== This Week ==
* GNU-659 (calculix LTO regression)
- Gathered perf stats for various opt levels and posted on ML
* LLVM-611 (tune blx lowering heuristic)
- Dave Green committed my patch which had to be reverted due
to build-bot failure with expensive checks turned on.
- Posted analysis of the issue upstream
* Misc
- Committed patch to add --strict option to bmk-scripts/spec2xxx-config.
== Next Week ==
- Continue ongoing tasks
Progress: (short week, three days)
* VIRT-65 [QEMU upstream maintainership]
+ sent patches to remove QEMU's support for KVM on AArch32 hosts
(which we deprecated when the kernel removed its end of the support)
* VIRT-364 [QEMU support for ARMv8.1-M extensions]
+ fp16 work is now in master
+ sent patches to add support for mps2-an386 and mps2-an500
boards (Cortex-M4 and Cortex-M7 setups similar to the existing
mps2-an385 Cortex-M3 board model)
thanks
-- PMM
== This Week ==
* GNU-659 (LTO calculix regression)
- Started upstream discussion
* LLVM-611 (Tune lowering heuristic to blx for code size)
- Addressing upstream feedback for patch to fold tLDRpci, tBLx -> tBL
in foldMemoryOperand hook
- Added new heuristic to ISel, for follow up patch.
* LLVM-639 (Better idiom for comparing 32-bit values on arm)
- Started working on patch
== Next Week ==
- Continue ongoing tasks
1 day off
== Progress ==
* GCC upstream validation:
- reported several regressions
- committed testcase fixes
* GCC
- patch for PR94758 (-mpurecode and cortex-m23) applied to
thunk/gcc-10/gcc-9. Bug closed, but opened several others to track
performance improvements with -mpurecode
- PR96768: switch tables for thumb-1 with -mpure-code. Patch sent,
discussion on-going
- PR96769: WIP
- PR71233: missing ACLE intrinsics. I will first have to update the list
* benchmarking:
- Scripts to run coremark on stm32 now merged, working on using them
in production
* misc:
- infra patches/reviews
- internal support, leading to GCC testsuite cleanups for cortex-m
== Next ==
* GCC/cortex-M testing improvements & fixes
* cortex-m benchmarking
Progress:
* VIRT-65 [QEMU upstream maintainership]
+ Just the usual code review and pull request work.
* VIRT-364 [QEMU support for ARMv8.1-M extensions]
+ I have all fp16 insns (both VFP and Neon) implemented now, and the
VFP patches have been reviewed. Just need to track down a few last
test failures and then I can send the Neon patches out for review.
thanks
-- PMM
== Progress ==
* GCC upstream validation:
- reported several regressions
- committed testcase fixes
* GCC
- sent patch for PR94758 (-mpurecode and cortex-m23)
* benchmarking:
- Scripts to run coremark on stm32 now merged, working on using them
in production
* misc:
- infra patches/reviews
== Next ==
* GCC/cortex-M testing improvements & fixes
* cortex-m benchmarking
Progress (short week, 2 days):
* VIRT-65 [QEMU upstream maintainership]
+ We got 5.1 out the door. I think this is the first release
in the many years I've been handling them where we managed to
avoid needing an rc4.
+ Spent a fair amount of time helping Paolo to debug his conversion
of the QEMU build system to Meson. We managed to land that today.
+ First of the KVM Forum programme committee meetings
* VIRT-364 [QEMU support for ARMv8.1-M extensions]
+ More progress with fp16 support work; another handful of insns done.
thanks
-- PMM
