# Issues #
* Don't have arm board to run native gdb testsuite (my pandaboard is
too slow to run native tests). Get a chromebook, but takes a lot
of time installing ubuntu on it. Network still doesn't work, don't
know how much time are still needed to fix networking issues.
# Progress #
* Linaro GDB
** TCWG-517, GDB AAarch64 record/replay. [3/10]
These bits are finally posted upstream. The review looks good.
** TCWG-805, aarch64 native debugging multi-arch support. [4/10]
*** Switch to PTRACE_GETREGSET and PTRACE_SETREGSET. Patches are done,
but takes much time to set up GDB native testing env, as linaro
tcwglab is still unreachable.
Install ubuntun 14.02 on chrombook. Done, but network driver
doesn't work.
** Give my input on GDB kernel awareness.
** Tweak pandaboard for native gdb testing. Done.
** AArch64 debugging on different ELs, a meeting is scheduled
in the sprint but I ask more details before that meeting.
* FSF GDB [1/10]
** Review a patch about arm abi detection for ELFOSABI_GNU. Approve it.
** Review a patch about compile feature in GDB.
* Misc [2/10]
** Go to London for France Visa application on Monday and get visa on
Thu.
# Plan #
* TCWG-517, GDB AAarch64 record/replay.
Address review comments and push them in.
* TCWG-805, aarch64 native debugging multi-arch support.
Set up chromebook and test my patches on it.
--
Yao
Holiday Thursday afternoon and Friday [3/10]
Benchmark automation - TCWG-360 [4/10]
* CPU2006 patches cleaned up, submitted
* A few other small fixes
Misc [3/10]
* ~1/10 getting a Juno running
=Plan=
Holiday next week, then:
Pick a direction on streaming vs listeners, write down the reasoning
Follow up (close out?) results storage and restricted targets
Write more tests, refactoring where appropriate
Investigate running benchmarks through ci.linaro.org
If time, and it looks sane, wrap a LAVA job around benchmark scripts
== Progress ==
LLDB development
-- Submitted and committed AArch64 hardware watchpoint/breakpoint
support [2/10] [TCWG-771]
-- Submitted Arm hardware watchpoint/breakpoint support [1/10] [TCWG-770]
-- Debugged arm watchpoints on Nexus 7 Android device. [2/10] [TCWG-651]
Miscellaneous [5/10]
-- Submitted AArch64 gdb record-replay patches
-- Travel to Islamabad for interview at French Embassy
-- Meetings, emails, discussions etc.
== Plan ==
LLDB development
-- Fix issues with hardware watchpoints and try to run tests on Nexus 7
-- Update and resubmit Arm hardware watchpoint patch.
== This week ==
* TCWG-619 (2/10)
- Disabling seccomp-bpf, works with LTO build for chromium for
multiple partitions (-flto-partitions={balanced, 1to1})
- disabling fPIC does not prevent the error - "branch out of range"
for single or no partition (--param lto-partitions=1 / -flto-partitions=none )
- The docs say there is option to disable PLT (-fno-plt),
https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html#Code-Gen-Options
but using it results in unrecognized command line option.
* PR49551 (2/10)
- Found out why r221297 prevents the ICE:
merge_decl() calls make_decl_rtl() upon condtion DECL_RTL_SET_P (olddecl),
which causes the ICE. Before r221297, in merge_decls() DECL_RTL_SET_P
(olddecl) was true which resulted in call to make_decl_rtl() causing
ICE. After r221297, DECL_RTL_SET_P (olddecl) is 0, which doesn't
result in call to make_decl_rtl() so no ICE is produced. However,
DECL_COMMON is still incorrectly set in merge_decls() because it
copies newdecl into olddecl.
- Will post upstream for review after final testing on x86 and ARM.
* Misc (6/10)
- Exams
Short week (2 days off)
== Progress ==
o Upstream GCC (3/6)
* Flags for Cortex-a53 erratum #84341 workaround:
- Backported into FSF 4.9 and 5 branches
* Some cleanup to avoid format-security warnings
* TCWG-786 (LRA and PIC): Look for potential improvements
- nothing found so far.
o Linaro GCC release (2/6)
* Reviewed FSF branch merge into 4.8 branch
* Look at differences between our actual source tarball, and
the ones that can be created by ABE.
o Misc (1/6)
* Various meetings
== Plan ==
- Continue ongoing tasks
Benchmark automation - TCWG-360 [7/10]
* Convinced benchmark scripts to run entirely within LAVA
* Did a couple of benchmark runs for Kugan
* Cleaned up my CPU2000 patches and committed to central repository
Misc [3/10]
=Plan=
Clean up, commit CPU2006 patches
Pick a direction on streaming vs listeners, write down the reasoning
Follow up (close out?) results storage and restricted targets
Write more tests, refactoring where appropriate
If time, and it looks sane, wrap a LAVA job around benchmark scripts
== Progress ==
* Upstream GCC (5/10)
- committed delaying constant splitting in arm after full testing
(TCWG-486)
- Spec benchmarking is still an issue (TCWG-779)
- Bernie agreed to kindly help benchmark it
- taking time due to long lava queue
- Started working on widening pass based on review comments (TCWG-555)
- Wilco from ARM provided some test-cases and looking into it as well
* GCC bootstrap/build issues (4/10)
- Looked at bare-metal
- aarch-elf-none issue
- gcc bootstrap issue with ARM
- middle-end error
- Thanks to all, all now fixed in mainline.
* Misc (1/10)
- gcc-patches, gcc-bugs list
- Meetings
== Plan ==
- Continue with gcc stage1 activities
- Look at register allocation
