On Thu, 1 Dec 2011, Martin Jambor wrote:
Hi,
PR 50622 is an omission in load_assign_lhs_subreplacements, which
should force a gimple operand on a RHS of a gimple assignment if both
sides are new replacements of scalar types which are not gimple
registers, because they are partially
On Thu, Dec 1, 2011 at 11:13 PM, William J. Schmidt
wschm...@linux.vnet.ibm.com wrote:
Greetings,
Bug 39976 reported a degradation to 200.sixtrack wherein a hot
single-block loop is broken into two blocks. Investigation showed the
cause to be a redundant PHI statement in the block, which the
Sorry for the slow response, still catching up.
Joseph S. Myers jos...@codesourcery.com writes:
This code for doing the increment in from_mode comes from the fix for
PR 37451 and the follow-up fix for PR 37782
http://gcc.gnu.org/ml/gcc-patches/2008-09/msg01070.html
Dodji Seketeli do...@redhat.com writes:
Jason Merrill ja...@redhat.com writes:
I guess let's check DECL_ORIGINAL_TYPE instead of TREE_TYPE for alias
templates.
Like the below that I am currently bootstrapping?
Finally this is what passed bootstrap and testing on
x86_64-unknown-linux-gnu
On Thu, Dec 01, 2011 at 12:36:08PM +0100, Jakub Jelinek wrote:
On Thu, Dec 01, 2011 at 09:58:08PM +1030, Alan Modra wrote:
The GOMP_task change fixes a similar potential problem. Bootstrapped
and regression tested powerpc-linux. OK to apply?
PR libgomp/51376
* task.c
On Fri, Dec 2, 2011 at 3:28 AM, Andrew Pinski pins...@gmail.com wrote:
On Thu, Dec 1, 2011 at 6:08 PM, Joseph S. Myers jos...@codesourcery.com
wrote:
Ping. This patch
http://gcc.gnu.org/ml/gcc-patches/2011-11/msg02405.html is pending
review.
From my point of view, reverting my patch is
Tom de Vries tom_devr...@mentor.com writes:
On 27/11/11 23:59, Eric Botcazou wrote:
No, DELETED_LABEL notes still work just fine. It depends on how you
remove the label and replace it with a note, and Tom isn't showing
what he did, so...
I agree that there is no obvious reason why just
On 12/01/11 13:01, Ramana Radhakrishnan wrote:
Sergio: Other than a few minor tweaks to the Changelog it largely
looks obvious to me.
Bernd, could you take another look at this since this is now shared
with the c6x backend ?
Doesn't look like it would cause problems. I have no idea what
This fixes a recent regression on the testcase of PR47259 where
we bring local a global register variable during WPA time and
then mangle its register name spec during ltrans. Oops.
It doesn't make sense to bring local global register variables.
Fixed by making varpool_externally_visible_p
On Tue, Nov 29, 2011 at 17:44, Kaz Kojima kkoj...@rr.iij4u.or.jp wrote:
Uli, could you please approve the libc patch?
Has the gcc patch been committed?
Hi All,
I'm trying to implement DImode shifts using ARM NEON instructions. This
wouldn't be difficult in itself, but making it play nice with the
existing implementation is causing me problems. I'd like a few
suggestions/pointers/comments to help me get this right, please.
The existing
Ulrich Drepper drep...@gmail.com wrote:
Has the gcc patch been committed?
Yes, it has been committed as revision 181825 on gcc trunk.
Regards,
kaz
On Tue, 29 Nov 2011, Hans-Peter Nilsson wrote:
On Tue, 29 Nov 2011, Richard Henderson wrote:
On 11/28/2011 08:49 PM, Hans-Peter Nilsson wrote:
On Sat, 26 Nov 2011, Richard Henderson wrote:
The m68k-linux failure for the various omp atomic tests
is due to the fact that
This rewrites SSA operand checking to basically check that the
operand status is the same as if you'd do an update_stmt on the
stmt. The current checking code does not properly verify that
all stmts are properly updated as it misses missing and swapped
operands for example (bugs like that I've
Hi,
On Fri, 2 Dec 2011, Richard Guenther wrote:
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2011-12-01 Jakub Jelinek ja...@redhat.com
PR debug/50317
* tree-ssa-dce.c (remove_dead_stmt): Add a debug stmt when removing
as unnecessary a store
On Tue, 29 Nov 2011, Vladimir Makarov wrote:
3. The patch rewrite the way of dealing with the secondary memory
moves in constraint pass of LRA. Previously we generated secondary
moves if macro SECONDARY_MEMORY_NEEDED says so. Unfortunately, the
macro is usually defined inaccurately.
On Fri, 2011-12-02 at 10:24 +0100, Richard Guenther wrote:
On Thu, Dec 1, 2011 at 11:13 PM, William J. Schmidt
wschm...@linux.vnet.ibm.com wrote:
Greetings,
Bug 39976 reported a degradation to 200.sixtrack wherein a hot
single-block loop is broken into two blocks. Investigation showed
On 1 Dec 2011, at 23:28, Iain Sandoe wrote:
now I'm slightly confused - do we need to preserve if across the
call or not?
erm. not well phrased.
I am trying to get a grasp on what determines the set of registers
that should be saved.
Initially, I was thinking that it was the
Hi,
On Fri, 2 Dec 2011, William J. Schmidt wrote:
- tree def = gimple_get_lhs (stmt);
+ /* If this is a PHI, we only want to consider it if all of its
+ arguments are SSA names (which are known to be defined in a
+ single place). This avoids errors when dealing with
On Fri, 2011-12-02 at 14:59 +0100, Michael Matz wrote:
Hi,
On Fri, 2 Dec 2011, William J. Schmidt wrote:
- tree def = gimple_get_lhs (stmt);
+ /* If this is a PHI, we only want to consider it if all of its
+ arguments are SSA names (which are known to be defined in a
Am Fri 02 Dec 2011 01:38:46 PM CET schrieb Richard Guenther
rguent...@suse.de:
This fixes a recent regression on the testcase of PR47259 where
we bring local a global register variable during WPA time and
then mangle its register name spec during ltrans. Oops.
indeed... (I am not in
OK.
Jason
This change implements a reversible iterator for multiset containers.
An iterator can either be partial, visiting only some of the items in the
container (in which case the start position is specified), or complete,
visiting all of them. The iterator caches the start position during its
This patch incorporates the support for AI95-0041. For the purposes of the
rules for allowing allocated unconstrained objects, any ancestor that has
a constrained partial view causes the rules to apply.
In addition, in a generic body, 3.10.2(27.2/2) is checked assuming that any
untagged
The iterator for visiting children of a node in a multiway tree must check the
value of the Parent parameter to ensure that it is non-null, and that it
actually designates a node in the tree.
There were also several instances where cursor values returned by iterator
operations were not
The side effect removal machinery may generate illegal Ada code to avoid the
usage of access types and 'reference in Alfa mode. Since this is legal code
with respect to theorem proving, do not emit the warning.
Tested on x86_64-pc-linux-gnu, committed on trunk
2011-12-02 Yannick Moy
This patch does not modify the functionality of the compiler. It
moves semantic routines from Sem_Util to Sem_Aux to have them
available in ASIS.
Tested on x86_64-pc-linux-gnu, committed on trunk
2011-12-02 Javier Miranda mira...@adacore.com
* sem_util.ads, sem_util.adb, sem_aux.ads,
This patch corrects a buffer issue which may lead to bogus expanded names at
link time. The problem is initiated while creating the external name of a
tagged type whose scope is an overloaded name. This places useless data in the
Homonym_Numbers buffer which is then reused when building the
The iterators for the multiway trees are now implemented as a type
hierarchy. Iterating over a tree is the same as iterating over a subtree
starting from the root, and so the tree iterator forwards the request to the
subtree iterator.
Tested on x86_64-pc-linux-gnu, committed on trunk
2011-12-02
On 12/02/2011 01:48 AM, Arnaud Charlet wrote:
The attached patch is necessary to let the gcc head
compile Ada for *-*-rtems*. Other than terminals.c,
the files impacted are RTEMS specific. OK to commit?
OK
Thanks. Committed.
--
Joel Sherrill, Ph.D. Director of Research
Hans-Peter Nilsson writes:
BTW, on the topic, I cringe whenever I
see futexes expressed as plain int, they absolutely have to
have at least natural alignment which is not always true e.g. in
structs. People, please keep the atomic types
target-overridable in libraries.
+1 for
Thanks, Andreas. You are right in that fully peeling a loop is done by
a different code path (peel_loops_completely() and earlier in the tree
unroller).
Teresa
On Fri, Dec 2, 2011 at 12:54 AM, Andreas Krebbel
kreb...@linux.vnet.ibm.com wrote:
On Thu, Dec 01, 2011 at 11:39:36PM -0800, Teresa
This removes the CALL_CANNOT_INLINE_P tree flag. As discussed in
the thread about the duplicate edge/gimple stmt flag.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
Richard.
2011-12-02 Richard Guenther rguent...@suse.de
* tree.h (CALL_CANNOT_INLINE_P): Remove.
*
Hello,
I have conflated the handling of two PRs here, because I think they
are related.
Consider this short example that illustrates the issue of PR
c++/51239:
struct S {};
templatetypename T, typename...
using head = T;
templatetypename... Ts
using x = headTs...;//#1
In
Hi,
I have been testing with this for almost a month. It is
my attempt to follow the changes I think Joseph made
to other rs6000 targets. If this change looks right,
I would like to commit it.
Test results have been posted for it.
Thanks.
2011-12-02 Joel Sherrill joel.sherr...@oarcorp.com
2011/12/2 Michael Zolotukhin michael.v.zolotuk...@gmail.com:
Hi,
This patch increases array sizes in tests from vect.exp suite, thus
enabling 256-bit vectorization where it's available.
Ok for trunk?
Shouldn't we add a variant for each testcase so that we still
excercise both 128-bit and
This is the 2nd piece of the stmt flag removal.
Bootstrapped on x86_64-unknown-linux-gnu, testing in progress.
I'll apply the short series tomorrow unless there are any objections
or comments.
Thanks,
Richard.
2011-12-02 Richard Guenther rguent...@suse.de
* cgraph.c
On Fri, 2 Dec 2011, Joel Sherrill wrote:
2011-12-02 Joel Sherrill joel.sherr...@oarcorp.com
* config/rs6000/rtems.h: Switch to using global_options_set
in SUBSUBTARGET_OVERRIDE_OPTIONS.
Is it deliberate that you are removing the first part of each if
condition (thus, no longer
On 12/02/2011 10:38 AM, Joseph S. Myers wrote:
On Fri, 2 Dec 2011, Joel Sherrill wrote:
2011-12-02 Joel Sherrilljoel.sherr...@oarcorp.com
* config/rs6000/rtems.h: Switch to using global_options_set
in SUBSUBTARGET_OVERRIDE_OPTIONS.
Is it deliberate that you are removing the first
gcc-patches-ow...@gcc.gnu.org wrote on 02/12/2011 06:23:25 PM:
Hi,
This patch increases array sizes in tests from vect.exp suite, thus
enabling 256-bit vectorization where it's available.
Ok for trunk?
--- a/gcc/testsuite/gcc.dg/vect/slp-24.c
+++ b/gcc/testsuite/gcc.dg/vect/slp-24.c
...
Hi,
this fixes a thinko in my recent fix for the PR: just be consistent with
our other inserters.
Testing x86_64-linux, will commit soon.
Thanks,
Paolo.
///
2011-12-02 Paolo Carlini paolo.carl...@oracle.com
* include/std/iomanip (put_money): Fix thinko, use __err
Sorry for the previous brain-fart! Here's a revised patch.
It seems like a fair amount of rip-up to avoid keeping the PHI state
around between blocks, so I just check to ensure the PHI definitions
occur in the same block before recording their equivalence.
lookup_avail_expr may return a constant
Shouldn't we add a variant for each testcase so that we still
excercise both 128-bit and 256-bit vectorization paths?
These tests are still good to test 128-bit vectorization, the changes
was made just to make sure that 256-bit vectorization is possible on
the tests.
Actually, It's just first
Hi,
On Fri, 2 Dec 2011, William J. Schmidt wrote:
It seems like a fair amount of rip-up to avoid keeping the PHI state
around between blocks, so I just check to ensure the PHI definitions
occur in the same block before recording their equivalence.
Then you should at least mix the BB
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 12/02/11 10:40, Michael Matz wrote:
Hi,
On Fri, 2 Dec 2011, William J. Schmidt wrote:
It seems like a fair amount of rip-up to avoid keeping the PHI
state around between blocks, so I just check to ensure the PHI
definitions occur in the
On 12/02/2011 11:57 AM, Joseph S. Myers wrote:
On Fri, 2 Dec 2011, Joel Sherrill wrote:
On 12/02/2011 10:38 AM, Joseph S. Myers wrote:
On Fri, 2 Dec 2011, Joel Sherrill wrote:
2011-12-02 Joel Sherrilljoel.sherr...@oarcorp.com
* config/rs6000/rtems.h: Switch to using
On Fri, 2011-12-02 at 10:52 -0700, Jeff Law wrote:
On 12/02/11 10:40, Michael Matz wrote:
Hi,
On Fri, 2 Dec 2011, William J. Schmidt wrote:
It seems like a fair amount of rip-up to avoid keeping the PHI
state around between blocks, so I just check to ensure the PHI
definitions
Bernd Schmidt ber...@codesourcery.com writes:
On 12/01/11 13:01, Ramana Radhakrishnan wrote:
Sergio: Other than a few minor tweaks to the Changelog it largely
looks obvious to me.
Bernd, could you take another look at this since this is now shared
with the c6x backend ?
Doesn't look like
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 12/02/11 11:10, William J. Schmidt wrote:
I see a mention of creating equivalences for redundant PHIs? Are
we just trying to determine that two PHIs are going to result in
the same value?
Jeff, see comment #37 in
Ramana Radhakrishnan ramana.radhakrish...@linaro.org writes:
Sergio: Other than a few minor tweaks to the Changelog it largely
looks obvious to me.
Hello Ramana,
Thanks for the review. Here is the updated version of the patch.
I asked Tom Tromey to commit it for me, since I don't have write
Denis Chertykov wrote:
2011/11/29 Georg-Johann Lay a...@gjlay.de:
Ian Lance Taylor wrote:
Georg-Johann Lay a...@gjlay.de writes:
So if a frontend can define address spaces and it is a generic feature, the
question is how to get the name of an address space in a generic, language
independent
http://gcc.gnu.org/viewcvs?view=revisionrevision=181933
Committed the following, obvious fix to a test case that assumed
sizeof(int) = 4:
Johann
--- trunk/gcc/testsuite/gcc.c-torture/execute/vector-subscript-1.c
2011/12/02
17:44:28181932
+++
Michael Zolotukhin michael.v.zolotuk...@gmail.com wrote on 02/12/2011
08:11:41 PM:
Please don't change initial values to 0, we want to check that
everything
works fine for non-zeros as well.
There are several other occasions in the patch.
Please check the update patch (attached).
This
This patch is part 1 of a cleanup needed to support namespace aliases.
When processing a namespace aliase, we cannot access its binding level
because it is NULL. Besides, it does not make sense to do anything
with it, since the binding level for the original namespace would've
been processed
Have you uploaded the revised patch?
David
http://codereview.appspot.com/5416043/diff/1/gcc/config/i386/i386.c
File gcc/config/i386/i386.c (right):
http://codereview.appspot.com/5416043/diff/1/gcc/config/i386/i386.c#newcode10927
gcc/config/i386/i386.c:10927: + is later renamed to
On Fri, 2011-12-02 at 11:28 -0700, Jeff Law wrote:
On 12/02/11 11:10, William J. Schmidt wrote:
I see a mention of creating equivalences for redundant PHIs? Are
we just trying to determine that two PHIs are going to result in
the same value?
Jeff, see comment #37 in
The second part of the cleanup prepares support for writing merge keys
and merge bodies for structures other than ASTs. The problem we are
having now is that when merging the bindings of namespace, the reader
creates an empty binding level, but after the merging is done, we go
on to create
Hi!
As reported by Michael, vect-116.c testcase fails with -mavx2, the
problem is that mulv32qi pattern computes wrong result, the second and
third quarters of the vector are swapped compared to what it should
contain. This is because we can't use vec_extract_even_odd for V32QI
when we prepared
On Fri, Dec 2, 2011 at 9:51 AM, Dodji Seketeli do...@redhat.com wrote:
After talking with you offline, we settled on using the existing
BOUND_TEMPLATE_TEMPLATE_PARM tree to represent this new construct.
The alias template and its unbound arguments are stored in the
TYPE_TEMPLATE_INFO of the
Denis Chertykov wrote:
2011/11/29 Georg-Johann Lay:
For devices with 8-bit SP reading the high byte SP_H of SP will get garbage.
The patch uses CLR instead of IN SP_H to read the high part of SP.
There are two issues with this patch:
== 1 ==
I cannot really test it because for devices
Hi!
As I found during investigation of PR51387, e.g. on the attached testcase
we generate pretty bad code (for f1 even with bigger N like 256 for avx2),
because after vectorization cunroll unrolls the loops completely and we
end up with lots of VEC_PACK_TRUNC_EXPR etc. expressions with VECTOR_CST
On Fri, 2011-12-02 at 13:08 -0600, William J. Schmidt wrote:
On Fri, 2011-12-02 at 11:28 -0700, Jeff Law wrote:
On 12/02/11 11:10, William J. Schmidt wrote:
I see a mention of creating equivalences for redundant PHIs? Are
we just trying to determine that two PHIs are going to
This patch to the libgo script merge.sh does the following:
* Permit specifying a revision on the command line.
* Add various new C files to the merge, reflecting their addition to the
library.
* Automatically converts the Unicode character · used in the master
library C code to _ before
On Fri, Dec 02, 2011 at 02:27:40PM +0100, Richard Guenther wrote:
This change seems wrong. We are turning valid gimple
# DEBUG D#2 = transfer.0 [with addres taken]
into invalid one
# DEBUG D#2 = transfer.0 [without address taken]
once you update that stmt with update_stmt you'll
This patch updates the Go library to the weekly.2011-11-01 release (I
know that was a month ago). This patch includes a change to the
reflection string for the empty interface type to match a similar change
to the other Go compiler; the change to gccgo was by Rémy Oudompheng.
The patch is too
Teresa Johnson tejohn...@google.com writes:
Interesting optimization. I would be concerned a little bit
about compile time, does it make a measurable difference?
The attached patch detects loops containing instructions that tend to
incur high LCP (loop changing prefix) stalls on Core i7, and
http://codereview.appspot.com/5416043/diff/6001/gcc/config/i386/i386.c
File gcc/config/i386/i386.c (right):
http://codereview.appspot.com/5416043/diff/6001/gcc/config/i386/i386.c#newcode10881
gcc/config/i386/i386.c:10881: + '_function_patch_epilogue'. The
backpointer section can be used to
Hi,
This addresses all of the Go compilation issues on the
head except one.
Ian.. Is this OK to commit? Or do you have suggestions
on how to make it more general?
Thanks.
2011-12-02 Joel Sherrill joel.sherr...@oarcorp.com
* runtime/go-signal.c: Add conditional on SIGPROF.
*
;
+/* Determine whether LOOP contains floating-point computation. */
+bool
+loop_has_FP_comp(struct loop *loop)
+{
+ rtx set, dest;
This probably should be extended to detect other long latency
operations in the future.
+
+ if (ix86_tune != PROCESSOR_COREI7_64
+ ix86_tune !=
On Fri, Dec 2, 2011 at 11:36 AM, Andi Kleen a...@firstfloor.org wrote:
Teresa Johnson tejohn...@google.com writes:
Interesting optimization. I would be concerned a little bit
about compile time, does it make a measurable difference?
The attached patch detects loops containing instructions
OK, one more version. This removes the basic block test and instead
implements Michael's suggestion:
On Fri, 2011-12-02 at 18:40 +0100, Michael Matz wrote:
But I wonder why it's not enough to just do a push/pop sequence on
avail_exprs_stack around your new PHI processing in
On Fri, Dec 2, 2011 at 11:36 AM, Andi Kleen a...@firstfloor.org wrote:
Teresa Johnson tejohn...@google.com writes:
Interesting optimization. I would be concerned a little bit
about compile time, does it make a measurable difference?
I haven't measured compile time explicitly, but I don't it
On Fri, 2 Dec 2011, Mikael Pettersson wrote:
but futexes must be 32-bit aligned (or at least not cross page
boundaries).
Don't mix up futexes with hardware-mandated atomic alignment
(except that preferably the letter should not be more strict).
Linux futexes must be 32-bit aligned *for all
I've committed this patch that adds support to gcov to distinguish unexecuted
exceptional lines. Normally unexecuted lines are marked by '#'. This patch
augments that so that lines of code that are only reachable by an exceptional
path (a catch clause in C++), are marked by '=' (half
On Fri, Dec 2, 2011 at 12:24 PM, Iyer, Balaji V balaji.v.i...@intel.com wrote:
Hello Everyone,
I found that patch 4 stepped on parts of patch 3. Here is the updated
patches (Please notice patch 4, 5 and 6 are replaced).
Thanks,
Balaji V. Iyer.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On 12/02/11 12:27, William J. Schmidt wrote:
Erm, wait. How are PHIs in different blocks going to have the
same incoming edges? (I was thinking of control dependence edges,
but these are just regular control flow incoming edges, right?) So
On Fri, 2011-12-02 at 13:46 -0700, Jeff Law wrote:
On 12/02/11 12:27, William J. Schmidt wrote:
Erm, wait. How are PHIs in different blocks going to have the
same incoming edges? (I was thinking of control dependence edges,
but these are just regular control flow incoming edges,
Found via Reinhold Bader's test suite: If a component is public, it
remains public even if the extended type has PRIVATE.
Build and regtested on x86-64-linux.
OK for the trunk?
Tobias
2011-12-02 Tobias Burnus bur...@net-b.de
PR fortran/51378
* symbol.c (gfc_find_component): Fix access
This patches fixes my previous MOVE_ALLOC patch. The standard states for
TO: It shall be polymorphic if FROM is polymorphic.
I somehow read this bijectively, but the it is actually allowed to have
a nonpolymorphic FROM with a polymorphic TO. Thanks for Damian for
finding this.
Build and
On 12/02/2011 11:18 AM, Jakub Jelinek wrote:
PR target/51387
* config/i386/sse.md (mulmode3 with VI1_AVX2 iterator): For
V32QImode use { 0,2,..,14,32,34,..,46,16,18,..,30,48,50,..,62 }
permutation instead of extract even permutation.
Ok.
r~
On Fri, Dec 2, 2011 at 2:35 PM, Iyer, Balaji V balaji.v.i...@intel.com wrote:
Hello H. J.,
Here are the fixed patches. I have numbered their order from 1 through 4.
The ChangeLog entries should be added to the beginning of
ChangeLog, not in the middle.
--
H.J.
This patch adds support for doing merge reads of lang_decl by
processing PPH_RECORD_START_MERGE_BODY records. I need it for
the next patch that will emit the merge keys for namespaces.
Tested on x86_64.
Diego.
* pph-in.c (pph_in_lang_decl_start): Rename from pph_in_ref_lang_decl.
On 12/02/2011 05:37 AM, Iain Sandoe wrote:
Richard: things that I did, intentionally, differently (and I'm not sure are
correct).
1. I saved the CR
2. Once the vrs are saved, I update the VRsave mask to reflect that.
Don't update VRsave. This gives the OS license to clobber those
Hi Richard,
On 2 Dec 2011, at 22:59, Richard Henderson wrote:
On 12/02/2011 05:37 AM, Iain Sandoe wrote:
Richard: things that I did, intentionally, differently (and I'm not
sure are correct).
1. I saved the CR
2. Once the vrs are saved, I update the VRsave mask to reflect that.
Don't
On 12/02/2011 04:25 AM, Bernd Schmidt wrote:
Doesn't look like it would cause problems. I have no idea what
builtin_frob_return_addr does but it appears to exist everywhere.
It's for adjusting the return address in magic ways. E.g. Sparc pc+8 for
structure returns, ARM low bit for thumb
This patch adds automatic dependency generation to the packages in
libgo. This saves me from having to fix all the dependencies each time
I import a new version of the library. As can be seen in this patch, I
really need to learn how to use autogen. Bootstrapped and ran Go
testsuite on
On Fri, 2 Dec 2011, Joel Sherrill wrote:
OK. I obviously read too much into the other uses.
I did not intend to change semantics just account for
the change making this not compile.
How does the the new version look?
This version is OK.
--
Joseph S. Myers
jos...@codesourcery.com
Hi,
PR51347 shows up a problem due to the TM IPA rework.
tree_function_versioning segfault because the cfg of old_decl (alias) is
NULL.
Indeed, an alias can get called but tm cg data are in the parent of the
alias.
Bootstrapped and regtested.
Thanks,
Patrick.
ChangeLog
2011-12-02 Patrick
On 27 November 2011 00:24, Gerald Pfeifer wrote:
On the way I spotted an odd reference to GCC. Looking at the
overall document, it occurs to me that
- newer versions of GCC are not covered, and
- references to GCC generally are of the form gcc-X.Y instead of GCC X.Y.
Is this something
I have updated the libgo library to the weekly.2011-11-02 release of the
master library. The only changes in this patch are to switch from using
the type os.Error to the new predeclared type error, and to add a new
errors package providing a couple of convenience functions. This causes
90 matches
Mail list logo