[VIRT-249 # SVE system mode emulation ]
Fixed a bug in last week's reorg, reported by Laurent.
Fixed a bug in our risu scripts for generating sve memory ops;
I was able to replicate Laurent's failure in risu, finally, and
then fix it. Still need to clean up the risu patch for submission.
[Upstream]
Reviewed mips mxu tcg patch set.
Reviewed s390 afp tcg patch set.
r~
== This Week ==
* TCWG-1234: Hoisting and register pressure (4/10)
- Came up with another heuristic to avoid spill
- Verified it does not regress other test-cases (testsuite + private benchmarks)
* TCWG-319: Use vrecpe / vrecps to implement division (1/10)
- Rebased patch and validated it
- Seems like the issue of bad data alignment is also now seen with
armhf. Investigating
it further.
* PR83661 (2/10)
- Finished with patch prototype
* Public Holiday (2/10)
* Misc (1/10)
- Committed a typo fix in r263819
- Meetings
== Next Week ==
- Continue ongoing tasks
o LLVM
* Machine Outliner on ARM prototype:
- stack fixup support on-going
- test coverage in machine IR
* Buildbots babysitting
o Misc
* Jira gardening
* Various meetings and discussions.
[VIRT-198 # QEMU: SVE Emulation Support ]
Adjusted patch set based on Peter's review and upstreaming.
We talked about a method to initialize cpu config registers.
[Upstream]
Review:
nanomips round 11, now mostly pushed to master.
MPS2 patch set
raspi frame buffer patch set
Partway through a round 6 of the do_syscall split.
r~
Investigated yet another sanitizer related buildbot failure. This time
it looks like latent problems in the compiler-rt/cfi implementation
and tests provoked by the introduction of the Arm LLD bot.
- cfi requires LTO so it requires a bot using LLD or a bot configured
to run gold with the LTO plugin. Raised upstream PR on cfi.
[Linaro Connect]
Made some more progress on a clang-tool example to use for my
presentation on how to build a clang tool. Learned quite a bit more
about ast-matchers and the clang-ast.
[LLD]
Committed support for Armv5 and Armv6 in LLD.
[Miscellaneous]
Quite a heavy patch review week.
Planned absences:
On holiday on Tuesday
Progress:
* VIRT-65 [QEMU upstream maintainership]
+ put some reviewed MPS device patches into master, did a v2 of the
rest and sent them out again; these are now in master
+ ditto with patches for missing AArch32-Hyp features
+ code review:
- rth's patchset fixing float/int-with-scaling conversion
- rth's patchset adding system mode support for SVE
- linux-user: netlink fixes
+ sent out a patchset that wires up the GICv2 virtualization
memory regions and IRQs on the remaining boards that need it;
got this into master
thanks
-- PMM
Note: annual leave tomorrow and Friday, as per my calendar
=== Work done during 3-day week ===
* TCWG-1379 / GCC PR85434 / CVE-2018-12886: rework needed
+ fix 2 more issues with retesting each time
* More knowledge transfer within Arm (writing a script to automate manual
task in GNU Arm Embedded Toolchain release to allow others to do it)
* TCWG-1428 / DSGHACK-25 (Support arithmetic on FileCheck regex variable):
address comments
+ continue to rework patch
* Enable weekly regression testing of LLVM release process jenkins job
* 2 days off annual leave
=== Plan for week 35 ===
* Enable regression testing of Linaro LLVM release process when scripts it
use change
* TCWG-1428 / DSGHACK-25 (Support arithmetic on FileCheck regex variable):
address comments
+ continue to rework patch
* TCWG-1379 / GCC PR85434 / CVE-2018-12886: rework needed
+ submit for external review
* Try to reproduce perf issue mentioned in week #30s weekly report on
latest perf
[VIRT-198 # QEMU: SVE Emulation Support ]
Fixed a double-rounding bug affecting scaled conversion
from integer to float.
[Upstream]
Posted a proposed fix for 128-bit atomic detection.
Posted v4 of linux-user syscall split.
Review of nanomips, round 9.
r~
* One day off
o GNU
* Committed Linaro version string and macros patch in ARM GCC 8 vendor branch
o LLVM
* libcxxabi exception address alignment test failure:
- Committed in trunk and merged into 7.0 branch
* Machine Outliner on ARM prototype:
- LR saving and stack fixup support on-going
* Buildbots babysitting
o Misc
* Jira gardening
* Various meetings and discussions.
Note: I'll be on holidays Thursday and Friday next week.
=== Work done during 3-day week ===
* TCWG-1379 / GCC PR85434 / CVE-2018-12886: rework needed
+ heavy testing
* TCWG-1337 / LLVM PR34170: get backport to LLVM 7 branch
* create design review for -mfpu=auto option in GAS
-> https://confluence.arm.com/display/CTP/Auto+FPU+for+GNU+as
* Knowledge transfer within Arm
* Misc:
+ GCC Linaro sync
* 2 days off taking care of my feverish child
=== Plan for week 34 ===
* Enable regression testing of Linaro release process
* TCWG-1428 (Support arithmetic on FileCheck regex variable): address
comments
+ continue to rework patch
* TCWG-1379 / GCC PR85434 / CVE-2018-12886: rework needed
+ investigate aarch64 glibc regression
* Try to reproduce perf issue mentioned in week #30s weekly report on
latest perf
