== Progress ==
* BFD Linker:
- GNU-629: non-contiguous memory support: found problems in the
testsuite when the flag is activated by default
* GCC:
- trying to find a way to add run more validation of -mpure-code on
v6m, to support a request to backport to gcc-9: need to patch newlib's
crt0.S which has some offending sequences
* GCC upstream validation:
- updated scripts to cope with the new git repo and release branches names
* misc:
- infra fixes / troubleshooting / reviews
- fixed problems with automatic bisection of gcc regressions
== Next ==
* Binutils: GNU-629: support non-contiguous memory regions in linker
* GCC-9: continue work for validation of -mpure-code on v6m
Morello
- Upstream LLD INPUT_SECTION_FLAGS that I implemented downstream.
Quite a few review comments to work through but I think I'm close to
getting something accepted. This will hopefully come in useful for LLD
with embedded systems.
- Wrote down some definitions for the new relocations.
LLD
- landed fix for clang-built-linux-812 linux allyesconfig problem.
- Quite a few upstream LLD and MC reviews
Other
Some research for the clang built linux conference.
== Progress ==
* Out of office for 2 days
* Reviewed EuroLLVM submissions
* A bit more investigation into Morello bare metal debugging
== Plan ==
* PR44157
* Morello bare metal debugging
Hello Linaro team,
I want to use your cross compiler version 6.3.1, but I miss the library
asound and its header (alsa/asoundlib.h)
What do I need to do to fix this ?
Best regards,
Markus Bollinger
Hi Ana,
I don't know anything about the POC myself, but I'm forwarding this so our
QEMU folks can answer.
Cheers,
Diana
---------- Forwarded message ---------
From: Ana Pazos <apazos(a)quicinc.com>
Date: Fri, 10 Jan 2020 at 19:08
Subject: about QEMU support for SVE2 POC
To: diana.picus(a)linaro.org <diana.picus(a)linaro.org>
Cc: Ana Pazos <apazos(a)quicinc.com>
Hello Diana,
We at Qualcomm are trying to find out the POC for QEMU support for SVE2.
We have been using QEMU with SVE support, and need to find out when SVE2
support will be added to QEMU.
I assume Linaro is involved in this task. Do you know the POC?
Thanks for the help!
Ana.
== Progress ==
* BFD Linker:
- GNU-629: non-contiguous memory support: sent updated versions with
testcases and doc.
* GCC:
- trying to find a way to add run more validation of -mpure-code on
v6m, to support a request to backport to gcc-9
- proposal to implement LLVM's -arm-assume-misaligned-load-store
rejected by the community
* GCC upstream validation:
- reported/checked a few issues
* misc:
- infra fixes / troubleshooting / reviews
- investigating problems with automatic bisection of gcc regressions
- fixed a long-standing bug in proot
== Next ==
* Binutils: GNU-629: support non-contiguous memory regions in linker
[VIRT-327 # Richard's upstream QEMU work ]
* TCG patch queue flush, including phase 1 increase for
number of mmu_idx, on which VHE is dependent.
* Random patch review.
* Capstone update.
* cputlb cleanup in prep for modeling ASIDs.
r~
Morello:
- Implemented INPUT_SECTION_FLAGS in LLD for the Morello toolchain,
now in review. Took most of spare engineering time this week. Will
attempt to upstream once committed in Morello
- Fixed pr44451 upstream, it was already fixed on Morello toolchain.
LLD:
- Made an attempt to fix the LLD linux kernel allyesconfig +
cortex-a53 erratum problem. Was not successful as kernel linker script
is just about as awkward as possible for this problem. I have a good
idea of what to try next. Will aim to fix early next week.
- Quite a few upstream reviews, some pending from last year.
Progress:
* VIRT-65 [QEMU upstream maintainership]
- caught up on the backlog of email that accumulated over the Christmas break
- investigated an intermittent failure of a test case on our BSD hosts
which had reached a failure rate that was seriously impeding the
flow of pull requests. Tracked it down to glib's g_source_ref/unref
not being thread-safe unless the GSource is attached to a GMainContext;
QEMU gets this right for its own use but a set of mock functions used
for a few testcases were not doing so, resulting in occasional races
where the GSource would get destroyed while it was still in use.
- code review:
+ Beata's "inject DABT for load/store insns KVM couldn't emulate" patch
+ a minor cubieboard cleanup patchset
+ i.MX RNGC device emulation patch
- annual review season, some time spent on that
thanks
-- PMM
== Progress ==
* Out of office until Tuesday
* LLVM 9.0.1
- Uploaded final binaries
- Still looking into PR44157 (CFI tests failing on armv7)
* Morello LLDB
- Trying to get bare metal debugging to work
* More ARM Code of Conduct courses etc
== Plan ==
* More Morello and PR44157
* Patch set cleaning up PIE/non-PIE linking
- Add new support for static pie.
* Another revision removing MMU_MODE_SUFFIX.
* Random patch review.
* Planning for Linux on ARM summit in Cambridge in February.
r~
To whom,
I’m seeing a failure on this bot:
--
Command Output (stderr):
--
/home/buildslave/buildslave/clang-cmake-armv7-quick/stage1/bin/llvm-objdump: error: '/home/buildslave/buildslave/clang-cmake-armv7-quick/llvm/llvm/test/tools/llvm-objdump/Inputs/macho-stabs-x86_64': can't find target: : error: unable to get target for 'x86_64---macho', see --version and --triple.
This is just a binary file verifier. Is it the case this bot doesn’t have a mach-o disassembler? And if so, is there a way to disable this test on this bot?
Thanks!
MDT
QEMU Tooling ([VIRT-252])
=========================
Extend gdbstub for SVE ([VIRT-281])
- posted [PATCH v4 00/21] gdbstub refactor and SVE support (+check-tcg
tweaks) Message-Id: <20191220120438.16114-1-alex.bennee(a)linaro.org>
[VIRT-281] <https://projects.linaro.org/browse/VIRT-281>
GSoC Mentoring Afermath ([VIRT-348])
- started working on re-base of [TCG code quality tracking]
- bit of a re-factor rethink required
[TCG code quality tracking]
<https://github.com/stsquad/qemu/tree/tcg/tbstats-and-perf-v10>
Upstream Work ([VIRT-109])
==========================
- posted [PULL 00/25] testing and logging updates Message-Id:
<20191219104934.866-1-alex.bennee(a)linaro.org>
- posted [PATCH v1 0/4] semihosting read console support Message-Id:
<20191218180029.6744-1-alex.bennee(a)linaro.org>
- posted [PATCH v2 0/5] semihosting read console support Message-Id:
<20191220132246.6759-1-alex.bennee(a)linaro.org>
[VIRT-109] <https://projects.linaro.org/browse/VIRT-109>
Completed Reviews [4/4]
=======================
{PATCH} Semihost SYS_READC implementation (v6)
Message-Id: <20191104204230.12249-1-keithp(a)keithp.com>
- CLOSING NOTE [2019-12-18 Wed 19:39]
Found some issues with deadlocks, fixed and sent my own series
including this patch.
Added: <2019-12-06 Fri>
[PATCH] docker: gtester is no longer used
Message-Id: <1576632611-55032-1-git-send-email-pbonzini(a)redhat.com>
- CLOSING NOTE [2019-12-20 Fri 12:20]
Queued to my tree
Added: <2019-12-18 Wed>
{Qemu-devel} {RFC PATCH} Implement qemu_thread_yield for posix, use it in mttcg to handle EXCP_YIELD
Message-Id: <20190717054655.14104-1-npiggin(a)gmail.com>
- CLOSING NOTE [2019-12-20 Fri 13:11]
Replied.
[PATCH v2 00/28] cputlb: Remove support for MMU_MODE*_SUFFIX
Message-Id: <20191216221158.29572-1-richard.henderson(a)linaro.org>
Absences
========
- Christmas and New Year holidays
Current Review Queue
====================
* [RFC PATCH 00/13] hw/timer/allwinner: Make it reusable
Message-Id: <20191219185127.24388-1-f4bug(a)amsat.org>
Added: <2019-12-19 Thu>
* [PATCH v2 00/14] chardev: Use QEMUChrEvent enum in IOEventHandler typedef
Message-Id: <20191218172009.8868-1-philmd(a)redhat.com>
Added: <2019-12-18 Wed>
* [RFC PATCH v3 000/132] Proof of concept for Meson integration
Message-Id: <1576155176-2464-1-git-send-email-pbonzini(a)redhat.com>
Added: <2019-12-12 Thu>
* {PATCH 0/2} tests/acceptance: Add boot vmlinux test
Message-Id: <20191206140012.15517-1-wainersm(a)redhat.com>
Added: <2019-12-06 Fri>
--
Alex Bennée
Progress (covers two weeks; lost half of last week to the cold that's
been going around the office here):
* VIRT-65 [QEMU upstream maintainership]
- code review
+ RTH's ARMv8.2-UAO patchset
+ RTH's ARMv8.1-PAN patchset
+ had a look at Huawei SDEI patchset; there are some things I don't
like about the approach but I don't currently have any better ideas :-(
Sent an email to the list to see if others have any suggestions.
+ Andrew Jones' patchset to pause the virtual timer when VM is paused
+ Some smmuv3 emulation fixes from Amazon
+ Support for emulating generic timers that run at platform-dependent
frequency (rather than insisting they always run at 62.5MHz)
- release work
+ herded a handful of key bugfixes into rc5; got rc5 and
then the final release out of the door
+ got new s390 box into the set of machines we test QEMU on
(IBM want to turn off the old one at the end of the year)
+ a couple of arm pull requests now 5.0 is open for new contributions
thanks
-- PMM
== Progress ==
* LLVM 9.0.1
- Had trouble with our TK1 machine
- Trying to build rc3 on one of the buildbots
- AArch64 rc3 looks fine
* Morello
- Got lldb-server working on android and it seems to behave fine
- Working on getting the lldb test-suite to work in remote mode with the
morello android simulator
- Running into all sorts of issues with it
== Plan ==
* More of the same
* Out of office: 25-26 December, 1-7 January
One cannot use objcopy --target=efi-app-aarch64 because it appears
that PE/COFF targets are not enabled for aarch64.
Making UEFI & Windows on Snapdragon work harder.
Given that Windows 10 is fully supported on aarch64, I assume it does
have PE/COFF.
Can someone please fix bfd to provide PE targets on aarch64?
--
Regards,
Dimitri.
== Progress ==
* GCC:
- -mpure-code on v6m: committed.
- investigated potential problem with -mno-unaligned-access. It was a
wrong-user-code case, but maybe it would be desirable to implement an
option like llvm's -arm-assume-misaligned-load-store
* BFD Linker:
- GNU-629: non-contiguous memory support: Now able to detect cases
where input sections change size during the linker iterations and when
linker-created stubs would cause overflows.
* GCC upstream validation:
- reported/checked a few issues
* misc:
- infra fixes / troubleshooting / reviews
== Next ==
* Binutils: GNU-629: support non-contiguous memory regions in linker
* Holidays: Dec 23rd-Jan 2nd
There’s been an ongoing failure on the clang-cmake-armv8-selfhost
builder for the last day or so; see e.g.:
http://lab.llvm.org:8011/builders/clang-cmake-armv7-selfhost/builds/2827/
This list is recorded as the contact information for that builder.
It is apparently crashing in a tblgen backend that I just added:
```
Stack dump:
0. Program arguments:
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/stage1/bin/clang-tblgen
-gen-clang-type-writer -I
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/llvm/clang/include
-I
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/llvm/clang/include/clang/AST
-I
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/llvm/llvm/include
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/llvm/clang/include/clang/AST/TypeProperties.td
--write-if-changed -o
/home/buildslave/buildslave/clang-cmake-armv7-selfhost/stage1/tools/clang/include/clang/AST/AbstractTypeWriter.inc
Segmentation fault (core dumped)
```
I would like to fix this, but I have no ability to reproduce it, and the
crash information in the log is quite minimal. Is there a way I can get
a shell on a system that can reproduce this, or at the very least get a
line number for where exactly it is crashing?
John.
The Arm GNU Toolchain for the A-profile Architecture
=====================================================
We are pleased to announce the Arm release of the pre-built GNU cross-toolchain
for the A-profile cores: GCC 9.2-2019.12.
Further information about the GNU Arm toolchain and the release packages is
available at Arm Developer site
https://developer.arm.com/open-source/gnu-toolchain/gnu-a/downloads
Features
========
* Based on GCC 9.2 (See https://gcc.gnu.org/gcc-9/changes.html for details).
Supported targets
=================
* On Windows(x86_64):
- AArch64 (bare-metal and Linux)
- AArch32 (bare-metal and Linux hard-float)
* On Linux(x86_64):
- AArch64 (bare-metal, Linux and Linux big-endian)
- AArch32 (bare-metal and Linux hard-float)
* On Linux(AArch64)
- AArch64 (bare-metal)
- AArch32 (bare-metal and Linux hard-float)
Changes since Arm release GCC 8.3-2019.03
=========================================
* Additional AArch64 hosted cross-toolchains for AArch64 (bare-metal) and
AArch32 (bare-metal and Linux hard-float)
* Additional Windows hosted cross-toolchains for AArch64 (Linux) and
AArch32 (Linux hard-float)
* Retired Linux(x86_64) toolchain for AArch64 (big-endian bare-metal) and
AArch32 (Linux soft-float)
* Changed toolchain naming convention to match standard target triplet
naming convention, with vendor name being none.
For example, arm-eabi is now arm-none-eabi.
* Fixed the Windows toolchain convention to correctly include mingw-w64
instead of mingw32
Host Requirements
==================
* x86-64 Linux: Ubuntu 16.04 LTS or later, or RHEL 7 or later
* AArch64 Linux: Ubuntu 18.04 LTS or later, or RHEL 8
* x86 Windows: Windows 10
Package Versions
=================
* GCC 9.2.1
* glibc 2.30
* binutils 2.33.1
* GDB 8.3.0
* libexpat 2.2.5
* Linux Kernel v4.20.13
* libgmp 4.3.2
* libisl 0.15
* libmpfr 3.1.6
* libmpc 1.0.3
* libiconv 1.15
Contact Arm
===========
For any questions, please use the Arm Communities forums at:
https://community.arm.com/tools/
Please report any bugs via the Linaro Bugzilla at:
https://bugs.linaro.org/
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
* One day off
o LLVM:
* Couple of failures with the bots
* Machine Outliner:
- Completed testcases
- Validation before re-submission
o Misc
* Various meetings and discussions.
[VIRT-327 # Richard's upstream QEMU work ]
Following up on the feedback from last week on VHE and PAN, posted a patch set
eliminating MMU_MODE*_SUFFIX, and the current limit of NB_MMU_IDX <= 12 that
went with that.
Some patch review.
r~