
- Misc (1/10)
* Looked at gcc bugs and closed old ones.
* Posted test-case patch.
== Plan ==
- Sha1 regressions
- Fixing assigned Bugs
