Hi Ramana,
The attached file is a reference patch to add more dwarf/unwind info
in epilogue. Please help to review.
Without the patch, dwarf check fail for the following cases when
enabling shrink-wrap:
tst ... L1 //simple_return
push ...
...
pop ... //.cfa_offset is not 0
L1:
bx lr //common simple_return
Thanks!
-Zhenqiang
== other ==
* upstream patch review
* target-arm pull request
* reviewed/committed some boot wrapper patches from Tixy to init the CCI
* rebased various qemu kvm-related patches on current upstream master,
sent out RFC v2 for review (matching kernel -v12 patches); this is
getting some useful review feedback
* added check for 'does KVM support vfp-d32' rather than just
assuming it on the QEMU side
* discussed API for userspace to tell KVM the base address of the
VGIC with Avi, Christoffer [todo: write QEMU patches to go with
C's proposed kernel patches]
KVM blueprint progress tracker:
http://ex.seabright.co.nz/helpers/backlog?group_by=topic&colour_by=state&pr…
-- PMM
Summary:
* Test shrink-wrap.
* Identify the root cause of dwarf check fail when enabling shrink-wrap.
Details:
1. Test shrink-wrap on PandaBoard with Precise
* Bootstrap is OK
* No make check regression
* No performance change for SPEC2000 INT.
2. Identify the root case for dwarf check fail on ARM, while they are
OK on X86 and MIPS when enabling shrink-wrap.
* "push" related INSNs are set as RTX_FRAME_RELATED_P in prologue; but
not all "pop" related INSNs are set as RTX_FRAME_RELATED_P.
* Prologue uses REG_FRAME_RELATED_EXPR notes to tell the dwarf backend
SP is adjusted; but no such kind of notes in epilogue.
* In addition, dwarf info for INSNs to adjust SP are not handled
correctly in epilogue.
Plans:
* Verify shrink-wrap related bugs
* Prepare Linaro binary toolchain 2012.10 release
Best regards!
-Zhenqiang
The Linaro Toolchain Working Group is pleased to announce the 2012.10
release of both Linaro GCC 4.7 and Linaro GCC 4.6.
Linaro GCC 4.7 2012.10 is the seventh release in the 4.7 series. Based
off the latest GCC 4.7.2+svn191881 release, it includes ARM-focused
performance improvements and bug fixes.
Interesting changes include:
* Updates to GCC 4.7.2+svn191881
* Backport changes to use VLD1 for Neon quad loads.
Fixes:
* LP 1053348: Missing binary files
Linaro GCC 4.6 2012.10 is the twentieth release in the 4.6 series.
Based off the latest GCC 4.6.3+svn191880 release, this is the seventh
release after entering maintenance.
Interesting changes include:
* Updates to 4.6.3+svn191880
* A fix to LP #1029454
The source tarballs are available from:
https://launchpad.net/gcc-linaro/+milestone/4.7-2012.10https://launchpad.net/gcc-linaro/+milestone/4.6-2012.10
Downloads are available from the Linaro GCC page on Launchpad:
https://launchpad.net/gcc-linaro
More information on the features and issues are available from the
release pages:
https://launchpad.net/gcc-linaro/4.7/4.7-2012.10https://launchpad.net/gcc-linaro/4.6/4.6-2012.10
Mailing list: http://lists.linaro.org/mailman/listinfo/linaro-toolchain
Bugs: https://bugs.launchpad.net/gcc-linaro/
Questions? https://ask.linaro.org/
Interested in commercial support? Inquire at support(a)linaro.org
--
Matthew Gretton-Dann
Linaro Toolchain Working Group
matthew.gretton-dann(a)linaro.org
== Progress ==
* Completed Linaro ramp up process.
* Obtained my SSH security exception
* Crosstool-ng :
- Repoduced the aarch64 builds (bare and linux)
- Updated gcc to Linaro 2012.10 and submitted the merge request.
* Cbuild autobuilder :
- Not able to spawn a job on seabright (lack of permission)
== Misc ==
* One day off.
== Progress ==
* builtin_bswap16:
* Backported missing patch from thunk to Linaro/4.7. One test still
failing in the hard-float + thumb configuration because it forces
armv6 and results in a unsupported configuration: vfp-1 + hard-float.
Resumed discussion about a fix I proposed 3 weeks ago.
* PGO/hot-cold partitioning/Spec2k:
* read doc about spec2k
* started looking at 254.gap causing an ICE.
* started studying qemu traces about 175.vpr crashing when compiled
using profiling info. Cold code returns from a function using a
clobbered lr value.
== Next ==
* builtin_bswap16:
* complete backport to linaro-gcc 4.7.
* PGO/hot-cold partitioning/Spec2k:
* investigate compiler crash and runtime failures.
== Progress ==
* Deputising for Michael
* 4.6 and 4.7 2012.10 releases
* Included extra AArch64 merge into gcc-lianro/4.7
* AArch64 support
* Did another merge
* Starting another one.
* HOT/COLD partitioning for PGO
* PINGed If/Then/Else block conversion patch
* Started reworking postreload-gcse.c patch
* Ran into issues with vanilla gcc-linaro/4.7 not being able to
build for 'arm-none-linux-gnueabihf' without some configury magic
* Admin
* Interviewing
== Next Week ==
* Do another AArch64 merge.
* HOT/COLD partitioning for PGO:
* Get two current patches accepted upstream
* Post question upstream about register allocation 'mistakes' I am seeing
* symbol_ref splitting
* Test code alignment hypothesis
* Test v2 patch.
== Future ==
* Look at Cards for Vectorization, PGO and LTO with Michael.
--
Matthew Gretton-Dann
Linaro Toolchain Working Group
matthew.gretton-dann(a)linaro.org
Hi Guys,
I just found an issue in kernel code and thought we must have a
compilation warning for such cases.
Submitted a bug here:
https://bugs.launchpad.net/gcc-linaro/+bug/1064218
Problem description:
-------------------------
When we have following declaration:
struct foo {
char array[5];
....
};
And have a definition like:
struct foo foo_abc = {
.array = "12345",
};
Problem here is: size of array is 5 bytes and so we can actually add a
string with four characters only to it, as '\0' will take
an additional space.
But with my definition of foo_abc.. i had 5 characters + '\0'... that
will make it 6 and that will overflow the array..
What will actually happen here?
- compiler will not add '\0' at all?
-or it will go outside of boundaries of array?
--
viresh
On 5 October 2012 12:10, Rob Herring <robherring2(a)gmail.com> wrote:
> I've been scratching my head with a "scheduling while atomic" bug I
> started seeing on 3.6. I can easily reproduce this problem when doing a
> wget on my system. It ultimately seems to be a combination of factors.
> The "scheduling while atomic" bug is triggered in do_alignment which
> gets triggered by this code in net/ipv4/af_inet.c, line 1356:
>
> id = ntohl(*(__be32 *)&iph->id);
> flush = (u16)((ntohl(*(__be32 *)iph) ^ skb_gro_len(skb)) | (id ^ IP_DF));
> id >>= 16;
>
> This code compiles into this using "gcc version 4.6.3 (Ubuntu/Linaro
> 4.6.3-1ubuntu5)":
>
> c02ac020: e8920840 ldm r2, {r6, fp}
> c02ac024: e6bfbf3b rev fp, fp
> c02ac028: e6bf6f36 rev r6, r6
> c02ac02c: e22bc901 eor ip, fp, #16384 ; 0x4000
> c02ac030: e0266008 eor r6, r6, r8
> c02ac034: e18c6006 orr r6, ip, r6
>
> which generates alignment faults on the ldm. These are silent until this
> commit is applied:
Hi Rob. I assume that iph is something like:
struct foo {
u32 x;
char id[8];
};
struct foo *iph;
GCC merged the two adjacent loads of x and id into one ldm. This is
an ARM specific optimisation done in load_multiple_sequence() and
enabled with -fpeephole2.
I think the assembly is correct - GCC knows that iph is aligned and
knows the offsets of both x and id. Happy to be corrected if I'm
wrong, but I think the assembly is valid given the C code.
-- Michael
== Progress ==
* builtin_bswap16:
* backporting of builtin_bswap16 for ARM to Linaro/4.7: one of the
tests fails. Waiting for Ramana's return to see if he can help me
identify the patch needed from trunk.
* Useless zero extensions: despite Ulrich support I didn't succeed
to rewrite the bswap16 pattern in a way that would avoid
generating useless zero extension. Asking for help on gcc list
shows that it would actually need a complete new optimization
pass, which has already been discussed several times.
* PGO/hot-cold partitioning/Spec2k:
* looking at spec2k build & run system
* Branch reviews: approved Matt's october merges for Linaro-gcc 4.6 and 4.7.
Briefly looked at aarch64-4.7 merge request.
== Next ==
* builtin_bswap16:
* try to complete backport to linaro-gcc 4.7.
* PGO/hot-cold partitioning/Spec2k:
* investigate compiler crash and runtime failures.
* setup Snowball
