[UM-2 # QEMU Upstream Work]
* A lot of patch review.
* Some target/i386 cleanup, as followup to said patch review.
* Collecting patches for tcg-next.
r~
== Progress ==
* GCC upstream validation:
- a few regressions to bisect. Fixed a minor testcase issue
* GCC
- MVE autovectorization: Working on vcmp. After some cleanup &
factorization, the cmp operators work on GCC vectors. I will now
resume work on auto-vectorization.
* Misc
- fixes in stm32 benchmarking harness
== Next ==
* MVE auto-vectorization/intrinsics improvements
* GCC/cortex-M testing improvements & fixes
* cortex-m benchmarking
* After proving that the aarch64 host problem with
booting s390x with virtio was lack of barriers for
the guest memory model, I've spent some time working
on a ld-acq/st-rel optimizer for tcg.
* Minor rev of tci rewrite.
* Some patch review on claudio's kvm/tcg split.
r~
Progress:
* QEMU-364 [QEMU support for ARMv8.1-M extensions]
+ sent out v2 of the MPS3-AN524 series (minor changes only)
+ discovered that I need to implement the SSE-300's System Counter
+ implemented the system counter, did some final testing and
reviewing, and sent out v1 of the MPS3-AN547 series to the list
+ once these patches are reviewed and in master the QEMU-413 epic
will be complete
-- PMM
Hello Linaro Toolchain Working Group,
libcxx-libcxxabi-libunwind-armv8-linux has been red for 5 days.
clang-cmake-aarch64-global-isel has been red for 10 days.
clang-cmake-aarch64-quick has been red for 10 days.
clang-cmake-armv7-global-isel has been red for 10 days.
clang-cmake-armv7-lnt has been red for 10 days.
clang-cmake-aarch64-lld has been red for 11 days.
clang-cmake-armv8-lld has been red for 11 days.
clang-cmake-thumbv7-full-sh has been red for 11 days.
clang-cmake-armv7-full has been red for 11 days.
Is anybody looking at this?
I also noticed that some of your workers are less reliable than others. For
example, inaro-tk1-09 worker was last time producing a green build 2 months
ago - http://lab.llvm.org:8011/#/workers/2?numbuilds=500
I'm removing the linaro-tk1-09 worker from the production buildbot for now.
Please feel free to connect it to the staging and make the builder reliably
green. Then it could be returned back to the production.
linaro-armv8-windows-msvc-01 and linaro-armv8-windows-msvc-02 have been red
for the least 4 months. I'm removing them from the production buildbot. You
can return them back to production after they are reliably green in the
staging.
Please feel free to ask if you have questions.
Thanks
Galina
== Progress ==
* GCC upstream validation:
- a few regressions to bisect. Fixed a minor testcase issue
- native validation in Linaro's lab: we still see a few random results
* GCC
- MVE autovectorization: Working on vcmp.
* Misc
- fixes in stm32 benchmarking harness
== Next ==
* MVE auto-vectorization/intrinsics improvements
* GCC/cortex-M testing improvements & fixes
* cortex-m benchmarking
Holidays next week, back Monday 22nd Feb
Progress:
* UM-2 [QEMU upstream maintainership]
+ sent out a small code-cleanup patchset for some arm display device
models to remove no-longer-needed support for non-32bpp outputs
+ tracked down (with the aid of Paolo) why our build system had started
building the documentation every time you run 'make'; sent a fix
+ code review todo queue status: 12 items, but the oldest was only
sent to the list on Monday...
* QEMU-364 [QEMU support for ARMv8.1-M extensions]
+ An MPS3 Cortex-M55 FPGA image is now publicly available: the AN547.
We'll provide a model of this in QEMU. Started working on the
patchseries to implement it (relatively small changes on top of
the SSE-300 work I've already done but not sent out, and the
mps3-an524 series I sent out for review last week). This is basically
code-complete but I need to do some more testing and ideally I'd
like the AN524 series to get reviewed before I send out another
40-patch series that would have to be based on top of that one.
thanks
-- PMM
Hello,
I've uploaded binaries for Windows on Arm:
081373cc76e88224020fea42eba2161d972f03bb83ebc055fb3cd4f2cfcdfb95
LLVM-12.0.0-rc1-woa64.exe
It was built from the current release/12.x branch (commit b46924ee) with
two patches applied on top:
- https://reviews.llvm.org/D96259
- https://reviews.llvm.org/D96498
It contains "clang;clang-tools-extra;lld;compiler-rt” projects.
Compiler-RT has sanitizers disabled (MEMPROF and XRAY use sanitizer
infrastucture). I.e.,
===
+ -DLLVM_ENABLE_PROJECTS="clang;clang-tools-extra;lld;compiler-rt" ^
+ -DCOMPILER_RT_BUILD_SANITIZERS=OFF ^
+ -DCOMPILER_RT_BUILD_MEMPROF=OFF ^
+ -DCOMPILER_RT_BUILD_XRAY=OFF ^
===
I've attached the changes we had to make to
llvm/utils/releases/build_llvm_package.bat. @Maxim Kuvyrkov is going to
clean them up and post a patch in the next couple of weeks (he's been doing
all the hard work, I'm just uploading).
This is the first time we're trying to provide a release for Windows on
Arm, so if anyone has any comments or questions, feel free to send an email
:)
Cheers,
Diana
On Thu, 28 Jan 2021 at 05:05, Tom Stellard via lldb-dev <
lldb-dev(a)lists.llvm.org> wrote:
> Hi,
>
> I've tagged the 12.0.0-rc1 release. Testers can begin testing and upload
> binaries.
>
> -Tom
>
> _______________________________________________
> lldb-dev mailing list
> lldb-dev(a)lists.llvm.org
> https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-dev
>
