=Progress=
memcpy regression on A9 - REQ-363 [3/10]
* Verified regression on a couple of targets, some investigation around cause
lowlevellock performance bugs - TCWG-435 [3/10]
* Upstreamed, some reworking based on review, obsessed about odd code
change & intermittent test failure
memset improvement - TCWG-156 [1/10]
* Started looking at glibc benchtests as a quicker way of doing
1st-round benchmarking
Meetings/mail/etc [3/10]
=Plan=
More of the above -
* Get some memset benchmarks running
* Carry on with lowlevellock patches (breaks m68k build, apparently)
* More memcpy investigation
== Progress ==
* GDB arm v8 record/replay
-- Investigated prologue analysis and stepping issues. [3/10] [TCWG-484]
-- Updates to JIRA cards.
-- Updates to patches stalled till next week.
* GDB arm v7 record/replay
-- Prepare SIMD instruction recording support patches for re
submission. [1/10]
-- Chromebook issues stalls Testing and submission till next week.
* GDB on Android improvements [3/10]
-- Discussions with community developers
-- Review of already submitted patches for catch syscall and exec events.
-- Making things work with available upstream patches.
* Miscellaneous [3/10]
-- Compassionate leave on Monday 9th June 2013.
-- Submit application for travel health certificate for UK travel in July.
-- Confirm travel and hotel booking after UK visa approval.
== Plan ==
* GDB arm v8 record/replay
-- Update all patches after upstream comments.
-- Investigate all remaining testsuite failures.
-- Update all related jira cards and create new for persisting test failures.
* GDB arm v7 record/replay
-- Update and resubmit upstream.
* GDB Remote vs Native feature parity (GDB on Android improvements) [CARD-321]
-- Prioritize todo list in light of currently active upstream development.
== This week ==
Out sick Tuesday and Wednesday (4/10)
Launchpad cleanup (6/10)
Launchpad 944572 - Infinite loop printing "warning: missing braces
around initializer" in g++
- fix released
Bugs marked "won't fix" after investigation:
Launchpad 683683 - run-init on omap3, omap4 in natty dies if busybox
is built with -marm
Launchpad 757427 - gconftool-2 segfaults on arm [shrinkwrap]
Launchpad 768384 - G++ gets stuck (100% cpu never finish) when
compiling C++ code with -O2/3
Launchpad 771675 - shrink-wrap optimization produce buggy code
Launchpad 625228 - Avoid unnecessarily saving $lr on Thumb-1
Launchpad 771832 - bogus reference to unused symbols
Launchpad 788841 - GCC ICE: internal compiler error: in push_minipool_fix
Launchpad 795131 - [armel, vfp] ICE in simplify_gen_subreg_concatn,
at lower-subreg.c:468
Launchpad 803798 - FTBFS on armel: error: unable to find a register
to spill in class 'VFP_LO_REGS'
Launchpad 807156 - GCC (g++):internal compiler error: in
extract_insn, at recog.c:2104
Launchpad 821455 - internal compiler error: in translate_clast, at
graphite-clast-to-gimple.c:1123
Launchpad 839959 - 4.5-2011.08 internal compiler error
Launchpad 841437 - fftw3 version 3.2.2-1 failed to build on armel
Launchpad 887158 - NULL pointer deref in mark_reg_pointer
Launchpad 890657 - Please merge the fix for upstream bug 45102
Launchpad 892880 - MIPS: link error against __unpack_d with
--enable-target-optspace
Launchpad 903178 - [4.6 regression/armhf] ICE, unable to find a
register to spill in class 'LO_REGS'
Launchpad 903347 - [armel/armhf] file fails to compile (indefinite loop?)
Launchpad 903951 - The kernel on staging-panda does not boot with the
11.12 released toolchain
Launchpad 937864 - Produces NEON code on armhf even when explicitly
asked not to
Launchpad 941676 - ppl ftbfs in precise on powerpc
Launchpad 943264 - ICE in arm_select_dominance_cc_mode, at
config/arm/arm.c:10625
Launchpad 1049614 - Assembler messages with -O0: Error: offset out of
range
== Next week ==
On vacation June 16th - June 20th
== Future ==
No Plans
= Progress ==
* Went to UK VISA office and submitted application and bio metrics (2/10)
* Tried Installing chrubuntu on chromebook. But failed and later found that
SD card issue. Given SD card back in office for replacement. Installed
crouton and other packages for building gcc and testing gcc testsuites
(1/10)
* Reload spill failure (bug fix) on armhf target (3/10)
Tested conservative patch that prevent ior_scc_scc patterns for thumb2
target using chrome book.
Thumb2 compare condition test cases fails. Found that these test cases
using constants for operands 2 and 5. Change conservative patch to allow
constant operands for thumb2 target in ior_scc_scc patterns. Testing is in
progress.
* LTO experiments (2/10)
Drilling down ICE with LTO on GCC 4.9 branch.
* Misc (1/10)
- 1-1 meetings (Ryan, Christophe) (0.5/10)
- AMD internal support work and meetings (0.5/10)
* Installed some packages and tested christophe libsantizer patch on
hardware (1/10)
== Plan ==
* Continue bug fixing.
* LTO bootstrap failure
== Progress ==
* Public holiday (2/10)
* Benchmarking and regression testing (TCWG-468) (8/10)
- Ran release benchmarking in chrome-book
- Set-up package build environment and ran ubutest
- spec2k benchmarking of patches that was included in 2014.06 release
on aarch64
== Plan ==
* Benchmarking.
* Upstream zero/sign extension elimination activities.
== Issues ==
* None.
== Progress ==
* Work out 4 patches to enhance loop2-invariant heuristics and send
out for community review (TCWG-469, 3/10).
* Test the rebased ccmp patches (TCWG-488, 2/10).
* Investigate how to optimize large constant. Patch is in testing
(TCWG-486, 5/10). Basic idea is:
- Do not split large constant when expanding.
- Improve cprop pass to check the rtx_cost when propagating constants.
== Plans ==
* Ping pending patches.
* Send out ccmp patches for review.
* Send out patches to optimize constant.
== Progress ==
* Kernel (CARD-1246 1/10)
- Helping LLVMLinux to test in LAVA
* Background (9/10)
- Code review, meetings, discussions, etc.
- Setting up Chromebooks again (up, crouton)
- Setting up APM boards (up)
- Setting up D01 boards (bricked 2)
== Plan ==
* work with guodong, fabo, tyler, to revive the D01s
* put crouton on llvm-chrome-test to be ready for release 3.5
* have a look at compiler-rt autoconf build
* keep helping LLVMLinux with LAVA
