== Progress ==
* Zero/sign extension elimination (TCWG-15) (2/10)
- Posted patch for comment
* benchmarking (TCWG-468) (1/10)
- Ran a53 benchmarks
* regressions (7/10)
- THUMB1 regression for ARM fenv
* Issue due to thumb1 not supporting mrc/mcr. Patch to fix this is
posted for review.
- Regression when allocating 128bit integer to VFP register
* When LRA assigns DImode value to TImode register, it is not
setting up it in the right place of TImode. Due to this, one of the
moves becomes dead. Patterns needs to be checked.
* VFP registers store big-endian values in little-endian format.
Hence, subreg for mode greater than word has to
be aware of this. As it is, aarch64_cannot_change_mode_class will need
the fix like done in ARM.
== Plan ==
* Benchmarking.
* Upstream zero/sign extension elimination activities.
= Progress ==
* LTO experiments (3/10)
Tested with various compiler revisions and in native X86 machines. Compare
errors in gcc 4.9 Linaro 14.05 version. Aarch64, LTO bootstrap passes in
trunk. On FSF GCC 4.9 branch encountering ICE. Opened PRs in GCC Bugzilla.
* Misc (2/10)
- 1-1 meetings (Ryan, Christophe and Maxim) (0.5/10)
- AMD internal support work and meetings (0.5/10)
Sick Leave on 2nd June (2/10)
UK VISA application processing (3/10)
== Plan ==
* UK VISA processing attend interview, biometrics.
* Continue bug fixing.
* LTO bootstrap failure
* Test libsanitizer patch for christophe.
=Progress=
lowlevellock performance bugs - TCWG-435 [1/10]
* All ready to go
cbuild benchmarking - TCWG-360 [4/10]
* Completed a draft implementation for spec2k
* Parked pending review
Meetings/mail/etc [5/10]
=Plan=
Send lowlevellock patches to list
Get back to benchmarking/improving cortex-strings memset
== This week ==
Backported Launchpad 1234146 - Bad Neon intrinsics code gen when using
ld4/st4 on AArch64 (2/10)
Bug Investigations (4/10)
Launchpad Bug 640521 - Incorrect function prologue with Thumb-1
high register variable
Launchpad Bug 1318831 - Invalid unpoisoning of stack redzones on ARM
Launchpad bug 771832 - bogus reference to unused symbols
Launchpad cleanup (4/10)
- 1305042 - Changed status to invalid
- 634731 - Marked as won't fix for older releases. Improved code
generated for 4.8 and 4.9
- 637882 - Marked as won't fix for older releases. Improved code
generated for 4.8 and 4.9
- 638687 - Changed status to invalid. Code for alloca is correct
and alignment is necessary
- 1312931 - Could not verify.
- 772085 - Marked as won't fix for older releases. Fixed in 4.8 and 4.9
- 1225317 - Marked as incomplete due to lack of test case
== Next week ==
Vacation
== Future ==
No Plans.
== Issues ==
* None
== Progress ==
* CARD-1162 : Linaro GCC 4.9 and CARD-1355 : stabilization and
optimization effort for ARMv8-a (8/10)
- Put the backporting script on https://git.linaro.org/toolchain/backflip.git
(documentation still needed)
- FSF branch merges for 4.9 and 4.8
- Backported requested fixes in 4.9
- Validate merges and some backports in cbuildv1
- Looked at the different results we have in cbuildv2 and cbuildv1
- Still some issues in Jenkins (build crash or are stuck)
- 40 backports stiil need validation
* Misc: (2/10)
o Various meetings.
== Next ==
* Complete FSF branch merges and release 4.9-2014.06
* Document backporting script usage
* Continue feedback and help with the validation
== Issues ==
* None.
== Progress ==
* One day off.
* Clean up and test loop2-invariant heuristic tune patches (TCWG-469, 6/10).
* Rebase conditional compare patches and split them into smaller ones
(TCWG-488, 2/10).
== Plans ==
* Send out loop2-invariant heuristic tuning patches for review.
* Rework on the conditional compare patches.
== Progress ==
* Toolchain (CARD-862 0/10)
- Some investigations on LLVM Linkers
* Background (10/10)
- Code review, meetings, discussions, etc.
- Cleaned up the rack, drown a map
- Jenkins Chromebooks with native Arch Linux
- Failures and problems ensued...
- More time wasted on APMs, but some progress was made
- All Chromebooks had I/O errors on SD cards, starting over
== Plan ==
Spend the weekend and the next week re-flashing Chromebooks
== Progress ==
* GCC trunk cross-validation (CARD-647) (6/10)
- recent LRA change to fix AArch64 compiler causes huge memory consumption.
- this has caused my validations to crash several servers, and I had
to stop them
until I could restart reliably.
- so far unable to restore them
- trying to test LRA tentative fix separately, but results are unreliable too
- tested Thomas Preud'homme patch (bswap-2.c), luckily before
everything became unstable.
* GCC 4.9 branch cross-validation (CARD-647) (1/10)
- validation scripts needed some adaptations to handle the 4.9-branch
- not running yet, until the trunk problems are solved
* Neon intrinsic tests (CARD-????) (1/10)
- submitted patch series containing 40 test files (total will be 136)
* AArch64 libsanitizer support (TCGW-58)
- qemu still needs a patch of mine which I thought had been
committed since February..
- once patched, the sanitizer tests now start executing but lead to
SEGV+timeout
* Misc (meetings, conf-calls, ...) (2/10)
== Next ==
* short week (Monday off, lots of external constraints)
* GCC cross-validation:
- hopefully fix unstability and restore the automatic validations of
trunk + start those of 4.9 branch
- try to actually start mail-driven robot to validate patches
* Neon intrinsics tests: handle feedback
* AArch64 libsanitizer support: look at why tests fail
