Monday 16th May 2011

This month's meetings

WorkingGroups/ToolChain/Meetings
<< <  2011 / 5 >  >>
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          

Attendees

Agenda

  • Review action items from last meeting
  • Release this week
    • 4.5 is clean on i686, x86_64, armv7
    • 4.6 is clean on i686, x86_64, armv7
      • A new bug in the testsuite - should it be reported?
      • gcc.dg/pr48335-2.c:19:30: internal compiler error: in expand_expr_addr_expr_1, at expr.c:6949
    • Pending branches:
      • ~rsandifo/gcc-linaro/vldN-vstN-vectorisation
      • ~rsandifo/gcc-linaro/vldN-vstN-vectorisation-4.6
      • ~ramana/gcc-linaro/neon-setlane4.6 - gfortran bessel -O3 failure
      • ~cltang/gcc-linaro/lp-748138-cfgrtl-fix-4_5 - fixes the reported fault

Action Items from this Meeting

  • ACTION: Michael to shift the call an hour later starting next week
  • ACTION: Michael to point Ramana at the logs for VFP vs NEON
  • ACTION: Ramana to check into the bintuils vmov patch and recommend on the upgrade
  • ACTION: Andrew to check inside CSL re: shrink-wrap clobbers to see what Bernd has done
  • ACTION: Peter to try today's hwpack and rootfs on vexpress and OMAP3

Action Items from Previous Meeting

Minutes

ACTION: Michael to shift the call an hour later starting next week

gcc.dg/pr48335-2.c:

  • Caused by a new patch
  • Michael to see if it is a regression and log

Significant difference between the A8+VFP and A9+NEON builds:

  • ACTION: Michael to point Ramana at the logs
  • The gfortran ones may be fixed by Matthew's recent patch to binutils
  • ACTION: Ramana to check into the patch and recommend on the upgrade
  • Richard notes that 2.20 may not support LTO and changing to 2.21 could have exposed problems

Strided load-store patches:

  • Waiting on review
  • Builds are looking good
  • Richard E: can you fetch past the end of memory?
    • Richard S: no, as the middle end knows the size of the objects and lengths
    • Does the vectoriser some times read one past the end of the array?

https://code.launchpad.net/~rsandifo/gcc-linaro/lp-758082/+merge/57828:

  • ACTION: Andrew to check inside CSL to see what Bernd has done
  • Want the same as what will go upstream
  • State of shrink wrap and upstreaming?
    • Is going upstream
    • A fair amount has been reviewed, but not the core shrink wrap part
    • About a month old

Planning:

  • TRs are at Cycles/1111/TechnicalTopics/Toolchain

  • Next step is to create the TR blueprints
  • and start creating engineering blueprints
  • Finish by the end of next week
  • Have done more investigation ahead of time this cycle
  • Blueprints should be reasonably well fleshed out
  • If not, have sufficient work items that can be re-purposed once the investigation is done

QEMU release:

  • Skipping this month as there are no changes
  • Will note this on the toolchain wiki page
  • ACTION: Peter to try today's hwpack and rootfs on vexpress and OMAP3

Ulrich noted that his holidays are coming up. GDB release is under way.

WorkingGroups/ToolChain/Meetings/Archive/2011-05-16 (last modified 2013-08-30 11:49:06)