== Progress ==
* [Thumb GlobalISel] Support G_SELECT [LLVM-529]
- Committed upstream
* [Thumb GlobalISel] Support control flow [LLVM-530]
- Committed support for branches
- Support for PHI in progress
* LLVM 8.0.0 Release for ARM & AArch64 [LLVM-526]
- rc2 uploaded, no changes from rc1 on our side
* Use new version of GCC on buildbots [LLVM-515]
- Restarted the buildbots, they are running GCC 7 now
* Buildbot babysitting [LLVM-158]
- Reproduced a couple of failures, ran a pre-commit for one of them
== Plan ==
* Continue LLVM-526 and LLVM-530
== This Week ==
* SVE ACLE (4/10)
- Committed svlsl patch to branch
* SVE (3/10)
- Studying GCC's vectorizer
- Started looking at PR88839
* Validation (2/10)
- GNU-156: Working on tcwg_gnu CI job.
* Misc (1/10)
- Meetings
- Filed PR89332 in GCC bugzilla.
== Next Week ==
- Continue ongoing tasks
[VIRT-242 # ARMv8.3-JSCVT ]
Posted v1 & v2.
[VIRT-339 # ARMv8.5-BTI, Branch Target Identification ]
Posted v3.
[VIRT-327 # Richard's upstream QEMU work ]
Posted v3 of arm tcg vector improvements. Now uses ssadd & usadd tcg opcodes
from AdvSIMD and not just SVE. This would have made it easier to find a bug in
my previous tcg backend patches.
Found the last two bugs in a conversion of target/hppa to decodetree, and
posted the patch set. Sent pull request for some other queued hppa patches.
Fix a minor RISU bug wrt SVE.
[LoAAS]
Travel to/from Cambridge for 2 days of Linux on ARM Summit.
r~
== Progress ==
* FDPIC
- (GNU-411) GDB: debugging problems with FDPIC support.
* GCC upstream validation:
- reported a few regressions, helping test patches
- dealing with some random results, still
* GCC:
- (GNU-99) rebased ubsan / bare-metal patches. No progress.
* misc (conf-calls, meetings, emails, ....)
- reviewing/submitted infra script patches
- debugging new benchmarking round-robin jobs
== Next ==
FDPIC:
- GCC: handle feedback on v4 patches
- GDB: update patches
- uclibc-ng: look at how to test fdpic mode with openadk
Infra:
- benchmarking jobs
QEMU Tooling ([VIRT-252])
=========================
QEMU plugin support ([VIRT-280])
- re-based [ldst demacro to v3]
- recollect stats to see if slowdown still there :todo
- awaiting v3 of plugin patches
[VIRT-280] https://projects.linaro.org/browse/VIRT-280
[ldst demacro to v3]
https://github.com/stsquad/qemu/tree/ldst/demacrofy-v3
Upstream Work ([VIRT-109])
==========================
- need to finish the re-work of [system test and misc arch] tests
:todo
- will be useful for test cases for plugins
- also more test cases queued up for system tests
- debugged an occasional failure in pauth-1 test case
- expected given PAuth architecture, test needs tightening up
- posted {PATCH v1 0/4} HWCAP_CPUID registers for aarch64 Message-Id:
<20190128173940.25813-1-alex.bennee(a)linaro.org>
- posted {PATCH v2 0/6} HWCAP_CPUID registers for aarch64 Message-Id:
<20190205190224.2198-1-alex.bennee(a)linaro.org>
- posted {PATCH v2 00/16} current testing/next queue (with build
fixes) Message-Id: <20190207183744.5054-1-alex.bennee(a)linaro.org>
- this fixes regression of check-tcg but also
- discovered the Debian Buster's QEMU arm64 packaging is broken
- spent time trying to able to join #debian's IRC channels (aws
hosts banned!)
- tried to revive the cross-tools packaging I'd started with Wookey
last Connect
[VIRT-109] https://projects.linaro.org/browse/VIRT-109
[the linux-user multiarch but it is not ready yet]
https://github.com/stsquad/qemu/tree/testing/next-with-linux-user
[system test and misc arch]
https://github.com/stsquad/qemu/tree/testing/enable-system-tcg-tests-v2
Other
=====
- Connect talk has been accepted, Friday 8.30 slot
Completed Reviews [1/1]
=======================
{Qemu-arm} {PATCH v6 00/73} per-CPU locks
Message-Id: <20190130004811.27372-1-cota(a)braap.org>
- CLOSING NOTE [2019-02-08 Fri 17:19]
Looks good, improves MTTCG scalability to 10-12 core now ;-)
Absences
========
- 15-18th Feb (long w/e)
- Connect BKK19 (1-5th April 2019)
- holiday after Connect
Current Review Queue
====================
* {RFC} arm: Allow system registers for KVM guests to be changed by QEMU code
Message-Id: <20181206151401.13455-1-peter.maydell(a)linaro.org>
* {RFC v2 00/38} Plugin support
Message-Id: <20181209193749.12277-1-cota(a)braap.org>
* {RFC QEMU v2 0/2} arm/virt: Account for guest pause time
Message-Id: <1543352837-21529-1-git-send-email-bijan.mottahedeh(a)oracle.com>
* {PATCH v12 00/25} Fixing record/replay and adding reverse debugging
Message-Id: <154935478086.7509.4349987292919289245.stgit@pasha-VirtualBox>
* {PATCH v2 00/18} OpenBSD: Enable qtesting
Message-Id: <20190129175403.18017-1-philmd(a)redhat.com>
* {PATCH v11 00/25} Fixing record/replay and adding reverse debugging
Message-Id: <20190131131520.23264.75724.stgit@pasha-VirtualBox>
--
Alex Bennée
Fosdem:
- Presented on LLD performance at LLVM devroom
https://fosdem.org/2019/schedule/track/llvm/
- Other interesting presentations from the devroom include:
-- Compiling the Linux kernel with LLVM tools
-- Lessons in TableGen
LLVM-523 LLD should synthesise .cantunwind .ARM.exidx sections for
sections without them
- Implementation and regression tests done. Will do a bit more testing
and post upstream on Monday.
LLVM-521 Taking the address of an ifunc in LLD
- Spent some time reviewing upstream patch that should fix the
remaining LLD ifunc address equivalence problems for all Targets.
Progress:
* VIRT-65 [QEMU upstream maintainership]
+ reviewed RTH's patchsets for BTI and MemTag support
* VIRT-268 [QEMU support for dual-core Cortex-M Musca board]
+ Implemented a first pass at a Musca board model, which
has the PPC and MPC infrastructure, plus RAM, and ROM for
the flash; need to add at least the UARTs and do some testing
* misc
+ sorting out travel arrangements, visa, etc for Connect
+ kicked off a discussion on whether there are aspects of
QEMU's role as the userspace component of a KVM VM that
Linaro should be putting more effort/attention into
thanks
-- PMM
== Progress ==
* Out of office on Monday
* [Thumb GlobalISel] Support G_GEP [LLVM-532]
- Committed upstream
* [Thumb GlobalISel] Support G_ICMP [LLVM-528]
- Committed upstream
* [Thumb GlobalISel] Support G_SELECT [LLVM-529]
- Almost ready to commit
* LLVM 8.0.0 Release for ARM & AArch64 [LLVM-526]
- rc2 in progress
* Minor changes to Jenkins jobs [LLVM-533]
- TCWG Regression Detection job: launch LLVM 7.0.0 release on xenial
containers (as opposed to 6.0.0 on trusty)
- Release job: forced parallelism to 3 on the TK1s, to reduce some
of the swapping; also posted a patch upstream to allow us to use
ninja, so we can limit only the number of link jobs (like we do on the
buildbots)
== Plan ==
* Continue LLVM-526 and LLVM-529
* Use new version of GCC on buildbots [LLVM-515]
- Restart the buildbots so they begin using GCC 7