== This week ==
* TCWG-619 (8/10)
PR65858
- workaround: add explicit casts in the file cld_generated_cjk_uni_prop_80.cc
- created reduced test case: http://pastebin.com/WJvKRjn2
- commit r222249 introduced the issue
- patch submitted upstream:
https://gcc.gnu.org/ml/gcc-patches/2015-04/msg02063.html
- Paolo Carlini pointed out it doesn't work with -w option, instead of testing
on warn_narrowing flag, he modified it to check on pedwarn()'s return
value, which
covers up for both -Wno-narrowing and -w, and committed with
this modification (r222700).
PR65837
- patch to lto-wrapper rejected.
- Another hack to pass driver's argv to lto-wrapper via another
environment variable
COLLECT_GCC_OPTIONS_DRIVER rejected by Richard.
- Ramana and Richard pointed out the proper fix would be to handle
target attributes in ARM backend.
- I could try extending Christian Bruel's target attribute patches for
fpu options
after they're checked in which would then fix the PR.
chromium LTO build
- Many instances of same error message during LINK chrome:
chrome.ltrans0.s:18399164: Error: Thumb2 branch out of range
chrome.ltrans0.s:9671243: Error: branch out of range
works fine for non-lto.
A similar issue, occurred for v8 LTO build - PR65778, but
it was marked resolved invalid.
- chromium builds with LTO for chrome component.
- ld.bfd segfault while building chromium with LTO not reproducible with master.
- non-LTO build undefined reference to libattr1 functions not reproducible with
ld.bfd master. This probably was sysroot/libc issue than issue with ld.bfd.
* PR49551 (2/10)
- reverting r221297 reproduces the ICE
- tried to fix with following patch:
http://pastebin.com/knuWeY0C
- fixes the ICE but following regressions observed on arm-linux-gnueabihf:
http://pastebin.com/E1zLtq8s
== Issues ==
- Can't login to cards.linaro.org (was able to before).
I get error - "You do not have permission to login".
== Next Week ==
- Investigate testsuite failures for PR49551
- Selectively enable LTO on components in chromium.
- exams preparation
* Half day leave on Thu, [1/10].
# Progress #
* arm/aarch64 gdb bug fixing. [3/10]
** TCWG-765, fails in gdb.base/coredump-filter.exp. No progress, as no
one reviews the patch.
** TCWG-672, fails in gdb.base/relativedebug.exp, fixed.
** TCWG-567, fails in gdb.base/break-idempotent.exp, fixed.
** TCWG-767, Propose to disable target-side evaluation of breakpoint
condition for software single step targets. Patch is posted.
* TCWG-805, aarch64 gdb multi-arch support, [2/10]
* TCWG-729, investigate GDB multi-target support, [2/10]
* TCWG-757, upstream patch review, [1/10]
* Misc [1/10]
** Meeting,
** Fill online forms for France visa application,
** Book travel to Grenoble.
# Plan #
* Public holiday on Monday.
* TCWG-805, aarch64 gdb multi-arch support,
* TCWG-729, investigate GDB multi-target support,
* other arm/aarch64 gdb test fails fixing.
--
Yao
benchmark automation - TCWG-360 [3/10]
* Shifted more generic patches from backport branch to review
* One complete run of backport benchmarking before lab went down
Misc [7/10]
* Most of this (4 or 5 tenths) is reaction to lab downtime
=Plan=
Benchmarking test suite
Fish out, finish off thoughts about benchmark access control
NB Next Monday (4th May) is a public holiday in UK
== Progress ==
LLDB development
-- Debugging test code for hardware breakpoint/watchpoint issues on
chromebook for ARM LLDB support [4/10] [TCWG-770]
-- Wrote native process linux helpers for arm and aarch64 hardware
breakpoints [2/10] [TCWG-770] [TCWG-771]
-- Fix, test, debug, cleanup and patch re-submission. [3/10]
-- LLDB ARM SysV ABI implementation. [TCWG-643]
-- LLDB AArch64 SysV ABI implementation. [TCWG-715]
-- Native process linux bug fixes [TCWG-651]
Miscellaneous [1/10]
-- Upstream reviews
-- Meetings, emails, discussions etc.
== Plan ==
LLDB development
-- At least submit lldb arm hardware breakpoint native process linux
helper functions
-- Resubmit register read/write bug fix after changes
-- Resubmit LLDB ARM SysV ABI implementation.
-- Resubmit LLDB AArch64 SysV ABI implementation.
-- Bug fixing, patch reviews etc
== This Week ==
* TCWG-619:
* abe sysroot/libc issue
While building chromium with development build of toolchain on
arm-linux-gnueabihf, I get the following error:
undefined reference to clock_gettime()
Builds fine with release toolchain build.
Symlinking libc in toolchain install directory to
sysroots/arm-linux-gnueabihf appears to work (which is what --tarbin
did).
* PR65778 - v8 fails to build with LTO
- Resolved Invalid
* PR65837 - target specific builtin not available
- created a reduced test case
- There's a weird issue: looks like for lto1, translation unit command
line options,
precede the default ones rather than overriding them.
Assume test.c contains a neon intrinsic:
arm-linux-gnueabihf-gcc -flto -c test.c -mfpu=neon
arm-linux-gnueabihf-gcc -flto test.o
results in target specific builtin not available,
because both -mfpu=neon and -mfpu=vfpv3-d16 is passed
and -mfpu=vfpv3-d16 overrides earlier -mfpu=neon
Passing -mfpu=neon at link time correctly overrides -mfpu=vfpv3-d16.
* PR6778 - ICE in varpool_node::get_constructor.
- Looks like PR65776 appears due to same issue as well.
- It appears error_mark_node gets streamed that causes ICE.
- Triage went wrong, thought it was an error in typeck2.c:check_narrowing().
* PR49551
- The C front-end c/c-decl.c:merge_decls still incorrectly marks the
variable as 'common' for the test-case:
int x = 5;
int x;
However it's not reproducible anymore on trunk (maybe latent?)
- I have an untested patch that resolves the issue in merge_decls().
* Issues
- abe built native gcc on x8_64-unknown-linux-gnu fails to build code
when passed -m32 option (bug 1508).
* Misc
- exams on 20th and 24th april.
* Next week
- Continue working on PR65837, PR65778
- Test patch for PR49551
== Issue ==
* none
== Progress ==
o Upstream GCC (8/10)
* [TCWG-785] ARM backend insn cleanup
- Validate and re-worked the patch
- About to be submitted upstream
* [TCWG-762] - GCC Maintenance
- PR64208 (iWMMXT LRA bug):
Found hardware to validate the fix (Thanks Riku :)
Validation ongoing
* Lot of time wasted due to lab compromisation
o Misc (2/10)
* Various meetings
* GCC git repository branches cleanup
== Plan ==
- Continue cleaning GCC git repo to prepare our migration
- Continue upstream pending work
== Progress ==
* Upstream GCC (9/10)
- TCWG-780 - Improve register allocation for
aarch64_*_sisd_or_int<mode>3 patterns.
- TCWG-486 - Optimize Constant Uses in Loops. Posted the arm part of
the patch that addresses review comments after full testing. Waiting for
the cprop review.
- TCWG-796 Zero/sign extension elimination with vrp. Posted the new set
of patches for review. Looking at the testcases H.J.L. pointed me to.
- TCWG-779 Vector rtx costs for AArch64. Posted the patch that
addresses review comments after full testing.
* Misc (1/10)
- gcc-patches, gcc-bugs list
- Meetings
== Plan ==
- Continue with gcc stage1 activities
- Monday public holiday
# Progress #
* arm/aarch64 gdb bug fixing. [5/10]
** PR 18208 (TCWG-765) fails in gdb.base/coredump-filter.exp. Workaround
tricky quote issue with dejagnu/tcl/ssh/bash. A RFC is posted out.
** Fails in gdb.base/relativedebug.exp (TCWG-672), patch is posted out
review.
** Fails in gdb.base/break-idempotent.exp. Skip some tests in it which
require more than one HW watchpoint registers, while arm-linux only
has one. Patch is posted out.
** TCWG-793, do experiment with showing LSB of address in thumb mode inside GDB,
in order to align with MIPS(microMIPS/MIPS16). It needs more work
than expected, and even out the scope of GDB. It is paused.
* FSF GDB patch review. [4/10]
** Review "all stop on top of non stop" V3, done.
** Review "memory-mapped register browsing" patch. Requested more
information from the author.
* Misc, meeting. [1/10] # Plan # * Fix test fails for arm and aarch64 gdb.
* Continue my HW watchpoint resource counting patches once linaro boxes
are online. These boxes don't have HW watchpoint support, and GDB
doesn't handle this case well.
* Upstream patches review.
--
Yao Qi
