== Progress ==
* Tuesday sick
* Automation Framework (CARD-1378 5/8)
- Fixing more crashed Junos
- Working on QEMU VMs on gateway (buildbot master, nagios)
- Finalising server purchase, building new rack
* Release 3.6 (TCWG-575 1/8)
- AArch64 built and tested, all green.
- ARM is broken, investigating...
* Background (2/8)
- Code review, meetings, discussions, etc.
- Jira farming
== Plan ==
* More Jira farming
* Setup the new server in the rack
* 3.6 benchmarking against 3.5
== Progress ==
* GCC trunk/4.9 monitoring (1/10)
- noticed a regression in the 4.9 branch
* AArch64 sanitizers (2/10)
- juno-01 back online, so I could rebuild & test llvm with asan tests enabled
- several failures reported, debugging...
* Neon intrinsics tests (1/10)
- pushed a series of patches, probably too late before stage 4.
* Backports (1/10)
- reviews
* Misc (5/10)
- meetings, conf-calls, emails, ....
- Connect preparation
== Next ==
* GCC trunk/4.9 monitoring
* AArch64 sanitizers
* Neon intrinsics tests
* cbuild2/abe: improve reporting, look at backport and tcwgweb scripts
The Linaro Toolchain Working Group (TCWG) is pleased to announce the 2015.01
engineering release of the Linaro GCC 4.9 source package.
Linaro GCC 4.9 2015.01 is the tenth Linaro GCC source package release in the
4.9 series. It is based on FSF GCC 4.9.3-pre+svn219502 and includes performance
improvements and bug fixes.
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.
Interesting changes in this GCC source package release include
* Linaro bugzilla PR fixed : #902
* Updates to GCC 4.9.3-pre+svn219502
* Backport of [AArch64] Support SISD variants of SCVTF,UCVTF
* Backport of [AArch64] Fix ICE in aarch64_float_const_representable_p
* Backport of [AArch64] Switch to sched-pressure by default.
* Backport of [AArch64] Add scheduler for ThunderX
* Backport of [AArch64] Remove crypto extension from default for cortex-a53,
cortex-a57
* Backport of [AArch64] doloop pattern for -fmodulo-sched
* Backport of [AArch32] Add execution tests of ARM REV intrinsics.
* Backport of [AArch32] Post-indexed addressing for NEON memory access
* Backport of [AArch32] Improve 64 bit division performance (serie)
* Backport of [AArch32] Revert 215321 backport.
* Backport of [AArch32/AArch64] Add ACLE 2.0 predefined macros
* Backport of PR tree-optimization/54742 - extend jump thread for finite state
automata
* Backport of PR target/61997 - cc1plus ICE with aarch64 target using PCH and
builtin functions
* Backport of PR target/63724 - Fix up BSL expander for floating point types
* Backport of [LRA] Relax one gcc_assert in lra-eliminate for fixed register
* Backport of Add clobber_reg function
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
* 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 ==
* LLD (3/10, TCWG-561, TCWG-570)
- Re-spin and re-submitted error handling patch and dependencies
- Submitted a patch for crash with sh_addralign == 0
* AArch64 ILP32 toolchain (2/10)
- Got a native ILP32 toolchain building on a hackbox
- Got a working hacking session on Juno with help from Dave
- Just need to combine the two
* Other stuff (1/10)
- Email, meetings, etc.
* Holiday Thursday and Friday (4/10)
== Issues ==
* None
== Plan ==
* Run make check on ILP32 glibc
* Connect presentation
* Try and get outstanding LLD patches committed
--
Will Newton
Toolchain Working Group, Linaro
== Issues ==
* none.
== Progress ==
* Give compilation courses at University (5/10)
* GCC 4.9 2015.01 (4/10)
- Gave some support to Michael for his backports
- Performed mine.
* Misc: (1/10)
- meetings.
- mails
== Plan ==
* Linaro GCC 4,9 2015.01 release
ABE benchmarking automation - TCWG-360 [8/10]
* Added some sanity checks, fixed some bugs in LAVA URL handling
* Some struggling with Jenkins, with support from Rob
* Compress builds before scp-ing
* Resurrected SPEC, fixed some problems I'd forgotten about
Misc [2/10]
* Email catchup, start-of-year meetings
* Brief look at single-thread optimization in glibc (TCWG-436)
* Connect prep (including immunization boosters, I'm still a little woozy)
=Plan=
Test SPEC with present state of automation
Keep struggling with Jenkins
Address benchmark storage
Actually do something with TCWG-436
Write up/think about where to go next with libm exercising
== Progress ==
* TCWG-486 (4/10)
- Ready to start benchmarking
- Discussed with Bernie on new benchmarking set-up
- Waiting for Ryan on permission
* https://bugs.linaro.org/show_bug.cgi?id=412 (4/10)
- Created a reduced test-case and filed upstream bug
- looked at GCSE code and ifcvt code/dumps
* gcc bugzilla and gcc-patches list (2/10)
- Review patches
- Looked at arm related bugs
== Plan ==
TCWG-486 and TCWG-555
== This week ==
* GCC Modularization Project
- Worked on header-file flattening tool to make it usable for
flattening headers.
Used it to flatten expr.h.
- fixed build issue with ggc-compile.c due to tree.h flattening.
- committed Michael's tree.h flattening patch to FSF trunk (r219402).
- built r219402 on all targets in config-list.mk with languages: all, go
- upstreamed patch to fix visium build on FSF trunk which broke due to r219402.
- fixed plugin/ggcplug.c test-case which was failing due to r219402.
- sent expr.h flattening patch upstream for review.
== Next Week ==
* Commit expr.h flattening patch to FSF trunk.
* Commit cfgloop.h flattening patch to FSF trunk.
* Try to commit df.h flattening patch to FSF trunk.
== Progress ==
* Automation Framework (CARD-1378 2/10)
- DNS requests (Hetzner, bridge), new APM
- Following up on purchase of new server, new rack, UPS
* Buildbots (TCWG-76 4/10)
- Enabling swap on dragon boards (local bots)
- Investigating AArch64 RT failure, XFAIL like ARM
- AARch64 full bot (Clang+LLVM+RT) green! Getting ARM to make it public
- Installing QEMU VMs for different buildmasters
* Background (4/10)
- Code review, meetings, discussions, etc.
- Email back-log...
- FOSDEM organisation (accepting talks, fixing Penta cock-ups)
== Plan ==
* Increase check-all verbosity on bots
* Finish installing all buildmaster QEMU images
* Setup new rack
* Probably more release 3.5.1 testing (RC3), CMake bot testing
* Help Will and Christophe setting up their bots
== Backlog ==
* Fix libc++abi NEON/VFPv3 unwinder
* FPU parser (PR20787)
* .fpu in assembler (PR20757)
* -mpfu in driver (PR20700)
== Progress ==
* LLD (5/10, TCWG-561, TCWG-570)
- Re-spin and re-submitted error handling patch
- Implemented some more relocations and added tests
* AArch64 ILP32 toolchain (2/10)
- Investigated OE for building a native toolchain
- Started on hacking ABE to build a "cross on native" toolchain
- Switched to hackboxes as the performance is better than LAVA
* Released newlib and binutils 2015.01 releases and updated release
procedure (2/10)
* Other stuff (1/10)
- Email, meetings, etc.
- Pushed new strcpy to cortex-strings
== Issues ==
* None
== Plan ==
* Get ABE building a native ILP32 toolchain on LP64
* More LLD work
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
* Back from holiday
* Public holiday Thursday (2/10)
* TCWG-486 (4/10)
- Simplified existing patch
- Discussed with Zhenqiang
* TCWG-555 (4/10)
- propagate wrap/overflow information to ssa
== Plan ==
TCWG-486 and TCWG-555
== Progress ==
* Holiday Monday to Wednesday (6/10)
* Public holiday Thursday (2/10)
* Catch up on email (1/10)
* LLD for AArch64 (1/10, TCWG-563)
- Rebased, tidied up and submitted error handling patch
== Issues ==
* None
== Plan ==
* Figure out how to get a native ILP32 toolchain
* More LLD work
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
1 day week:
* Automation Framework (CARD-1378 1/2)
- Resetting APMs and Junos that fell during the holidays
* Release 3.5.1 (TCWG-477 1/2)
- Testing RC2, all good
== Plan ==
* improve buildbots (swap, verbose, libs)
* fix libc++abi NEON/VFPv3 unwinder
* purchase new builder, prepare lab for it
== This week ==
* GCC Modularization Project
- started working on flattening rtl.h
- submitted expr.h flattening patch upstream for review.
- worked with Michael on tree.h flattening patch.
== Next week ==
- investigate bug 905
- modify expr.h, cfgloop.h patches to avoid conflicts with Michael's
tree.h flattening patches and submit upstream.
== Progress ==
* 2015.01 release
- reviewed Yvan's backports
* GCC trunk/4.9 monitoring (2/10)
- still tracking cause of random "interrupted system call" errors
* AArch64 sanitizers (1/10)
- LLVM+compiler-rt build seems OK, but still unable to run the tests on Juno
* Neon intrinsics tests (3/10)
- continued conversion to GCC testsuite
- fixed checks for Neon QC flags
- patch reviews
* Misc (4/10)
- meetings, conf-calls, emails, ....
== Next ==
* GCC trunk/4.9 monitoring
* AArch64 sanitizers
* Neon intrinsics tests
* cbuild2/abe: look at backport and tcwgweb
On holiday until January 5h
ABE benchmarking automation - TCWG-360 [6/10]
* TCWG/LAVA lab interactions working with just a couple hacks
* Jenkins job created (thanks Rob) and beginning to work
Misc - [4/10]
* Included a quick look at Will's single-thread performance patches -
unfortunately, no time to actually do anything with them
=Plan=
Benchmarking storage/automation
Look though Will's single-thread performance patches
Switch to TCWG Junos
Write up/think about where to go next with libm exercising
Test benchmark repeatability
=Absences=
On holiday Monday 22nd Dec to Friday 2nd Jan
== Progress ==
* Friday off
* Automation Framework (CARD-1378 2/8)
- Writing up best practices, lab structure, etc
- Removing port forwards, private keys from the lab
- IT ticket to remove tcwg-sysadmin from public git
* Toolchain (CARD-862 1/8)
- Fixed compiler-rt problem on AArch64 build
- Fixing libcxxabi problem on ARM unwind
* Release 3.5.1 (TCWG-477 1/8)
- Finished RC1 testing on ARM and AArch64
* Buildbots (TCWG-76 2/8)
- Found commit that broke lnt bot, green again
- Setting up internal bots (4 ARM, 1 ARM64)
* Background (2/8)
- Code review, meetings, discussions, etc.
== Plan ==
* Continue 3.5.1 release (RC2, etc)
* Fix a few remaining problems with Compiler-RT / libc++abi
Holiday until January...
== Progress ==
* Holiday Thursday and Friday (4/10)
* LLD for AArch64 (2/10, TCWG-563)
- Completed patch to improve relocation range checking
- llvm mailing lists down so no reviews happening
* ILP32 work (3/10)
- Several test builds on Juno, still cannot get a working native toolchain
* Email, meetings, other stuff (1/10)
== Issues ==
* Cable power level issue causing broadband problems
== Plan ==
* On holiday until in January, but I will check email periodically
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
QEMU kernel debugging setup [3/10] [TCWG-568]
-- Finally figured out the steps to debug kernel with ARM cortex A9 core.
Kernel debugging to investigate limitation of ARM watchpoint
implementation [2/10] [TCWG-569]
-- Debug kernel code with gdb to investigate problems.
Miscellaneous [1/10]
-- Meetings, Emails etc
-- Wiki updates
Sick Leave [4/10]
== Plan ==
QEMU kernel debugging setup [TCWG-568]
-- Figure out steps to debug kernel with ARM cortex A15 core.
More Kernel debugging to investigate limitation of ARM watchpoint
implementation [TCWG-569]
-- Get A15 core running on QEMU and debug kernel hw debug
implementation on both A9 and A15.
Updates to GDB wiki pages with kernel debugging howto.
I've completed renaming Cbuildv2 to Abe, as well as modifying all the
board files for remote testing to match what the new DNS server is
using. Please stop using the existing cbuild2 repository, as all changes
are now done in the new repository. I'll leave the old repository for a
while, I know there are branches there people depend on. For those, I'm
sorry for the pain (having just gone through that myself), but please
migrate any important branches to the new repository. URLs for that are
here:
https://git.linaro.org/toolchain/abe.git
- rob -
ABE benchmarking - TCWG-360 [7/10]
* Finished post-move LAVA lab workarounds
* Should be flexible enough for TCWG lab too, but we'll see
* Refactored sources - now more robust and maintainable
lowlevellock.h comments - CARD-341 [1/10]
* A further round of changes, got an OK
Raised a bug on GCC re questionable generation of pow calls - TCWG-558 [0/10]
* Upstream fixed it really fast
Misc - [2/10]
=Plan=
Switch to TCWG Junos, perhaps start using gprof
Think about where to go next with libm exercising
Look though Will's single-thread performance patches
Test benchmark repeatability
Port benchmarking scripts to ABE repo
Start on storage/automation, if Rob has time
=Absences=
On holiday Monday 22nd Dec to Friday 2nd Jan
== Issues ==
* none.
== Progress ==
* GCC 4.9 2014.12 (2/10)
- Released
* Bugzilla #902 (ICE in split_reg on armeb-linux-gnueabihf) (6/10)
- reduced test case
- identified the commit that introduced the bug
- identified the trunk revision that fixed it which is
not backportable on our branch as it
* Misc: (2/10)
- Revisions management script
- Various meetings.
== Plan ==
* Bugzilla #902
* List the backports needed for 2015.01
* Libunwind
