Hi,
It seems to me that your current toolchain releases [1] don't default to a
multiarch layout. Am I looking in the right place? Do you anticipate enabling
multiarch in the future? Is doing this currently blocked by limitations in
tools such as crosstool-ng?
1. http://releases.linaro.org/13.01/components/toolchain/binaries
Thanks,
Christopher
--
Employee of Qualcomm Innovation Center, Inc.
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by
the Linux Foundation
As of revision 174322, AArch64 back-end is compiled by default on all
architectures.
This means it'll be built and tested on all platforms and it'll get heavy
bombardment of the changes in the front- and middle-end without breaking
for too long.
Thanks Tim for the hard work!
cheers,
--renato
== Progress ==
* Boehm GC AArch64 support:
- Added information to the libatomic_ops patch, stiil under review
- backported AArch64 support to the GCC boehm-gc, test on-going
* Libunwind AArch64 support:
- started.
* Aarch64 porting meeting:
- request to move to eglibc 2.17 (done by Zhenqiang)
* LCA13:
- hotel and plane booked
== Next ==
* Validate boehm-gc GCC integration
* Libunwind AArch64 support
== This Week ===
* Joined Linaro, had welcome meeting with Matt
* Set up my office at home
* Laptop purchased and had it ready for development
* Exchanged emails with various people for getting started and read a lot
of administrative material on the wiki pages.
* Registered for the connect, figured out visa process and sent request to
Arwen for required documents.
* Attended Weekly toolchain call, stand-up call and 1:1 with Matt.
* Started work on GDB Testing in various configurations for analysis of
current GDB status on ARM.
* Build GDB on x86_64 machine and had test suite run in local and remote
configuration.
* Suffered from a bad run of cough n fever.
== Next week ==
* Continue with Analysis of GDB status on ARM, hopefully get to access a
pandaboard remotely and setup QEMU as well.
* Apply for Hong Kong Visa for the connect.
* Visit local market to resolve pending office setup and hardware
requirements.
* More Wiki reading and getting to know different processes at Linaro.
* 5th February public holiday in Pakistan.
Summary:
* Linaro toolchain binary 2013.01 release
* Move to using EGLIBC 2.17 in aarch64 build
Details:
1. Linaro toolchain binary.
* Document on how to do release test.
* Help Bero on 2013.01 release.
* Investigate lp:1107659. Crosstool-ng upstream has a workaround fix.
2. Move to using EGLIBC 2.17 in aarch64 build.
* Update Linux version to 3.7.0 since eglibc 2.17 configure
"checking for kernel header at least 3.7.0".
* Migrate local patches for eglibc 2.16 to 2.17 and update the config to 2.17
* Package build is OK. Test on simulation is ongoing.
3. Collect SPEC performance for different branch costs.
* The total results wave less than 0.5%.
* Three cases have 1-2% regression for smaller branch cost. Will
analyze the detail impact for the three cases.
Plan:
* Investigate coremark performance impact for different branch costs.
* Commit EGLIBC 2.17 changes
Planed leaves:
* Feb. 9 - 15: Chinese Spring Festival.
Best Regards!
-Zhenqiang
== Progress ==
* Backport
Tested and backported patch written to CSE optmize compares.
http://gcc.gnu.org/ml/gcc-patches/2013-02/msg00015.html
Ran GCC regression suite with ARMV8 foundation model for aarch64-none-elf
* Implementing GC sections support in bintuils
Had a quick look at changes need to be added.
Reading about relocations that needed to considered for
elf_backend_gc_sweep_hook
* Tried using Cbuild for buildign and testing gcc4.7 aarch64 branch.
ssh keys are not in sync. I am getting permission denied.
Cheking with David and Andy in valdation team.
Misc
* Attend tool chain weekly meet, and stand up call
* Attend usual AMD meetings
* Performed one x86_64 AMD task
* Booked tickets, hotel and got invite letter for Linaro connect.
== Next week ==
* Complete Cbuild experiments to test gcc svn aarch64 4.7 branch.
* Continue gc section support. Implement elf_backend_gc_sweep_hook
== Progress ==
* 64-bits ops in Neon: No news from upstream.
* vectorizer cost model: added possibility to tune the problem by
processor type.
* smin-umin: regression actually caused by a previous patch brought
when merging bswap16 support. Investigation on-going.
* managed to run spec2k on local board. Need to upload tools compiled
for hf platform.
* tcpanda heat problems: GCC testsuite passed.
* hot-cold partitioning: posted a follow-up to google patches on gcc-patches.
* looked at some codecs needing improvements (Neon intrinsics usage)
* feedback on lava/cbuild integration
* hotel/plane booked for Connect.
== Next ==
* handle 64-bits bitops in Neon feedback from upstream if any
* smin-umin: check bench results with probable culprit patch reverted.
Will need investigation on trunk too.
* vectorizer cost model: check bench results
* get more codecs sample codes.
Activity:
* calls and meetings
* last bits and pieces of ARM QEMU patches sent prior to upstream
1.4 hardfreeze
* continuing fixes for running a 3.8-rc4 vexpress kernel on QEMU:
* refactored the way we handle SYS_CFG vexpress registers so it's
possible to reasonably add support for more of them
* put together a basic implementation of the various regs
we're missing (oscillator, voltage, and some misc)
* ideally this would be done via variable-length qdev array
properties; started looking at how hard those would be to implement
* investigated linux-user bug where an edge case in use of
get_user_u32() had accidentally been broken; sent patch
* all travel/hotels now booked for Connect
-- PMM
== Progress ==
* Welcomed Omair to the team working on GDB
* Starting investigating cortex-a9 bootstrap failure
* Blueprints updating
* Preparing for two new starters next week
== Next week ==
* Welcome two new members to the team.
* Finish cortex-a9 bootstrapping investigations
== Future ==
* Run HOT/COLD partitioning benchmarks
* Analyse ARM results
* On x86_64 to see what the actual benefit we could get
* fix-gcc-multiarch-testing
* Come up with strawman proposal for updating testsuite to handle
testing with varying command-line options.
--
Matthew Gretton-Dann
Toolchain Working Group, Linaro
== Progress ==
* Maintenance
- Fixing ARM buildbots, poking people to fix bugs, keeping them green
- http://llvm.org/viewvc/llvm-project?view=rev&revision=173510
* Cost Model
- Fixing some bugs on the generic code
- http://llvm.org/viewvc/llvm-project?view=rev&revision=173691
- Adding some simple free cast (plus some infrastructure)
- http://llvm.org/viewvc/llvm-project?view=rev&revision=173849
* LLVM
- Investigating APFloat issue on Chromebook (bad libraries?)
- Clang miscompiles and show same synthoms, will play with options next
week
- AArch64 back-end in, to be built by default
* LAVA
- Got three last errors due to include path ('bits/predefs.h' file not
found)
- libc6-dev + libstdc++-dev have no effect, problem doesn't show on
buildbots
- Testing heating problem with multiple images (only 12.02 is good)
- Testing other boards, other images (with Dave)
* Friday Holiday
== Plan ==
* Try a bit more on the APFloat issue in Chromebook, but I think that's
just bad distro (ChrUbuntu), since no one else has this problem. Has anyone
put any Linaro image on a Chromebook?
* Continue working on getting faster builds on LAVA (quad-core origen,
Arndale, etc) with Dave Pigot.
* Continue micro-benchmarking the vectorization and updating the
cost-model. Start discussing the side-effects that are not modelled at all.
