o LLVM
* Buildbots babysitting:
- Investigating armv7 bots failures, having hard time to reproduce the issue
* Machine Outliner on ARM prototype:
- worked on LTO integration
- more stack fixups to handle
o Misc
* Various meetings and discussions.
QEMU Tooling ([VIRT-252])
=========================
- thoughts after reviewing {Qemu-devel} {RFC PATCH v2 0/7} QEMU binary
instrumentation prototype Message-Id:
<152819515565.30857.16834004920507717324.stgit@pasha-ThinkPad-T60>
- we want rich tooling but don't want to leak internals to plugins
- the proposed helper per insn is pretty limited/hacky
- I think we could expand exiting trace support with plugins
- plugin hook to trace points
- plugin either filters trace points or does in-situ analysis
- existing trace infrastructure used to export data
[VIRT-252] https://projects.linaro.org/browse/VIRT-252
Upstream Work ([VIRT-109])
==========================
- investigating CI failure in master:
- The [Travis builds have been broken for a while]
- meanwhile it looks like [atomic_8 issues have broken 32 bit
builds]
- reviewed {PATCH 0/6} i386 + x86_64 mttcg Message-Id:
<20180903171831.15446-1-cota(a)braap.org>
- posted {RFC PATCH 0/4} Add Nios II cross-compiler and enable
tests/tcg Message-Id:
<8346c1bb-9cb6-4c08-66a2-b5e5a31903d4(a)vivier.eu>
[VIRT-109] https://projects.linaro.org/browse/VIRT-109
[Travis builds have been broken for a while]
https://travis-ci.org/qemu/qemu/builds
[atomic_8 issues have broken 32 bit builds]
https://app.shippable.com/github/qemu/qemu/dashboard
Other Tasks
===========
- Continued working on "My Other Machine is Virtual" talk for YVR18
- finished main talk, iterating and practising now
- also made a few [fixes and enhacements for tlb tracking]
- Administrava
- caught up with expense claims for SynQuacer
- organising for KVM Forum 2018
- organising/travel to Connect
[fixes and enhacements for tlb tracking]
https://github.com/stsquad/qemu/tree/misc/dfilter-and-trace-tweaks-v2
Completed Reviews [4/4]
=======================
{PATCH 0/6} target/arm: More sve-ish fixes
Message-Id: <20180810193129.1556-1-richard.henderson(a)linaro.org>
- CLOSING NOTE [2018-09-07 Fri 16:36]
Already in target-arm.next
{Qemu-devel} {PATCH 0/3} synchronization profiler
Message-Id: <20180813171132.21939-3-cota(a)braap.org>
- CLOSING NOTE [2018-09-07 Fri 17:41]
Already merged
{PATCH 0/6} i386 + x86_64 mttcg
Message-Id: <20180903171831.15446-1-cota(a)braap.org>
- CLOSING NOTE [2018-09-10 Mon 10:18]
Quick pass, looks OK but a question about __thread w.r.t tcg globals
{PATCH v2 0/2} softfloat tests based on berkeley's testfloat
Message-Id: <20180908191735.22861-1-cota(a)braap.org>
- CLOSING NOTE [2018-09-10 Mon 12:27]
Looks like an improvement on IBM test suite but needs a few more
tweaks.
Absences
========
- YVR18 Connect (17th-21st September 2018)
- KVM Forum 2018 (24th-26th October 2018)
Current Review Queue
====================
* {PATCH v3 0/3} softfloat tests based on berkeley's testfloat
Message-Id: <20180913213910.28189-1-cota(a)braap.org>
* {PATCH v3 00/13} i386 + x86_64 mttcg
Message-Id: <20180911202823.21657-1-cota(a)braap.org>
* {PATCH v6 00/25} Fixing record/replay and adding reverse debugging
Message-Id: <20180912081747.3228.21861.stgit@pasha-VirtualBox>
* {Qemu-devel} {PATCH v2 00/11} convert CPU list to RCU
Message-Id: <374f82bc-1680-a59d-aa71-31e88e4936a2(a)redhat.com>
* {Qemu-arm} {PATCH 00/15} gdbstub: support for the multiprocess extension
Message-Id: <20180901124639.19735-1-luc.michel(a)greensocs.com>
* {Qemu-devel} {PATCH v4 0/5} Acceptance/functional tests
Message-Id: <20180530184156.15634-1-crosa(a)redhat.com>
--
Alex Bennée
Last week:
Holiday.
3 days at GNU Cauldron. Highlights:
* Talked with arm folk re the sve simd abi,
* Talked with arm folk re valgrind support for sve,
* Talked with riscv folk re their nascent vector extension,
* Peter Sewell's memory object semantics for defacto c.
This week:
Caught up on email backlog.
Queued a few patch sets for first 3.1 tcg-next pull.
Posted an rfc for adjusting arm id system regs vs feature bits.
r~
Hi Mathias,
I am not able to reproduce error in
https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a7…
Do you have any instructions for reproducing it locally.
This is what I have done:
Downloaded https://www.python.org/ftp/python/2.7.4/Python-2.7.4.tgz
Applied following patch to build for aarch64:
diff -r 84cef4f1999a -r 05e8999a3901 Modules/_ctypes/libffi/fficonfig.py.in
--- a/Modules/_ctypes/libffi/fficonfig.py.in Mon Apr 29 16:09:39 2013 -0400
+++ b/Modules/_ctypes/libffi/fficonfig.py.in Tue Apr 30 01:00:34 2013 +0200
@@ -28,6 +28,7 @@
'PA': ['src/pa/linux.S', 'src/pa/ffi.c'],
'PA_LINUX': ['src/pa/linux.S', 'src/pa/ffi.c'],
'PA_HPUX': ['src/pa/hpux32.S', 'src/pa/ffi.c'],
+ 'AARCH64' : ['src/aarch64/ffi.c', 'src/aarch64/sysv.S'],
}
ffi_sources += ffi_platforms['@TARGET@']
./configure
Also tried ./configure --enable-pydebug and --disable-optimization
Also tried changing to -O0 manually in the make file
make
./python Lib/test/regrtest.py -v test_ctypes
Used gcc version 8.2.1 20180907 (GCC)
Results are OK.
Thanks,
Kugan
o LLVM
* Buildbots babysitting
* LLVM Kernel CI babysitting
* Machine Outliner on ARM prototype still on-going
o Misc
* More Jira gardening
* Various meetings and discussions.
Upstream Work ([VIRT-109])
==========================
- reviewed {Qemu-devel} {RFC PATCH v2 0/7} QEMU binary instrumentation
prototype Message-Id:
<152819515565.30857.16834004920507717324.stgit@pasha-ThinkPad-T60>
- reviewed {PATCH 0/6} qht improvements for 3.1 Message-Id:
<20180817232923.28899-1-cota(a)braap.org>
- investigating CI failure in master:
- The [Travis builds have been broken for a while]
- meanwhile it looks like [atomic_8 issues have broken 32 bit
builds]
[VIRT-109] https://projects.linaro.org/browse/VIRT-109
[Travis builds have been broken for a while]
https://travis-ci.org/qemu/qemu/builds
[atomic_8 issues have broken 32 bit builds]
https://app.shippable.com/github/qemu/qemu/dashboard
Other Tasks
===========
- Continued working on "My Other Machine is Virtual" talk for YVR18
- finished main talk, iterating and practising now
- also made a few [fixes and enhacements for tlb tracking]
- Administrava
- caught up with expense claims for SynQuacer
- organising for KVM Forum 2018
[fixes and enhacements for tlb tracking]
https://github.com/stsquad/qemu/tree/misc/dfilter-and-trace-tweaks-v2
Completed Reviews [2/2]
=======================
{Qemu-devel} {RFC PATCH v2 0/7} QEMU binary instrumentation prototype
Message-Id: <152819515565.30857.16834004920507717324.stgit@pasha-ThinkPad-T60>
- CLOSING NOTE [2018-09-07 Fri 15:40]
Some interesting ideas but I think the interface needs to be a bit
richer and expose the tcg code generation to the plugins.
{PATCH 0/6} qht improvements for 3.1
Message-Id: <20180817232923.28899-1-cota(a)braap.org>
- CLOSING NOTE [2018-09-07 Fri 16:35]
Good improvements to test coverage, one minor tweak needed
{PATCH 0/6} target/arm: More sve-ish fixes
Message-Id: <20180810193129.1556-1-richard.henderson(a)linaro.org>
- CLOSING NOTE [2018-09-07 Fri 16:36]
Already in target-arm.next
{Qemu-devel} {PATCH 0/3} synchronization profiler
Message-Id: <20180813171132.21939-3-cota(a)braap.org>
- CLOSING NOTE [2018-09-07 Fri 17:41]
Already merged
Absences
========
- Holiday (16th-28th August 2018)
- YVR18 Connect (17th-21st September 2018)
- KVM Forum 2018 (24th-26th October 2018)
Current Review Queue
====================
* {Qemu-arm} {PATCH 00/15} gdbstub: support for the multiprocess extension
Message-Id: <20180901124639.19735-1-luc.michel(a)greensocs.com>
* {PATCH 0/6} i386 + x86_64 mttcg
Message-Id: <20180903171831.15446-1-cota(a)braap.org>
* {Qemu-devel} {PATCH v4 0/5} Acceptance/functional tests
Message-Id: <20180530184156.15634-1-crosa(a)redhat.com>
* {Qemu-arm} {PATCH 00/10} target/arm: Some pieces of support for 32-bit Hyp mode
Message-Id: <20180814124254.5229-1-peter.maydell(a)linaro.org>
* {PATCH 00/20} target/arm: sve system mode patches
Message-Id: <20180809042206.15726-1-richard.henderson(a)linaro.org>
* {PATCH v4 00/14} fp-test + hardfloat
Message-Id: <1528768140-17894-1-git-send-email-cota(a)braap.org>
--
Alex Bennée
[Linaro Connect]
- Wrote my presentation for the official track (on using Clang
libtooling and ASTMatchers)
- Started working on hackroom presentation on profile guided
optimisation in LLVM.
In both cases writing a presentation on something you don't know much
about forces you do a lot of learning.
Plans for next week:
- Finish hackroom presentation.
Spare time will be spent on reviews and llvm/clang bugs
=== Work done during this 3-day week ===
* TCWG-1428 (Support arithmetic on FileCheck regex variable):
+ continue to rework patch, feature complete and most testcase pass
+ sent another external email to hammer a few details
* Arm new starter welcome and related faff
* GNU Tools Cauldron + travel from Thursday to Sunday
=== Plan for week 37 ===
* TCWG-1428 (Support arithmetic on FileCheck regex variable):
+ fix last bugs, start testing and cleanup
* Try to reproduce perf issue mentioned in week #30's weekly report on
latest perf
* One day off
o LLVM
* Machine Outliner on ARM prototype:
- fixup and testcases work still on-going
- first benchmark results
* Buildbots babysitting
o Misc
* More Jira gardening
* Various meetings and discussions.
Other Tasks
===========
- Continued working on "My Other Machine is Virtual" talk for YVR18
- got working heatmaps for coverage
- writing up TCG memory tracing
Absences
========
- Holiday (16th-28th August 2018)
- YVR18 Connect (17th-21st September 2018)
- KVM Forum 2018 (24th-26th October 2018)
Current Review Queue
====================
* {PATCH 0/6} qht improvements for 3.1
Message-Id: <20180817232923.28899-1-cota(a)braap.org>
* {Qemu-devel} {PATCH 0/3} synchronization profiler
Message-Id: <20180813171132.21939-3-cota(a)braap.org>
* {PATCH 0/6} target/arm: More sve-ish fixes
Message-Id: <20180810193129.1556-1-richard.henderson(a)linaro.org>
* {Qemu-arm} {PATCH 00/10} target/arm: Some pieces of support for 32-bit Hyp mode
Message-Id: <20180814124254.5229-1-peter.maydell(a)linaro.org>
* {PATCH 00/20} target/arm: sve system mode patches
Message-Id: <20180809042206.15726-1-richard.henderson(a)linaro.org>
* {PATCH v4 00/14} fp-test + hardfloat
Message-Id: <1528768140-17894-1-git-send-email-cota(a)braap.org>
--
Alex Bennée
* very short week: off Monday/Tuesday
== Progress ==
* FDPIC
- uClibc-ng: patches committed upstream
- GCC: handling feedback on v2 patches
* GCC upstream validation:
- trunk build broken for a while on RHEL6 (gcc-4.4.7)
- fixed upstream just before I reported the problem
- my scripts now use gcc-5.4.0
* misc (conf-calls, meetings, emails, ....)
- book hotel/plane for GNU Cauldron and Connect
== Next ==
FDPIC:
- GCC: handle v2 patches feedback
- uclibc-ng: look at how to test fdpic mode with openadk
- Connect: prepare hacking room presentation
- GNU Cauldron 2018
=== Work done during this week ===
* TCWG-1379 / GCC PR85434 / CVE-2018-12886: rework needed
+ finished testing and submitted for external review
* TCWG-1428 (Support arithmetic on FileCheck regex variable): address comments
+ continue to rework patch
* Regression testing of LLVM release process:
+ Add support to build from a branch (incl. trunk)
+ Make weekly regression testing from trunk
+ start enabling regression testing of Linaro LLVM release process
when scripts it use change
* Prepare for becoming a line manager next week
=== Plan for week 36 ===
* Welcome new starter
* Finish enabling regression testing of Linaro LLVM release process
when scripts it use change
* DSGHACK-25 (Support arithmetic on FileCheck regex variable): address comments
+ continue to rework patch
* Try to reproduce perf issue mentioned in week #30s weekly report on
latest perf
[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
4 day week, out on holiday Wednesday.
[TCWG Jira]
- Recommendations made on what to do with outstanding epics.
- Will be making the changes on Monday.
[Linaro Connect]
- Research for presentation on libtooling.
- Outline of presentation written and made a start on the first of the examples.
- Will be doing a further presentation in the hack-room on LLVM
profiling and code-size
Involved in more than the usual amount of reviews.