Summary:
* Send out shrink-wrap related patches for review.
* Investigate how to add a new tree code.
Details:
1. Fix lp:1157050 and lp:1107659 for Linaro toolchain binaries.
2. Test shrink-wrap dwarf/unwind info and send out the patches for review.
3. Make progress on conditional compare support. With some hard-codes
in gcc, it can gimplify a small case to conditional compare on tree
and no assert until expand.
Plan:
* Investigate how to expand conditional compare GIMPLE to RTL and emit asm.
Planned leaves:
* March 31.
* April 4-8.
Best Regards!
-Zhenqiang
Hi all,
I'd like to run the binutils testsuite in a native configuration
regularly to find any regressions. Do we have any infrastructure for
doing this?
It looks like LAVA would be useful but it looks complex and I am not
sure if we use that already for running toolchain tests (I think there
was a Connect session about that?).
Thanks,
--
Will Newton
Toolchain Working Group, Linaro
== Progress ==
* Disable-peeling:
- benchmarking locally on snowball
- failed to build reference branch on cbuild, for an unknown reason
* Turnoff 64bits ops in Neon:
- committed upstream in 4.9 (now stage 1)
- benches of backport into Linaro-4.7 OK: no significant diff,
except +12.6% on spec2k's crafty)
- Filed merge request.
* Vectorizer cost model:
- benches of backport inot Linaro-4.7 OK (no significant diff, as expected).
- Filed merge request.
* Revert-coalesce:
- benches ran in cbuild, but no *-diff.txt file generated.
* Libsanitizer:
- testsuite results OK. All unsupported tests are normal.
* Neon intrinsics codegen:
- got internal feedback
* GDB reverse debug:
- filed 2 bug reports after internal requests.
* Misc:
- filed several bug reports against cbuild to track the problems I observed.
== Next ==
* Disable-peeling:
- analyze bench results
- understand/'fix what's wrong with cbuild
* Turnoff 64bits ops in Neon:
- backport to Linaro-4.8
* Revert-coalesce:
- manually compute bench differences if unable to have cbuild do it.
* Libsanitizer:
- propose patch upstream.
== Progress ==
Knocked out more GDB remote configuration failures that were failing due to
timeouts.
Had 1:1 with Matt for direction on GDB improvement plan.
Looked around for Ireland visa process and filled out visa application.
== Plan ==
Complete review of remote timeout based test case failures.
Start work on the GDB improvement blueprint in JIRA from this week.
Get remaining documentation for Ireland visa and complete application.
== Progress ==
- Looking at TLS relocs and coming up with test cases for garbage
collecting them.
- Patch for missedout testcases in gc section is up streamed. Thanks to Marcus.
== Plan ==
- Complete gc section test cases for TLS relocs.
- Understand and evaluate gprof support work for Aarch64
Misc
-----
Leave on Wednesday and Thursday (27th and 28th)
== Progress ==
- Worked on VRP based zero/sign extension elimination at tree level
- Some regression test cases are failing; investigating them
== Plan ==
- Analyse the reason for test case failure and fix them
- Extend the code to handle all the possible cases (currently
implemented for subset to get the framework working)
== Issues ==
* None
== Progress ==
* Libunwind AArch64 support:
- basic implementation for building lib and testsuite done
- but even a simple test case doesn't work for the moment
== Plan ==
* Libunwind AArch64 support:
- continue AArch64 specific implementation and informtation gathering
== Progress ==
* EuroLLVM paper review
* Disabled EH tests on ARM for now
* Fixed TSVC: NEON vs. VFP VMUL.f32 (and IEEE 754 correctness)
* Changed FeatureNEONForFP to false on EABI ARM targets unless unsafe-math
* Trying Arndale again, with u-boot patched, not good
* Broken Panda buildbots, trying to fix
== Issues ==
Got ill on Tuesday and only worked partially, if at all, the rest of the
week
Hopefully will be better next week...
== Plan ==
* Fix our panda buildbot, maybe add more
* Final try on Arndale GCC bootstrap before giving up
* Continue fixing test-suite failures (only 5 left)
Progress:
* VIRT-4 [Guest migration support for KVM]
** VIRT-51: vexpress migration:
** wrote and submitted patches for all the devices that need fixes
** tracked down a painful bug where the guest just hung on vmload
to a missing state-save for a timer device, submitted patch
** remaining work for this item is just responding to patch review
-- PMM
