Thursday 1st December 2011

This month's meetings

<< <  2012 / 1 >  >>
Mon Tue Wed Thu Fri Sat Sun
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          



  • Stand up call



  • Ran some timer tests on the Panda for inconsistent timing
  • membase isn't due to that
  • Updating mechr() to go into GLIBC
  • Reading into the specs of ucontext, seem reasonable enough
  • Read through the v8 instruction documentation


  • Building OpenEmbedded Core with the Linaro toolchain

  • Got an older version working with the CSL one
  • Has bitrotted, is old, are complaints on the mailing list


  • Working on 64 bit shift and 32 to 64 bit extends without doing this in the core registers
  • Next is testing both they apply in interesting cases and correct
  • Will do the FSF merges today


  • Working on register pressure estimation in SMS based on feedback from upstream
  • Sent a merge request, will resend so it can be used for testing past that


  • Looking into peeling and alignment issues
  • Fixed a bug in over promotion of shifts, might be the same as Michael was looking at
  • ACTION: Michael to forward that to Ira to check


  • Running more and more benchmarks across different toolchains
  • Learning as she goes, such as with/without vectoriser not changing
  • Working on how the report should look
  • Will share with rest of the group
  • ACTION: Michael to put on Monday's agenda


  • Getting KVM QEMU working on a fast model (fiddle, document, repeat)
  • Now have a wiki page that shows most of what needs to be done
  • Can show A9 guest in QEMU on A15 host on fast model :)

  • Has imported the KVM patchset to rebase
  • Found out that it doesn't support Thumb-2 or VFP at the moment...


  • Used Installgen to create Windows installer packages an experiment
  • Backporting the upstream patches into the Linaro branch
  • Next step is how we test under Windows


  • Checking the 7.4 branch which is coming up
  • Found some problems on S390 especially on remote debugging
  • ...which is worthwhile towards what we're doing
  • Remote core file generation patch set is finished
    • exporting the whole of /proc over the protocol
    • and removes the need for any other extensions
    • Hope to post later today


  • Looking at register pressure so far
  • Found a fault so far, results look reasonable


  • Binary toolchains
  • Mucking about with alignment
  • Getting ready for presentation

WorkingGroups/ToolChain/Meetings/Archive/2011-12-01 (last modified 2013-08-30 11:48:31)