Monday 22nd April 2013

This month's meetings

WorkingGroups/ToolChain/Meetings
<< <  2012 / 12 >  >>
Mon Tue Wed Thu Fri Sat Sun
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
31            

EtherPad

Attendees

  • Matthew Gretton-Dann
  • James Elliott

GCC Team

  • Christophe Lyon
  • Yvan Roux
  • Venkat Kumar
  • Kugan Vivekarandarajah
  • Zhenqiang Chen on vacation

LLVM + Libs & Dev Tools:-

  • Renato Golin
  • Will Newton
  • Mans Rullgard
  • Omair Javaid
  • Peter Maydell

From ARM:-

  • Richard Earnshaw
  • Ramana Radhakrishnan (absent)

From Canonical

  • Mathias

Agenda

  • Review activity last week
  • Discuss plans for coming week
  • Raise any issues holding up progress

Action Items from this Meeting

* TBA

Action Items from Previous Meeting

  • [Matt / Christophe] Discuss Cbuild disk full cleanup

Minutes

GCC Team

Christophe

  • From JAMES - Thanks for getting confirmation of the closeout on 323 Neon Intrinsics!

Progress

  • Disable-peeling: waiting for reference job results. Some issues caused by failed attempt to kill posted jobs, resulted in killed databases. Hardware back on line now.
  • Libsanitizer:
    • identified the reason for different behaviour on board and on simulator
    • proposed a workaround on gcc-patches (qemu-user has a few limitations, so it's desirable to skip some libsanitizer tests)
    • in the mean time libsanitizer developers received similar concerns from android, and are considering a runtime environment variable to control use of stderr
  • Neon intrinsics:
    • resumed investigation on vuzp crc
  • Internal support

Plan

  • Peeling: analyze results when available
  • Revert-coalesce: same
  • Libsanitizer: reach agreement with upstream
  • Neon intrinsics: continue with vuzp crc

Issues

  • none

Yvan

Progress

  • Libunwind AArch64 support:
    • Implemented signal frame detection routine.
    • Resuming with signal is failing, debug ongoing.

Plan

  • Libunwind AArch64 support:
    • Fix and submit upstream

Issues

  • No access to ARMv8 documentation
  • ACTION: James to sort out with Philip (DONE 22nd April)

Venkat

Progress

  • Chromebook received!

Very short week (3 days gone in meetings), was doing some AMD internal tasks and attend local meetings.

  • gprof support work for Aarch64.
  • Working on GCC side to support gprof for Aarch64

Plan

  • Continue gprof support work for Aarch64

Issues

  • No access to ARMv8 documentation
  • ACTION: James to sort out with Philip (DONE 22nd April)

Kugan

Progress

- http://cards.linaro.org/browse/TCWG-13

  • - performance regression is due to alignment of function. - there is an increase in runtime of core_state_transition even

though there is no difference in the code generated with the patch

  • - Adding nops seems to improve the locality and performance; it gets

better than without the patch for THUMB2.

- Get spec2000 results for http://cards.linaro.org/browse/TCWG-14 to decide on the next step

  • - Couldn?t get SPEC2000 results with CBUILD - set-up spec benchmarks in chromebook and now running locally.

- https://blueprints.launchpad.net/gcc-linaro/+spec/better-end-of-loop-counter-optimisation

  • - Initial investigation shows that the code generated is same as expected.

Plan

- http://cards.linaro.org/browse/TCWG-13 - follow it up - Get spec2000 results for http://cards.linaro.org/browse/TCWG-14 to decide on the next step - Look for improvement in VRP for zero/sign extension

Issues

  • None

Zhenqiang Chen

Progress

  • ARM internal training and R/M toolchain related work.
  • Investigate Linaro toolchain 4.8 build issues.

Details: 1. Fix several linaro toolchain 4.8 binary build issues:

  • nls patch need be updated to add (char *) when assigning the

result of xmalloc to a char*.

  • gcc build pass-2 need build libbacktrace (get patch from

crosstool-ng upstream).

  • gcc build pass-2 build with "-j4" fail. Seams build order issue. A

workaround is to remove "-j4".

  • Mingw32 confiugre fail due to missing ISL. A workaround is to add

"--without-isl"

Plan

* Work with Bero to release 4.8. * Swith to ISL/CLooG for future release.

Outside Linaro

Points raised

Mathias

  • last week gcc 4.8, don’t keep change log, caused pain with merging, others do. ACTION: Mathew

==Richard==

  • Raised issue about Jira card states being out of date.
  • ACTION: James to chase with Kate - they should have been updated after last TSC meeting.

Ramana

  • Not present

Matthew

Progress / Announcements / Updates / Comments to team

LLVM Team

Renato

  • Thanks for supporting Aditya on LLVM Linux build over the weekend.

Progress

  • Running around JM/lencode bug - caused by a codegen opt (ICMP fold) that had repercussions only on A9 and A15 code generation - spent three days trying to reduce the case when the problem fixed itself

miraculously >:(

  • Planning for the future - Agreeing on short-term plans for Q2
  • Buildbot - Working on self-hosting bot - Moved local buildmaster to hackbox
  • Investigating LLVMLinux - Building Android kernel with LLVM - Investigating breakage in Debug mode

Plan

  • Continue self-hosting bot
  • Try running a CBuild benchmark with LLVM - need Matts support.
  • Start putting up together the infrastructure for release 3.3
  • Try to extract useful information from perf database

Issues

Needs more stable boards for LLVM for buildbots..

LIBS & DEV TOOLS:-

Will Newton

Progress

  • Further work on glibc memcpy IFUNC patch based on review.
  • Ported libdwarf to aarch64 and submitted upstream.
  • Disabled gold build in binutils cbuild job and created a gold job.
  • Investigated binutils testsuite failure on precise.
  • On leave Friday.

Plan

  • Submit a patch for binutils testsuite failure on precise.
  • Complete work on glibc memcpy iFUNC patch.
  • Follow up binutils IFUNC patch.
  • AArch64 IFUNC next...

Issues

* None.

Mans Rullgard

Progress

  • continued glibc performance, ran spec 2k,

1.5 % spent in lib c rather than lib m, sqrt fn not using. Function list pushed on wiki.linaro.org - need to add Spec 2k6 results

Plan

  • Spec 2k6 data,
  • qemu 32bit extns - ACTION Mathew to email with details

Issues

  • No access to ARMv8 documentation
  • ACTION: James to sort out with Philip (DONE 22nd April)

Omair Javaid

Progress

  • Completed investigation of GDB dwarf test suite failures on ARM.
    • Almost all failures will be fixed with small updates to test cases assembly language files.
  • Checked-out GDB 7.6 branch and did comparison of dwarf test case failures on arm. Some test cases have been updated already and dont need a patch.
  • Started filtering GDB test suite results on ARM for possible failures due to assembly language incompatibility.
  • Tried some experimentation with cbuild, and ec2 instances only to find out they are not connected with lab.
  • Completed Ireland Visa process. Spent Friday and most part of Thursday for the documents attestation, fee submission and postage stuff.

*** Still No blue-print available to log work in JIRA.

Plan

  • Filter all arm assembly language compatibility updates required to gdb test suite and create a patch.
  • Get an intro to gdb patch submission process for a possible fix to gdb test suite for arm assembly compatibility.
  • Start looking into gdb.mi test suite failures on arm.
  • Fill up the arm native vs arm remote comparison sheet.
  • Figure out a way to use hackbox for gdb testing on arm.

Issues

  • time lag to remote lab boards local chromebook needed.
  • ACTION James to send Chromebook this week.

Peter Maydell

Progress

  • qemu maintenance
  • * sent arm-devs and target-arm pullreqs now we're in softfreeze
  • VIRT-4
  • * received Arndale board, confirmed it works (took several
    • hours mostly due to bonkers power switch design)
  • * VIRT-49
  • ** making progress; updated card with a list of sub-subtasks

Plan

  • keep pushing on with VIRT-49
  • finish config of arndale board, test running KVM on it
  • book travel/hotel for Connect Dublin
  • office move Fri 26/Mon 29

Issues

Outside Linaro

Points raised

Mathias

LLVM question, discussion stable LLVM branch, would Linaro be part of effort? NO,

===Richard ===

  • No,

Ramana

  • Absent

Mathias

  • No

Matthew

Progress / Announcements / Updates / Comments to team

  • None

WorkingGroups/ToolChain/Meetings/Archive/2013-04-22 (last modified 2013-08-30 11:49:30)