Re: [x86, 2/n] Replace builtins with vector extensions

[Uros Bizjak]

On Tue, Nov 4, 2014 at 9:31 PM, Marc Glisse  wrote:
> Ping?

Uh, yes, LGTM.

(I was under impression that I already OK'd this relatively
non-controversial patch. The effect of having too much open tasks in
parallel, I guess.)

Thanks,
Uros.

> https://gcc.gnu.org/ml/gcc-patches/2014-10/msg01808.html
>
>
> On Sat, 18 Oct 2014, Marc Glisse wrote:
>
>> Hello,
>>
>> this time, +-* for 128 bit integer vectors. I am using an unsigned type so
>> the compiler knows that we expect wrapping. I don't know why Intel's
>> description of mullo insists that the multiplication is signed, that only
>> matters for the high part...
>>
>> Next parts (waiting for approval for this one) should be:
>> - same thing with 256 and 512 bit integer vectors
>> - & | ^ (integer only)
>>
>> Maybe (or it can wait until the next release):
>> - < > == abs min max (integer only)
>>
>>
>> 2014-10-20  Marc Glisse  
>>
>> * config/i386/emmintrin.h (__v2du, __v4su, __v8hu, __v16qu): New
>> typedefs.
>> (_mm_add_epi8, _mm_add_epi16, _mm_add_epi32, _mm_add_epi64,
>> _mm_sub_epi8, _mm_sub_epi16, _mm_sub_epi32, _mm_sub_epi64,
>> _mm_mullo_epi16): Use vector extensions instead of builtins.
>> * config/i386/smmintrin.h (_mm_mullo_epi32): Likewise.
>>
>>
>
> --
> Marc Glisse

RE: [PATCH][6/n] Merge from match-and-simplify, make forwprop fold all stmts

[Zhenqiang Chen]

The patch leads to big regression for float operators on target without hard
fpu support due to register shuffle.

Please refer https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63743 for more
detail.

Thanks!
-Zhenqiang

> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Jeff Law
> Sent: Saturday, October 25, 2014 1:48 AM
> To: Richard Biener; gcc-patches@gcc.gnu.org
> Cc: Jakub Jelinek
> Subject: Re: [PATCH][6/n] Merge from match-and-simplify, make forwprop
> fold all stmts
> 
> On 10/24/14 07:16, Richard Biener wrote:
> >
> > This patch makes GIMPLE forwprop fold all statements, following
> > single-use SSA edges only (as suggested by Jeff and certainly how this
> > will regress the least until we replace manual simplification code
> > that does not restrict itself this way).
> >
> > forwprop is run up to 4 times at the moment (once only for -Og, not at
> > all for -O0), which still seems reasonable.  IMHO the forwprop pass
> > immediately after inlining is somewhat superfluous, it was added there
> > just for its ADDR_EXPR propagation.  We should eventually split this
> > pass into two.
> >
> > Note that just folding what we propagated into (like the SSA
> > propagators do during substitute-and-fold phase) will miss cases where
> > we propagate into a stmt feeding the one we could simplify.  Unless we
> > always fold all single-use (and their use) stmts we have to fold
> > everything from time to time.  Changing how / when we fold stuff is
> > certainly sth to look after with fold_stmt now being able to follow
> > SSA edges.
> >
> > Bootstrapped on x86_64-unknown-linux-gnu, testing still in progress.
> >
> >  From earlier testing I remember I need to adjust a few testcases that
> > don't expect the early folding - notably two strlenopt cases
> > (previously XFAILed but then PASSed again).
> >
> > I also expect to massage the single-use heuristic as I get to merging
> > the patterns I added for the various forwprop manual pattern matchings
> > to trunk (a lot of them do not restrict themselves this way).
> >
> > Does this otherwise look ok?
> >
> > Thanks,
> > Richard.
> >
> > 2014-10-24  Richard Biener  
> >
> > * tree-ssa-forwprop.c: Include tree-cfgcleanup.h and
tree-into-ssa.h.
> > (lattice): New global.
> > (fwprop_ssa_val): New function.
> > (fold_all_stmts): Likewise.
> > (pass_forwprop::execute): Finally fold all stmts.
> Seems reasonable.  After all, we can iterate on the single-use heuristic.
> 
> jeff
> 

Re: [AArch64, Patch] Restructure arm_neon.h vector types's implementation(Take 2).

[Tejas Belagod]

On 03/11/14 16:49, Marcus Shawcroft wrote:

On 1 October 2014 09:26, Tejas Belagod  wrote:

Hi,

Returning to this old thread,

https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02285.html

here is a patch after a few rounds of review comments, specifically:

https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02248.html
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02285.html
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00566.html
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00324.html

Sorry for the delay in respinning this patch.

Tested on aarch64-none-elf. OK for trunk?

Thanks,
Tejas.

Changelog:

2014-10-01  Tejas Belagod  

 * config/aarch64/aarch64-builtins.c
 (aarch64_build_scalar_type): Remove.
 (aarch64_scalar_builtin_types, aarch64_simd_type,
 aarch64_simd_type, aarch64_mangle_builtin_scalar_type,
 aarch64_mangle_builtin_vector_type,
 aarch64_mangle_builtin_type, aarch64_simd_builtin_std_type,
 aarch64_lookup_simd_builtin_type, aarch64_simd_builtin_type,
 aarch64_init_simd_builtin_types,
 aarch64_init_simd_builtin_scalar_types): New.
 (aarch64_init_simd_builtins): Refactor.
 (aarch64_init_crc32_builtins): Fixup with qualifier.
 * config/aarch64/aarch64-protos.h
 (aarch64_mangle_builtin_type): Export.
 * config/aarch64/aarch64-simd-builtin-types.def: New.
 * config/aarch64/aarch64.c (aarch64_simd_mangle_map): Remove.
 (aarch64_mangle_type): Refactor.
 * config/aarch64/arm_neon.h: Declare vector types based on
 internal types.
 * config/aarch64/t-aarch64: Update dependency.


OK /Marcus



Thanks. Rebased, retested on aarch64-none-elf and committed as r217114.

Tejas.

Re: [AArch64, Patch] Restructure arm_neon.h vector types's implementation(Take 2).

[Tejas Belagod]

On 05/11/14 08:28, Tejas Belagod wrote:

On 03/11/14 16:49, Marcus Shawcroft wrote:

On 1 October 2014 09:26, Tejas Belagod  wrote:

Hi,

Returning to this old thread,

https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02285.html

here is a patch after a few rounds of review comments, specifically:

https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02248.html
https://gcc.gnu.org/ml/gcc-patches/2014-06/msg02285.html
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00566.html
https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00324.html

Sorry for the delay in respinning this patch.

Tested on aarch64-none-elf. OK for trunk?

Thanks,
Tejas.

Changelog:

2014-10-01  Tejas Belagod  

  * config/aarch64/aarch64-builtins.c
  (aarch64_build_scalar_type): Remove.
  (aarch64_scalar_builtin_types, aarch64_simd_type,
  aarch64_simd_type, aarch64_mangle_builtin_scalar_type,
  aarch64_mangle_builtin_vector_type,
  aarch64_mangle_builtin_type, aarch64_simd_builtin_std_type,
  aarch64_lookup_simd_builtin_type, aarch64_simd_builtin_type,
  aarch64_init_simd_builtin_types,
  aarch64_init_simd_builtin_scalar_types): New.
  (aarch64_init_simd_builtins): Refactor.
  (aarch64_init_crc32_builtins): Fixup with qualifier.
  * config/aarch64/aarch64-protos.h
  (aarch64_mangle_builtin_type): Export.
  * config/aarch64/aarch64-simd-builtin-types.def: New.
  * config/aarch64/aarch64.c (aarch64_simd_mangle_map): Remove.
  (aarch64_mangle_type): Refactor.
  * config/aarch64/arm_neon.h: Declare vector types based on
  internal types.
  * config/aarch64/t-aarch64: Update dependency.


OK /Marcus



Thanks. Rebased, retested on aarch64-none-elf and committed as r217114.

Tejas.


Sorry, forgot to attach the rebased patch - here you go.
diff --git a/gcc/config/aarch64/aarch64-builtins.c 
b/gcc/config/aarch64/aarch64-builtins.c
index 527445c..c0881e6 100644
--- a/gcc/config/aarch64/aarch64-builtins.c
+++ b/gcc/config/aarch64/aarch64-builtins.c
@@ -363,257 +363,335 @@ static GTY(()) tree 
aarch64_builtin_decls[AARCH64_BUILTIN_MAX];
 #define NUM_DREG_TYPES 6
 #define NUM_QREG_TYPES 6
 
-/* Return a tree for a signed or unsigned argument of either
-   the mode specified by MODE, or the inner mode of MODE.  */
-tree
-aarch64_build_scalar_type (machine_mode mode,
-  bool unsigned_p,
-  bool poly_p)
-{
-#undef INT_TYPES
-#define INT_TYPES \
-  AARCH64_TYPE_BUILDER (QI) \
-  AARCH64_TYPE_BUILDER (HI) \
-  AARCH64_TYPE_BUILDER (SI) \
-  AARCH64_TYPE_BUILDER (DI) \
-  AARCH64_TYPE_BUILDER (EI) \
-  AARCH64_TYPE_BUILDER (OI) \
-  AARCH64_TYPE_BUILDER (CI) \
-  AARCH64_TYPE_BUILDER (XI) \
-  AARCH64_TYPE_BUILDER (TI) \
-
-/* Statically declare all the possible types we might need.  */
-#undef AARCH64_TYPE_BUILDER
-#define AARCH64_TYPE_BUILDER(X) \
-  static tree X##_aarch64_type_node_p = NULL; \
-  static tree X##_aarch64_type_node_s = NULL; \
-  static tree X##_aarch64_type_node_u = NULL;
-
-  INT_TYPES
-
-  static tree float_aarch64_type_node = NULL;
-  static tree double_aarch64_type_node = NULL;
-
-  gcc_assert (!VECTOR_MODE_P (mode));
-
-/* If we've already initialised this type, don't initialise it again,
-   otherwise ask for a new type of the correct size.  */
-#undef AARCH64_TYPE_BUILDER
-#define AARCH64_TYPE_BUILDER(X) \
-  case X##mode: \
-if (unsigned_p) \
-  return (X##_aarch64_type_node_u \
- ? X##_aarch64_type_node_u \
- : X##_aarch64_type_node_u \
- = make_unsigned_type (GET_MODE_PRECISION (mode))); \
-else if (poly_p) \
-   return (X##_aarch64_type_node_p \
- ? X##_aarch64_type_node_p \
- : X##_aarch64_type_node_p \
- = make_unsigned_type (GET_MODE_PRECISION (mode))); \
-else \
-   return (X##_aarch64_type_node_s \
- ? X##_aarch64_type_node_s \
- : X##_aarch64_type_node_s \
- = make_signed_type (GET_MODE_PRECISION (mode))); \
-break;
+/* Internal scalar builtin types.  These types are used to support
+   neon intrinsic builtins.  They are _not_ user-visible types.  Therefore
+   the mangling for these types are implementation defined.  */
+const char *aarch64_scalar_builtin_types[] = {
+  "__builtin_aarch64_simd_qi",
+  "__builtin_aarch64_simd_hi",
+  "__builtin_aarch64_simd_si",
+  "__builtin_aarch64_simd_sf",
+  "__builtin_aarch64_simd_di",
+  "__builtin_aarch64_simd_df",
+  "__builtin_aarch64_simd_poly8",
+  "__builtin_aarch64_simd_poly16",
+  "__builtin_aarch64_simd_poly64",
+  "__builtin_aarch64_simd_poly128",
+  "__builtin_aarch64_simd_ti",
+  "__builtin_aarch64_simd_uqi",
+  "__builtin_aarch64_simd_uhi",
+  "__builtin_aarch64_simd_usi",
+  "__builtin_aarch64_simd_udi",
+  "__builtin_aarch64_simd_ei",
+  "__builtin_aarch64_simd_oi",
+  "__builtin_aarch64_simd_ci",
+  "__builtin_aarch64_simd_xi",
+  NUL

RE: [Ping] [PATCH, 9/10] aarch64: generate conditional compare instructions

[Zhenqiang Chen]

I had retested all the ccmp patches.

Bootstrap and no make check regression on X86-64.
Bootstrap and no make check regression on AARCH64 qemu.

OK for trunk?

Thanks!
-Zhenqiang

> -Original Message-
> From: gcc-patches-ow...@gcc.gnu.org [mailto:gcc-patches-
> ow...@gcc.gnu.org] On Behalf Of Zhenqiang Chen
> Sent: Monday, October 27, 2014 3:50 PM
> To: 'Richard Henderson'
> Cc: gcc-patches@gcc.gnu.org
> Subject: RE: [Ping] [PATCH, 9/10] aarch64: generate conditional compare
> instructions
> 
> 
> 
> > -Original Message-
> > From: Richard Henderson [mailto:r...@redhat.com]
> > Sent: Sunday, October 12, 2014 4:46 AM
> > To: Zhenqiang Chen; gcc-patches@gcc.gnu.org
> > Subject: Re: [Ping] [PATCH, 9/10] aarch64: generate conditional
> > compare instructions
> >
> > On 09/22/2014 11:46 PM, Zhenqiang Chen wrote:
> > > +static bool
> > > +aarch64_convert_mode (rtx* op0, rtx* op1, int unsignedp) {
> > > +  enum machine_mode mode;
> > > +
> > > +  mode = GET_MODE (*op0);
> > > +  if (mode == VOIDmode)
> > > +mode = GET_MODE (*op1);
> > > +
> > > +  if (mode == QImode || mode == HImode)
> > > +{
> > > +  *op0 = convert_modes (SImode, mode, *op0, unsignedp);
> > > +  *op1 = convert_modes (SImode, mode, *op1, unsignedp);
> > > +}
> > > +  else if (mode != SImode && mode != DImode)
> > > +return false;
> > > +
> > > +  return true;
> > > +}
> >
> > Hum.  I'd rather not replicate too much of the expander logic here.
> >
> > We could avoid that by using struct expand_operand,
> > create_input_operand et al, then expand_insn.  That does require that
> > the target hooks be given trees rather than rtl as input.
> 
> I had tried to use tree/gimple as input. But the codes was more complexity
> than current one.
> 
> And comments in https://gcc.gnu.org/ml/gcc-patches/2014-
> 06/msg02027.html
> " I suspect it might be better to just hoist any preparation operations
above
> the entire conditional compare sequence, so that by the time we start the
> ccmp expansion we're dealing with operands that are in the 'natural' sizes
for
> the machine (breaking up the conditional compare sequence for what are
> almost side-effect operations sounds like a source of potential bugs).
This
> would also ensure that the back-end can safely re-order at least some
> comparison operations if this leads a workable conditional compare
> sequence. "
> 
> I think the mode conversion (to SImode or DImode) is target dependent.
> 
> Thanks!
> -Zhenqiang
> 
> 
> 

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/03/2014 05:27 PM, Marek Polacek wrote:

Another shot at optimizing redundant UBSAN_NULL statements.

This time we walk the dominator tree - that should result in
more effective optimization - and keep a list of UBSAN_NULL
statements that dominate the current block, see the comment
before sanopt_optimize_walker.


Marek,

A general question - have you considered coding this as a dataflow loop 
instead of dominator walk?  That would allow to also remove checks for 
variables defined via PHI nodes provided that all arguments of PHI have 
already been checked.


-Y

Re: [PATCH RFC]Pair load store instructions using a generic scheduling fusion pass

[Bin.Cheng]

On Sat, Nov 1, 2014 at 4:29 AM, Jeff Law  wrote:
> On 09/30/14 03:22, Bin Cheng wrote:
>
>>
>> 2014-09-30  Bin Cheng
>> Mike Stump
>>
>> * timevar.def (TV_SCHED_FUSION): New time var.
>> * passes.def (pass_sched_fusion): New pass.
>> * config/arm/arm.c (TARGET_SCHED_FUSION_PRIORITY): New.
>> (extract_base_offset_in_addr, fusion_load_store): New.
>> (arm_sched_fusion_priority): New.
>> (arm_option_override): Disable scheduling fusion on non-armv7
>> processors by default.
>> * sched-int.h (struct _haifa_insn_data): New field.
>> (INSN_FUSION_PRIORITY, FUSION_MAX_PRIORITY, sched_fusion): New.
>> * sched-rgn.c (rest_of_handle_sched_fusion): New.
>> (pass_data_sched_fusion, pass_sched_fusion): New.
>> (make_pass_sched_fusion): New.
>> * haifa-sched.c (sched_fusion): New.
>> (insn_cost): Handle sched_fusion.
>> (priority): Handle sched_fusion by calling target hook.
>> (enum rfs_decision): New enum value.
>> (rfs_str): New element for RFS_FUSION.
>> (rank_for_schedule): Support sched_fusion.
>> (schedule_insn, max_issue, prune_ready_list): Handle sched_fusion.
>> (schedule_block, fix_tick_ready): Handle sched_fusion.
>> * common.opt (flag_schedule_fusion): New.
>> * tree-pass.h (make_pass_sched_fusion): New.
>> * target.def (fusion_priority): New.
>> * doc/tm.texi.in (TARGET_SCHED_FUSION_PRIORITY): New.
>> * doc/tm.texi: Regenerated.
>> * doc/invoke.texi (-fschedule-fusion): New.
>>
>> gcc/testsuite/ChangeLog
>> 2014-09-30  Bin Cheng
>>
>> * gcc.target/arm/ldrd-strd-pair-1.c: New test.
>> * gcc.target/arm/vfp-1.c: Improve scanning string.
>>
>>
>> sched-fusion-20140929.txt
>>
>>
>> Index: gcc/doc/tm.texi
>> ===
>> --- gcc/doc/tm.texi (revision 215662)
>> +++ gcc/doc/tm.texi (working copy)
>> @@ -6677,6 +6677,29 @@ This hook is called by tree reassociator to determ
>>   parallelism required in output calculations chain.
>>   @end deftypefn
>>
>> +@deftypefn {Target Hook} void TARGET_SCHED_FUSION_PRIORITY (rtx_insn
>> *@var{insn}, int @var{max_pri}, int *@var{fusion_pri}, int *@var{pri})
>> +This hook is called by scheduling fusion pass.  It calculates fusion
>> +priorities for each instruction passed in by parameter.  The priorities
>> +are returned via pointer parameters.
>> +
>> +@var{insn} is the instruction whose priorities need to be calculated.
>> +@var{max_pri} is the maximum priority can be returned in any cases.
>> +@var{fusion_pri} is the pointer parameter through which @var{insn}'s
>> +fusion priority should be calculated and returned.
>> +@var{pri} is the pointer parameter through which @var{insn}'s priority
>> +should be calculated and returned.
>> +
>> +Same @var{fusion_pri} should be returned for instructions which should
>> +be scheduled together.  Different @var{pri} should be returned for
>> +instructions with same @var{fusion_pri}.  All instructions will be
>> +scheduled according to the two priorities.  @var{fusion_pri} is the major
>> +sort key, @var{pri} is the minor sort key.  All priorities calculated
>> +should be between 0 (exclusive) and @var{max_pri} (inclusive).  To avoid
>> +false dependencies, @var{fusion_pri} of instructions which need to be
>> +scheduled together should be smaller than @var{fusion_pri} of irrelevant
>> +instructions.
>> +@end deftypefn
>> +
>>   @node Sections
>>   @section Dividing the Output into Sections (Texts, Data, @dots{})
>>   @c the above section title is WAY too long.  maybe cut the part between
>
> So I think we need to clarify that this hook is useful when fusing to
> related insns, but which don't have a data dependency.  Somehow we need to
> describe that the insns to be fused should have the same (or +-1) priority.
> It may be useful to use code from the ARM implementation to show how to use
> this to pair up loads as an example.
>
> It may also be useful to refer to the code which reorders insns in the ready
> queue for cases where we want to fuse two truly independent insns.
>
>
>>
>> +  if (sched_fusion)
>> +{
>> +  /* The instruction that has the same fusion priority as the last
>> +instruction is the instruction we picked next.  If that is not
>> +the case, we sort ready list firstly by fusion priority, then
>> +by priority, and at last by INSN_LUID.  */
>> +  int a = INSN_FUSION_PRIORITY (tmp);
>> +  int b = INSN_FUSION_PRIORITY (tmp2);
>> +  int last = -1;
>> +
>> +  if (last_nondebug_scheduled_insn
>> + && !NOTE_P (last_nondebug_scheduled_insn)
>> + && BLOCK_FOR_INSN (tmp)
>> +  == BLOCK_FOR_INSN (last_nondebug_scheduled_insn))
>> +   last = INSN_FUSION_PRIORITY (last_nondebug_scheduled_insn);
>> +
>> +  if (a != last && b != last)
>> +   {
>> +   

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 12:19:22PM +0300, Yury Gribov wrote:
> On 11/03/2014 05:27 PM, Marek Polacek wrote:
> >Another shot at optimizing redundant UBSAN_NULL statements.
> >
> >This time we walk the dominator tree - that should result in
> >more effective optimization - and keep a list of UBSAN_NULL
> >statements that dominate the current block, see the comment
> >before sanopt_optimize_walker.
> 
> Marek,
> 
> A general question - have you considered coding this as a dataflow loop
> instead of dominator walk?  That would allow to also remove checks for
> variables defined via PHI nodes provided that all arguments of PHI have
> already been checked.

I'd be afraid that we'd turn sanopt into another var-tracking that way,
with possibly huge hash tables being copied on write, merging of the tables
etc., with big memory and time requirements, having to add --param limits
to give up if the sum size of the tables go over certain limit.  Or can you
explain how this problem is different from the var-tracking problem?

The way Marek has coded it up is pretty cheap optimization.

BTW, as discussed privately with Marek last time, we probably want to
optimize UBSAN_NULL (etc.) only if -fno-sanitize-recover=null (etc.)
or if location_t is the same, otherwise such optimizations lead to only
one problem being reported instead of all of them.

Jakub

Re: [PATCH, i386] Add prefixes avoidance tuning for silvermont target

[Ilya Enkovich]

Hi,

Having stage1 close to end, may we make some decision regarding this
patch? Having a couple of working variants, may we choose and use one
of them?

Thanks,
Ilya

2014-07-15 17:38 GMT+04:00 Uros Bizjak :
> On Tue, Jul 15, 2014 at 1:01 PM, Ilya Enkovich  wrote:
>> On 15 Jul 10:42, Uros Bizjak wrote:
>>> On Tue, Jul 15, 2014 at 10:25 AM, Ilya Enkovich  
>>> wrote:
>>>
>>> > Also fully restrict xmm8-15 does not seem right.  It is just costly
>>> > but not fully disallowed.
>>> 
>>>  As said earlier, you can try "Ya*x" as a constraint.
>>> >>>
>>> >>> I tried it. It does not seem to affect allocation much. I do not see
>>> >>> any gain on targeted tests.
>>> >>
>>> >> Strange, because the documentation claims:
>>> >>
>>> >> '*'
>>> >>  Says that the following character should be ignored when choosing
>>> >>  register preferences.  '*' has no effect on the meaning of the
>>> >>  constraint as a constraint, and no effect on reloading.  For LRA
>>> >>  '*' additionally disparages slightly the alternative if the
>>> >>  following character matches the operand.
>>> >>
>>> >> Let me rethink this a bit. Prehaps we could reconsider Jakub's
>>> >> proposal with "Ya,!x" (with two alternatives). IIRC this approach was
>>> >> needed for some MMX alternatives, where we didn't want RA to allocate
>>> >> a MMX register when the value could be passed in integer regs, but the
>>> >> value was still allowed in MMX register.
>>> >
>>> > That's is what my patch already does, but with '?' instead of '!'.
>>>
>>> Yes, I know. The problem is, that Ya*x type conditional allocation
>>> worked OK in the past for "not preferred, but still alowed regclass"
>>> registers, There are several patterns in i386.md that live by this
>>> premise, including movsf_internal and movdf_internal. If this approach
>>> doesn't work anymore, then we have to either figure out what is the
>>> reason, or invent a new strategy that will be applicable to all cases.
>>>
>>> Can you please post a small test that illustrates the case where Ya,!x
>>> works, but Ya*x doesn't?
>>
>> It's hard to compose a small testcase which will have SSE4 instructions 
>> generated with required register usage.  I use tcpjumbo test from TCPmark 
>> for initial check of how my patch works.  This test has a lot of pmovzxwd 
>> instructions generated and many of them use xmm8-15.  I tried two versions 
>> of a simple patch which modifies only pmovzxwd instruction.
>>
>> Patch1:
>>
>> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
>> index d907353..6b03b72 100644
>> --- a/gcc/config/i386/sse.md
>> +++ b/gcc/config/i386/sse.md
>> @@ -11852,10 +11852,10 @@
>> (set_attr "mode" "OI")])
>>
>>  (define_insn "sse4_1_v4hiv4si2"
>> -  [(set (match_operand:V4SI 0 "register_operand" "=x")
>> +  [(set (match_operand:V4SI 0 "register_operand" "=Yr,!x")
>> (any_extend:V4SI
>>   (vec_select:V4HI
>> -   (match_operand:V8HI 1 "nonimmediate_operand" "xm")
>> +   (match_operand:V8HI 1 "nonimmediate_operand" "Yr,!xm")
>> (parallel [(const_int 0) (const_int 1)
>>(const_int 2) (const_int 3)]]
>>"TARGET_SSE4_1"
>>
>> Patch2:
>>
>> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
>> index d907353..b3721c4 100644
>> --- a/gcc/config/i386/sse.md
>> +++ b/gcc/config/i386/sse.md
>> @@ -11852,10 +11852,10 @@
>> (set_attr "mode" "OI")])
>>
>>  (define_insn "sse4_1_v4hiv4si2"
>> -  [(set (match_operand:V4SI 0 "register_operand" "=x")
>> +  [(set (match_operand:V4SI 0 "register_operand" "=Yr*x")
>> (any_extend:V4SI
>>   (vec_select:V4HI
>> -   (match_operand:V8HI 1 "nonimmediate_operand" "xm")
>> +   (match_operand:V8HI 1 "nonimmediate_operand" "Yr*xm")
>> (parallel [(const_int 0) (const_int 1)
>>(const_int 2) (const_int 3)]]
>>"TARGET_SSE4_1"
>>
>>
>> Here are results of looking for pmovzxwd in resulting binaries:
>> #objdump -d tcpjumbo-orig | grep pmovzxwd | grep 
>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>> 76
>> #objdump -d tcpjumbo-patch1 | grep pmovzxwd | grep 
>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>> 0
>> #objdump -d tcpjumbo-patch2 | grep pmovzxwd | grep 
>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>> 76
>>
>> Therefore I make a conclusion that Yr*x does not really differ much from x.
>
> Just FTR:
>
> Using "Yr,*x" is also a viable option:
>
> #objdump -d tcpjumbo-patch3 | grep pmovzxwd | grep
> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
> 0
>
> I believe that the above is the way to go with LRA. Vladimir, what do you 
> think?
>
> Uros.

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/05/2014 12:33 PM, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 12:19:22PM +0300, Yury Gribov wrote:

On 11/03/2014 05:27 PM, Marek Polacek wrote:

Another shot at optimizing redundant UBSAN_NULL statements.

This time we walk the dominator tree - that should result in
more effective optimization - and keep a list of UBSAN_NULL
statements that dominate the current block, see the comment
before sanopt_optimize_walker.


Marek,

A general question - have you considered coding this as a dataflow loop
instead of dominator walk?  That would allow to also remove checks for
variables defined via PHI nodes provided that all arguments of PHI have
already been checked.


I'd be afraid that we'd turn sanopt into another var-tracking that way,
with possibly huge hash tables being copied on write, merging of the tables
etc., with big memory and time requirements, having to add --param limits
to give up if the sum size of the tables go over certain limit.


Sure, that would be slower.  I was just curious whether you considered 
alternatives (looks like you did).



The way Marek has coded it up is pretty cheap optimization.


Right.


BTW, as discussed privately with Marek last time, we probably want to
optimize UBSAN_NULL (etc.) only if -fno-sanitize-recover=null (etc.)
or if location_t is the same, otherwise such optimizations lead to only
one problem being reported instead of all of them.


Are you going to work on ASan soon?  I could rebase my patches on top of 
Marek's infrastructure.


-Y

Re: [PATCH, i386] Add prefixes avoidance tuning for silvermont target

[Uros Bizjak]

On Wed, Nov 5, 2014 at 10:35 AM, Ilya Enkovich  wrote:
> Hi,
>
> Having stage1 close to end, may we make some decision regarding this
> patch? Having a couple of working variants, may we choose and use one
> of them?

I propose to wait for Vlad for an update about his plans on register
preference algorythm that would fix this (and other "Ya*r"-type
issues).

In the absence of the fix, we'll go with "Yr,*x".

Uros.

>
> Thanks,
> Ilya
>
> 2014-07-15 17:38 GMT+04:00 Uros Bizjak :
>> On Tue, Jul 15, 2014 at 1:01 PM, Ilya Enkovich  
>> wrote:
>>> On 15 Jul 10:42, Uros Bizjak wrote:
 On Tue, Jul 15, 2014 at 10:25 AM, Ilya Enkovich  
 wrote:

 > Also fully restrict xmm8-15 does not seem right.  It is just costly
 > but not fully disallowed.
 
  As said earlier, you can try "Ya*x" as a constraint.
 >>>
 >>> I tried it. It does not seem to affect allocation much. I do not see
 >>> any gain on targeted tests.
 >>
 >> Strange, because the documentation claims:
 >>
 >> '*'
 >>  Says that the following character should be ignored when choosing
 >>  register preferences.  '*' has no effect on the meaning of the
 >>  constraint as a constraint, and no effect on reloading.  For LRA
 >>  '*' additionally disparages slightly the alternative if the
 >>  following character matches the operand.
 >>
 >> Let me rethink this a bit. Prehaps we could reconsider Jakub's
 >> proposal with "Ya,!x" (with two alternatives). IIRC this approach was
 >> needed for some MMX alternatives, where we didn't want RA to allocate
 >> a MMX register when the value could be passed in integer regs, but the
 >> value was still allowed in MMX register.
 >
 > That's is what my patch already does, but with '?' instead of '!'.

 Yes, I know. The problem is, that Ya*x type conditional allocation
 worked OK in the past for "not preferred, but still alowed regclass"
 registers, There are several patterns in i386.md that live by this
 premise, including movsf_internal and movdf_internal. If this approach
 doesn't work anymore, then we have to either figure out what is the
 reason, or invent a new strategy that will be applicable to all cases.

 Can you please post a small test that illustrates the case where Ya,!x
 works, but Ya*x doesn't?
>>>
>>> It's hard to compose a small testcase which will have SSE4 instructions 
>>> generated with required register usage.  I use tcpjumbo test from TCPmark 
>>> for initial check of how my patch works.  This test has a lot of pmovzxwd 
>>> instructions generated and many of them use xmm8-15.  I tried two versions 
>>> of a simple patch which modifies only pmovzxwd instruction.
>>>
>>> Patch1:
>>>
>>> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
>>> index d907353..6b03b72 100644
>>> --- a/gcc/config/i386/sse.md
>>> +++ b/gcc/config/i386/sse.md
>>> @@ -11852,10 +11852,10 @@
>>> (set_attr "mode" "OI")])
>>>
>>>  (define_insn "sse4_1_v4hiv4si2"
>>> -  [(set (match_operand:V4SI 0 "register_operand" "=x")
>>> +  [(set (match_operand:V4SI 0 "register_operand" "=Yr,!x")
>>> (any_extend:V4SI
>>>   (vec_select:V4HI
>>> -   (match_operand:V8HI 1 "nonimmediate_operand" "xm")
>>> +   (match_operand:V8HI 1 "nonimmediate_operand" "Yr,!xm")
>>> (parallel [(const_int 0) (const_int 1)
>>>(const_int 2) (const_int 3)]]
>>>"TARGET_SSE4_1"
>>>
>>> Patch2:
>>>
>>> diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
>>> index d907353..b3721c4 100644
>>> --- a/gcc/config/i386/sse.md
>>> +++ b/gcc/config/i386/sse.md
>>> @@ -11852,10 +11852,10 @@
>>> (set_attr "mode" "OI")])
>>>
>>>  (define_insn "sse4_1_v4hiv4si2"
>>> -  [(set (match_operand:V4SI 0 "register_operand" "=x")
>>> +  [(set (match_operand:V4SI 0 "register_operand" "=Yr*x")
>>> (any_extend:V4SI
>>>   (vec_select:V4HI
>>> -   (match_operand:V8HI 1 "nonimmediate_operand" "xm")
>>> +   (match_operand:V8HI 1 "nonimmediate_operand" "Yr*xm")
>>> (parallel [(const_int 0) (const_int 1)
>>>(const_int 2) (const_int 3)]]
>>>"TARGET_SSE4_1"
>>>
>>>
>>> Here are results of looking for pmovzxwd in resulting binaries:
>>> #objdump -d tcpjumbo-orig | grep pmovzxwd | grep 
>>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>>> 76
>>> #objdump -d tcpjumbo-patch1 | grep pmovzxwd | grep 
>>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>>> 0
>>> #objdump -d tcpjumbo-patch2 | grep pmovzxwd | grep 
>>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm13\|xmm14\|xmm15" | wc -l
>>> 76
>>>
>>> Therefore I make a conclusion that Yr*x does not really differ much from x.
>>
>> Just FTR:
>>
>> Using "Yr,*x" is also a viable option:
>>
>> #objdump -d tcpjumbo-patch3 | grep pmovzxwd | grep
>> "xmm8\|xmm9\|xmm10\|xmm11\|xmm12\|xmm1

Re: [gofrontend-dev] [PATCH 4/4] Gccgo port to s390[x] -- part II

[Dominik Vogt]

On Tue, Nov 04, 2014 at 08:16:51PM -0800, Ian Taylor wrote:
> I committed the change to go-test.exp.  Thanks.
> 
> The other changes are not OK.  As described in
> gcc/testsuite/go.test/test/README.gcc, the files in
> gcc/testsuite/go.test/test are an exact copy of the master Go
> testsuite.  Any changes must be made to the master Go testsuite first.

I understand that, but I'm unsure how to handle a set of patches
that all depend on each other but refer to three different
reposiories.  So I posted this patch intentionally in the wrong
place, not knowing how to do it in a better way.

> I don't know what's up with the complex number change. In general the
> Go compiler and libraries go to some effort to produce the same
> answers on all platforms.  We need to understand why we get different
> answers on s390 (you may understand the differences, but I don't).  I
> won't change the tests without a clear understanding of why we are
> changing them.

It's actually not a Go specific problem, the same deviation occurs
in C code too.  The cause is that constant folding is done with a
higher precision and may yield a different result than the run
time calculations.  There is a Gcc bug report for that "issue":
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60181

> The nilptr test doesn't run on some other platforms when using
> gccgo--search for "nilptr" in go-test.exp.  If you want to work out a
> way to change the master Go testsuite such that the nilptr test passes
> on more platforms, that would be great.

I don't have the slightest clue how this could be done in a
platform independent way because the test heavily depends on the
target's memory map layout.

> The way to do it is not by
> copying the test.  If the test needs to be customized, add additional
> files that use // +build lines to pick which files is built.  Move
> them into a directory, like method4.go or other tests that use
> "rundir".

I'll check that.

Ciao

Dominik ^_^  ^_^

-- 

Dominik Vogt
IBM Germany

Re: Small multiplier support in Cortex-M0/1/+

[Ramana Radhakrishnan]

On Tue, Oct 21, 2014 at 11:01 AM, Hale Wang  wrote:
> Hi,
>
> Some configurations of the Cortex-M0 and Cortex-M1 come with a high latency
> multiplier. This patch adds support for such configurations.
>
> Small multiplier means using add/sub/shift instructions to replace the mul
> instruction for the MCU that has no fast multiplier.
>
> The following strategies are adopted in this patch:
> 1. Define new CPUs as
> -mcpu=cortex-m0.small-multiply,cortex-m0plus.small-multiply,cortex-m1.small-
> multiply to support small multiplier.
> 2. -Os means size is preferred. A threshold of 5 is set which means it will
> prevent spliting if ending up with more than 5 instructions. As for non-OS,
> there will be no such a limit.
>
> Some test cases are also added in the testsuite to verify this function.
>
> Is it ok for trunk?

This is OK .

Ramana
>
> Thanks and Best Regards,
> Hale Wang
>
> gcc/ChangeLog:
>
> 2014-08-29  Hale Wang  
>
> * config/arm/arm-cores.def: Add support for
> -mcpu=cortex-m0.small-multiply,cortex-m0plus.small-multiply,
> cortex-m1.small-multiply.
> * config/arm/arm-tables.opt: Regenerate.
> * config/arm/arm-tune.md: Regenerate.
> * config/arm/arm.c: Update the rtx-costs for MUL.
> * config/arm/bpabi.h: Handle
> -mcpu=cortex-m0.small-multiply,cortex-m0plus.small-multiply,
> cortex-m1.small-multiply.
> * doc/invoke.texi: Document
> -mcpu=cortex-m0.small-multiply,cortex-m0plus.small-multiply,
> cortex-m1.small-multiply.
> * testsuite/gcc.target/arm/small-multiply-m0-1.c: New test case.
> * testsuite/gcc.target/arm/small-multiply-m0-2.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m0-3.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m0plus-1.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m0plus-2.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m0plus-3.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m1-1.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m1-2.c: Likewise.
> * testsuite/gcc.target/arm/small-multiply-m1-3.c: Likewise.
>
> ===
> diff --git a/gcc/config/arm/arm-cores.def b/gcc/config/arm/arm-cores.def
> index a830a83..af4b373 100644
> --- a/gcc/config/arm/arm-cores.def
> +++ b/gcc/config/arm/arm-cores.def
> @@ -137,6 +137,11 @@ ARM_CORE("cortex-m1",  cortexm1, cortexm1,
> 6M, FL_LDSCHED, v6m)
>  ARM_CORE("cortex-m0",  cortexm0, cortexm0, 6M,
> FL_LDSCHED, v6m)
>  ARM_CORE("cortex-m0plus",  cortexm0plus, cortexm0plus, 6M,
> FL_LDSCHED, v6m)
>
> +/* V6M Architecture Processors for small-multiply implementations.  */
> +ARM_CORE("cortex-m1.small-multiply",   cortexm1smallmultiply, cortexm1,
> 6M, FL_LDSCHED | FL_SMALLMUL, v6m)
> +ARM_CORE("cortex-m0.small-multiply",   cortexm0smallmultiply, cortexm0,
> 6M, FL_LDSCHED | FL_SMALLMUL, v6m)
> +ARM_CORE("cortex-m0plus.small-multiply",cortexm0plussmallmultiply,
> cortexm0plus,6M, FL_LDSCHED | FL_SMALLMUL, v6m)
> +
>  /* V7 Architecture Processors */
>  ARM_CORE("generic-armv7-a",genericv7a, genericv7a, 7A,
> FL_LDSCHED, cortex)
>  ARM_CORE("cortex-a5",  cortexa5, cortexa5, 7A,
> FL_LDSCHED, cortex_a5)
> diff --git a/gcc/config/arm/arm-tables.opt b/gcc/config/arm/arm-tables.opt
> index bc046a0..bd65bd2 100644
> --- a/gcc/config/arm/arm-tables.opt
> +++ b/gcc/config/arm/arm-tables.opt
> @@ -241,6 +241,15 @@ EnumValue
>  Enum(processor_type) String(cortex-m0plus) Value(cortexm0plus)
>
>  EnumValue
> +Enum(processor_type) String(cortex-m1.small-multiply)
> Value(cortexm1smallmultiply)
> +
> +EnumValue
> +Enum(processor_type) String(cortex-m0.small-multiply)
> Value(cortexm0smallmultiply)
> +
> +EnumValue
> +Enum(processor_type) String(cortex-m0plus.small-multiply)
> Value(cortexm0plussmallmultiply)
> +
> +EnumValue
>  Enum(processor_type) String(generic-armv7-a) Value(genericv7a)
>
>  EnumValue
> diff --git a/gcc/config/arm/arm-tune.md b/gcc/config/arm/arm-tune.md
> index 954cab8..8b5c778 100644
> --- a/gcc/config/arm/arm-tune.md
> +++ b/gcc/config/arm/arm-tune.md
> @@ -25,6 +25,7 @@
> arm1176jzs,arm1176jzfs,mpcorenovfp,
> mpcore,arm1156t2s,arm1156t2fs,
> cortexm1,cortexm0,cortexm0plus,
> +
> cortexm1smallmultiply,cortexm0smallmultiply,cortexm0plussmallmultiply,
> genericv7a,cortexa5,cortexa7,
> cortexa8,cortexa9,cortexa12,
> cortexa15,cortexr4,cortexr4f,
> diff --git a/gcc/config/arm/arm.c b/gcc/config/arm/arm.c
> index 93b989d..5062c85 100644
> --- a/gcc/config/arm/arm.c
> +++ b/gcc/config/arm/arm.c
> @@ -751,6 +751,8 @@ static int thumb_call_reg_needed;
>  #define FL_ARCH8  (1 << 24)   /* Architecture 8.  */
>  #define FL_CRC32  (1 << 25)  /* ARMv8 CRC32 instructions.
> */
>
> +#define FL_SMALLMUL   (1 << 26)   /* Small mu

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Marek Polacek]

On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:
> Are you going to work on ASan soon?  I could rebase my patches on top of
> Marek's infrastructure.

I'm not going to work on ASan today or tomorrow, but it'd be nice to
get this ASan opt in in this stage1.

So if you can rebase your patch, I think that will be appreciated.

Marek

Re: [gofrontend-dev] [PATCH 1/4] Gccgo port to s390[x] -- part II

[Dominik Vogt]

On Tue, Nov 04, 2014 at 02:39:34PM -0800, Ian Taylor wrote:
> Note that libgo/runtime/runtime.c now refers to S390_HAVE_STCKF.  It's
> not obvious to me that that is defined anywhere.  Perhaps it is in a
> later patch in this series--I haven't looked.

This chunk is broken but harmless (because S390_HAVE_STCKF is
never defined anyway).  The code needs to check at runtime whether
the stckf instruction is available.  I'll provide a patch for that
later.  In the mean time there's no immediate need to back out the
flawed chunk.

Ciao

Dominik ^_^  ^_^

-- 

Dominik Vogt
IBM Germany

RE: [PATCH][MIPS] Fix P5600 memory cost

[Matthew Fortune]

> The patch below fixes the memory cost for P5600.
> 
> ChangeLog:
> 2014-11-05 Prachi Godbole 
> 
> * config/mips/mips.c (mips_rtx_cost_data): Fix memory_letency cost for
> p5600.

Please follow these instructions to add yourself to MAINTAINERS in the
write-after-approval section now that you have write access to GCC:

https://gcc.gnu.org/svnwrite.html#authenticated

OK with fixes to the changelog entry:

latency not latency. Remember to tab in the changelog entry and split the
line as it will exceed 80 chars. Also two spaces between the date/name and
name/email. E.g.

2014-11-05  Prachi Godbole  

* config/mips/mips.c (mips_rtx_cost_data): Fix memory_latency cost for
p5600.

Thanks,
Matthew

> diff --git a/gcc/config/mips/mips.c b/gcc/config/mips/mips.c
> index af6a913..558ba2f 100644
> --- a/gcc/config/mips/mips.c
> +++ b/gcc/config/mips/mips.c
> @@ -1193,7 +1193,7 @@ static const struct mips_rtx_cost_data
>  COSTS_N_INSNS (8),/* int_div_si */
>  COSTS_N_INSNS (8),/* int_div_di */
> 2,/* branch_cost */
> -  10 /* memory_latency */
> +   4 /* memory_latency */
>}
>  };
>  ^L

[PATCH, libfortran] PR 47007, 61847 Locale failures in libgfortran

[Janne Blomqvist]

Hi,

the attached patch fixes a few locale related failures in libgfortran,
in the case where the POSIX 2008 extended locale functionality and
extensions strto{f,d,ld}_l are present.

These failures typically occur when libgfortran is used from a program
which has set the locale with setlocale(), and the locale uses a
different decimal separator than the C locale. The patch fixes this by
creating a C locale which is then used by strto{f,d,ld}_l, and also is
installed as the per-thread locale when starting a formatted IO, then
reset to the previous value when the IO is done. I have chosen to not
fallback to calling setlocale() in case the POSIX 2008 locale stuff
isn't available, as that could create nasty hard to debug race
conditions in a multi-threaded program.

(I think Jerry's proposed patch which checks the locale for the
decimal separator is still useful as a fallback in case the POSIX 2008
locale stuff isn't available)

Regtested on x86_64-unknown-linux-gnu, Ok for trunk?

2014-11-06  Janne Blomqvist  

PR libfortran/47007
PR libfortran/61847
* config.h.in: Regenerated.
* configure: Regenerated.
* configure.ac (AC_CHECK_HEADERS_ONCE): Check for locale.h.
(AC_CHECK_FUNCS_ONCE): Check for newlocale, freelocale, uselocale,
strtof_l, strtod_l, strtold_l.
* io/io.h (locale.h): Include if present.
(c_locale): New variable.
(gfc_strtof): Move macro from libgfortran.h, use strtof_l if present.
(gfc_strtod): Likewise.
(gfc_strtold): Likewise.
(st_parameter_dt): Add old_locale member.
* io/transfer.c (data_transfer_init): Set thread locale to
c_locale if doing formatted transfer.
(finalize_transfer): Reset thread locale to previous.
* io/unit.c (c_locale): New variable.
(init_units): Init c_locale.
(close_units): Free c_locale.
* libgfortran.h (gfc_strto{f,d,ld}): Move macros to io/io.h.


-- 
Janne Blomqvist
diff --git a/libgfortran/configure.ac b/libgfortran/configure.ac
index b3150f4..5550380 100644
--- a/libgfortran/configure.ac
+++ b/libgfortran/configure.ac
@@ -255,7 +255,7 @@ AC_CHECK_TYPES([ptrdiff_t])
 # check header files (we assume C89 is available, so don't check for that)
 AC_CHECK_HEADERS_ONCE(unistd.h sys/time.h sys/times.h sys/resource.h \
 sys/types.h sys/stat.h sys/wait.h floatingpoint.h ieeefp.h fenv.h fptrap.h \
-fpxcp.h pwd.h complex.h) 
+fpxcp.h pwd.h complex.h locale.h) 
 
 GCC_HEADER_STDINT(gstdint.h)
 
@@ -290,7 +290,8 @@ else
strcasestr getrlimit gettimeofday stat fstat lstat getpwuid vsnprintf dup \
getcwd localtime_r gmtime_r getpwuid_r ttyname_r clock_gettime \
readlink getgid getpid getppid getuid geteuid umask getegid \
-   secure_getenv __secure_getenv mkostemp strnlen strndup strtok_r)
+   secure_getenv __secure_getenv mkostemp strnlen strndup strtok_r newlocale \
+   freelocale uselocale strtof_l strtod_l strtold_l)
 fi
 
 # Check strerror_r, cannot be above as versions with two and three arguments 
exist
diff --git a/libgfortran/io/io.h b/libgfortran/io/io.h
index 1e0d092..a638daf 100644
--- a/libgfortran/io/io.h
+++ b/libgfortran/io/io.h
@@ -32,6 +32,10 @@ see the files COPYING3 and COPYING.RUNTIME respectively.  If 
not, see
 
 #include 
 
+#ifdef HAVE_LOCALE_H
+#include 
+#endif
+
 /* Forward declarations.  */
 struct st_parameter_dt;
 typedef struct stream stream;
@@ -40,6 +44,35 @@ struct format_data;
 typedef struct fnode fnode;
 struct gfc_unit;
 
+#ifdef HAVE_NEWLOCALE
+/* We have POSIX 2008 extended locale stuff.  */
+extern locale_t c_locale;
+#endif
+
+#ifdef __MINGW32__
+extern float __strtof (const char *, char **);
+#define gfc_strtof __strtof
+extern double __strtod (const char *, char **);
+#define gfc_strtod __strtod
+extern long double __strtold (const char *, char **);
+#define gfc_strtold __strtold
+#else
+#ifdef HAVE_STRTOF_L
+#define gfc_strtof(nptr, endptr) strtof_l(nptr, endptr, c_locale)
+#else
+#define gfc_strtof strtof
+#endif
+#ifdef HAVE_STRTOD_L
+#define gfc_strtod(nptr, endptr) strtod_l(nptr, endptr, c_locale)
+#else
+#define gfc_strtod strtod
+#endif
+#ifdef HAVE_STRTOLD_L
+#define gfc_strtold(nptr, endptr) strtold_l(nptr, endptr, c_locale)
+#else
+#define gfc_strtold strtold
+#endif
+#endif
 
 /* Macros for testing what kinds of I/O we are doing.  */
 
@@ -450,6 +483,9 @@ typedef struct st_parameter_dt
  char *line_buffer;
  struct format_data *fmt;
  namelist_info *ionml;
+#ifdef HAVE_NEWLOCALE
+ locale_t old_locale;
+#endif
  /* Current position within the look-ahead line buffer.  */
  int line_buffer_pos;
  /* Storage area for values except for strings.  Must be
diff --git a/libgfortran/io/transfer.c b/libgfortran/io/transfer.c
index dc1b6f4..4706865 100644
--- a/libgfortran/io/transfer.c
+++ b/libgfortran/io/transfer.c
@@ -2874,7 +2874,12 @@ data_transfer_init (st_parameter_dt *dtp, int read_flag)
   if (dtp->u.p.current_unit->flags.form == FORM_FORMATTED
   && ((cf & (IOPARM_DT_LIST_FORMAT

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:
> On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:
> > Are you going to work on ASan soon?  I could rebase my patches on top of
> > Marek's infrastructure.
> 
> I'm not going to work on ASan today or tomorrow, but it'd be nice to
> get this ASan opt in in this stage1.
> 
> So if you can rebase your patch, I think that will be appreciated.

Note, the algorithm we were discussing with Honza for the
"is there any possibility of a freeing call on the path between a dominating
and dominated ASAN_CHECK"
problem was to compute it lazily; have flags for asan per-bb:
1) bb might contain a !nonfreeing_call_p
2) there is a bb with flag 1) set in some path between imm(bb) and bb
3) flag whether 2) has been computed already
4) some temporary "being visited" flag
and the algorithm:
1) when walking a bb, if you encounter a !nonfreeing_call_p call, either
   immediately nuke recorded earlier ASAN_CHECKs from the current bb,
   or use gimple_uids for lazily doing that; but in any case, record
   the flag 1) for the current bb
2) if you are considering ASAN_CHECK in a different bb than ASAN_CHECK
   it is dominating, check the 2) flag on the current bb, then on
   get_immediate_dominator (bb) etc. until you reach the bb with the
   dominating bb, if the 2) flag is set on any of them, don't optimize;
   if the 2) flag is not computed on any of these (i.e. flag 3) unset),
   then compute it recursively; set the 4) flag on a bb, for incoming
   edges if the src bb is not the imm(bb) of the original bb, and does
   not have 4) flag set: if it has 1) set, use 1, if it has 3) flag set,
   use 2), otherwise recurse (and or the result); unset 4) flag before
   returning; or so.

For tsan, pretty much the same thing, just with different 1)/2)/3)
flags and different test for that (instead of !nonfreeing_call_p
we are interested in: uses atomics or calls that might use atomics
or other pthread_* synchronization primitives).

Jakub

Re: [PATCH, PR63742][ARM] Fix arm *movhi_insn_arch4 pattern for big-endian

[Ramana Radhakrishnan]

On 05/11/14 07:09, Yangfei (Felix) wrote:

Hi,

 This patch fixes PR63742 by improving arm *movhi_insn_arch4 pattern to 
make it works under big-endian.
 The idea is simple: Use movw for certain const source operand instead of 
ldrh.  And exclude the const values which cannot be handled by mov/mvn/movw.
 I am doing regression test for this patch.  Assuming no issue pops up, OK 
for trunk?


So, doesn't that makes the bug latent for architectures older than 
armv6t2 and big endian and only fixed this in ARM state ? I'd prefer a 
complete solution please. What about *thumb2_movhi_insn in thumb2.md ?





Index: gcc/ChangeLog
===
--- gcc/ChangeLog   (revision 216838)
+++ gcc/ChangeLog   (working copy)
@@ -1,3 +1,12 @@
+2014-11-05  Felix Yang  
+   Shanyao Chen  
+


I'm assuming you have copyright assignments sorted.



+   PR target/63742
+   * config/arm/predicates.md (arm_hi_operand): New predicate.
+   (arm_movw_immediate_operand): Similarly.
+   * config/arm/arm.md (*movhi_insn_arch4): Use arm_hi_operand instead of
+   general_operand and add "movw" to the output template.
+
  2014-10-29  Richard Sandiford  

* addresses.h, alias.c, asan.c, auto-inc-dec.c, bt-load.c, builtins.c,
Index: gcc/config/arm/predicates.md
===
--- gcc/config/arm/predicates.md(revision 216838)
+++ gcc/config/arm/predicates.md(working copy)
@@ -144,6 +144,12 @@
(and (match_code "const_int")
 (match_test "INTVAL (op) == 0")))

+(define_predicate "arm_movw_immediate_operand"
+  (and (match_test "TARGET_32BIT && arm_arch_thumb2")
+   (ior (match_code "high")


I don't see why you need to check for "high" here ?


+(and (match_code "const_int")
+ (match_test "(INTVAL (op) & 0x) == 0")
+
  ;; Something valid on the RHS of an ARM data-processing instruction
  (define_predicate "arm_rhs_operand"
(ior (match_operand 0 "s_register_operand")
@@ -211,6 +217,11 @@
(ior (match_operand 0 "arm_rhs_operand")
 (match_operand 0 "arm_not_immediate_operand")))

+(define_predicate "arm_hi_operand"
+  (ior (match_operand 0 "arm_rhsm_operand")
+   (ior (match_operand 0 "arm_not_immediate_operand")
+(match_operand 0 "arm_movw_immediate_operand"
+
  (define_predicate "arm_di_operand"
(ior (match_operand 0 "s_register_operand")
 (match_operand 0 "arm_immediate_di_operand")))
Index: gcc/config/arm/arm.md
===
--- gcc/config/arm/arm.md   (revision 216838)
+++ gcc/config/arm/arm.md   (working copy)
@@ -6285,8 +6285,8 @@

  ;; Pattern to recognize insn generated default case above
  (define_insn "*movhi_insn_arch4"
-  [(set (match_operand:HI 0 "nonimmediate_operand" "=r,r,m,r")
-   (match_operand:HI 1 "general_operand"  "rIk,K,r,mi"))]
+  [(set (match_operand:HI 0 "nonimmediate_operand" "=r,r,r,m,r")
+   (match_operand:HI 1 "arm_hi_operand" "rIk,K,j,r,mi"))]
"TARGET_ARM
 && arm_arch4
 && (register_operand (operands[0], HImode)
@@ -6294,16 +6294,18 @@
"@
 mov%?\\t%0, %1\\t%@ movhi
 mvn%?\\t%0, #%B1\\t%@ movhi
+   movw%?\\t%0, %1\\t%@ movhi
 str%(h%)\\t%1, %0\\t%@ movhi
 ldr%(h%)\\t%0, %1\\t%@ movhi"
[(set_attr "predicable" "yes")
-   (set_attr "pool_range" "*,*,*,256")
-   (set_attr "neg_pool_range" "*,*,*,244")
+   (set_attr "pool_range" "*,*,*,*,256")
+   (set_attr "neg_pool_range" "*,*,*,*,244")
 (set_attr_alternative "type"
   [(if_then_else (match_operand 1 "const_int_operand" 
"")
  (const_string "mov_imm" )
  (const_string "mov_reg"))
(const_string "mvn_imm")
+  (const_string "mov_imm")
(const_string "store1")
(const_string "load1")])]
  )



Ramana

Re: [PATCH, libfortran] PR 47007, 61847 Locale failures in libgfortran

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 12:48:01PM +0200, Janne Blomqvist wrote:
> @@ -3528,6 +3533,11 @@ finalize_transfer (st_parameter_dt *dtp)
>if ((dtp->common.flags & IOPARM_DT_HAS_SIZE) != 0)
>  *dtp->size = dtp->u.p.size_used;
>  
> +#ifdef HAVE_USELOCALE
> +  if (dtp->u.p.old_locale != (locale_t) 0)
> +uselocale (dtp->u.p.old_locale);
> +#endif

I wonder if you shouldn't clear dtp->u.p.old_locale here too, so
that uselocale isn't called again.

Jakub

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Marek Polacek]

On Wed, Nov 05, 2014 at 11:50:20AM +0100, Jakub Jelinek wrote:
> On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:
> > On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:
> > > Are you going to work on ASan soon?  I could rebase my patches on top of
> > > Marek's infrastructure.
> > 
> > I'm not going to work on ASan today or tomorrow, but it'd be nice to
> > get this ASan opt in in this stage1.
> > 
> > So if you can rebase your patch, I think that will be appreciated.
> 
> Note, the algorithm we were discussing with Honza for the
> "is there any possibility of a freeing call on the path between a dominating
> and dominated ASAN_CHECK"

Right.  Let me see then if I can implement the following soon, maybe
it makes sense to rebase Yuri's patch only on top of this algorithm.

> problem was to compute it lazily; have flags for asan per-bb:
> 1) bb might contain a !nonfreeing_call_p
> 2) there is a bb with flag 1) set in some path between imm(bb) and bb
> 3) flag whether 2) has been computed already
> 4) some temporary "being visited" flag
> and the algorithm:
> 1) when walking a bb, if you encounter a !nonfreeing_call_p call, either
>immediately nuke recorded earlier ASAN_CHECKs from the current bb,
>or use gimple_uids for lazily doing that; but in any case, record
>the flag 1) for the current bb
> 2) if you are considering ASAN_CHECK in a different bb than ASAN_CHECK
>it is dominating, check the 2) flag on the current bb, then on
>get_immediate_dominator (bb) etc. until you reach the bb with the
>dominating bb, if the 2) flag is set on any of them, don't optimize;
>if the 2) flag is not computed on any of these (i.e. flag 3) unset),
>then compute it recursively; set the 4) flag on a bb, for incoming
>edges if the src bb is not the imm(bb) of the original bb, and does
>not have 4) flag set: if it has 1) set, use 1, if it has 3) flag set,
>use 2), otherwise recurse (and or the result); unset 4) flag before
>returning; or so.
> 
> For tsan, pretty much the same thing, just with different 1)/2)/3)
> flags and different test for that (instead of !nonfreeing_call_p
> we are interested in: uses atomics or calls that might use atomics
> or other pthread_* synchronization primitives).

Marek

Re: [PATCH][AARCH64]Add ACLE arch-related predefined macros

[Marcus Shawcroft]

On 31 October 2014 14:37, Renlin Li  wrote:
> Hi all,
>
> This is a simple patch to add arch-related macros defined ACLE 2.0.
>
> aarch64-none-elf target is tested on the model, no new issues.
>
> Is this Okay for trunk?
>
> gcc/ChangeLog:
>
> 2014-10-31  Renlin Li  
>
> * config/aarch64/aarch64.c (aarch64_architecture_version): New.
> (processor): New architecture_version field.
> (aarch64_override_options): Initialize aarch64_architecture_version.
> * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define __ARM_ARCH,
> __ARM_ARCH_PROFILE, aarch64_arch_name macro.

OK /Marcus

Re: [4.9 RFA PATCH, RTL-optimization]: Backport recent AND-alignment alias fixes to 4.9 branch

[Richard Biener]

On Wed, 5 Nov 2014, Uros Bizjak wrote:

> Ping for [1], quoted below.

Ok.

Thanks,
Richard.

> [1] https://gcc.gnu.org/ml/gcc-patches/2014-10/msg03189.html
> 
> Thanks,
> Uros.
> 
> On Thu, Oct 30, 2014 at 11:38 AM, Uros Bizjak  wrote:
> > Hello!
> >
> > I would like to backport recent alias fixes to correctly handle memory
> > references with AND-alignment to 4.9 branch. These patches fix
> > hundreds of failures in gfortran testsuite on alpha-linux-gnu due to
> > invalid aliasing of AND-aligned memory references of two QImode flags.
> >
> > These patches were baking for a couple of weeks in the mainline
> > without problems. Modulo removal of old and unnecessary functionality,
> > these changes affect only alpha target.
> >
> > 2014-10-30  Uros Bizjak  
> >
> > Backport from mainline:
> > 2014-10-20  Uros Bizjak  
> >
> > * varasm.c (const_alias_set): Remove.
> > (init_varasm_once): Remove initialization of const_alias_set.
> > (build_constant_desc): Do not set alias set to const_alias_set.
> >
> > Backport from mainline:
> > 2014-10-14  Uros Bizjak  
> >
> > PR rtl-optimization/63475
> > * alias.c (true_dependence_1): Always use get_addr to extract
> > true address operands from x_addr and mem_addr.  Use extracted
> > address operands to check for references with alignment ANDs.
> > Use extracted address operands with find_base_term and
> > base_alias_check. For noncanonicalized operands call canon_rtx with
> > extracted address operand.
> > (write_dependence_1): Ditto.
> > (may_alias_p): Ditto.  Remove unused calls to canon_rtx.
> >
> > Backport from mainline:
> > 2014-10-10  Uros Bizjak  
> >
> > PR rtl-optimization/63483
> > * alias.c (true_dependence_1): Do not exit early for MEM_READONLY_P
> > references when alignment ANDs are involved.
> > (write_dependence_p): Ditto.
> > (may_alias_p): Ditto.
> >
> > The complete backport was tested on alpha-linux-gnu,
> > alphaev68-linux-gnu and x86_64-linux-gnu on 4.9 branch.
> >
> > OK for branch?
> >
> > Uros.
> 
> 

-- 
Richard Biener 
SUSE LINUX GmbH, GF: Jeff Hawn, Jennifer Guild, Felix Imendoerffer, HRB 21284
(AG Nuernberg)
Maxfeldstrasse 5, 90409 Nuernberg, Germany

Re: [PATCH, libfortran] PR 47007, 61847 Locale failures in libgfortran

[Janne Blomqvist]

On Wed, Nov 5, 2014 at 1:07 PM, Jakub Jelinek  wrote:
> On Wed, Nov 05, 2014 at 12:48:01PM +0200, Janne Blomqvist wrote:
>> @@ -3528,6 +3533,11 @@ finalize_transfer (st_parameter_dt *dtp)
>>if ((dtp->common.flags & IOPARM_DT_HAS_SIZE) != 0)
>>  *dtp->size = dtp->u.p.size_used;
>>
>> +#ifdef HAVE_USELOCALE
>> +  if (dtp->u.p.old_locale != (locale_t) 0)
>> +uselocale (dtp->u.p.old_locale);
>> +#endif
>
> I wonder if you shouldn't clear dtp->u.p.old_locale here too, so
> that uselocale isn't called again.

Sure. In principle this shouldn't be needed, since IIRC the entire
dtp->u.p structure is set to 0 when starting an IO operation,  but
OTOH the cost is insignificant.


-- 
Janne Blomqvist

[ARM] RFA: Use new rtl iterators in arm_find_sub_rtx_with_code

[Richard Sandiford]

I think these functions only want to iterate over instruction patterns
rather than whole instructions (which would include things like
REG_EQUAL notes), since only the patterns are relevant for finding
dependencies.  There's then no need to check for null rtxes.

Tested by making sure there were no code changes for gcc.dg, gcc.c-torture
and g++.dg for plain arm-linux-gnueabi and aarch64-linux-gnu.  Ramana also
asked me to try -mcpu=cortex-a7, -mcpu=cortex-a9, -mcpu=arm9tdmi and
-mcpu=cortex-a15.  There were differences in:

   gcc.c-torture/execute/20060110-2.c
   gcc.c-torture/execute/ashrdi-1.c and
   gcc.dg/tree-ssa/pr24627.c

for -mcpu=cortex-a7 and no differences for the other combinations.
The A7 differences were due to the way that arm_get_set_operands handles
multi-set instructions such as:

(set (reg:CC_C 100 cc)
(compare:CC_C (plus:SI (reg:SI 8 r8 [orig:121 a ] [121])
(reg:SI 0 r0 [orig:122 b ] [122]))
(reg:SI 8 r8 [orig:121 a ] [121])))
(set (reg:SI 2 r2 [orig:120 D.4117 ] [120])
(plus:SI (reg:SI 8 r8 [orig:121 a ] [121])
(reg:SI 0 r0 [orig:122 b ] [122])))

for_each_rtx iterates over the subrtxes in forward order, so
arm_get_set_operands would pick the set of CC.  The new iterator pushes
the contents of a PARALLEL onto a stack and pulls them in reverse order,
so arm_get_set_operands would pick the set of r2.  This means that after
the patch the code sees a producer/consumer relationship that it
previously missed.  I think the new behaviour is what was intended.

This code shouldn't really be relying on a particular iteration order
though.  There's a dependency if any SET in the potential producer sets
a register used by the potential consumer.  I think any fix for that
should be done separately from the iterator rewrite.

OK to install?

Thanks,
Richard


gcc/
* config/arm/aarch-common.c: Include rtl-iter.h.
(search_term, arm_find_sub_rtx_with_search_term): Delete.
(arm_find_sub_rtx_with_code): Use FOR_EACH_SUBRTX_VAR.
(arm_get_set_operands): Pass the insn pattern rather than the
insn itself.
(arm_no_early_store_addr_dep): Likewise.

Index: gcc/config/arm/aarch-common.c
===
--- gcc/config/arm/aarch-common.c   2014-10-25 09:42:00.631168827 +0100
+++ gcc/config/arm/aarch-common.c   2014-10-25 09:51:24.212872553 +0100
@@ -30,6 +30,7 @@
 #include "tree.h"
 #include "c-family/c-common.h"
 #include "rtl.h"
+#include "rtl-iter.h"
 
 /* In ARMv8-A there's a general expectation that AESE/AESMC
and AESD/AESIMC sequences of the form:
@@ -68,13 +69,6 @@ aarch_crypto_can_dual_issue (rtx_insn *p
   return 0;
 }
 
-typedef struct
-{
-  rtx_code search_code;
-  rtx search_result;
-  bool find_any_shift;
-} search_term;
-
 /* Return TRUE if X is either an arithmetic shift left, or
is a multiplication by a power of two.  */
 bool
@@ -96,68 +90,32 @@ static rtx_code shift_rtx_codes[] =
   { ASHIFT, ROTATE, ASHIFTRT, LSHIFTRT,
 ROTATERT, ZERO_EXTEND, SIGN_EXTEND };
 
-/* Callback function for arm_find_sub_rtx_with_code.
-   DATA is safe to treat as a SEARCH_TERM, ST.  This will
-   hold a SEARCH_CODE.  PATTERN is checked to see if it is an
-   RTX with that code.  If it is, write SEARCH_RESULT in ST
-   and return 1.  Otherwise, or if we have been passed a NULL_RTX
-   return 0.  If ST.FIND_ANY_SHIFT then we are interested in
-   anything which can reasonably be described as a SHIFT RTX.  */
-static int
-arm_find_sub_rtx_with_search_term (rtx *pattern, void *data)
-{
-  search_term *st = (search_term *) data;
-  rtx_code pattern_code;
-  int found = 0;
-
-  gcc_assert (pattern);
-  gcc_assert (st);
-
-  /* Poorly formed patterns can really ruin our day.  */
-  if (*pattern == NULL_RTX)
-return 0;
-
-  pattern_code = GET_CODE (*pattern);
-
-  if (st->find_any_shift)
-{
-  unsigned i = 0;
-
-  /* Left shifts might have been canonicalized to a MULT of some
-power of two.  Make sure we catch them.  */
-  if (arm_rtx_shift_left_p (*pattern))
-   found = 1;
-  else
-   for (i = 0; i < ARRAY_SIZE (shift_rtx_codes); i++)
- if (pattern_code == shift_rtx_codes[i])
-   found = 1;
-}
-
-  if (pattern_code == st->search_code)
-found = 1;
-
-  if (found)
-st->search_result = *pattern;
-
-  return found;
-}
-
-/* Traverse PATTERN looking for a sub-rtx with RTX_CODE CODE.  */
+/* Traverse PATTERN looking for a sub-rtx with RTX_CODE CODE.
+   If FIND_ANY_SHIFT then we are interested in anything which can
+   reasonably be described as a SHIFT RTX.  */
 static rtx
 arm_find_sub_rtx_with_code (rtx pattern, rtx_code code, bool find_any_shift)
 {
-  search_term st;
-  int result = 0;
+  subrtx_var_iterator::array_type array;
+  FOR_EACH_SUBRTX_VAR (iter, array, pattern, NONCONST)
+{
+  rtx x = *iter;
+  if (f

[ARM] RFA: Use new rtl iterators in arm_cannot_copy_insn

[Richard Sandiford]

Tested in the same way as the aarch-common.c patch.  OK to install?

Thanks,
Richard


gcc/
* config/arm/arm.c (arm_note_pic_base): Delete.
(arm_cannot_copy_insn_p): Use FOR_EACH_SUBRTX.

Index: gcc/config/arm/arm.c
===
--- gcc/config/arm/arm.c2014-11-05 11:48:55.030053470 +
+++ gcc/config/arm/arm.c2014-11-05 11:48:57.406073646 +
@@ -13157,16 +13157,6 @@ tls_mentioned_p (rtx x)
 
 /* Must not copy any rtx that uses a pc-relative address.  */
 
-static int
-arm_note_pic_base (rtx *x, void *date ATTRIBUTE_UNUSED)
-{
-  if (GET_CODE (*x) == UNSPEC
-  && (XINT (*x, 1) == UNSPEC_PIC_BASE
- || XINT (*x, 1) == UNSPEC_PIC_UNIFIED))
-return 1;
-  return 0;
-}
-
 static bool
 arm_cannot_copy_insn_p (rtx_insn *insn)
 {
@@ -13175,7 +13165,16 @@ arm_cannot_copy_insn_p (rtx_insn *insn)
   if (recog_memoized (insn) == CODE_FOR_tlscall)
 return true;
 
-  return for_each_rtx (&PATTERN (insn), arm_note_pic_base, NULL);
+  subrtx_iterator::array_type array;
+  FOR_EACH_SUBRTX (iter, array, PATTERN (insn), ALL)
+{
+  const_rtx x = *iter;
+  if (GET_CODE (x) == UNSPEC
+ && (XINT (x, 1) == UNSPEC_PIC_BASE
+ || XINT (x, 1) == UNSPEC_PIC_UNIFIED))
+   return true;
+}
+  return false;
 }
 
 enum rtx_code

[ARM] RFA: Use new rtl iterators in arm_tls_referenced_p

[Richard Sandiford]

Tested in the same way as the aarch-common.c patch.  OK to install?

Thanks,
Richard


gcc/
* config/arm/arm.c: Include rtl-iter.h.
(arm_tls_referenced_p_1): Delete.
(arm_tls_referenced_p): Use FOR_EACH_SUBRTX.

Index: gcc/config/arm/arm.c
===
--- gcc/config/arm/arm.c2014-11-02 19:59:27.588237213 +
+++ gcc/config/arm/arm.c2014-11-05 11:48:55.030053470 +
@@ -82,6 +82,7 @@
 #include "gimple-expr.h"
 #include "builtins.h"
 #include "tm-constrs.h"
+#include "rtl-iter.h"
 
 /* Forward definitions of types.  */
 typedef struct minipool_nodeMnode;
@@ -8078,25 +8079,6 @@ thumb_legitimize_reload_address (rtx *x_
   return NULL;
 }
 
-/* Test for various thread-local symbols.  */
-
-/* Helper for arm_tls_referenced_p.  */
-
-static int
-arm_tls_operand_p_1 (rtx *x, void *data ATTRIBUTE_UNUSED)
-{
-  if (GET_CODE (*x) == SYMBOL_REF)
-return SYMBOL_REF_TLS_MODEL (*x) != 0;
-
-  /* Don't recurse into UNSPEC_TLS looking for TLS symbols; these are
- TLS offsets, not real symbol references.  */
-  if (GET_CODE (*x) == UNSPEC
-  && XINT (*x, 1) == UNSPEC_TLS)
-return -1;
-
-  return 0;
-}
-
 /* Return TRUE if X contains any TLS symbol references.  */
 
 bool
@@ -8105,7 +8087,19 @@ arm_tls_referenced_p (rtx x)
   if (! TARGET_HAVE_TLS)
 return false;
 
-  return for_each_rtx (&x, arm_tls_operand_p_1, NULL);
+  subrtx_iterator::array_type array;
+  FOR_EACH_SUBRTX (iter, array, x, ALL)
+{
+  const_rtx x = *iter;
+  if (GET_CODE (x) == SYMBOL_REF && SYMBOL_REF_TLS_MODEL (x) != 0)
+   return true;
+
+  /* Don't recurse into UNSPEC_TLS looking for TLS symbols; these are
+TLS offsets, not real symbol references.  */
+  if (GET_CODE (x) == UNSPEC && XINT (x, 1) == UNSPEC_TLS)
+   iter.skip_subrtxes ();
+}
+  return false;
 }
 
 /* Implement TARGET_LEGITIMATE_CONSTANT_P.

[AArch64] RFA: Use new rtl iterators in arm_cannot_copy_insn

[Richard Sandiford]

This is part of a series to remove uses of for_each_rtx from the ports.

Tested by making sure there were no code changes for gcc.dg, gcc.c-torture
and g++.dg for aarch64-linux-gnu.  OK to install?

Thanks,
Richard


gcc/
* config/aarch64/aarch64.c: Include rtl-iter.h.
(aarch64_tls_operand_p_1): Delete.
(aarch64_tls_operand_p): Use FOR_EACH_SUBRTX.

Index: gcc/config/aarch64/aarch64.c
===
--- gcc/config/aarch64/aarch64.c2014-11-02 19:59:26.977231633 +
+++ gcc/config/aarch64/aarch64.c2014-11-05 11:48:59.982095520 +
@@ -2791,28 +2791,23 @@ aarch64_output_mi_thunk (FILE *file, tre
   reload_completed = 0;
 }
 
-static int
-aarch64_tls_operand_p_1 (rtx *x, void *data ATTRIBUTE_UNUSED)
-{
-  if (GET_CODE (*x) == SYMBOL_REF)
-return SYMBOL_REF_TLS_MODEL (*x) != 0;
-
-  /* Don't recurse into UNSPEC_TLS looking for TLS symbols; these are
- TLS offsets, not real symbol references.  */
-  if (GET_CODE (*x) == UNSPEC
-  && XINT (*x, 1) == UNSPEC_TLS)
-return -1;
-
-  return 0;
-}
-
 static bool
 aarch64_tls_referenced_p (rtx x)
 {
   if (!TARGET_HAVE_TLS)
 return false;
-
-  return for_each_rtx (&x, aarch64_tls_operand_p_1, NULL);
+  subrtx_iterator::array_type array;
+  FOR_EACH_SUBRTX (iter, array, x, ALL)
+{
+  const_rtx x = *iter;
+  if (GET_CODE (x) == SYMBOL_REF && SYMBOL_REF_TLS_MODEL (x) != 0)
+   return true;
+  /* Don't recurse into UNSPEC_TLS looking for TLS symbols; these are
+TLS offsets, not real symbol references.  */
+  if (GET_CODE (x) == UNSPEC && XINT (x, 1) == UNSPEC_TLS)
+   iter.skip_subrtxes ();
+}
+  return false;
 }
 
 

Re: [PATCH, libfortran] PR 47007, 61847 Locale failures in libgfortran

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 01:48:32PM +0200, Janne Blomqvist wrote:
> On Wed, Nov 5, 2014 at 1:07 PM, Jakub Jelinek  wrote:
> > On Wed, Nov 05, 2014 at 12:48:01PM +0200, Janne Blomqvist wrote:
> >> @@ -3528,6 +3533,11 @@ finalize_transfer (st_parameter_dt *dtp)
> >>if ((dtp->common.flags & IOPARM_DT_HAS_SIZE) != 0)
> >>  *dtp->size = dtp->u.p.size_used;
> >>
> >> +#ifdef HAVE_USELOCALE
> >> +  if (dtp->u.p.old_locale != (locale_t) 0)
> >> +uselocale (dtp->u.p.old_locale);
> >> +#endif
> >
> > I wonder if you shouldn't clear dtp->u.p.old_locale here too, so
> > that uselocale isn't called again.
> 
> Sure. In principle this shouldn't be needed, since IIRC the entire
> dtp->u.p structure is set to 0 when starting an IO operation,  but
> OTOH the cost is insignificant.

If finalize transfer is guaranteed to be called exactly once on
a structure on which dtp->u.p.old_locale has been set, then it is ok as is,
though clearing it after such an restoration operation is a good style;
you could e.g. then assert it is NULL when the done with a particular dtp
structure in debugging builds etc.

Jakub

Re: [PARCH 1/2, x86, PR63534] Fix darwin bootstrap

[Eric Botcazou]

> Now if your argument is that IRA/LRA handle this, that's fine, a pointer
> to that code would be appreciated so that it can be quickly audited.
> Certainly the old local-alloc/global-alloc had magic for setjmp/longjmp
> and maybe IRA/LRA does too, but it's better to be sure than just assume.

See ira-lives.c:1217 and below.

-- 
Eric Botcazou

Re: The nvptx port [8/11+] Write undefined decls.

[Bernd Schmidt]

On 10/22/2014 08:11 PM, Jeff Law wrote:

I'm not going to insist you do this in the same way as the PA.  That was
a different era -- we had significant motivation to make things work in
such a way that everything could be buried in the pa specific files.
That sometimes led to less than optimal approaches to fix certain problems.


So... is this patch approved?


Bernd

Re: [Patch, testsuite] [AArch64,ARM] support bswap tests on aarch64_be

[Christophe Lyon]

On 31 October 2014 18:15, Ramana Radhakrishnan
 wrote:
> On Wed, Oct 29, 2014 at 1:22 PM, Christophe Lyon
>  wrote:
>> Hi,
>>
>> Following discussions after Thomas's patches improving bswap support
>> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg01279.html
>>
>> I noticed that:
>> * the associated tests weren't executed on aarch64_be
>> * ARM targets older than v6 do not support the needed instructions.
>>
>> The attached patch changes check_effective_target_bswap():
>> - accept aarch64*-*-* instead of aarch64-*-*
>> - when target is arm*-*-*, check __ARM_ARCH >= 6
>>
>> 2014-10-29  Christophe Lyon  
>>
>> * lib/target-supports.exp (check_effective_target_bswap): Update
>> conditions for AArch64 and ARM targets.
>>
>> OK?
>
> The ARM (AArch32) changes are ok.
>
> Ramana
>
Thank you Ramana.

Marcus, I believe the AArch64 part is obvious. OK?

Christophe.

>>
>> Christophe.

Re: [Patch ARM-AArch64/testsuite v3 00/21] Neon intrinsics executable tests

[Christophe Lyon]

ping?


On 26 October 2014 17:50, Christophe Lyon  wrote:
> On 24 October 2014 10:07, Marcus Shawcroft  wrote:
>> On 21 October 2014 14:02, Christophe Lyon  wrote:
>>> This patch series is an updated version of the series I sent here:
>>> https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00022.html
>>>
>>> I addressed comments from Marcus and Richard, and decided to skip
>>> support for half-precision variants for the time being. I'll post
>>> dedicated patches later.
>>>
>>> Compared to v2:
>>> - the directory containing the new tests is named
>>>   gcc.target/aarch64/adv-simd instead of
>>>   gcc.target/aarch64/neon-intrinsics.
>>> - the driver is named adv-simd.exp instead of neon-intrinsics.exp
>>> - the driver is guarded against the new test parallelization framework
>>> - the README file uses 'Advanced SIMD (Neon)' instead of 'Neon'
>>>
>>
>> Thank you Christophe.  Please commit all 21 patches in the series.
>>
> Thanks, I have committed the whole series.
>
> I've just realized afterwards that the tests aren't guarded against
> targets not supporting Neon.
>
> How about adding the attached small patch?
>
> (ChangeLog incorrectly formatted :-()
> 2014-10-26  Christophe Lyon  
>
> gcc.target/aarch64/advsimd-intrinsics/advsimd-intrinsics.exp: Skip
> tests if target does not support Neon.
>
>
> Christophe
>
>> /Marcus

Re: [Patch, testsuite] [AArch64,ARM] support bswap tests on aarch64_be

[Marcus Shawcroft]

On 5 November 2014 12:08, Christophe Lyon  wrote:
> On 31 October 2014 18:15, Ramana Radhakrishnan
>  wrote:
>> On Wed, Oct 29, 2014 at 1:22 PM, Christophe Lyon
>>  wrote:
>>> Hi,
>>>
>>> Following discussions after Thomas's patches improving bswap support
>>> https://gcc.gnu.org/ml/gcc-patches/2014-09/msg01279.html
>>>
>>> I noticed that:
>>> * the associated tests weren't executed on aarch64_be
>>> * ARM targets older than v6 do not support the needed instructions.
>>>
>>> The attached patch changes check_effective_target_bswap():
>>> - accept aarch64*-*-* instead of aarch64-*-*
>>> - when target is arm*-*-*, check __ARM_ARCH >= 6
>>>
>>> 2014-10-29  Christophe Lyon  
>>>
>>> * lib/target-supports.exp (check_effective_target_bswap): Update
>>> conditions for AArch64 and ARM targets.
>>>
>>> OK?
>>
>> The ARM (AArch32) changes are ok.
>>
>> Ramana
>>
> Thank you Ramana.
>
> Marcus, I believe the AArch64 part is obvious. OK?

Yep, it looks obvious to me.  /Marcus

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/05/2014 02:23 PM, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 11:50:20AM +0100, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:

Are you going to work on ASan soon?  I could rebase my patches on top of
Marek's infrastructure.


I'm not going to work on ASan today or tomorrow, but it'd be nice to
get this ASan opt in in this stage1.

So if you can rebase your patch, I think that will be appreciated.


Note, the algorithm we were discussing with Honza for the
"is there any possibility of a freeing call on the path between a dominating
and dominated ASAN_CHECK"


Right.  Let me see then if I can implement the following soon, maybe
it makes sense to rebase Yuri's patch only on top of this algorithm.


The algorithm looks like should_hoist_expr_to_dom in gcse.c btw.

BTW have you considered relaxing the non-freeing restriction to not drop 
accesses to globals and stack variables? I wonder if we could win 
something there.


-Y

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 03:16:49PM +0300, Yury Gribov wrote:
> On 11/05/2014 02:23 PM, Marek Polacek wrote:
> >On Wed, Nov 05, 2014 at 11:50:20AM +0100, Jakub Jelinek wrote:
> >>On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:
> >>>On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:
> Are you going to work on ASan soon?  I could rebase my patches on top of
> Marek's infrastructure.
> >>>
> >>>I'm not going to work on ASan today or tomorrow, but it'd be nice to
> >>>get this ASan opt in in this stage1.
> >>>
> >>>So if you can rebase your patch, I think that will be appreciated.
> >>
> >>Note, the algorithm we were discussing with Honza for the
> >>"is there any possibility of a freeing call on the path between a dominating
> >>and dominated ASAN_CHECK"
> >
> >Right.  Let me see then if I can implement the following soon, maybe
> >it makes sense to rebase Yuri's patch only on top of this algorithm.
> 
> The algorithm looks like should_hoist_expr_to_dom in gcse.c btw.
> 
> BTW have you considered relaxing the non-freeing restriction to not drop
> accesses to globals and stack variables? I wonder if we could win something
> there.

Wouldn't it break most uses of __asan_poison_memory_region ?

Jakub

Re: [RTL, Patch] Int div by constant compilation enhancement

[Eric Botcazou]

> 2014-11-03  Alex Velenko  
> 
>  * simplify-rtx.c (simplify_binary_operation_1): Div check added.
>  * rtl.h (SUBREG_P): New macro added.

Present tense in Change entries:

* rtl.h (SUBREG_P): New macro.
* simplify-rtx.c (simplify_binary_operation_1): Simplify consecutive
right shifts in combination with a low-part operation.

Can't the 'c1 == size (M2) - size (M1)' condition be relaxed?

-- 
Eric Botcazou

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/05/2014 03:21 PM, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 03:16:49PM +0300, Yury Gribov wrote:

On 11/05/2014 02:23 PM, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 11:50:20AM +0100, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:

Are you going to work on ASan soon?  I could rebase my patches on top of
Marek's infrastructure.


I'm not going to work on ASan today or tomorrow, but it'd be nice to
get this ASan opt in in this stage1.

So if you can rebase your patch, I think that will be appreciated.


Note, the algorithm we were discussing with Honza for the
"is there any possibility of a freeing call on the path between a dominating
and dominated ASAN_CHECK"


Right.  Let me see then if I can implement the following soon, maybe
it makes sense to rebase Yuri's patch only on top of this algorithm.


The algorithm looks like should_hoist_expr_to_dom in gcse.c btw.

BTW have you considered relaxing the non-freeing restriction to not drop
accesses to globals and stack variables? I wonder if we could win something
there.


Wouldn't it break most uses of __asan_poison_memory_region ?


Most probably but I wonder if we should ask people to simply do asm 
volatile with memory clobber in this case?  And we probably shouldn't 
call the whole thing is_nonfreeing anyway.


-Y

Re: [PATCH 2/n] OpenMP 4.0 offloading infrastructure: LTO streaming

[Ilya Verbin]

On 03 Nov 10:24, Jakub Jelinek wrote:
> On Tue, Oct 28, 2014 at 10:30:47PM +0300, Ilya Verbin wrote:
> > @@ -474,6 +475,13 @@ cgraph_node::create (tree decl)
> >gcc_assert (TREE_CODE (decl) == FUNCTION_DECL);
> >  
> >node->decl = decl;
> > +
> > +  if (lookup_attribute ("omp declare target", DECL_ATTRIBUTES (decl)))
> > +{
> > +  node->offloadable = 1;
> > +  g->have_offload = true;
> > +}
> 
> I wonder if we shouldn't optimize here and call lookup_attribute only
> if there is a chance that the attribute might be present, so guard with
> flag_openmp (and flag_openacc later on?).  During LTO the cgraph nodes
> are streamed in and supposedly the flag offloadable too.
> 
> > @@ -2129,8 +2141,12 @@ symbol_table::compile (void)
> >  fprintf (stderr, "Performing interprocedural optimizations\n");
> >state = IPA;
> >  
> > +  /* OpenMP offloading requires LTO infrastructure.  */
> > +  if (!in_lto_p && flag_openmp && g->have_offload)
> > +flag_generate_lto = 1;
> 
> On the other side, do you need flag_openmp here?  Supposedly g->have_offload
> would already been set if needed.

Done, flag_openmp moved from symbol_table::compile to cgraph_node::create and
varpool_node::get_create.  OK for trunk?

Maybe also with this change?

diff --git a/gcc/omp-low.c b/gcc/omp-low.c
index 4e9ed25..beae5b5 100644
--- a/gcc/omp-low.c
+++ b/gcc/omp-low.c
@@ -1653,8 +1653,7 @@ scan_sharing_clauses (tree clauses, omp_context *ctx)
  if (OMP_CLAUSE_CODE (c) == OMP_CLAUSE_MAP
  && DECL_P (decl)
  && is_global_var (maybe_lookup_decl_in_outer_ctx (decl, ctx))
- && lookup_attribute ("omp declare target",
-  DECL_ATTRIBUTES (decl)))
+ && varpool_node::get_create (decl)->offloadable)
break;
  if (OMP_CLAUSE_CODE (c) == OMP_CLAUSE_MAP
  && OMP_CLAUSE_MAP_KIND (c) == OMP_CLAUSE_MAP_POINTER)
@@ -1794,8 +1793,7 @@ scan_sharing_clauses (tree clauses, omp_context *ctx)
  decl = OMP_CLAUSE_DECL (c);
  if (DECL_P (decl)
  && is_global_var (maybe_lookup_decl_in_outer_ctx (decl, ctx))
- && lookup_attribute ("omp declare target",
-  DECL_ATTRIBUTES (decl)))
+ && varpool_node::get_create (decl)->offloadable)
break;
  if (DECL_P (decl))
{

Thanks,
  -- Ilya


---

diff --git a/gcc/cgraph.c b/gcc/cgraph.c
index 9a47ba2..a491886 100644
--- a/gcc/cgraph.c
+++ b/gcc/cgraph.c
@@ -70,6 +70,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "tree-dfa.h"
 #include "profile.h"
 #include "params.h"
+#include "context.h"
 
 /* FIXME: Only for PROP_loops, but cgraph shouldn't have to know about this.  
*/
 #include "tree-pass.h"
@@ -474,6 +475,14 @@ cgraph_node::create (tree decl)
   gcc_assert (TREE_CODE (decl) == FUNCTION_DECL);
 
   node->decl = decl;
+
+  if (flag_openmp
+  && lookup_attribute ("omp declare target", DECL_ATTRIBUTES (decl)))
+{
+  node->offloadable = 1;
+  g->have_offload = true;
+}
+
   node->register_symbol ();
 
   if (DECL_CONTEXT (decl) && TREE_CODE (DECL_CONTEXT (decl)) == FUNCTION_DECL)
diff --git a/gcc/cgraph.h b/gcc/cgraph.h
index 377adce..4988f2d 100644
--- a/gcc/cgraph.h
+++ b/gcc/cgraph.h
@@ -463,6 +463,13 @@ public:
   /* Set when init priority is set.  */
   unsigned in_init_priority_hash : 1;
 
+  /* Set when symbol needs to be streamed into LTO bytecode for LTO, or in case
+ of offloading, for separate compilation for a different target.  */
+  unsigned need_lto_streaming : 1;
+
+  /* Set when symbol can be streamed into bytecode for offloading.  */
+  unsigned offloadable : 1;
+
 
   /* Ordering of all symtab entries.  */
   int order;
diff --git a/gcc/cgraphunit.c b/gcc/cgraphunit.c
index 3e76bf0..83ab419 100644
--- a/gcc/cgraphunit.c
+++ b/gcc/cgraphunit.c
@@ -218,6 +218,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "tree-nested.h"
 #include "gimplify.h"
 #include "dbgcnt.h"
+#include "lto-section-names.h"
 
 /* Queue of cgraph nodes scheduled to be added into cgraph.  This is a
secondary queue used during optimization to accommodate passes that
@@ -2049,7 +2050,18 @@ ipa_passes (void)
 targetm.asm_out.lto_start ();
 
   if (!in_lto_p)
-ipa_write_summaries ();
+{
+  if (g->have_offload)
+   {
+ section_name_prefix = OFFLOAD_SECTION_NAME_PREFIX;
+ ipa_write_summaries (true);
+   }
+  if (flag_lto)
+   {
+ section_name_prefix = LTO_SECTION_NAME_PREFIX;
+ ipa_write_summaries (false);
+   }
+}
 
   if (flag_generate_lto)
 targetm.asm_out.lto_end ();
@@ -2129,8 +2141,12 @@ symbol_table::compile (void)
 fprintf (stderr, "Performing interprocedural optimizations\n");
   state = IPA;
 
+  /* Offloading requires LTO infrastructure.  */
+  if (!in_lto_p && g->have_offload)
+flag_generate_lto = 1;
+
   /* If LTO is enabled, i

Re: [PATCH 2/n] OpenMP 4.0 offloading infrastructure: LTO streaming

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 03:46:55PM +0300, Ilya Verbin wrote:
> Maybe also with this change?
> 
> diff --git a/gcc/omp-low.c b/gcc/omp-low.c
> index 4e9ed25..beae5b5 100644
> --- a/gcc/omp-low.c
> +++ b/gcc/omp-low.c
> @@ -1653,8 +1653,7 @@ scan_sharing_clauses (tree clauses, omp_context *ctx)
> if (OMP_CLAUSE_CODE (c) == OMP_CLAUSE_MAP
> && DECL_P (decl)
> && is_global_var (maybe_lookup_decl_in_outer_ctx (decl, ctx))
> -   && lookup_attribute ("omp declare target",
> -DECL_ATTRIBUTES (decl)))
> +   && varpool_node::get_create (decl)->offloadable)
>   break;
> if (OMP_CLAUSE_CODE (c) == OMP_CLAUSE_MAP
> && OMP_CLAUSE_MAP_KIND (c) == OMP_CLAUSE_MAP_POINTER)
> @@ -1794,8 +1793,7 @@ scan_sharing_clauses (tree clauses, omp_context *ctx)
> decl = OMP_CLAUSE_DECL (c);
> if (DECL_P (decl)
> && is_global_var (maybe_lookup_decl_in_outer_ctx (decl, ctx))
> -   && lookup_attribute ("omp declare target",
> -DECL_ATTRIBUTES (decl)))
> +   && varpool_node::get_create (decl)->offloadable)
>   break;
> if (DECL_P (decl))
>   {

That looks reasonable (of course if the other patch is committed).

> --- a/gcc/cgraph.c
> +++ b/gcc/cgraph.c
> @@ -70,6 +70,7 @@ along with GCC; see the file COPYING3.  If not see
>  #include "tree-dfa.h"
>  #include "profile.h"
>  #include "params.h"
> +#include "context.h"
>  
>  /* FIXME: Only for PROP_loops, but cgraph shouldn't have to know about this. 
>  */
>  #include "tree-pass.h"
> @@ -474,6 +475,14 @@ cgraph_node::create (tree decl)
>gcc_assert (TREE_CODE (decl) == FUNCTION_DECL);
>  
>node->decl = decl;
> +
> +  if (flag_openmp
> +  && lookup_attribute ("omp declare target", DECL_ATTRIBUTES (decl)))
> +{
> +  node->offloadable = 1;
> +  g->have_offload = true;
> +}
> +
>node->register_symbol ();

LGTM.

Jakub

Re: [gofrontend-dev] [PATCH 1/4] Gccgo port to s390[x] -- part II

[Dominik Vogt]

On Wed, Nov 05, 2014 at 11:31:28AM +0100, Dominik Vogt wrote:
> On Tue, Nov 04, 2014 at 02:39:34PM -0800, Ian Taylor wrote:
> > Note that libgo/runtime/runtime.c now refers to S390_HAVE_STCKF.  It's
> > not obvious to me that that is defined anywhere.  Perhaps it is in a
> > later patch in this series--I haven't looked.

The attached patch fixes this (by using stckf unconditionally).

Ciao

Dominik ^_^  ^_^

-- 

Dominik Vogt
IBM Germany
ChangeLog
2014-11-05  Dominik Vogt  

* libgo/runtime/runtime.c (runtime_cputicks): s390: use stckf
unconditionally
>From 9b787b6190b5dd12e5be42e4c65a6907ca99bb59 Mon Sep 17 00:00:00 2001
From: Dominik Vogt 
Date: Wed, 5 Nov 2014 13:38:23 +0100
Subject: [PATCH 1/2] libgo: Use stckf unconditionally on s390[x].

---
 libgo/runtime/runtime.c | 12 ++--
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/libgo/runtime/runtime.c b/libgo/runtime/runtime.c
index abc1aff..496e77b 100644
--- a/libgo/runtime/runtime.c
+++ b/libgo/runtime/runtime.c
@@ -195,12 +195,12 @@ runtime_cputicks(void)
   asm("rdtsc" : "=a" (low), "=d" (high));
   return (int64)(((uint64)high << 32) | (uint64)low);
 #elif defined (__s390__) || defined (__s390x__)
-  uint64 clock;
-#ifdef S390_HAVE_STCKF
-  asm("stckf\t%0" : "=Q" (clock) : : );
-#else
-  clock = 0;
-#endif
+  uint64 clock = 0;
+  /* stckf may not write the return variable in case of a clock error, so make
+ it read-write to prevent that the initialisation is optimised out.
+ Note: Targets below z9-109 will crash when executing store clock fast, i.e.
+ we don't support Go for machines older than that.  */
+  asm volatile(".insn s,0xb27c,%0" /* stckf */ : "+Q" (clock) : : "cc" );
   return (int64)clock;
 #else
   // FIXME: implement for other processors.
-- 
1.8.4.2

Re: [PARCH 1/2, x86, PR63534] Fix darwin bootstrap

[Evgeny Stupachenko]

On Tue, Nov 4, 2014 at 1:40 AM, Jeff Law  wrote:
> On 11/01/14 06:39, Evgeny Stupachenko wrote:
>>
>> When PIC register is pseudo there is nothing special about it's value
>> that setjmp can hurt. So if the pseudo register lives across
>> setjmp_receiver RA should care about correct allocation (in case it is
>> not saved/restored, it should go on stack). gcc.dg tests and specs
>> I've tested behave like this.
>
> If the allocator picked a call-clobbered register for the PIC register, then
> we're obviously OK since the setjmp has to be expected to clobber the PIC
> register.
>
> But if the PIC register is in a call-saved register, then it's going to be
> assumed to not be clobbered across calls and I don't believe that is
> guaranteed for builtin setjmp/longjmp.  Those restore SP, FP and an ARGP,
> but not anything else by default.

I still don't see what is special for PIC register here. PIC pseudo
now behave as every other pseudo register.
If we assume that setjmp can change a pseudo register value we need
IRA/LRA "magic" for each pseudo register.

I believe that when we had EBX fixed, IRA/LRA don't save/restore it anywhere.
Therefore we had to care about EBX value in special cases like
setjmp/non-local goto.
Now RA cares about PIC pseudo as well as about correct allocation for
any pseudo register.

>
> So the callee might have clobbered the call saved hard register, expecting
> to restore its value in its epilogue.  But due to the longjmp, that epilogue
> never gets called and thus the call-saved register won't have the right
> value in the receiver.
>
> Now if your argument is that IRA/LRA handle this, that's fine, a pointer to
> that code would be appreciated so that it can be quickly audited. Certainly
> the old local-alloc/global-alloc had magic for setjmp/longjmp and maybe
> IRA/LRA does too, but it's better to be sure than just assume.
>
>>
>> The initial problem comes from non-local goto as it tries to emit
>> pseudo PIC register after reload.
>
> ?  You mean it emits a reference to the pseudo into RTL?  That would
> indicate that the allocators never put the pseudo into a hard register?!?
> RTL dumps with a few pointers to key insns would help here.

Correct, that is why Darwin crashes with ICE on non-local goto.
We still have:

(define_insn_and_split "nonlocal_goto_receiver"
  [(unspec_volatile [(const_int 0)] UNSPECV_NLGR)]
  "TARGET_MACHO && !TARGET_64BIT && flag_pic"
  "#"
  "&& reload_completed"
  [(const_int 0)]
{
  if (crtl->uses_pic_offset_table)
{
  rtx xops[3];
  rtx label_rtx = gen_label_rtx ();
  rtx tmp;

  /* Get a new pic base.  */
  emit_insn (gen_set_got_labelled (pic_offset_table_rtx, label_rtx));
.
Here for MAC only we are trying to use pseudo PIC:
pic_offset_table_rtx when reload_completed.

>
> jeff
>
>

Re: [PARCH 1/2, x86, PR63534] Fix darwin bootstrap

[Evgeny Stupachenko]

We don't emit extra SET_GOT. That is beneficial.
As for stack usage, that is RA to decide which register is more
beneficial to put on stack.

On Sat, Nov 1, 2014 at 8:33 PM, Mike Stump  wrote:
> On Nov 1, 2014, at 5:39 AM, Evgeny Stupachenko  wrote:
>> When PIC register is pseudo there is nothing special about it's value
>> that setjmp can hurt. So if the pseudo register lives across
>> setjmp_receiver RA should care about correct allocation (in case it is
>> not saved/restored, it should go on stack).
>
> So, why is consuming more stack space beneficial?

Re: The nvptx port [10/11+] Target files

[Bernd Schmidt]

On 11/04/2014 05:51 PM, Bernd Schmidt wrote:

On 11/04/2014 05:48 PM, Richard Henderson wrote:

On 10/28/2014 03:56 PM, Bernd Schmidt wrote:

+nvptx_ptx_type_from_mode (enum machine_mode mode, bool promote)
+{
+  switch (mode)
+{
+case BLKmode:
+  return ".b8";
+case BImode:
+  return ".pred";
+case QImode:
+  if (promote)
+return ".u32";
+  else
+return ".u8";
+case HImode:
+  return ".u16";


Promote here too?  Or does this have nothing to do with


+static enum machine_mode
+arg_promotion (enum machine_mode mode)
+{
+  if (mode == QImode || mode == HImode)
+return SImode;
+  return mode;
+}


No, these are different problems - the one in arg promotion is purely
about K&R C and trying to match untyped function decls with calls, while
the type_from_mode bit was about some ptx ideosyncracy. Although I
forget what the problem was, that code is more than a year old - I'll
see if I can get rid of this.


Err, no, it's quite necessary. From the manual "The .u8, .s8 and .b8 
instruction types are restricted to ld, st and cvt instructions." This 
means that if the compiler generates reasonable-looking code along the 
lines of


.reg .u8 %r70;
mov.u8 %r70,48;

you get

ptxas 2211-1.o, line 191; error   : Arguments mismatch for 
instruction 'mov'


Now, one _could_ write .cvt.u8.u32 for the load immediate, but then one 
would also have to write .cvt.u8.u8 for register-register moves, and 
that's starting to look iffy. I don't really want to rely on the ptx 
assembler to do the right thing for "conversions" from one type to itself.



Bernd

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/05/2014 03:34 PM, Yury Gribov wrote:

On 11/05/2014 03:21 PM, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 03:16:49PM +0300, Yury Gribov wrote:

On 11/05/2014 02:23 PM, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 11:50:20AM +0100, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 11:29:19AM +0100, Marek Polacek wrote:

On Wed, Nov 05, 2014 at 12:54:37PM +0300, Yury Gribov wrote:

Are you going to work on ASan soon?  I could rebase my patches on
top of
Marek's infrastructure.


I'm not going to work on ASan today or tomorrow, but it'd be nice to
get this ASan opt in in this stage1.

So if you can rebase your patch, I think that will be appreciated.


Note, the algorithm we were discussing with Honza for the
"is there any possibility of a freeing call on the path between a
dominating
and dominated ASAN_CHECK"


Right.  Let me see then if I can implement the following soon, maybe
it makes sense to rebase Yuri's patch only on top of this algorithm.


The algorithm looks like should_hoist_expr_to_dom in gcse.c btw.

BTW have you considered relaxing the non-freeing restriction to not drop
accesses to globals and stack variables? I wonder if we could win
something
there.


Wouldn't it break most uses of __asan_poison_memory_region ?


Most probably but I wonder if we should ask people to simply do asm
volatile with memory clobber in this case?  And we probably shouldn't
call the whole thing is_nonfreeing anyway.


Added Kostya to maybe comment on this.

-Y

Re: [Patch ARM-AArch64/testsuite v3 00/21] Neon intrinsics executable tests

[Ramana Radhakrishnan]

On Sun, Oct 26, 2014 at 4:50 PM, Christophe Lyon
 wrote:
> On 24 October 2014 10:07, Marcus Shawcroft  wrote:
>> On 21 October 2014 14:02, Christophe Lyon  wrote:
>>> This patch series is an updated version of the series I sent here:
>>> https://gcc.gnu.org/ml/gcc-patches/2014-07/msg00022.html
>>>
>>> I addressed comments from Marcus and Richard, and decided to skip
>>> support for half-precision variants for the time being. I'll post
>>> dedicated patches later.
>>>
>>> Compared to v2:
>>> - the directory containing the new tests is named
>>>   gcc.target/aarch64/adv-simd instead of
>>>   gcc.target/aarch64/neon-intrinsics.
>>> - the driver is named adv-simd.exp instead of neon-intrinsics.exp
>>> - the driver is guarded against the new test parallelization framework
>>> - the README file uses 'Advanced SIMD (Neon)' instead of 'Neon'
>>>
>>
>> Thank you Christophe.  Please commit all 21 patches in the series.
>>
> Thanks, I have committed the whole series.
>
> I've just realized afterwards that the tests aren't guarded against
> targets not supporting Neon.
>
> How about adding the attached small patch?
>
> (ChangeLog incorrectly formatted :-()
> 2014-10-26  Christophe Lyon  
>
> gcc.target/aarch64/advsimd-intrinsics/advsimd-intrinsics.exp: Skip
> tests if target does not support Neon.

Ok by me ...

Ramana

>
>
> Christophe
>
>> /Marcus

Re: [Patch ARM-AArch64/testsuite v3 00/21] Neon intrinsics executable tests

[Marcus Shawcroft]

On 26 October 2014 16:50, Christophe Lyon  wrote:

> I've just realized afterwards that the tests aren't guarded against
> targets not supporting Neon.
>
> How about adding the attached small patch?

+if {[istarget arm*-*-*]
+&& ![check_effective_target_arm_neon_ok]} then {
+  return
+}
+

Umm, first thought was that this is a bit of a hack but having just
discussed it with Ramana we don;t have a better alternative to hand,
so OK.

/Marcus

Re: [PATCH 3/n] OpenMP 4.0 offloading infrastructure: offload tables

[Ilya Verbin]

On 08 Oct 11:23, Jakub Jelinek wrote:
> LGTM, with the requested var/section renames.
> Would like if Honza and/or Richard had a look at the cgraph/LTO stuff
> in the patch though.

Since patch 2 was updated, this patch also should be updated.
Now the offload_vars array is filled in varpool_node::get_create .

Richard, is it OK for trunk?

Thanks,
  -- Ilya


---

diff --git a/gcc/Makefile.in b/gcc/Makefile.in
index f31af05..3db30bf 100644
--- a/gcc/Makefile.in
+++ b/gcc/Makefile.in
@@ -2303,6 +2303,7 @@ GTFILES = $(CPP_ID_DATA_H) $(srcdir)/input.h 
$(srcdir)/coretypes.h \
   $(srcdir)/tree-profile.c $(srcdir)/tree-nested.c \
   $(srcdir)/tree-parloops.c \
   $(srcdir)/omp-low.c \
+  $(srcdir)/omp-low.h \
   $(srcdir)/targhooks.c $(out_file) $(srcdir)/passes.c $(srcdir)/cgraphunit.c \
   $(srcdir)/cgraphclones.c \
   $(srcdir)/tree-phinodes.c \
diff --git a/gcc/cgraphunit.c b/gcc/cgraphunit.c
index 83ab419..bafbadb 100644
--- a/gcc/cgraphunit.c
+++ b/gcc/cgraphunit.c
@@ -218,6 +218,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "tree-nested.h"
 #include "gimplify.h"
 #include "dbgcnt.h"
+#include "omp-low.h"
 #include "lto-section-names.h"
 
 /* Queue of cgraph nodes scheduled to be added into cgraph.  This is a
diff --git a/gcc/doc/tm.texi b/gcc/doc/tm.texi
index 5036d4f..6a5a031 100644
--- a/gcc/doc/tm.texi
+++ b/gcc/doc/tm.texi
@@ -11205,6 +11205,12 @@ If defined, this function returns an appropriate 
alignment in bits for an atomic
 ISO C11 requires atomic compound assignments that may raise floating-point 
exceptions to raise exceptions corresponding to the arithmetic operation whose 
result was successfully stored in a compare-and-exchange sequence.  This 
requires code equivalent to calls to @code{feholdexcept}, @code{feclearexcept} 
and @code{feupdateenv} to be generated at appropriate points in the 
compare-and-exchange sequence.  This hook should set @code{*@var{hold}} to an 
expression equivalent to the call to @code{feholdexcept}, @code{*@var{clear}} 
to an expression equivalent to the call to @code{feclearexcept} and 
@code{*@var{update}} to an expression equivalent to the call to 
@code{feupdateenv}.  The three expressions are @code{NULL_TREE} on entry to the 
hook and may be left as @code{NULL_TREE} if no code is required in a particular 
place.  The default implementation leaves all three expressions as 
@code{NULL_TREE}.  The @code{__atomic_feraiseexcept} function from 
@code{libatomic} may be of use as part of the code generated in 
@code{*@var{update}}.
 @end deftypefn
 
+@deftypefn {Target Hook} void TARGET_RECORD_OFFLOAD_SYMBOL (tree)
+Used when offloaded functions are seen in the compilation unit and no named
+sections are available.  It is called once for each symbol that must be
+recorded in the offload function and variable table.
+@end deftypefn
+
 @defmac TARGET_SUPPORTS_WIDE_INT
 
 On older ports, large integers are stored in @code{CONST_DOUBLE} rtl
diff --git a/gcc/doc/tm.texi.in b/gcc/doc/tm.texi.in
index 5674e6c..cadf05d 100644
--- a/gcc/doc/tm.texi.in
+++ b/gcc/doc/tm.texi.in
@@ -8167,6 +8167,8 @@ and the associated definitions of those functions.
 
 @hook TARGET_ATOMIC_ASSIGN_EXPAND_FENV
 
+@hook TARGET_RECORD_OFFLOAD_SYMBOL
+
 @defmac TARGET_SUPPORTS_WIDE_INT
 
 On older ports, large integers are stored in @code{CONST_DOUBLE} rtl
diff --git a/gcc/gengtype.c b/gcc/gengtype.c
index e48b448..06c37d5 100644
--- a/gcc/gengtype.c
+++ b/gcc/gengtype.c
@@ -1843,7 +1843,7 @@ open_base_files (void)
   "tree-ssa.h", "reload.h", "cpp-id-data.h", "tree-chrec.h",
   "except.h", "output.h",  "cfgloop.h",
   "target.h", "ipa-prop.h", "lto-streamer.h", "target-globals.h",
-  "ipa-inline.h", "dwarf2out.h", NULL
+  "ipa-inline.h", "dwarf2out.h", "omp-low.h", NULL
 };
 const char *const *ifp;
 outf_p gtype_desc_c;
diff --git a/gcc/lto-cgraph.c b/gcc/lto-cgraph.c
index 45655ba..6c442cf 100644
--- a/gcc/lto-cgraph.c
+++ b/gcc/lto-cgraph.c
@@ -57,6 +57,7 @@ along with GCC; see the file COPYING3.  If not see
 #include "context.h"
 #include "pass_manager.h"
 #include "ipa-utils.h"
+#include "omp-low.h"
 
 /* True when asm nodes has been output.  */
 bool asm_nodes_output = false;
@@ -1057,6 +1058,49 @@ read_string (struct lto_input_block *ib)
   return str;
 }
 
+/* Output function/variable tables that will allow libgomp to look up offload
+   target code.  OFFLOAD_FUNCS is filled in expand_omp_target, OFFLOAD_VARS is
+   filled in ipa_passes.  In WHOPR (partitioned) mode during the WPA stage both
+   OFFLOAD_FUNCS and OFFLOAD_VARS are filled by input_offload_tables.  */
+
+void
+output_offload_tables (void)
+{
+  if (vec_safe_is_empty (offload_funcs) && vec_safe_is_empty (offload_vars))
+return;
+
+  struct lto_simple_output_block *ob
+= lto_create_simple_output_block (LTO_section_offload_table);
+
+  for (unsigned i = 0; i < vec_safe_length (offload_funcs); i++)
+{
+  streamer_write_enum (ob->main_stream, LTO_symtab_tags,
+  

Re: [PATCH][AArch64][4.8] LINK_SPEC changes for Cortex-A53 erratum 835769 workaround

[Marcus Shawcroft]

On 22 October 2014 15:20, Kyrill Tkachov  wrote:
> Hi all,
>
> This is the 4.8 backport of the LINK_SPEC changes to pass down the linker
> option
> --fix-cortex-a53-835769
>
> Bootstrapped and tested on aarch64-none-linux-gnu.
> This depends on the patches under review at:
> https://gcc.gnu.org/ml/gcc-patches/2014-10/msg01757.html
> and
> https://gcc.gnu.org/ml/gcc-patches/2014-10/msg01758.html
>
> Ok for that branch after the prerequisites go in?
>
>
> Thanks,
> Kyrill
>
> 2014-10-22  Kyrylo Tkachov  
>
> * config/aarch64/aarch64-elf-raw.h (CA53_ERR_835769_SPEC): Define.
> (LINK_SPEC): Include CA53_ERR_835769_SPEC.
> * config/aarch64/aarch64-linux.h (CA53_ERR_835769_SPEC): Define.
> (LINK_SPEC): Include CA53_ERR_835769_SPEC.

OK /Marcus

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Jakub Jelinek]

On Wed, Nov 05, 2014 at 04:13:01PM +0300, Yury Gribov wrote:
> >>Wouldn't it break most uses of __asan_poison_memory_region ?
> >
> >Most probably but I wonder if we should ask people to simply do asm
> >volatile with memory clobber in this case?  And we probably shouldn't
> >call the whole thing is_nonfreeing anyway.
> 
> Added Kostya to maybe comment on this.

Well, right now !nonfreeing_call_p is any non-builtin call or
non-leaf builtin call (i.e. builtin that might call functions in the
current CU), or free/tm_free/realloc/stack_restore.
So, by this definition __asan_poison_memory_region is also a
!nonfreeing_call_p.  Where would you like to see the volatile with
memory clobber?  You might very well just call some function that
does the free for you.
  *p = 1;
  foo (p);
  *p = 2;
and foo (p) could asm volatile ("" : : : "memory"); somewhere
and free (p) somewhere else.

If in the future we e.g. IPA-prop propagate the nonfreeing_call_p
property through the callgraph (as in, if the function you call
is non-overridable and you know the flag for it, use it), things
would still work unless you LTOed libasan together with your app
(to make that work you'd probably want to add asm volatile to
those calls, but doing it now would be very premature, or
make __asan_*poison_memory_region a builtin that would be handled
explicitly).

Anyway, what I mean, ATM most of the calls are still going to be considered
possibly freeing, and it will be pretty much the same calls that are
considered as potentially calling __asan_poison_memory_region, so the
optimization wouldn't change much.

Jakub

Re: [Ping] [PATCH, 9/10] aarch64: generate conditional compare instructions

[Richard Henderson]

On 11/05/2014 10:05 AM, Zhenqiang Chen wrote:
> I had retested all the ccmp patches.
> 
> Bootstrap and no make check regression on X86-64.
> Bootstrap and no make check regression on AARCH64 qemu.
> 
> OK for trunk?

No patch?  Or what is it that you're wanting approval for?


r~

Re: [PATCH] Optimize UBSAN_NULL checks, add sanopt.c

[Yury Gribov]

On 11/05/2014 04:23 PM, Jakub Jelinek wrote:

On Wed, Nov 05, 2014 at 04:13:01PM +0300, Yury Gribov wrote:

Wouldn't it break most uses of __asan_poison_memory_region ?


Most probably but I wonder if we should ask people to simply do asm
volatile with memory clobber in this case?  And we probably shouldn't
call the whole thing is_nonfreeing anyway.


Added Kostya to maybe comment on this.


Well, right now !nonfreeing_call_p is any non-builtin call or
non-leaf builtin call (i.e. builtin that might call functions in the
current CU), or free/tm_free/realloc/stack_restore.
So, by this definition __asan_poison_memory_region is also a
!nonfreeing_call_p.  Where would you like to see the volatile with
memory clobber?

>

You might very well just call some function that
does the free for you.
   *p = 1;
   foo (p);
   *p = 2;
and foo (p) could asm volatile ("" : : : "memory"); somewhere
and free (p) somewhere else.


I was thinking about e.g. removing check for the second access in

extern int x[];

void foo (int i) {
x[i] = 1;
foo (p);
x[i] = 2;
}

because accessability of &a[i] obviously can't be changed by any call to 
free () inside foo ().  But you are probably right that 
__asan_poison_memory could potentially be called inside foo (however 
rare it is) which would preclude this sort of optimization.



If in the future we e.g. IPA-prop propagate the nonfreeing_call_p
property through the callgraph (as in, if the function you call
is non-overridable and you know the flag for it, use it),


FYI we tried this on SPEC and some other apps but saw no performance 
improvements.


-Y

Re: [libcc1, build] Enable libcc1 on Solaris

[Phil Muldoon]

On 03/11/14 16:54, Rainer Orth wrote:
> I noticed that the new libcc1 wasn't built on Solaris.  This happens
> because socketpair doesn't live in libc, but in libsocket instead.  To
> deal with this, I've copied the libgo (and libjava) code to detect the
> need for libsocket and libnsl.  Once the build was attempted, two
> failures had to be dealt with:
>
> * FD_ZERO and friends need  for a memset declaration.
>
> * On Solaris 10, AF_LOCAL isn't defined in system headers, while AF_UNIX
>   is.  In both libgo and libjava, there are unconditional uses of
>   AF_UNIX, so I've followed their lead.
>
> Those changes allowed libcc1.so to build.
>
> Bootstrapped without regressions on i386-pc-solaris2.1[01] and
> x86_64-unknown-linux-gnu, ok for mainline?
>
> Btw., MAINTAINERS doesn't currently list a libcc1 maintainer.  I believe
> it should.
>
> Rainer
>
>
> 2014-10-31  Rainer Orth  
>
> * configure.ac (libcc1_cv_lib_sockets): Check for -lsocket -lnsl.
> * configure: Regenerate.
> * connection.cc: Include .
> * libcc1.cc (libcc1_compile): Use AF_UNIX instead of AF_LOCAL.

The configure change is fine.  Also the include.  (From a libcc
authorship point of view).  I am not aware of the history of AF_UNIX
over AF_LOCAL so I have no comment on that.  Please await permission
from a GCC maintainer (I am not one).

Cheers

Phil

[PATCH] PR 63721 IPA ICF cause atomic-comp-swap-release-acquire.c ICE

[Jiong Wang]

the same ICE will happen on x86-64, if compile with -O2 -fPIC.

the reason is for the following two functions, they are identical, so 
IPA-ICF

pass try to transform the second function to call the first one directly.

int
atomic_compare_exchange_STRONG_RELEASE_ACQUIRE (int a, int b)
{
  return __atomic_compare_exchange (&v, &a, &b,
STRONG, __ATOMIC_RELEASE,
__ATOMIC_ACQUIRE);
}

int
atomic_compare_exchange_n_STRONG_RELEASE_ACQUIRE (int a, int b)
{
  return __atomic_compare_exchange_n (&v, &a, b,
  STRONG, __ATOMIC_RELEASE,
  __ATOMIC_ACQUIRE);
}

while during this transformation, looks like there are something wrong 
with the
function argument handling. take "a" for example, because later there 
are "&a",

so it's marked as addressable. while after transformation, if we turn the
second function into

int atomic_compare_exchange_n_STRONG_RELEASE_ACQUIRE (int a, int b)
{
  return atomic_compare_exchange_STRONG_RELEASE_ACQUIRE (a, b)
}

then argument a is no longer addressable.

so, in cgraph_node::release_body, when making the wrapper, except 
clearing the

function body, we should also clear the addressable flag for function args
because they are decided by the function body which is cleared.

bootstrap ok on x86-64 and no regression.
bootstrap ok on aarch64 juno.
ICE gone away on arm & x86-64

ok for trunk?

gcc/

  PR  tree-optimization/63721
  * cgraph.c (cgraph_node::release_body): Clear addressable flag for function 
args.
diff --git a/gcc/cgraph.c b/gcc/cgraph.c
index d430bc5..7ac0b2a 100644
--- a/gcc/cgraph.c
+++ b/gcc/cgraph.c
@@ -1668,6 +1668,7 @@ release_function_body (tree decl)
 void
 cgraph_node::release_body (bool keep_arguments)
 {
+  tree arg_p;
   ipa_transforms_to_apply.release ();
   if (!used_as_abstract_origin && symtab->state != PARSING)
 {
@@ -1676,6 +1677,10 @@ cgraph_node::release_body (bool keep_arguments)
   if (!keep_arguments)
 	DECL_ARGUMENTS (decl) = NULL;
 }
+
+  for (arg_p = DECL_ARGUMENTS (decl); arg_p; arg_p = DECL_CHAIN (arg_p))
+TREE_ADDRESSABLE (arg_p) = 0;
+
   /* If the node is abstract and needed, then do not clear DECL_INITIAL
  of its associated function function declaration because it's
  needed to emit debug info later.  */

libstdc++ new deque failures

[David Edelsohn]

Jonathan,

I still am seeing new failures in the libstdc++ deque testsuite as of
last night.  I don't know if you still are working through the fallout
from the earlier patches, but I wanted to make you aware.

AIX defaults to 32 bit. A template was not initialized for int?

FAIL: 23_containers/deque/requirements/dr438/insert_neg.cc  (test for
errors, line 1943)
FAIL: 23_containers/deque/requirements/dr438/insert_neg.cc (test for
excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/bits/stl_deque.h:1931:25:
error: no matching function for call to
'std::deque::_M_fill_insert(std::deque::iterator&, int&, int&)'

FAIL: 23_containers/deque/requirements/dr438/assign_neg.cc  (test for
errors, line 1859)
FAIL: 23_containers/deque/requirements/dr438/assign_neg.cc (test for
excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/bits/stl_deque.h:1847:25:
error: no matching function for call to
'std::deque::_M_fill_assign(int&, int&)'

FAIL: 23_containers/deque/requirements/dr438/constructor_1_neg.cc
(test for errors, line 1792)
FAIL: 23_containers/deque/requirements/dr438/constructor_1_neg.cc
(test for excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/bits/stl_deque.h:1780:22:
error: no matching function for call to 'std::deque
>::_M_fill_initialize(int&)'

FAIL: 23_containers/deque/requirements/dr438/constructor_2_neg.cc
(test for errors, line 1792)
FAIL: 23_containers/deque/requirements/dr438/constructor_2_neg.cc
(test for excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/bits/stl_deque.h:1780:22:
error: no matching function for call to
'std::deque >
::_M_fill_initialize(char&)'


And these are not related to deque, but appear to be additional issues
in the libstdc++ implementation:

FAIL: 20_util/tuple/comparison_operators/overloaded.cc (test for excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/tuple:102:12:
error: 'constexpr std::_Head_base<_Idx, _Head,
false>::_Head_base(const std::_Head_base<_Idx, _Head, false>&) [with
long unsigned int _Idx = 0ul; _Head = std::nullptr_t]' conflicts with
a previous declaration
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/tuple:102:12:
error: 'constexpr std::_Head_base<_Idx, _Head,
false>::_Head_base(const std::_Head_base<_Idx, _Head, false>&) [with
long unsigned int _Idx = 0ul; _Head = std::nullptr_t]' conflicts with
a previous declaration

FAIL: 20_util/tuple/creation_functions/tuple_cat.cc (test for excess errors)
Excess errors:
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/tuple:102:12:
error: 'constexpr std::_Head_base<_Idx, _Head,
false>::_Head_base(const std::_Head_base<_Idx, _Head, false>&) [with
long unsigned int _Idx = 6ul; _Head = std::nullptr_t]' conflicts with
a previous declaration
/tmp/20141104/powerpc-ibm-aix7.1.0.0/libstdc++-v3/include/tuple:102:12:
error: 'constexpr std::_Head_base<_Idx, _Head,
false>::_Head_base(const std::_Head_base<_Idx, _Head, false>&) [with
long unsigned int _Idx = 6ul; _Head = std::nullptr_t]' conflicts with
a previous declaration

If you would prefer that I open Bugzilla issues, let me know.

Thanks, David

[PATCH, committed] Fix AIX testsuite failures

[David Edelsohn]

The appended patch XFAILs or adjusts testcases to avoid spurious
warnings on AIX.

Bootstrapped on powerpc-ibm-aix7.1.0.0

Thanks, David

* gcc.dg/torture/pr59166.c: XFAIL on AIX.
* g++.dg/ext/visitibility/anon1.C: XFAIL on AIX.
* g++.dg/opt/pr60002.C: XFAIL on AIX.
* g++.dg/torture/pr63419.C: Ignore non-standard ABI warning.
* g++.dg/ipa/ipa-icf-5.C: Require visibility support.

Index: torture/pr59166.c
===
--- torture/pr59166.c   (revision 217109)
+++ torture/pr59166.c   (working copy)
@@ -1,5 +1,6 @@
 /* PR rtl-optimization/59166 */
 /* { dg-additional-options "-fcompare-debug" } */
+/* { dg-xfail-if "" { powerpc-ibm-aix* } { "*" } { "" } } */

 int a, b, c, f, g;

Index: ext/visibility/anon1.C
===
--- ext/visibility/anon1.C  (revision 217109)
+++ ext/visibility/anon1.C  (working copy)
@@ -3,6 +3,7 @@

 // { dg-do compile }
 // { dg-final { scan-assembler-not "globl.*_ZN.*1fEv" } }
+// { dg-xfail-if "" { powerpc-ibm-aix* } { "*" } { "" } }

 namespace
 {
Index: opt/pr60002.C
===
--- opt/pr60002.C   (revision 217109)
+++ opt/pr60002.C   (working copy)
@@ -1,6 +1,7 @@
 // PR tree-optimization/60002
 // { dg-do compile }
 // { dg-options "-O2 -fcompare-debug -fdeclone-ctor-dtor -fipa-cp-clone" }
+// { dg-xfail-if "" { powerpc-ibm-aix* } { "*" } { "" } }

 struct A {};

Index: torture/pr63419.C
===
--- torture/pr63419.C   (revision 217109)
+++ torture/pr63419.C   (working copy)
@@ -1,5 +1,7 @@
 // { dg-do compile }
 // { dg-additional-options "-Wno-psabi" }
+// Ignore warning on some powerpc-linux configurations.
+// { dg-prune-output "non-standard ABI extension" }

 typedef float __m128 __attribute__ ((__vector_size__ (16)));
 const int a = 0;
Index: ipa/ipa-icf-5.C
===
--- ipa/ipa-icf-5.C (revision 217109)
+++ ipa/ipa-icf-5.C (working copy)
@@ -1,4 +1,5 @@
 /* { dg-do compile } */
+/* { dg-require-visibility "" } */
 /* { dg-options "-O2 -fdump-ipa-icf" } */

 struct test

Re: [PATCH, committed] Fix AIX testsuite failures

[Rainer Orth]

Hi David,

> The appended patch XFAILs or adjusts testcases to avoid spurious
> warnings on AIX.
>
> Bootstrapped on powerpc-ibm-aix7.1.0.0
>
> Thanks, David
>
> * gcc.dg/torture/pr59166.c: XFAIL on AIX.
> * g++.dg/ext/visitibility/anon1.C: XFAIL on AIX.
> * g++.dg/opt/pr60002.C: XFAIL on AIX.
> * g++.dg/torture/pr63419.C: Ignore non-standard ABI warning.
> * g++.dg/ipa/ipa-icf-5.C: Require visibility support.
>
> Index: torture/pr59166.c
> ===
> --- torture/pr59166.c   (revision 217109)
> +++ torture/pr59166.c   (working copy)
> @@ -1,5 +1,6 @@
>  /* PR rtl-optimization/59166 */
>  /* { dg-additional-options "-fcompare-debug" } */
> +/* { dg-xfail-if "" { powerpc-ibm-aix* } { "*" } { "" } } */

please omit the default args to dg-xfail-if/dg-skip/if ({ "*" } { "" })
here and in several other testcases: they have been unnecessary for
quite some time.  Also, please include an explanation why you are
xfailing/skipping the test in the comment field so others can understand
what's happening and eventually add their targets to the list if
appropriate.

Thanks.
Rainer

-- 
-
Rainer Orth, Center for Biotechnology, Bielefeld University

Re: [PATCHv5][Kasan] Allow to override Asan shadow offset from command line

[Kirill Yukhin]

Hello,
On 24 Oct 17:56, Yury Gribov wrote:
...
> +const struct test_data_t test_data[] = {
> +  { STRTOL,  "-0x8000", 0, -0x8000L,  0  },
...
> +  switch (test_data[i].fun)
> + {
> + case STRTOL:
> +   res = strtol (test_data[i].nptr, 0, test_data[i].base);
> +   break;
As far as we might have `long long int' on 32-bit, `res' will fail
to compare with corresponding `test_data[i].base'.

Tiny patch fixes it.

--
Thanks, K

diff --git a/libiberty/testsuite/test-strtol.c 
b/libiberty/testsuite/test-strtol.c
index 96d6871..6faf81b 100644
--- a/libiberty/testsuite/test-strtol.c
+++ b/libiberty/testsuite/test-strtol.c
@@ -132,7 +132,8 @@ run_tests (const struct test_data_t *test_data, size_t 
ntests)
   switch (test_data[i].fun)
{
case STRTOL:
- res = strtol (test_data[i].nptr, 0, test_data[i].base);
+ res = (unsigned long) strtol (test_data[i].nptr,
+   0, test_data[i].base);
  break;
case STRTOUL:
  res = strtoul (test_data[i].nptr, 0, test_data[i].base);

Re: nvptx offloading patches [1/n]

[Michael Matz]

Hi,

On Tue, 4 Nov 2014, Jeff Law wrote:

> They still need to agree on the layout of the structure.  And assuming 
> it'll always be memcpy perhaps isn't wise.  Consider the possibility 
> that one day (perhaps soon) the host and GPU may share address space & 
> memory.

Not only soon, there is already hardware out that does exactly that.  
HSA.


Ciao,
Michael.

RE: [PATCH, i686] Fix for asan test failures with -m32 happened after EBX enabling in PIC mode

[Zamyatin, Igor]

> > Hi!
> >
> > Following patch (moving initialization of pic_offset_table_rtx
> earlier) fixes failures for asan tests on 32 bits in PIC mode mentioned here -
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63534#c48
> >
> >
> > Bootstrapped/regtested on x86_64, i686
> >
> > Is it ok for trunk?
> >
> > ChangeLog:
> >
> > 2014-10-30  Igor Zamyatin  
> >
> > * function.c (assign_parms): Move init of pic_offset_table_rtx
> > from here to...
> > * cfgexpand.c (expand_used_vars): ...here.
> The patch is probably fine.  However, it would be good to have the analysis
> why you want to move initialization of the PIC register earlier.

Asan (and anybody else can) emits global variable(s) in expand_used_vars  
during function expanding while pic reg is currently initialized later, during 
expand_function_start in assign_parms thus to be late in asan case in PIC mode.

So to avoid such cases we put pic reg initialization in the beginning of 
expand_used_vars. This seems to be early enough.

Thanks,
Igor

Re: [patch,gomp-4_0-branch] acc nested function support

[David Malcolm]

On Tue, 2014-11-04 at 16:45 -0800, Cesar Philippidis wrote:
> Here's an updated version of my nested function patch.
> 
> David, I tweaked the gimple class hierarchy a little bit. Here's what
> the updated class diagram looks like:
> 
>  + gimple_statement_omp
>  |   |layout: GSS_OMP.  Used for code GIMPLE_OMP_SECTION
>  |   |
>  |   + gimple_statement_omp_parallel_layout
>  |   |   |layout: GSS_OMP_PARALLEL_LAYOUT
>  |   |   |
>  |   |   + gimple_statement_omp_targetreg
>  |   |   |
>  |   |   + gimple_statement_oacc_kernels
>  |   |   |code: GIMPLE_OACC_KERNELS
>  |   |   |
>  |   |   + gimple_statement_oacc_parallel
>  |   |   |code: GIMPLE_OACC_PARALLEL
>  |   |   |
>  |   |   + gimple_statement_omp_target
>  |   |code: GIMPLE_OMP_TARGET
> 
> Basically, I've introduced gimple_statement_omp_targetreg and made
> GIMPLE_OACC_{PARALLEL,KERNELS} and GIMPLE_OMP_TARGET inherit it. This
> seems to work out pretty good. It cleans up both
> {lower,expand}_oacc_offload in omp-low.c and allows OpenACC kernel and
> parallel regions to be treated as OpenMP target regions in
> tree-nested.c. Are these changes to gimple.h OK?

I'm not a reviewer, so it's not directly up to me, but if it simplifies
the code then it seems reasonable.  I'm interested in Jakub's opinion.

> Thomas, assuming these gimple changes are OK, should I commit this
> change to gomp-4_0-branch, or do you want to include this patch with
> your middle end trunk submission?

[gomp4] Remove unused BUILT_IN_OMP_SET_NUM_THREADS (was: various OpenACC/PTX built-ins and a reduction tweak)

[Thomas Schwinge]

Hi!

On Thu, 18 Sep 2014 20:43:20 +0200, I wrote:
> On Tue, 16 Sep 2014 17:32:54 -0700, Cesar Philippidis 
>  wrote:
> > The patch [...]

> > --- a/gcc/omp-builtins.def
> > +++ b/gcc/omp-builtins.def
> > @@ -236,6 +236,3 @@ DEF_GOMP_BUILTIN (BUILT_IN_GOMP_TARGET_UPDATE, 
> > "GOMP_target_update",
> >   BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR, ATTR_NOTHROW_LIST)
> >  DEF_GOMP_BUILTIN (BUILT_IN_GOMP_TEAMS, "GOMP_teams",
> >   BT_FN_VOID_UINT_UINT, ATTR_NOTHROW_LIST)
> > -
> > -DEF_GOMP_BUILTIN (BUILT_IN_OMP_SET_NUM_THREADS, "omp_set_num_threads",
> > - BT_FN_VOID_INT, ATTR_CONST_NOTHROW_LEAF_LIST)
> 
> To avoid confusion: that has been added to gomp-4_0-branch earlier, and
> is now reverted to the trunk state.

I have now actually removed this; r217135:

commit d2579456a7b9008ba19cabc88393f83334324bdd
Author: tschwinge 
Date:   Wed Nov 5 15:38:02 2014 +

Remove unused BUILT_IN_OMP_SET_NUM_THREADS.

gcc/
* omp-builtins.def (BUILT_IN_OMP_SET_NUM_THREADS): Remove.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217135 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog.gomp   | 4 
 gcc/omp-builtins.def | 3 ---
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git gcc/ChangeLog.gomp gcc/ChangeLog.gomp
index ce98a18..ae1afd0 100644
--- gcc/ChangeLog.gomp
+++ gcc/ChangeLog.gomp
@@ -1,3 +1,7 @@
+2014-11-05  Thomas Schwinge  
+
+   * omp-builtins.def (BUILT_IN_OMP_SET_NUM_THREADS): Remove.
+
 2014-11-03  Cesar Philippidis  
 
* builtins.def (DEF_GOACC_BUILTIN): Revert erroneous checkin.
diff --git gcc/omp-builtins.def gcc/omp-builtins.def
index 698dc79..08b825c 100644
--- gcc/omp-builtins.def
+++ gcc/omp-builtins.def
@@ -236,6 +236,3 @@ DEF_GOMP_BUILTIN (BUILT_IN_GOMP_TARGET_UPDATE, 
"GOMP_target_update",
  BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR, ATTR_NOTHROW_LIST)
 DEF_GOMP_BUILTIN (BUILT_IN_GOMP_TEAMS, "GOMP_teams",
  BT_FN_VOID_UINT_UINT, ATTR_NOTHROW_LIST)
-
-DEF_GOMP_BUILTIN (BUILT_IN_OMP_SET_NUM_THREADS, "omp_set_num_threads",
- BT_FN_VOID_INT, ATTR_CONST_NOTHROW_LEAF_LIST)


Grüße,
 Thomas


pgp3S6lYSv5Cl.pgp
Description: PGP signature

[gomp4] Remove unused OACC_WAIT (was: acc dealloc map)

[Thomas Schwinge]

Hi!

On Mon, 20 Oct 2014 13:30:23 -0700, Cesar Philippidis  
wrote:
> 2014-10-20  Cesar Philippidis  
> 
>   gcc/
>   * gimplify.c [...]
>   (gimplify_expr): Remove OACC_WAIT, since it handled directly by the
>   front ends.

In r217136, I have now completely removed it:

commit 1f7efc1d102a69676af101767271664a5788664e
Author: tschwinge 
Date:   Wed Nov 5 15:43:08 2014 +

Remove unused OACC_WAIT.

gcc/
* tree.def (OACC_WAIT): Remove.  Update all users.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217136 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog.gomp  | 2 ++
 gcc/doc/generic.texi| 5 -
 gcc/gimplify.c  | 2 --
 gcc/tree-pretty-print.c | 5 -
 gcc/tree.def| 4 
 gcc/tree.h  | 3 ---
 6 files changed, 2 insertions(+), 19 deletions(-)

diff --git gcc/ChangeLog.gomp gcc/ChangeLog.gomp
index ae1afd0..5b2bade 100644
--- gcc/ChangeLog.gomp
+++ gcc/ChangeLog.gomp
@@ -1,5 +1,7 @@
 2014-11-05  Thomas Schwinge  
 
+   * tree.def (OACC_WAIT): Remove.  Update all users.
+
* omp-builtins.def (BUILT_IN_OMP_SET_NUM_THREADS): Remove.
 
 2014-11-03  Cesar Philippidis  
diff --git gcc/doc/generic.texi gcc/doc/generic.texi
index e756cf3..a638b87 100644
--- gcc/doc/generic.texi
+++ gcc/doc/generic.texi
@@ -2058,7 +2058,6 @@ edge.  Rethrowing the exception is represented using 
@code{RESX_EXPR}.
 @tindex OACC_UPDATE
 @tindex OACC_ENTER_DATA
 @tindex OACC_EXIT_DATA
-@tindex OACC_WAIT
 @tindex OACC_CACHE
 @tindex OMP_PARALLEL
 @tindex OMP_FOR
@@ -2115,10 +2114,6 @@ Represents @code{#pragma acc enter data [clause1 @dots{} 
clauseN]}.
 
 Represents @code{#pragma acc exit data [clause1 @dots{} clauseN]}.
 
-@item OACC_WAIT
-
-Represents @code{#pragma acc wait [(num @dots{})]}.
-
 @item OACC_CACHE
 
 Represents @code{#pragma acc cache (var @dots{})}.
diff --git gcc/gimplify.c gcc/gimplify.c
index bdf4f4a..bfd7f66 100644
--- gcc/gimplify.c
+++ gcc/gimplify.c
@@ -4425,7 +4425,6 @@ is_gimple_stmt (tree t)
 case OACC_UPDATE:
 case OACC_ENTER_DATA:
 case OACC_EXIT_DATA:
-case OACC_WAIT:
 case OACC_CACHE:
 case OMP_PARALLEL:
 case OMP_FOR:
@@ -8755,7 +8754,6 @@ gimplify_expr (tree *expr_p, gimple_seq *pre_p, 
gimple_seq *post_p,
  && code != OACC_UPDATE
  && code != OACC_ENTER_DATA
  && code != OACC_EXIT_DATA
- && code != OACC_WAIT
  && code != OACC_CACHE
  && code != OMP_CRITICAL
  && code != OMP_FOR
diff --git gcc/tree-pretty-print.c gcc/tree-pretty-print.c
index 6f80e80..f311ed9 100644
--- gcc/tree-pretty-print.c
+++ gcc/tree-pretty-print.c
@@ -2546,11 +2546,6 @@ dump_generic_node (pretty_printer *buffer, tree node, 
int spc, int flags,
   dump_omp_clauses (buffer, OACC_EXIT_DATA_CLAUSES (node), spc, flags);
   break;
 
-case OACC_WAIT:
-  pp_string (buffer, "#pragma acc wait");
-  dump_omp_clauses (buffer, OACC_WAIT_CLAUSES (node), spc, flags);
-  break;
-
 case OACC_CACHE:
   pp_string (buffer, "#pragma acc cache");
   dump_omp_clauses (buffer, OACC_CACHE_CLAUSES(node), spc, flags);
diff --git gcc/tree.def gcc/tree.def
index 1af7d81..871a7fb 100644
--- gcc/tree.def
+++ gcc/tree.def
@@ -1163,10 +1163,6 @@ DEFTREECODE (OACC_ENTER_DATA, "oacc_enter_data", 
tcc_statement, 1)
Operand 0: OACC_EXIT_DATA_CLAUSES: List of clauses.  */
 DEFTREECODE (OACC_EXIT_DATA, "oacc_exit_data", tcc_statement, 1)
 
-/* OpenACC - #pragma acc wait [clause1 ... clauseN]
-   Operand 0: OACC_WAIT_CLAUSES: List of clauses.  */
-DEFTREECODE (OACC_WAIT, "oacc_wait", tcc_statement, 1)
-
 /* OpenACC - #pragma acc cache [clause1 ... clauseN]
Operand 0: OACC_CACHE_CLAUSES: List of clauses.  */
 DEFTREECODE (OACC_CACHE, "oacc_cache", tcc_statement, 1)
diff --git gcc/tree.h gcc/tree.h
index ba5fc83..c91e716 100644
--- gcc/tree.h
+++ gcc/tree.h
@@ -1199,9 +1199,6 @@ extern void protected_set_expr_location (tree, 
location_t);
 #define OACC_UPDATE_CLAUSES(NODE) \
   TREE_OPERAND (OACC_UPDATE_CHECK (NODE), 0)
 
-#define OACC_WAIT_CLAUSES(NODE) \
-  TREE_OPERAND (OACC_WAIT_CHECK (NODE), 0)
-
 #define OACC_CACHE_CLAUSES(NODE) \
   TREE_OPERAND (OACC_CACHE_CHECK (NODE), 0)
 


Grüße,
 Thomas


pgpP9UQS9kgi8.pgp
Description: PGP signature

Re: [gofrontend-dev] [PATCH 4/4] Gccgo port to s390[x] -- part II

[Ian Taylor]

On Wed, Nov 5, 2014 at 2:05 AM, Dominik Vogt  wrote:
> On Tue, Nov 04, 2014 at 08:16:51PM -0800, Ian Taylor wrote:
>> I committed the change to go-test.exp.  Thanks.
>>
>> The other changes are not OK.  As described in
>> gcc/testsuite/go.test/test/README.gcc, the files in
>> gcc/testsuite/go.test/test are an exact copy of the master Go
>> testsuite.  Any changes must be made to the master Go testsuite first.
>
> I understand that, but I'm unsure how to handle a set of patches
> that all depend on each other but refer to three different
> reposiories.  So I posted this patch intentionally in the wrong
> place, not knowing how to do it in a better way.

Changes to the master Go repository must follow the procedure
described at http://golang.org/doc/contribute.html.


>> I don't know what's up with the complex number change. In general the
>> Go compiler and libraries go to some effort to produce the same
>> answers on all platforms.  We need to understand why we get different
>> answers on s390 (you may understand the differences, but I don't).  I
>> won't change the tests without a clear understanding of why we are
>> changing them.
>
> It's actually not a Go specific problem, the same deviation occurs
> in C code too.  The cause is that constant folding is done with a
> higher precision and may yield a different result than the run
> time calculations.  There is a Gcc bug report for that "issue":
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60181

So far it doesn't sound appropriate to change the Go testsuite for
this.  If the immediate goal is simply to get the s390 tests to pass,
let's change go-test.exp to xfail the test unless and until somebody
figures out the whole issue.

Ian

[jit] Use ISALPHA and ISALNUM rather than writing our own

[David Malcolm]

On Tue, 2014-11-04 at 14:39 -0700, Jeff Law wrote:
> On 11/04/14 09:57, David Malcolm wrote:
> >>> +#define IS_ASCII_DIGIT(CHAR) \
> >>> +  ((CHAR) >= '0' && (CHAR) <='9')
> >>> +
> >>> +#define IS_ASCII_ALNUM(CHAR) \
> >>> +  (IS_ASCII_ALPHA (CHAR) || IS_ASCII_DIGIT (CHAR))
> >> Can't we rely on the C library to give us equivalents?
> >
> > I've been burned in the past by the C library using locales, in
> > particular the two lowercase "i" variants in Turkish.
> >
> > These macros are used by gcc_jit_context_new_function to enforce C's
> > naming restrictions, to avoid errors from the assembler.  The comment I
> > put there was:
> >
> >/* The assembler can only handle certain names, so for now, enforce
> >   C's rules for identifiers upon the name.
> >   Eventually we'll need some way to interact with e.g. C++ name 
> > mangling.  */
> >
> > Am I right in thinking that for the assembler we need to enforce the C
> > naming rules specifically on *ASCII*.
> >
> > (clearly another comment is needed here).
> I guess you've got to do it somewhere.  Presumably there isn't something 
> already in GCC that enforces an input character set?  I guess I just 
> dislike seeing something that feels like it ought to already be available.

It turns out that locale-independent tests for this did already exist in
libiberty, in safe-ctype.h, so I've committed this to the jit branch:

gcc/jit/ChangeLog.jit:
* libgccjit.c: Include safe-ctype.h from libiberty.
(IS_ASCII_ALPHA): Delete.
(IS_ASCII_DIGIT): Delete.
(IS_ASCII_ALNUM): Delete.
(gcc_jit_context_new_function): Replace use of IS_ASCII_ALPHA and
IS_ASCII_ALNUM with ISALPHA and ISALNUM respectively, from
libiberty.
---
 gcc/jit/ChangeLog.jit | 10 ++
 gcc/jit/libgccjit.c   | 24 +++-
 2 files changed, 17 insertions(+), 17 deletions(-)

diff --git a/gcc/jit/ChangeLog.jit b/gcc/jit/ChangeLog.jit
index 90fccdb..3d6361c 100644
--- a/gcc/jit/ChangeLog.jit
+++ b/gcc/jit/ChangeLog.jit
@@ -1,3 +1,13 @@
+2014-11-05  David Malcolm  
+
+   * libgccjit.c: Include safe-ctype.h from libiberty.
+   (IS_ASCII_ALPHA): Delete.
+   (IS_ASCII_DIGIT): Delete.
+   (IS_ASCII_ALNUM): Delete.
+   (gcc_jit_context_new_function): Replace use of IS_ASCII_ALPHA and
+   IS_ASCII_ALNUM with ISALPHA and ISALNUM respectively, from
+   libiberty.
+
 2014-10-30  David Malcolm  
 
* dummy-frontend.c (jit_langhook_init): Remove some dead code.
diff --git a/gcc/jit/libgccjit.c b/gcc/jit/libgccjit.c
index 286a85e..d9f259e 100644
--- a/gcc/jit/libgccjit.c
+++ b/gcc/jit/libgccjit.c
@@ -22,24 +22,12 @@ along with GCC; see the file COPYING3.  If not see
 #include "system.h"
 #include "coretypes.h"
 #include "opts.h"
+#include "safe-ctype.h"
 
 #include "libgccjit.h"
 #include "jit-common.h"
 #include "jit-recording.h"
 
-#define IS_ASCII_ALPHA(CHAR) \
-  (\
-((CHAR) >= 'a' && (CHAR) <='z')\
-|| \
-((CHAR) >= 'A' && (CHAR) <= 'Z')   \
-  )
-
-#define IS_ASCII_DIGIT(CHAR) \
-  ((CHAR) >= '0' && (CHAR) <='9')
-
-#define IS_ASCII_ALNUM(CHAR) \
-  (IS_ASCII_ALPHA (CHAR) || IS_ASCII_DIGIT (CHAR))
-
 struct gcc_jit_context : public gcc::jit::recording::context
 {
   gcc_jit_context (gcc_jit_context *parent_ctxt) :
@@ -589,13 +577,15 @@ gcc_jit_context_new_function (gcc_jit_context *ctxt,
   RETURN_NULL_IF_FAIL (return_type, ctxt, loc, "NULL return_type");
   RETURN_NULL_IF_FAIL (name, ctxt, loc, "NULL name");
   /* The assembler can only handle certain names, so for now, enforce
- C's rules for identiers upon the name.
- Eventually we'll need some way to interact with e.g. C++ name mangling.  
*/
+ C's rules for identiers upon the name, using ISALPHA and ISALNUM
+ from safe-ctype.h to ignore the current locale.
+ Eventually we'll need some way to interact with e.g. C++ name
+ mangling.  */
   {
 /* Leading char: */
 char ch = *name;
 RETURN_NULL_IF_FAIL_PRINTF2 (
-   IS_ASCII_ALPHA (ch) || ch == '_',
+   ISALPHA (ch) || ch == '_',
ctxt, loc,
"name \"%s\" contains invalid character: '%c'",
name, ch);
@@ -603,7 +593,7 @@ gcc_jit_context_new_function (gcc_jit_context *ctxt,
 for (const char *ptr = name + 1; (ch = *ptr); ptr++)
   {
RETURN_NULL_IF_FAIL_PRINTF2 (
- IS_ASCII_ALNUM (ch) || ch == '_',
+ ISALNUM (ch) || ch == '_',
  ctxt, loc,
  "name \"%s\" contains invalid character: '%c'",
  name, ch);
-- 
1.7.11.7

[gomp4] OpenACC Fortran testsuite: Expect some things to work by now.

[Thomas Schwinge]

Hi!

Committed to gomp-4_0-branch in r217137:

commit 83c3ae92fb16c23a782f012a49dd7aa1fcd01287
Author: tschwinge 
Date:   Wed Nov 5 15:54:45 2014 +

OpenACC Fortran testsuite: Expect some things to work by now.

gcc/testsuite/
* gfortran.dg/goacc/data-tree.f95: Remove dg-prune-output directive.
* gfortran.dg/goacc/kernels-tree.f95: Likewise.
* gfortran.dg/goacc/loop-tree-1.f90: Likewise.
* gfortran.dg/goacc/parallel-tree.f95: Likewise.
* gfortran.dg/goacc/private-1.f95: Likewise.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217137 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/testsuite/ChangeLog.gomp  | 8 
 gcc/testsuite/gfortran.dg/goacc/data-tree.f95 | 1 -
 gcc/testsuite/gfortran.dg/goacc/kernels-tree.f95  | 1 -
 gcc/testsuite/gfortran.dg/goacc/loop-tree-1.f90   | 1 -
 gcc/testsuite/gfortran.dg/goacc/parallel-tree.f95 | 1 -
 gcc/testsuite/gfortran.dg/goacc/private-1.f95 | 1 -
 6 files changed, 8 insertions(+), 5 deletions(-)

diff --git gcc/testsuite/ChangeLog.gomp gcc/testsuite/ChangeLog.gomp
index 4dd6b83..25be821 100644
--- gcc/testsuite/ChangeLog.gomp
+++ gcc/testsuite/ChangeLog.gomp
@@ -1,3 +1,11 @@
+2014-11-05  Thomas Schwinge  
+
+   * gfortran.dg/goacc/data-tree.f95: Remove dg-prune-output directive.
+   * gfortran.dg/goacc/kernels-tree.f95: Likewise.
+   * gfortran.dg/goacc/loop-tree-1.f90: Likewise.
+   * gfortran.dg/goacc/parallel-tree.f95: Likewise.
+   * gfortran.dg/goacc/private-1.f95: Likewise.
+
 2014-11-04  Cesar Philippidis  
 
* gfortran.dg/goacc/routine-1.f90: New test.
diff --git gcc/testsuite/gfortran.dg/goacc/data-tree.f95 
gcc/testsuite/gfortran.dg/goacc/data-tree.f95
index a5c012a..32c50fd 100644
--- gcc/testsuite/gfortran.dg/goacc/data-tree.f95
+++ gcc/testsuite/gfortran.dg/goacc/data-tree.f95
@@ -12,7 +12,6 @@ program test
   !$acc end data
 
 end program test
-! { dg-prune-output "unimplemented" }
 ! { dg-final { scan-tree-dump-times "pragma acc data" 1 "original" } } 
 
 ! { dg-final { scan-tree-dump-times "if" 1 "original" } }
diff --git gcc/testsuite/gfortran.dg/goacc/kernels-tree.f95 
gcc/testsuite/gfortran.dg/goacc/kernels-tree.f95
index 73f172c..7585a16 100644
--- gcc/testsuite/gfortran.dg/goacc/kernels-tree.f95
+++ gcc/testsuite/gfortran.dg/goacc/kernels-tree.f95
@@ -12,7 +12,6 @@ program test
   !$acc end kernels
 
 end program test
-! { dg-prune-output "unimplemented" }
 ! { dg-final { scan-tree-dump-times "pragma acc kernels" 1 "original" } } 
 
 ! { dg-final { scan-tree-dump-times "if" 1 "original" } }
diff --git gcc/testsuite/gfortran.dg/goacc/loop-tree-1.f90 
gcc/testsuite/gfortran.dg/goacc/loop-tree-1.f90
index 14779b6..47ff77e 100644
--- gcc/testsuite/gfortran.dg/goacc/loop-tree-1.f90
+++ gcc/testsuite/gfortran.dg/goacc/loop-tree-1.f90
@@ -35,7 +35,6 @@ program test
   !$acc end parallel
 
 end program test
-! { dg-prune-output "sorry" }
 ! { dg-final { scan-tree-dump-times "pragma acc loop" 5 "original" } } 
 
 ! { dg-final { scan-tree-dump-times "collapse\\(2\\)" 1 "original" } } 
diff --git gcc/testsuite/gfortran.dg/goacc/parallel-tree.f95 
gcc/testsuite/gfortran.dg/goacc/parallel-tree.f95
index f004702..48061b1 100644
--- gcc/testsuite/gfortran.dg/goacc/parallel-tree.f95
+++ gcc/testsuite/gfortran.dg/goacc/parallel-tree.f95
@@ -15,7 +15,6 @@ program test
   !$acc end parallel
 
 end program test
-! { dg-prune-output "unimplemented" }
 ! { dg-final { scan-tree-dump-times "pragma acc parallel" 1 "original" } } 
 
 ! { dg-final { scan-tree-dump-times "if" 1 "original" } }
diff --git gcc/testsuite/gfortran.dg/goacc/private-1.f95 
gcc/testsuite/gfortran.dg/goacc/private-1.f95
index 5aeee3b..54c027d 100644
--- gcc/testsuite/gfortran.dg/goacc/private-1.f95
+++ gcc/testsuite/gfortran.dg/goacc/private-1.f95
@@ -31,7 +31,6 @@ program test
   end do
   !$acc end parallel
 end program test
-! { dg-prune-output "unimplemented" }
 ! { dg-final { scan-tree-dump-times "pragma acc parallel" 3 "omplower" } }
 ! { dg-final { scan-tree-dump-times "private\\(i\\)" 3 "omplower" } }
 ! { dg-final { scan-tree-dump-times "private\\(j\\)" 2 "omplower" } }


Grüße,
 Thomas


pgpInHIpZoXdg.pgp
Description: PGP signature

Re: [PATCH AVX512] Fix dg.torture tests with avx512

[Ilya Tocar]

On 03 Nov 11:21, Jakub Jelinek wrote:
> On Fri, Oct 31, 2014 at 11:17:07AM +0100, Uros Bizjak wrote:
> > I'd like to ask Jakub for a review of the above two parts, other parts
> > are OK with a rename (as mentioned above).
> 
> Looks ok to me.  Where the ICEs discovered just by normal make check or only
> with GCC_TEST_RUN_EXPENSIVE ?  If the latter, can you promote one of the
> permutations that caused the ICEs to normal tests?  If not and
> GCC_TEST_RUN_EXPENSIVE has not been tested, can you try that?
>

This was discovered without GCC_TEST_RUN_EXPENSIVE, but I've tested it
with it enabled, and didn't see any fails.

I've committed version below.

---
 gcc/config/i386/i386.c | 59 --
 gcc/config/i386/sse.md | 54 ++---
 2 files changed, 98 insertions(+), 15 deletions(-)

diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
index c528599..aaffe9d 100644
--- a/gcc/config/i386/i386.c
+++ b/gcc/config/i386/i386.c
@@ -45943,6 +45943,42 @@ expand_vec_perm_pshufb (struct expand_vec_perm_d *d)
{
  if (!TARGET_AVX512BW)
return false;
+
+ /* If vpermq didn't work, vpshufb won't work either.  */
+ if (d->vmode == V8DFmode || d->vmode == V8DImode)
+   return false;
+
+ vmode = V64QImode;
+ if (d->vmode == V16SImode
+ || d->vmode == V32HImode
+ || d->vmode == V64QImode)
+   {
+ /* First see if vpermq can be used for
+V16SImode/V32HImode/V64QImode.  */
+ if (valid_perm_using_mode_p (V8DImode, d))
+   {
+ for (i = 0; i < 8; i++)
+   perm[i] = (d->perm[i * nelt / 8] * 8 / nelt) & 7;
+ if (d->testing_p)
+   return true;
+ target = gen_reg_rtx (V8DImode);
+ if (expand_vselect (target, gen_lowpart (V8DImode, d->op0),
+ perm, 8, false))
+   {
+ emit_move_insn (d->target,
+ gen_lowpart (d->vmode, target));
+ return true;
+   }
+ return false;
+   }
+
+ /* Next see if vpermd can be used.  */
+ if (valid_perm_using_mode_p (V16SImode, d))
+   vmode = V16SImode;
+   }
+ /* Or if vpermps can be used.  */
+ else if (d->vmode == V16SFmode)
+   vmode = V16SImode;
  if (vmode == V64QImode)
{
  /* vpshufb only works intra lanes, it is not
@@ -45962,6 +45998,9 @@ expand_vec_perm_pshufb (struct expand_vec_perm_d *d)
   if (vmode == V8SImode)
 for (i = 0; i < 8; ++i)
   rperm[i] = GEN_INT ((d->perm[i * nelt / 8] * 8 / nelt) & 7);
+  else if (vmode == V16SImode)
+for (i = 0; i < 16; ++i)
+  rperm[i] = GEN_INT ((d->perm[i * nelt / 16] * 16 / nelt) & 15);
   else
 {
   eltsz = GET_MODE_SIZE (GET_MODE_INNER (d->vmode));
@@ -46000,8 +46039,14 @@ expand_vec_perm_pshufb (struct expand_vec_perm_d *d)
emit_insn (gen_avx512bw_pshufbv64qi3 (target, op0, vperm));
   else if (vmode == V8SFmode)
emit_insn (gen_avx2_permvarv8sf (target, op0, vperm));
-  else
+  else if (vmode == V8SImode)
emit_insn (gen_avx2_permvarv8si (target, op0, vperm));
+  else if (vmode == V16SFmode)
+   emit_insn (gen_avx512f_permvarv16sf (target, op0, vperm));
+  else if (vmode == V16SImode)
+   emit_insn (gen_avx512f_permvarv16si (target, op0, vperm));
+  else
+   gcc_unreachable ();
 }
   else
 {
@@ -46055,21 +46100,21 @@ expand_vec_perm_1 (struct expand_vec_perm_d *d)
{
case V64QImode:
  if (TARGET_AVX512BW)
-   gen = gen_avx512bw_vec_dupv64qi;
+   gen = gen_avx512bw_vec_dupv64qi_1;
  break;
case V32QImode:
  gen = gen_avx2_pbroadcastv32qi_1;
  break;
case V32HImode:
  if (TARGET_AVX512BW)
-   gen = gen_avx512bw_vec_dupv32hi;
+   gen = gen_avx512bw_vec_dupv32hi_1;
  break;
case V16HImode:
  gen = gen_avx2_pbroadcastv16hi_1;
  break;
case V16SImode:
  if (TARGET_AVX512F)
-   gen = gen_avx512f_vec_dupv16si;
+   gen = gen_avx512f_vec_dupv16si_1;
  break;
case V8SImode:
  gen = gen_avx2_pbroadcastv8si_1;
@@ -46082,18 +46127,18 @@ expand_vec_perm_1 (struct expand_vec_perm_d *d)
  break;
case V16SFmode:
  if (TARGET_AVX512F)
-   gen = gen_avx512f_vec_dupv16sf;
+   gen = gen_avx512f_vec_dupv16sf_1;
  break;
case V8SFmode:
  gen = gen_avx2_vec_dupv8sf_1;
  break;
case V8DFmode:
   

Re: [gomp4] OpenACC / C++

[Thomas Schwinge]

Hi!

On Wed, 15 Oct 2014 11:21:05 -0500, James Norris  
wrote:
> This patch adds OpenACC support to C++ in the gomp4 branch.

I found a few missing pieces; applied to gomp-4_0-branch in r217139:

commit 09b8ef34550c377610c7a01aa2057fc8297e1b0a
Author: tschwinge 
Date:   Wed Nov 5 15:58:09 2014 +

OpenACC C++: Add a few missing pieces.

gcc/cp/
* parser.c (cp_parser_omp_clause_name): Also look for "pcopy",
"pcopyin", "pcopyout", "pcreate".  Look for "wait" instead of
"WAIT".

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217139 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/cp/ChangeLog.gomp |  6 ++
 gcc/cp/parser.c   | 14 +-
 2 files changed, 15 insertions(+), 5 deletions(-)

diff --git gcc/cp/ChangeLog.gomp gcc/cp/ChangeLog.gomp
index 89f3f22..024e6a5 100644
--- gcc/cp/ChangeLog.gomp
+++ gcc/cp/ChangeLog.gomp
@@ -1,3 +1,9 @@
+2014-11-05  Thomas Schwinge  
+
+   * parser.c (cp_parser_omp_clause_name): Also look for "pcopy",
+   "pcopyin", "pcopyout", "pcreate".  Look for "wait" instead of
+   "WAIT".
+
 2014-11-03  Cesar Philippidis  
 
* parser.c (cp_parser_omp_clause_name): Also consider CPP_KEYWORD
diff --git gcc/cp/parser.c gcc/cp/parser.c
index e3afdca..3ef2de7 100644
--- gcc/cp/parser.c
+++ gcc/cp/parser.c
@@ -27535,13 +27535,17 @@ cp_parser_omp_clause_name (cp_parser *parser)
result = PRAGMA_OMP_CLAUSE_PARALLEL;
  else if (!strcmp ("present", p))
result = PRAGMA_OMP_CLAUSE_PRESENT;
- else if (!strcmp ("present_or_copy", p))
+ else if (!strcmp ("present_or_copy", p)
+  || !strcmp ("pcopy", p))
result = PRAGMA_OMP_CLAUSE_PRESENT_OR_COPY;
- else if (!strcmp ("present_or_copyin", p))
+ else if (!strcmp ("present_or_copyin", p)
+  || !strcmp ("pcopyin", p))
result = PRAGMA_OMP_CLAUSE_PRESENT_OR_COPYIN;
- else if (!strcmp ("present_or_copyout", p))
+ else if (!strcmp ("present_or_copyout", p)
+  || !strcmp ("pcopyout", p))
result = PRAGMA_OMP_CLAUSE_PRESENT_OR_COPYOUT;
- else if (!strcmp ("present_or_create", p))
+ else if (!strcmp ("present_or_create", p)
+  || !strcmp ("pcreate", p))
result = PRAGMA_OMP_CLAUSE_PRESENT_OR_CREATE;
  else if (!strcmp ("private", p))
result = PRAGMA_OMP_CLAUSE_PRIVATE;
@@ -27587,7 +27591,7 @@ cp_parser_omp_clause_name (cp_parser *parser)
result = PRAGMA_CILK_CLAUSE_VECTORLENGTH;
  break;
case 'w':
- if (!strcmp ("WAIT", p))
+ if (!strcmp ("wait", p))
result = PRAGMA_OMP_CLAUSE_WAIT;
  break;
}


Grüße,
 Thomas


pgpvIJUHZcURD.pgp
Description: PGP signature

[gomp4] libgomp OpenACC testsuite: Remove two obsolete test cases.

[Thomas Schwinge]

Hi!

Applied to gomp-4_0-branch in r217140:

commit 3838e13dedee9217b067cf2ab4b3fb5bb7d5cf68
Author: tschwinge 
Date:   Wed Nov 5 16:02:30 2014 +

libgomp OpenACC testsuite: Remove two obsolete test cases.

libgomp/
* testsuite/libgomp.oacc-c/goacc_kernels.c: Remove file.
* testsuite/libgomp.oacc-c/goacc_parallel.c: Remove file.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217140 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 libgomp/ChangeLog.gomp|  5 
 libgomp/testsuite/libgomp.oacc-c/goacc_kernels.c  | 28 ---
 libgomp/testsuite/libgomp.oacc-c/goacc_parallel.c | 28 ---
 3 files changed, 5 insertions(+), 56 deletions(-)

diff --git libgomp/ChangeLog.gomp libgomp/ChangeLog.gomp
index bd8c119..d24ef43 100644
--- libgomp/ChangeLog.gomp
+++ libgomp/ChangeLog.gomp
@@ -1,3 +1,8 @@
+2014-11-05  Thomas Schwinge  
+
+   * testsuite/libgomp.oacc-c/goacc_kernels.c: Remove file.
+   * testsuite/libgomp.oacc-c/goacc_parallel.c: Remove file.
+
 2014-11-04  Cesar Philippidis  
 
* testsuite/libgomp.oacc-fortran/routine-1.f90: New test.
diff --git libgomp/testsuite/libgomp.oacc-c/goacc_kernels.c 
libgomp/testsuite/libgomp.oacc-c/goacc_kernels.c
deleted file mode 100644
index 683fefa..000
--- libgomp/testsuite/libgomp.oacc-c/goacc_kernels.c
+++ /dev/null
@@ -1,28 +0,0 @@
-/* { dg-do run } */
-/* { dg-skip-if "" { *-*-* } { "*" } { "-DACC_DEVICE_TYPE_host=1" } } */
-
-#include "libgomp_g.h"
-
-extern void abort ();
-
-volatile int i;
-
-void
-f (void *data)
-{
-  if (i != -1)
-abort ();
-  i = 42;
-}
-
-int main(void)
-{
-  i = -1;
-  GOACC_kernels (0, f, (const void *) 0,
-0, (void *) 0, (void *) 0, (void *) 0,
-1, 1, 1, -2, -1);
-  if (i != 42)
-abort ();
-
-  return 0;
-}
diff --git libgomp/testsuite/libgomp.oacc-c/goacc_parallel.c 
libgomp/testsuite/libgomp.oacc-c/goacc_parallel.c
deleted file mode 100644
index 232ce8a..000
--- libgomp/testsuite/libgomp.oacc-c/goacc_parallel.c
+++ /dev/null
@@ -1,28 +0,0 @@
-/* { dg-do run } */
-/* { dg-skip-if "" { *-*-* } { "*" } { "-DACC_DEVICE_TYPE_host=1" } } */
-
-#include "libgomp_g.h"
-
-extern void abort ();
-
-volatile int i;
-
-void
-f (void *data)
-{
-  if (i != -1)
-abort ();
-  i = 42;
-}
-
-int main(void)
-{
-  i = -1;
-  GOACC_parallel (0, f, (const void *) 0,
- 0, (void *) 0, (void *) 0, (void *) 0,
- 1, 1, 1, -2, -1);
-  if (i != 42)
-abort ();
-
-  return 0;
-}


Grüße,
 Thomas


pgpIYTfkoAS9U.pgp
Description: PGP signature

Re: [PING][PATCH] Don't call fatal_error before error reporting has been initialized.

[Ilya Tocar]

Ping.

On 20 Oct 19:25, Ilya Tocar wrote:
> Same in collect2.
> 
> On 09 Oct 15:40, Ilya Tocar wrote:
> > Ping.
> > 
> > On 29 Sep 18:02, Ilya Tocar wrote:
> > > Hi,
> > > 
> > > Currently if call to atexit (lto_wrapper_cleanup) fails we
> > > won't report error as we haven't initialized error-reporting
> > > infrastructure. This patch moves this call after diagnostic_initialize.
> > > I hope that we can't  exit inside diagnostic_initialize. Otherwise we
> > > won't cleanup after it.
> > > Ok for trunk?
> > >
> 
> ---
>  gcc/collect2.c| 6 +++---
>  gcc/lto-wrapper.c | 6 +++---
>  2 files changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/gcc/collect2.c b/gcc/collect2.c
> index c54e6fb..b0784e8 100644
> --- a/gcc/collect2.c
> +++ b/gcc/collect2.c
> @@ -955,9 +955,6 @@ main (int argc, char **argv)
>signal (SIGCHLD, SIG_DFL);
>  #endif
>  
> -  if (atexit (collect_atexit) != 0)
> -fatal_error ("atexit failed");
> -
>/* Unlock the stdio streams.  */
>unlock_std_streams ();
>  
> @@ -965,6 +962,9 @@ main (int argc, char **argv)
>  
>diagnostic_initialize (global_dc, 0);
>  
> +  if (atexit (collect_atexit) != 0)
> +fatal_error ("atexit failed");
> +
>/* Do not invoke xcalloc before this point, since locale needs to be
>   set first, in case a diagnostic is issued.  */
>  
> diff --git a/gcc/lto-wrapper.c b/gcc/lto-wrapper.c
> index 8033b15..d97f617 100644
> --- a/gcc/lto-wrapper.c
> +++ b/gcc/lto-wrapper.c
> @@ -879,13 +879,13 @@ main (int argc, char *argv[])
>  
>xmalloc_set_program_name (progname);
>  
> -  if (atexit (lto_wrapper_cleanup) != 0)
> -fatal_error ("atexit failed");
> -
>gcc_init_libintl ();
>  
>diagnostic_initialize (global_dc, 0);
>  
> +  if (atexit (lto_wrapper_cleanup) != 0)
> +fatal_error ("atexit failed");
> +
>if (signal (SIGINT, SIG_IGN) != SIG_IGN)
>  signal (SIGINT, fatal_signal);
>  #ifdef SIGHUP
> -- 
> 1.8.3.1
> 

Re: [debug-early] emit locals early patchset

[Martin Jambor]

On Tue, Oct 28, 2014 at 03:57:43PM +0100, Richard Biener wrote:
> On Tue, Oct 28, 2014 at 1:00 AM, Aldy Hernandez  wrote:
> > Gentlemen!
> >
> > My apologies for the big patch.  In getting locals emitted early (parameters
> > and locally scoped variables), I ran into many things which were in need of
> > surgery, many of which couldn't happen without the other.  Consequently, I
> > ended up fixing everything such that we are now back to no guality.exp
> > failures for any language.
> >
> > [Curiously, I hadn't noticed locals were not being dumped early because they
> > were being picked up by the late dwarf pass.  Fixing this oversight is what
> > caused this entire patch.]
> >
> > There are a lot of changes here, and I would greatly appreciate feedback, so
> > let me at least explain what's going on at a high level...
> >
> > 1. Changes to gen_subprogram_die() to handle early generation of locals, and
> > amending location information on the second pass.  Everything else in this
> > patch, basically stems from this change.
> >
> > 2. Changes to gen_variable_die() to handle multiple passes (early/late dwarf
> > generation).
> >
> > A lot of this is complicated by the fact that old_die's are cached and keyed
> > by `tree', but an abstract instance and an inline instance share trees,
> > while dwarf2out_abstract_function() sets DECL_ABSTRACT_P behind the scenes.
> >
> > The current support (and my changes) maintain this shared and delicate
> > design.  I wonder whether we could simplify a lot of this code by unsharing
> > these trees, but this may be beyond the scope of this work. Richi perhaps
> > you can comment?
> 
> I think that the abstract and inline instances are cases that are _only_
> generated "early" - that is, they don't contain any locations or whatever
> and thus do not need to ameded in the late dwarf pass.  So I'd simply
> generate dwarf for them and not remember their DIEs.

Please have a look at PR 63722, which is a complaint I got from people
exploring possiblities of using dwarf to analyze what gcc did to Linux
kernel and they discovered that DIEs of  abstract origins do not have
their DW_AT_inline set correctly for functions that were IPA-SRAed but
not inlined.  On the original testcase before reduction I've seen the
same problem also when using -fno-ipa-sra -fno-ipa-cp so it is not
just my passes but also probably ipa-split ;-)

In any case, if we want abstract origins to have this field set
correctly in abstract origins, we need to have the possibility to
modify it after IPA.

Martin

Re: [gomp4] Rationalise thread-local variables in libgomp OpenACC support

[Thomas Schwinge]

Hi!

On Wed, 29 Oct 2014 15:20:17 +, Julian Brown  
wrote:
> On Tue, 28 Oct 2014 11:16:19 +
> Julian Brown  wrote:
> > This patch rationalises TLS support by moving all thread-local
> > variables into a single structure. Because this meant interfering with
> > how per-thread/per-device initialisation was done, I took the
> > opportunity to tidy up a couple of other bits along the way.

> --- a/libgomp/plugin-nvptx.c
> +++ b/libgomp/plugin-nvptx.c
> @@ -749,7 +718,12 @@ PTX_get_num_devices (void)
>int n;
>CUresult r;
>  
> -  assert (PTX_inited);
> +  /* This function will be called before the plugin has been initialized in
> + order to enumerate available devices, but CUDA API routines can't be 
> used
> + until cuInit has been called.  Just call it now (but don't yet do any
> + further initialization).  */
> +  if (!PTX_inited)
> +cuInit (0);
>  
>r = cuDeviceGetCount (&n);
>if (r!= CUDA_SUCCESS)

> --- a/libgomp/target.c
> +++ b/libgomp/target.c
> @@ -1172,18 +1184,24 @@ gomp_find_available_plugins (void)

> +  for (i = 0; i < current_device.get_num_devices_func (); i++)
> +{
> +[...]

Committed to gomp-4_0-branch in r217141:

commit af80306c854b504fb0df6ac7d88d2472989ffbb7
Author: tschwinge 
Date:   Wed Nov 5 16:07:20 2014 +

libgomp plugins: Remove the (OpenACC) avail function.

This idea has been to use this to see whether a class of devices is 
generally
available, without determining the exact number of devices that are 
available.
Such a distinction probably doesn't make too much sense; in the following, 
we'd
like to use one device anyway.

Inline the few users.

libgomp/
* target.h (struct ACC_dispatch_t): Remove avail_func.  Update all
users.
* oacc-host.c (openacc_avail): Likewise.
* plugin-nvptx.c (PTX_avail, openacc_avail): Likewise.
* libgomp.texi (acc_get_num_devices): Revise.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217141 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 libgomp/ChangeLog.gomp |  6 ++
 libgomp/libgomp.texi   | 10 +-
 libgomp/oacc-host.c| 13 -
 libgomp/oacc-init.c|  6 +++---
 libgomp/plugin-nvptx.c | 20 
 libgomp/target.c   |  1 -
 libgomp/target.h   |  3 ---
 7 files changed, 14 insertions(+), 45 deletions(-)

diff --git libgomp/ChangeLog.gomp libgomp/ChangeLog.gomp
index d24ef43..d65ce71 100644
--- libgomp/ChangeLog.gomp
+++ libgomp/ChangeLog.gomp
@@ -1,5 +1,11 @@
 2014-11-05  Thomas Schwinge  
 
+   * target.h (struct ACC_dispatch_t): Remove avail_func.  Update all
+   users.
+   * oacc-host.c (openacc_avail): Likewise.
+   * plugin-nvptx.c (PTX_avail, openacc_avail): Likewise.
+   * libgomp.texi (acc_get_num_devices): Revise.
+
* testsuite/libgomp.oacc-c/goacc_kernels.c: Remove file.
* testsuite/libgomp.oacc-c/goacc_parallel.c: Remove file.
 
diff --git libgomp/libgomp.texi libgomp/libgomp.texi
index 9530a2b..66c678f 100644
--- libgomp/libgomp.texi
+++ libgomp/libgomp.texi
@@ -187,12 +187,12 @@ API routines for target platforms.
 @node acc_get_num_devices
 @section @code{acc_get_num_devices} -- Get number of devices for given device 
type
 @table @asis
-item @emph{Description}
-This routine returns a value, between 0 and @emph{n}, indicating the
-number of devices available for the given device type. It determines
-the number of devices in a @emph{passive} manner. In other words, it
+@item @emph{Description}
+This routine returns a value indicating the
+number of devices available for the given device type.  It determines
+the number of devices in a @emph{passive} manner.  In other words, it
 does not alter the state within the runtime environment aside from
-possibly initializing an uninitialized device. This aspect allows
+possibly initializing an uninitialized device.  This aspect allows
 the routine to be called without concern for altering the interaction
 with an attached accelerator device.
 
diff --git libgomp/oacc-host.c libgomp/oacc-host.c
index fc3e77c..f7a9fba 100644
--- libgomp/oacc-host.c
+++ libgomp/oacc-host.c
@@ -153,16 +153,6 @@ device_get_table (struct mapping_table **table)
   return 0;
 }
 
-STATIC bool
-openacc_avail (void)
-{
-#ifdef DEBUG
-  fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
-#endif
-
-  return 1;
-}
-
 STATIC void *
 openacc_open_device (int n)
 {
@@ -414,9 +404,6 @@ static struct gomp_device_descr host_dispatch =
   .get_device_num_func = openacc_get_device_num,
   .set_device_num_func = openacc_set_device_num,
 
-  /* Device available.  */
-  .avail_func = openacc_avail,
-
   .exec_func = openacc_parallel,
 
   .register_async_cleanup_func = openacc_register_async_cleanup,
diff --git libgomp/oacc-init.c libgomp/oacc-init.c
index 6f72578..0e0fee9 100644
--- libgomp/oacc-init.c
+++ libgomp/oacc-init.c
@@ -97,7 +97,7 @@ resolve_device (acc

[PATCH] Correctly check dg-require-effective-target in avx512 tests.

[Ilya Tocar]

Hi,

Currently we only check for dg-require-effective-target avx512vl in
avx512vl tests. We should also check for avx512dq/avx512bw.
Patch bwllow does this.
Ok for trunk?

2014-11-05  Ilya Tocar  

* gcc.target/i386/avx512vl-vandnpd-2.c: Fix
dg-require-effective-target cehck.
* gcc.target/i386/avx512vl-vandnps-2.c: Ditto.
* gcc.target/i386/avx512vl-vandpd-2.c: Ditto.
* gcc.target/i386/avx512vl-vandps-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcastf32x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcastf32x4-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcastf64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcasti32x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcasti32x4-2.c: Ditto.
* gcc.target/i386/avx512vl-vbroadcasti64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtpd2qq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtpd2uqq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtps2qq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtps2uqq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtqq2pd-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtqq2ps-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvttpd2qq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvttpd2uqq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvttps2qq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvttps2uqq-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtuqq2pd-2.c: Ditto.
* gcc.target/i386/avx512vl-vcvtuqq2ps-2.c: Ditto.
* gcc.target/i386/avx512vl-vdbpsadbw-2.c: Ditto.
* gcc.target/i386/avx512vl-vextractf64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vextracti64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vfpclasspd-2.c: Ditto.
* gcc.target/i386/avx512vl-vfpclassps-2.c: Ditto.
* gcc.target/i386/avx512vl-vinsertf64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vinserti64x2-2.c: Ditto.
* gcc.target/i386/avx512vl-vmovdqu16-2.c: Ditto.
* gcc.target/i386/avx512vl-vmovdqu8-2.c: Ditto.
* gcc.target/i386/avx512vl-vorpd-2.c: Ditto.
* gcc.target/i386/avx512vl-vorps-2.c: Ditto.
* gcc.target/i386/avx512vl-vpabsb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpabsw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpackssdw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpacksswb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpackusdw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpackuswb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddsb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddsw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddusb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddusw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpaddw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpalignr-2.c: Ditto.
* gcc.target/i386/avx512vl-vpavgb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpavgw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpblendmb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpblendmw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpbroadcastb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpbroadcastw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpeqb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpequb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpequw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpeqw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpgtb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpgtub-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpgtuw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpgtw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpub-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpuw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpcmpw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpermi2w-2.c: Ditto.
* gcc.target/i386/avx512vl-vpermt2w-2.c: Ditto.
* gcc.target/i386/avx512vl-vpermw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaddubsw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaddwd-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaxsb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaxsw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaxub-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmaxuw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpminsb-2.c: Ditto.
* gcc.target/i386/avx512vl-vpminsw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpminub-2.c: Ditto.
* gcc.target/i386/avx512vl-vpminuw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovb2m-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovd2m-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovm2b-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovm2d-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovm2q-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovm2w-2.c: Ditto.
* gcc.target/i386/avx512vl-vpmovq2m-2.c

[gomp4] OpenACC documentation updates.

[Thomas Schwinge]

Hi!

Applied to gomp-4_0-branch in r217142:

commit 0c5178ff5207bf1ede83070629c7d76fbbdf1afb
Author: tschwinge 
Date:   Wed Nov 5 16:12:51 2014 +

OpenACC documentation updates.

gcc/
* invoke.texi: Update for OpenACC.
* sourcebuild.texi: Likewise.
gcc/fortran/
* gfortran.texi: Update for OpenACC.
* intrinsic.texi: Likewise.
* invoke.texi: Likewise.
libgomp/
* libgomp.texi: Update for OpenACC.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217142 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog.gomp |  3 +++
 gcc/doc/invoke.texi|  4 ++--
 gcc/doc/sourcebuild.texi   |  2 +-
 gcc/fortran/ChangeLog.gomp |  6 ++
 gcc/fortran/gfortran.texi  | 38 ++
 gcc/fortran/intrinsic.texi | 31 ++-
 gcc/fortran/invoke.texi|  7 ++-
 libgomp/ChangeLog.gomp |  2 ++
 libgomp/libgomp.texi   | 10 ++
 9 files changed, 90 insertions(+), 13 deletions(-)

diff --git gcc/ChangeLog.gomp gcc/ChangeLog.gomp
index 5b2bade..fc624c8 100644
--- gcc/ChangeLog.gomp
+++ gcc/ChangeLog.gomp
@@ -1,5 +1,8 @@
 2014-11-05  Thomas Schwinge  
 
+   * invoke.texi: Update for OpenACC.
+   * sourcebuild.texi: Likewise.
+
* tree.def (OACC_WAIT): Remove.  Update all users.
 
* omp-builtins.def (BUILT_IN_OMP_SET_NUM_THREADS): Remove.
diff --git gcc/doc/invoke.texi gcc/doc/invoke.texi
index 4cd4f4a..0fe875b 100644
--- gcc/doc/invoke.texi
+++ gcc/doc/invoke.texi
@@ -1872,8 +1872,8 @@ freestanding and hosted environments.
 @item -fopenacc
 @opindex fopenacc
 @cindex OpenACC accelerator programming
-Enable handling of OpenACC directives @code{#pragma acc} in C.
-When @option{-fopenacc} is specified, the
+Enable handling of OpenACC directives @code{#pragma acc} in C/C++ and
+@code{!$acc} in Fortran.  When @option{-fopenacc} is specified, the
 compiler generates accelerated code according to the OpenACC Application
 Programming Interface v2.0 @w{@uref{http://www.openacc.org/}}.  This option
 implies @option{-pthread}, and thus is only supported on targets that
diff --git gcc/doc/sourcebuild.texi gcc/doc/sourcebuild.texi
index 5d1625d..d27fac0 100644
--- gcc/doc/sourcebuild.texi
+++ gcc/doc/sourcebuild.texi
@@ -89,7 +89,7 @@ The Go runtime library.  The bulk of this library is mirrored 
from the
 @uref{http://code.google.com/@/p/@/go/, master Go repository}.
 
 @item libgomp
-The GNU OpenMP runtime library.
+The GNU OpenACC and OpenMP runtime library.
 
 @item libiberty
 The @code{libiberty} library, used for portability and for some
diff --git gcc/fortran/ChangeLog.gomp gcc/fortran/ChangeLog.gomp
index 5f2e9ba..98e3971 100644
--- gcc/fortran/ChangeLog.gomp
+++ gcc/fortran/ChangeLog.gomp
@@ -1,3 +1,9 @@
+2014-11-05  Thomas Schwinge  
+
+   * gfortran.texi: Update for OpenACC.
+   * intrinsic.texi: Likewise.
+   * invoke.texi: Likewise.
+
 2014-11-04  Cesar Philippidis  
 
* gfortran.h (ST_OACC_ROUTINE): New statement enum.
diff --git gcc/fortran/gfortran.texi gcc/fortran/gfortran.texi
index 41d6559..c3e7518 100644
--- gcc/fortran/gfortran.texi
+++ gcc/fortran/gfortran.texi
@@ -474,7 +474,8 @@ The GNU Fortran compiler is able to compile nearly all
 standard-compliant Fortran 95, Fortran 90, and Fortran 77 programs,
 including a number of standard and non-standard extensions, and can be
 used on real-world programs.  In particular, the supported extensions
-include OpenMP, Cray-style pointers, and several Fortran 2003 and Fortran
+include OpenACC, OpenMP, Cray-style pointers, and several Fortran 2003
+and Fortran
 2008 features, including TR 15581.  However, it is still under
 development and has a few remaining rough edges.
 
@@ -531,7 +532,8 @@ The current status of the support is can be found in the
 @ref{Fortran 2003 status}, @ref{Fortran 2008 status} and
 @ref{TS 29113 status} sections of the documentation.
 
-Additionally, the GNU Fortran compilers supports the OpenMP specification
+Additionally, the GNU Fortran compilers supports the OpenACC specification
+(version 2.0, @url{http://www.openacc.org/}), and OpenMP specification
 (version 4.0, @url{http://openmp.org/@/wp/@/openmp-specifications/}).
 
 @node Varying Length Character Strings
@@ -963,7 +965,8 @@ module.
 @cindex statement, @code{ISO_FORTRAN_ENV}
 @code{USE} statement with @code{INTRINSIC} and @code{NON_INTRINSIC}
 attribute; supported intrinsic modules: @code{ISO_FORTRAN_ENV},
-@code{ISO_C_BINDING}, @code{OMP_LIB} and @code{OMP_LIB_KINDS}.
+@code{ISO_C_BINDING}, @code{OMP_LIB} and @code{OMP_LIB_KINDS},
+and @code{OPENACC}.
 
 @item
 Renaming of operators in the @code{USE} statement.
@@ -1358,6 +1361,7 @@ without warning.
 * Hollerith constants support::
 * Cray pointers::
 * CONVERT specifier::
+* OpenACC::
 * OpenMP::
 * Argument list functions::
 @end menu
@@ -1873,6 +1877,32 @@ carries a significant speed overhead.  If speed in t

[gomp4] libgomp testsuite: OpenACC C++ testing (was: [2/3] OpenACC 2.0 support for libgomp - new tests)

[Thomas Schwinge]

Hi!

Applied to gomp-4_0-branch in r217143:

commit a78a06124f4047ec46a85e539e83640cc973aec1
Author: tschwinge 
Date:   Wed Nov 5 16:16:14 2014 +

libgomp testsuite: OpenACC C++ testing.

libgomp/
* testsuite/libgomp.oacc-c++/c++.exp: Enable
libgomp.oacc-c-c++-common testing.
* testsuite/libgomp.oacc-c/c.exp: Likewise.
* testsuite/libgomp.oacc-c/abort-2.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/abort-2.c: ... this.
* testsuite/libgomp.oacc-c/abort.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/abort.c: ... this.
* testsuite/libgomp.oacc-c/acc_on_device-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/acc_on_device-1.c: ... this.
* testsuite/libgomp.oacc-c/clauses-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/clauses-1.c: ... this.
* testsuite/libgomp.oacc-c/clauses-2.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/clauses-2.c: ... this.
* testsuite/libgomp.oacc-c/context-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/context-1.c: ... this.
* testsuite/libgomp.oacc-c/context-2.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/context-2.c: ... this.
* testsuite/libgomp.oacc-c/context-3.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/context-3.c: ... this.
* testsuite/libgomp.oacc-c/context-4.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/context-4.c: ... this.
* testsuite/libgomp.oacc-c/data-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/data-1.c: ... this.
* testsuite/libgomp.oacc-c/data-2.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/data-2.c: ... this.
* testsuite/libgomp.oacc-c/data-3.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/data-3.c: ... this.
* testsuite/libgomp.oacc-c/deviceptr-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/deviceptr-1.c: ... this.
* testsuite/libgomp.oacc-c/if-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/if-1.c: ... this.
* testsuite/libgomp.oacc-c/kernels-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/kernels-1.c: ... this.
* testsuite/libgomp.oacc-c/lib-1.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-1.c: ... this.
* testsuite/libgomp.oacc-c/lib-10.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-10.c: ... this.
* testsuite/libgomp.oacc-c/lib-11.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-11.c: ... this.
* testsuite/libgomp.oacc-c/lib-12.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-12.c: ... this.
* testsuite/libgomp.oacc-c/lib-13.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-13.c: ... this.
* testsuite/libgomp.oacc-c/lib-14.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-14.c: ... this.
* testsuite/libgomp.oacc-c/lib-15.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-15.c: ... this.
* testsuite/libgomp.oacc-c/lib-16.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-16.c: ... this.
* testsuite/libgomp.oacc-c/lib-17.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-17.c: ... this.
* testsuite/libgomp.oacc-c/lib-18.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-18.c: ... this.
* testsuite/libgomp.oacc-c/lib-19.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-19.c: ... this.
* testsuite/libgomp.oacc-c/lib-2.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-2.c: ... this.
* testsuite/libgomp.oacc-c/lib-20.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-20.c: ... this.
* testsuite/libgomp.oacc-c/lib-21.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-21.c: ... this.
* testsuite/libgomp.oacc-c/lib-22.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-22.c: ... this.
* testsuite/libgomp.oacc-c/lib-23.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-23.c: ... this.
* testsuite/libgomp.oacc-c/lib-24.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-24.c: ... this.
* testsuite/libgomp.oacc-c/lib-25.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-25.c: ... this.
* testsuite/libgomp.oacc-c/lib-26.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-26.c: ... this.
* testsuite/libgomp.oacc-c/lib-27.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-27.c: ... this.
* testsuite/libgomp.oacc-c/lib-28.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-28.c: ... this.
* testsuite/libgomp.oacc-c/lib-29.c: Rename to...
* testsuite/libgomp.oacc-c-c++-common/lib-

Re: [PATCH] Correctly check dg-require-effective-target in avx512 tests.

[Uros Bizjak]

On Wed, Nov 5, 2014 at 5:14 PM, Ilya Tocar  wrote:
> Hi,
>
> Currently we only check for dg-require-effective-target avx512vl in
> avx512vl tests. We should also check for avx512dq/avx512bw.
> Patch bwllow does this.
> Ok for trunk?
>
> 2014-11-05  Ilya Tocar  
>
> * gcc.target/i386/avx512vl-vandnpd-2.c: Fix
> dg-require-effective-target cehck.
> * gcc.target/i386/avx512vl-vandnps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vandpd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vandps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcastf32x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcastf32x4-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcastf64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcasti32x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcasti32x4-2.c: Ditto.
> * gcc.target/i386/avx512vl-vbroadcasti64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtpd2qq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtpd2uqq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtps2qq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtps2uqq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtqq2pd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtqq2ps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvttpd2qq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvttpd2uqq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvttps2qq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvttps2uqq-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtuqq2pd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vcvtuqq2ps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vdbpsadbw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vextractf64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vextracti64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vfpclasspd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vfpclassps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vinsertf64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vinserti64x2-2.c: Ditto.
> * gcc.target/i386/avx512vl-vmovdqu16-2.c: Ditto.
> * gcc.target/i386/avx512vl-vmovdqu8-2.c: Ditto.
> * gcc.target/i386/avx512vl-vorpd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vorps-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpabsb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpabsw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpackssdw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpacksswb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpackusdw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpackuswb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddsb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddsw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddusb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddusw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpaddw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpalignr-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpavgb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpavgw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpblendmb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpblendmw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpbroadcastb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpbroadcastw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpeqb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpequb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpequw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpeqw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpgtb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpgtub-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpgtuw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpgtw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpub-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpuw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpcmpw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpermi2w-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpermt2w-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpermw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaddubsw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaddwd-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaxsb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaxsw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaxub-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmaxuw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpminsb-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpminsw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpminub-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpminuw-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmovb2m-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmovd2m-2.c: Ditto.
> * gcc.target/i386/avx512vl-vpmovm2b

Enable -fextended-identifiers by default

[Joseph Myers]

As proposed at ,
this patch enables -fextended-identifiers by default for all standard
versions including this feature (all C++ versions, C99 and above for
C, but not C90 / C94 / gnu89 / preprocessing assembler).  It adds a
couple of tests for areas where I previously noted testsuite coverage
for extended identifiers was lacking, removes -fextended-identifiers
from existing tests, adds -g to various such tests to verify that
extended identifiers don't break debug info generation and removes the
test that was only there to verify that the feature was off by
default.

The current state of the feature may not correspond exactly to any
particular checklist from 2004/5 (see bug 9449) of what was wanted
before enabling the feature by default, but I don't think it's any
worse than plenty of other features supported by default before every
corner case is fully functional, and think problems can readily be
fixed incrementally.

The following aspects of extended identifiers could still do with more
work (and should be straightforward):

* C -aux-info (output should use UCNs).

* ObjC -gen-decls (output should use UCNs; associated diagnostics from
  the ObjC front end should use extended characters or UCNs as
  appropriate to the locale, via using %qE or identifier_to_locale).

* Use DW_AT_use_UTF8 in DWARF-3 debug info for compilation units built
  with extended identifiers enabled (or unconditionally).

* cpplib diagnostics (outputting characters or UCNs as appropriate
  depending on the locale, as done for identifiers in non-cpplib
  diagnostics).

* C++ test for UCN linking with C and extern "C".

* Check GDB support / file issues for support if needed.

* Actual UTF-8 in identifiers (?).  (Be careful about not affecting
  performance for the normal fast path of lexing identifiers, if
  possible.)

The following may be trickier:

* cpplib spelling preservation (required to diagnose macro
  redefinition with different spellings of the same identifier in the
  definition or argument names; different spellings of the name of the
  macro itself are OK, however; also required for correct handling of
  multiple stringizing in C++); correct output for -d (UCNs), DWARF
  debug info for macros (UCNs), PCH and PCH tests.  (Spelling
  preservation is the issue that needs fixing to remove references to
  corner cases in the documentation of -std=c99 and -std=c11 and in
  c99status.html.)  The idea would be to add a second pointer to
  cpp_identifier that stores the original spelling (whether for
  extended identifiers only, or for all identifiers); this does not
  enlarge cpp_token because the resulting larger cpp_identifier
  structure is no bigger than cpp_string.

* C++ translation of extended characters (including $@` and various
  control characters) to UCNs in phase 1 (note diagnostics thus
  needed, but not for C++11, for control characters in strings /
  character constants as those UCNs invalid); a likely implementation
  approach is to do translation when identifiers / strings / character
  constants are lexed, together with errors for stray $@` / control
  characters in program as not being valid UCNs in identifiers ($ only
  if not accepted in identifiers); note that this translation should
  not take place inside raw string literals.

Bootstrapped with no regressions on x86_64-unknown-linux-gnu.  Applied
to mainline.

libcpp:
2014-11-05  Joseph Myers  

PR preprocessor/9449
* init.c (lang_defaults): Enable extended identifiers for C++ and
C99-based standards.

gcc:
2014-11-05  Joseph Myers  

PR preprocessor/9449
* doc/cpp.texi (Character sets, Tokenization)
(Implementation-defined behavior): Don't refer to UCNs in
identifiers requiring -fextended-identifiers.
* doc/cppopts.texi (-fextended-identifiers): Document as enabled
by default for C99 and later and C++.
* doc/invoke.texi (-std=c99, -std=c11): Don't refer to extended
identifiers needing -fextended-identifiers.

gcc/testsuite:
2014-11-05  Joseph Myers  

PR preprocessor/9449
* lib/target-supports.exp (check_effective_target_ucn_nocache):
Don't use -fextended-identifiers.
* c-c++-common/cpp/normalize-3.c, c-c++-common/cpp/ucnid-2011-1.c,
g++.dg/cpp/ucn-1.C, g++.dg/cpp/ucnid-1.C, g++.dg/other/ucnid-1.C,
gcc.dg/cpp/normalize-1.c, gcc.dg/cpp/normalize-2.c,
gcc.dg/cpp/normalize-4.c: Don't use -fextended-identifiers.
* gcc.dg/cpp/ucnid-1.c: Don't use -fextended-identifiers.  Use
-g3.
* gcc.dg/cpp/ucnid-10.c, gcc.dg/cpp/ucnid-2.c,
gcc.dg/cpp/ucnid-3.c, gcc.dg/cpp/ucnid-4.c, gcc.dg/cpp/ucnid-5.c,
gcc.dg/cpp/ucnid-7.c, gcc.dg/cpp/ucnid-9.c,
gcc.dg/cpp/warn-normalized-1.c, gcc.dg/cpp/warn-normalized-2.c,
gcc.dg/cpp/warn-normalized-3.c: Don't use -fextended-identifiers.
* gcc.dg/ucnid-1.c, gcc.dg/ucnid-2.c, gcc.dg

[gomp4] OpenACC cache directive for C.

[Thomas Schwinge]

Hi!

In r217145, I applied Jim's patch to gomp-4_0-branch:

commit 4361f9b6b2c74c2961c3a5290a4945abe2d7a444
Author: tschwinge 
Date:   Wed Nov 5 16:26:47 2014 +

OpenACC cache directive for C.

gcc/c-family/
* c-pragma.c (oacc_pragmas): Add "cache".
gcc/c/
* c-parser.c (c_parser_omp_variable_list): Handle
OMP_NO_CLAUSE_CACHE.
(c_parser_oacc_cache): New function.
(c_parser_omp_construct): Use it for PRAGMA_OACC_CACHE.
libgomp/
* testsuite/libgomp.oacc-c/cache-1.c: New file.
* testsuite/libgomp.oacc-c++/cache-1.C: Likewise.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217145 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/c-family/ChangeLog.gomp  |  4 ++
 gcc/c-family/c-pragma.c  |  1 +
 gcc/c/ChangeLog.gomp |  7 +++
 gcc/c/c-parser.c | 50 
 libgomp/ChangeLog.gomp   |  5 ++
 libgomp/testsuite/libgomp.oacc-c++/cache-1.C | 50 
 libgomp/testsuite/libgomp.oacc-c/cache-1.c   | 69 
 7 files changed, 186 insertions(+)

diff --git gcc/c-family/ChangeLog.gomp gcc/c-family/ChangeLog.gomp
index 2f8c8a6..5f3b641 100644
--- gcc/c-family/ChangeLog.gomp
+++ gcc/c-family/ChangeLog.gomp
@@ -1,3 +1,7 @@
+2014-11-05  James Norris  
+
+   * c-pragma.c (oacc_pragmas): Add "cache".
+
 2014-11-03  Cesar Philippidis  
 
* c-pragma.c (oacc_pragmas): Add entries for PRAGMA_OACC_ENTER_DATA
diff --git gcc/c-family/c-pragma.c gcc/c-family/c-pragma.c
index a28727e..e98b555 100644
--- gcc/c-family/c-pragma.c
+++ gcc/c-family/c-pragma.c
@@ -1181,6 +1181,7 @@ static vec registered_pp_pragmas;
 
 struct omp_pragma_def { const char *name; unsigned int id; };
 static const struct omp_pragma_def oacc_pragmas[] = {
+  { "cache", PRAGMA_OACC_CACHE },
   { "data", PRAGMA_OACC_DATA },
   { "enter", PRAGMA_OACC_ENTER_DATA },
   { "exit", PRAGMA_OACC_EXIT_DATA },
diff --git gcc/c/ChangeLog.gomp gcc/c/ChangeLog.gomp
index f4e2010..7acd7b3 100644
--- gcc/c/ChangeLog.gomp
+++ gcc/c/ChangeLog.gomp
@@ -1,3 +1,10 @@
+2014-11-05  James Norris  
+
+   * c-parser.c (c_parser_omp_variable_list): Handle
+   OMP_NO_CLAUSE_CACHE.
+   (c_parser_oacc_cache): New function.
+   (c_parser_omp_construct): Use it for PRAGMA_OACC_CACHE.
+
 2014-11-03  Cesar Philippidis  
 
* c-parser.c (c_parser_oacc_enter_exit_data): New function.
diff --git gcc/c/c-parser.c gcc/c/c-parser.c
index 6ac1ace..410b19f 100644
--- gcc/c/c-parser.c
+++ gcc/c/c-parser.c
@@ -10053,6 +10053,14 @@ c_parser_omp_variable_list (c_parser *parser,
{
  switch (kind)
{
+   case OMP_NO_CLAUSE_CACHE:
+ if (c_parser_peek_token (parser)->type != CPP_OPEN_SQUARE)
+   {
+ c_parser_error (parser, "expected %<[%>");
+ t = error_mark_node;
+ break;
+   }
+ /* FALL THROUGH.  */
case OMP_CLAUSE_MAP:
case OMP_CLAUSE_FROM:
case OMP_CLAUSE_TO:
@@ -10091,6 +10099,29 @@ c_parser_omp_variable_list (c_parser *parser,
  t = error_mark_node;
  break;
}
+
+ if (kind == OMP_NO_CLAUSE_CACHE)
+   {
+ mark_exp_read (low_bound);
+ mark_exp_read (length);
+
+ if (TREE_CODE (low_bound) != INTEGER_CST
+ && !TREE_READONLY (low_bound))
+   {
+ error_at (clause_loc,
+   "%qD is not a constant", low_bound);
+ t = error_mark_node;
+   }
+
+ if (TREE_CODE (length) != INTEGER_CST
+ && !TREE_READONLY (length))
+   {
+ error_at (clause_loc,
+   "%qD is not a constant", length);
+ t = error_mark_node;
+   }
+   }
+
  t = tree_cons (low_bound, length, t);
}
  break;
@@ -11864,6 +11895,21 @@ c_parser_omp_structured_block (c_parser *parser)
 }
 
 /* OpenACC 2.0:
+   # pragma acc cache (variable-list) new-line
+
+   LOC is the location of the #pragma token.
+*/
+
+static tree
+c_parser_oacc_cache (location_t loc __attribute__((unused)), c_parser *parser)
+{
+  c_parser_omp_var_list_parens (parser, OMP_NO_CLAUSE_CACHE, NULL);
+  c_parser_skip_to_pragma_eol (parser);
+
+  return NULL_TREE;
+}
+
+/* OpenACC 2.0:
# pragma acc data oacc-data-clause[optseq] new-line
  structured-block
 
@@ -14506,6 +14552,10 @@ c_parser_omp_construct (c_parser *parser)
 
   switch (p_kind)
 {
+case PRAGMA_OACC_CACHE:
+  strcpy (p_name, "#p

[gomp4] OpenACC cache directive maintenance (was: [PATCH 4/6] [GOMP4] OpenACC 1.0+ support in fortran front-end)

[Thomas Schwinge]

Hi!

On Fri, 24 Jan 2014 20:33:35 +0100, I wrote:
> On Thu, 23 Jan 2014 22:04:45 +0400, Ilmir Usmanov  
> wrote:
> > Subject: [PATCH 4/6] OpenACC GENERIC nodes

> > --- a/gcc/tree-core.h
> > +++ b/gcc/tree-core.h
> > @@ -216,12 +216,18 @@ enum omp_clause_code {

> > +  /* Internal structure to hold OpenACC cache directive's variable-list.
> > + #pragma acc cache (variable-_ist).  */
> > +  OACC_NO_CLAUSE_CACHE,
> 
> Hmm, yeah, while *_NO_CLAUSE_* perhaps isn't the most beautiful approach,
> I think it's fine at least for now.

In r217146, I applied the following to gomp-4_0-branch:

commit e8e44b733808997d06c0cdf9bf5756ce03530f42
Author: tschwinge 
Date:   Wed Nov 5 16:35:30 2014 +

OpenACC cache directive maintenance.

gcc/c/
* c-parser.c (c_parser_oacc_cache): Generate OACC_CACHE.
* c-typeck.c (c_finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
gcc/cp/
* parser.c (cp_parser_oacc_cache): Generate OACC_CACHE.
* semantics.c (finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
gcc/
* gimplify.c (gimplify_oacc_cache): New function.
(gimplify_expr): Use it for OACC_CACHE.
(gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses): Handle
OMP_CLAUSE__CACHE_.

gcc/c/
* c-parser.c (c_parser_omp_variable_list) :
Remove explicit mark_exp_read invocations.
gcc/cp/
* parser.c (cp_parser_omp_var_list_no_open) :
Remove explicit mark_exp_read invocations.

gcc/
* tree-core.h (enum omp_clause_code): Move OMP_NO_CLAUSE_CACHE
next to, and handle it like a data clause.  Rename it to
OMP_CLAUSE__CACHE_.  Update all users.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217146 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog.gomp |  9 +
 gcc/c/ChangeLog.gomp   |  8 
 gcc/c/c-parser.c   | 23 +++
 gcc/c/c-typeck.c   |  1 +
 gcc/cp/ChangeLog.gomp  |  6 ++
 gcc/cp/parser.c| 24 +++-
 gcc/cp/semantics.c |  1 +
 gcc/fortran/trans-openmp.c |  2 +-
 gcc/gimplify.c | 25 ++---
 gcc/omp-low.c  |  4 ++--
 gcc/tree-core.h|  8 
 gcc/tree-pretty-print.c| 11 +++
 gcc/tree.c |  6 +++---
 gcc/tree.def   |  5 +++--
 gcc/tree.h |  2 +-
 15 files changed, 98 insertions(+), 37 deletions(-)

diff --git gcc/ChangeLog.gomp gcc/ChangeLog.gomp
index fc624c8..2c2b349 100644
--- gcc/ChangeLog.gomp
+++ gcc/ChangeLog.gomp
@@ -1,5 +1,14 @@
 2014-11-05  Thomas Schwinge  
 
+   * gimplify.c (gimplify_oacc_cache): New function.
+   (gimplify_expr): Use it for OACC_CACHE.
+   (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses): Handle
+   OMP_CLAUSE__CACHE_.
+
+   * tree-core.h (enum omp_clause_code): Move OMP_NO_CLAUSE_CACHE
+   next to, and handle it like a data clause.  Rename it to
+   OMP_CLAUSE__CACHE_.  Update all users.
+
* invoke.texi: Update for OpenACC.
* sourcebuild.texi: Likewise.
 
diff --git gcc/c/ChangeLog.gomp gcc/c/ChangeLog.gomp
index 7acd7b3..70278b9 100644
--- gcc/c/ChangeLog.gomp
+++ gcc/c/ChangeLog.gomp
@@ -1,3 +1,11 @@
+2014-11-05  Thomas Schwinge  
+
+   * c-parser.c (c_parser_oacc_cache): Generate OACC_CACHE.
+   * c-typeck.c (c_finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
+
+   * c-parser.c (c_parser_omp_variable_list) :
+   Remove explicit mark_exp_read invocations.
+
 2014-11-05  James Norris  
 
* c-parser.c (c_parser_omp_variable_list): Handle
diff --git gcc/c/c-parser.c gcc/c/c-parser.c
index 410b19f..40d4314 100644
--- gcc/c/c-parser.c
+++ gcc/c/c-parser.c
@@ -10053,7 +10053,7 @@ c_parser_omp_variable_list (c_parser *parser,
{
  switch (kind)
{
-   case OMP_NO_CLAUSE_CACHE:
+   case OMP_CLAUSE__CACHE_:
  if (c_parser_peek_token (parser)->type != CPP_OPEN_SQUARE)
{
  c_parser_error (parser, "expected %<[%>");
@@ -10100,11 +10100,8 @@ c_parser_omp_variable_list (c_parser *parser,
  break;
}
 
- if (kind == OMP_NO_CLAUSE_CACHE)
+ if (kind == OMP_CLAUSE__CACHE_)
{
- mark_exp_read (low_bound);
- mark_exp_read (length);
-
  if (TREE_CODE (low_bound) != INTEGER_CST
  && !TREE_READONLY (low_bound))
{
@@ -11901,12 +11898,22 @@ c_parser_omp_structured_block (c_parser *parser)
 */
 
 static tree
-c_parser_oacc_cache (location_t loc __attribute__((unused)), c_parser *parser)
+c_parser_oacc_cache (location_t loc, c_parser *parser)
 {
-  c_parser_omp_var_list_parens (parser, OMP_NO_CLAUSE_CACHE, NULL);
+  tree stmt, clauses;
+
+  clauses = c_

[PING 2] Enhance array types debug info. for Ada

[Pierre-Marie de Rodat]

Hello,

Ping for https://gcc.gnu.org/ml/gcc-patches/2014-10/msg00694.html

Thanks in advance!
Regards,

--
Pierre-Marie de Rodat

Re: [gomp4] OpenACC cache directive maintenance

[Thomas Schwinge]

Hi Cesar!

On Wed, 05 Nov 2014 17:36:46 +0100, I wrote:
> In r217146, I applied the following to gomp-4_0-branch:
> 
> commit e8e44b733808997d06c0cdf9bf5756ce03530f42
> Author: tschwinge 
> Date:   Wed Nov 5 16:35:30 2014 +
> 
> OpenACC cache directive maintenance.
> 
>   gcc/c/
>   * c-parser.c (c_parser_oacc_cache): Generate OACC_CACHE.
>   * c-typeck.c (c_finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
>   gcc/cp/
>   * parser.c (cp_parser_oacc_cache): Generate OACC_CACHE.
>   * semantics.c (finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
>   gcc/
>   * gimplify.c (gimplify_oacc_cache): New function.
>   (gimplify_expr): Use it for OACC_CACHE.
>   (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses): Handle
>   OMP_CLAUSE__CACHE_.

> --- gcc/c/c-parser.c
> +++ gcc/c/c-parser.c

>  static tree
> -c_parser_oacc_cache (location_t loc __attribute__((unused)), c_parser 
> *parser)
> +c_parser_oacc_cache (location_t loc, c_parser *parser)
>  {
> -  c_parser_omp_var_list_parens (parser, OMP_NO_CLAUSE_CACHE, NULL);
> +  tree stmt, clauses;
> +
> +  clauses = c_parser_omp_var_list_parens (parser, OMP_CLAUSE__CACHE_, NULL);
> +  clauses = c_finish_omp_clauses (clauses);
> +
>c_parser_skip_to_pragma_eol (parser);
>  
> -  return NULL_TREE;
> +  stmt = make_node (OACC_CACHE);
> +  TREE_TYPE (stmt) = void_type_node;
> +  OACC_CACHE_CLAUSES (stmt) = clauses;
> +  SET_EXPR_LOCATION (stmt, loc);
> +  add_stmt (stmt);
> +
> +  return stmt;
>  }

> --- gcc/c/c-typeck.c
> +++ gcc/c/c-typeck.c
> @@ -12204,6 +12204,7 @@ c_finish_omp_clauses (tree clauses)
>   case OMP_CLAUSE_MAP:
>   case OMP_CLAUSE_TO:
>   case OMP_CLAUSE_FROM:
> + case OMP_CLAUSE__CACHE_:
> t = OMP_CLAUSE_DECL (c);
> if (TREE_CODE (t) == TREE_LIST)
>   {

[The "same" for C++.]

I also tried to make this work for Fortran, but didn't manage to (in
a reasonable amount of time, which has not been a lot that I allocated)
;-) -- would you please have a look at this (but it's not urgent).

commit 7cf14ddb307a5b271e098f3a3fdb0452f6036f91
Author: Thomas Schwinge 
Date:   Wed Nov 5 09:16:12 2014 +0100

cache: Fortran experimenting.
---
 gcc/fortran/frontend-passes.c |  3 ++-
 gcc/fortran/openmp.c  | 12 ++--
 gcc/fortran/trans-openmp.c| 18 --
 gcc/testsuite/gfortran.dg/goacc/cache-1.f95   |  1 -
 gcc/testsuite/gfortran.dg/goacc/coarray.f95   |  1 -
 gcc/testsuite/gfortran.dg/goacc/cray.f95  |  2 --
 gcc/testsuite/gfortran.dg/goacc/loop-1.f95|  1 -
 gcc/testsuite/gfortran.dg/goacc/parameter.f95 |  1 -
 8 files changed, 16 insertions(+), 23 deletions(-)

diff --git gcc/fortran/frontend-passes.c gcc/fortran/frontend-passes.c
index 97a9164..729629e 100644
--- gcc/fortran/frontend-passes.c
+++ gcc/fortran/frontend-passes.c
@@ -2190,7 +2190,8 @@ gfc_code_walker (gfc_code **c, walk_code_fn_t codefn, 
walk_expr_fn_t exprfn,
  gfc_omp_namelist *n;
  static int list_types[]
= { OMP_LIST_ALIGNED, OMP_LIST_LINEAR, OMP_LIST_DEPEND,
-   OMP_LIST_MAP, OMP_LIST_TO, OMP_LIST_FROM };
+   OMP_LIST_MAP, OMP_LIST_TO, OMP_LIST_FROM,
+   OMP_LIST_CACHE };
  size_t idx;
  WALK_SUBEXPR (co->ext.omp_clauses->if_expr);
  WALK_SUBEXPR (co->ext.omp_clauses->final_expr);
diff --git gcc/fortran/openmp.c gcc/fortran/openmp.c
index 959798a..167331a 100644
--- gcc/fortran/openmp.c
+++ gcc/fortran/openmp.c
@@ -3102,6 +3102,7 @@ resolve_omp_clauses (gfc_code *code, locus *where,
  case OMP_LIST_MAP:
  case OMP_LIST_TO:
  case OMP_LIST_FROM:
+ case OMP_LIST_CACHE:
for (; n != NULL; n = n->next)
  {
if (n->expr)
@@ -4594,13 +4595,6 @@ resolve_oacc_loop(gfc_code *code)
 }
 
 
-static void
-resolve_oacc_cache (gfc_code *code)
-{
-  gfc_error ("Sorry, !$ACC cache unimplemented yet at %L", &code->loc);
-}
-
-
 void
 gfc_resolve_oacc_declare (gfc_namespace *ns)
 {
@@ -4675,6 +4669,7 @@ gfc_resolve_oacc_directive (gfc_code *code, gfc_namespace 
*ns ATTRIBUTE_UNUSED)
 case EXEC_OACC_ENTER_DATA:
 case EXEC_OACC_EXIT_DATA:
 case EXEC_OACC_WAIT:
+case EXEC_OACC_CACHE:
   resolve_omp_clauses (code, &code->loc, code->ext.omp_clauses, NULL,
   true);
   break;
@@ -4683,9 +4678,6 @@ gfc_resolve_oacc_directive (gfc_code *code, gfc_namespace 
*ns ATTRIBUTE_UNUSED)
 case EXEC_OACC_LOOP:
   resolve_oacc_loop (code);
   break;
-case EXEC_OACC_CACHE:
-  resolve_oacc_cache (code);
-  break;
 default:
   break;
 }
diff --git gcc/fortran/trans-openmp.c gcc/fortran/trans-openmp.c
index 7dd4498..e39e903 100644
--- gcc/fortran/trans-openmp.c
+++ gcc/fortran/trans-openmp.c
@@ -1806,9 +1806,9 @@ gfc_trans_omp_clauses (stmtbl

[PATCH] Fix libbacktrace and libiberty tests fail on sanitized GCC due to wrong link options.

[Maxim Ostapenko]

Hi,

When I ran Asan tests under Asan-bootstrapped GCC 5.0, I've noted, that tests 
for libiberty and libbacktrace fail to link with sanitized libbacktrace.a and 
libiberty.a because of
missing -static-libasan -fsanitize=address linker flags.

This patch adds necessary flags to provide a linkage of these tests in 
bootstrap-asan case.

I've checked that regression tests pass with disabled bootstrap, normal 
bootstrap (stage1, stage3) and Asan-bootstrap (stage 1, stage3) on 
x86_64-unknown-linux-gnu.

Does the patch look sane?

-Maxim

libiberty/ChangeLog:

2014-11-05  Max Ostapenko  

	* testsuite/Makefile.in (LIBCFLAGS): Add LDFLAGS.

ChangeLog:

2014-11-05  Max Ostapenko  

	* Makefile.tpl (EXTRA_HOST_EXPORTS): New variables.
	(EXTRA_BOOTSTRAP_FLAGS): Likewise.
	(check-[+module+]): Add EXTRA_HOST_EXPORTS and EXTRA_BOOTSTRAP_FLAGS.
	* Makefile.in: Regenerate.

diff --git a/Makefile.in b/Makefile.in
index 4564dbe..62c8301 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -832,6 +832,14 @@ POSTSTAGE1_FLAGS_TO_PASS = \
 	$(LTO_FLAGS_TO_PASS) \
 	"`echo 'ADAFLAGS=$(BOOT_ADAFLAGS)' | sed -e s'/[^=][^=]*=$$/XFOO=/'`"
 
+@if gcc-bootstrap
+EXTRA_HOST_EXPORTS = if [ $(current_stage) != stage1 ]; then \
+		   $(POSTSTAGE1_HOST_EXPORTS) \
+		 fi ;
+
+EXTRA_BOOTSTRAP_FLAGS = CC="$$CC" CXX="$$CXX" LDFLAGS="$$LDFLAGS"
+@endif gcc-bootstrap
+
 # Flags to pass down to makes which are built with the target environment.
 # The double $ decreases the length of the command line; those variables
 # are set in BASE_FLAGS_TO_PASS, and the sub-make will expand them.  The
@@ -3597,9 +3605,9 @@ check-bfd:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS) $(EXTRA_HOST_EXPORTS) \
 	(cd $(HOST_SUBDIR)/bfd && \
-	  $(MAKE) $(FLAGS_TO_PASS)  check)
+	  $(MAKE) $(FLAGS_TO_PASS)  $(EXTRA_BOOTSTRAP_FLAGS) check)
 
 @endif bfd
 
@@ -4471,9 +4479,9 @@ check-opcodes:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS) $(EXTRA_HOST_EXPORTS) \
 	(cd $(HOST_SUBDIR)/opcodes && \
-	  $(MAKE) $(FLAGS_TO_PASS)  check)
+	  $(MAKE) $(FLAGS_TO_PASS)  $(EXTRA_BOOTSTRAP_FLAGS) check)
 
 @endif opcodes
 
@@ -5345,9 +5353,9 @@ check-binutils:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS) $(EXTRA_HOST_EXPORTS) \
 	(cd $(HOST_SUBDIR)/binutils && \
-	  $(MAKE) $(FLAGS_TO_PASS)  check)
+	  $(MAKE) $(FLAGS_TO_PASS)  $(EXTRA_BOOTSTRAP_FLAGS) check)
 
 @endif binutils
 
@@ -5775,9 +5783,9 @@ check-bison:
 	@if [ '$(host)' = '$(target)' ] ; then \
 	  r=`${PWD_COMMAND}`; export r; \
 	  s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	  $(HOST_EXPORTS) \
+	  $(HOST_EXPORTS)  \
 	  (cd $(HOST_SUBDIR)/bison && \
-	$(MAKE) $(FLAGS_TO_PASS)  check); \
+	$(MAKE) $(FLAGS_TO_PASS)  check)
 	fi
 
 @endif bison
@@ -6217,7 +6225,7 @@ check-cgen:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS)  \
 	(cd $(HOST_SUBDIR)/cgen && \
 	  $(MAKE) $(FLAGS_TO_PASS)  check)
 
@@ -6658,7 +,7 @@ check-dejagnu:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS)  \
 	(cd $(HOST_SUBDIR)/dejagnu && \
 	  $(MAKE) $(FLAGS_TO_PASS)  check)
 
@@ -7099,7 +7107,7 @@ check-etc:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS)  \
 	(cd $(HOST_SUBDIR)/etc && \
 	  $(MAKE) $(FLAGS_TO_PASS)  check)
 
@@ -7542,9 +7550,9 @@ check-fastjar:
 	@if [ '$(host)' = '$(target)' ] ; then \
 	  r=`${PWD_COMMAND}`; export r; \
 	  s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	  $(HOST_EXPORTS) \
+	  $(HOST_EXPORTS)  \
 	  (cd $(HOST_SUBDIR)/fastjar && \
-	$(MAKE) $(FLAGS_TO_PASS)  check); \
+	$(MAKE) $(FLAGS_TO_PASS)  check)
 	fi
 
 @endif fastjar
@@ -8430,9 +8438,9 @@ check-fixincludes:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPORTS) $(EXTRA_HOST_EXPORTS) \
 	(cd $(HOST_SUBDIR)/fixincludes && \
-	  $(MAKE) $(FLAGS_TO_PASS)  check)
+	  $(MAKE) $(FLAGS_TO_PASS)  $(EXTRA_BOOTSTRAP_FLAGS) check)
 
 @endif fixincludes
 
@@ -8845,9 +8853,9 @@ check-flex:
 	@if [ '$(host)' = '$(target)' ] ; then \
 	  r=`${PWD_COMMAND}`; export r; \
 	  s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	  $(HOST_EXPORTS) \
+	  $(HOST_EXPORTS)  \
 	  (cd $(HOST_SUBDIR)/flex && \
-	$(MAKE) $(FLAGS_TO_PASS)  check); \
+	$(MAKE) $(FLAGS_TO_PASS)  check)
 	fi
 
 @endif flex
@@ -9733,9 +9741,9 @@ check-gas:
 	@: $(MAKE); $(unstage)
 	@r=`${PWD_COMMAND}`; export r; \
 	s=`cd $(srcdir); ${PWD_COMMAND}`; export s; \
-	$(HOST_EXPORTS) \
+	$(HOST_EXPO

[gomp4] Testing of C/C++ OpenACC cache directive (was: OpenACC cache directive for C)

[Thomas Schwinge]

Hi!

On Wed, 05 Nov 2014 17:29:19 +0100, I wrote:
> In r217145, I applied Jim's patch to gomp-4_0-branch:

>   * testsuite/libgomp.oacc-c/cache-1.c: New file.
>   * testsuite/libgomp.oacc-c++/cache-1.C: Likewise.

Applied to gomp-4_0-branch in r217147:

commit 267cffc4105255a8372fb5788fdcbb54560f493b
Author: tschwinge 
Date:   Wed Nov 5 16:46:23 2014 +

Testing of C/C++ OpenACC cache directive.

gcc/testsuite/
* c-c++-common/goacc/cache-1.c: New file.
libgomp/
* testsuite/libgomp.oacc-c/cache-1.c: Remove directives that are
expected to fail, and rename the file to...
* testsuite/libgomp.oacc-c-c++-common/cache-1.c: ... this.
* testsuite/libgomp.oacc-c++/cache-1.C: Remove file.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217147 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/testsuite/ChangeLog.gomp   |  2 +
 gcc/testsuite/c-c++-common/goacc/cache-1.c | 88 ++
 libgomp/ChangeLog.gomp |  7 ++
 libgomp/testsuite/libgomp.oacc-c++/cache-1.C   | 50 
 .../testsuite/libgomp.oacc-c-c++-common/cache-1.c  | 48 
 libgomp/testsuite/libgomp.oacc-c/cache-1.c | 69 -
 6 files changed, 145 insertions(+), 119 deletions(-)

diff --git gcc/testsuite/ChangeLog.gomp gcc/testsuite/ChangeLog.gomp
index 25be821..1faf0fa 100644
--- gcc/testsuite/ChangeLog.gomp
+++ gcc/testsuite/ChangeLog.gomp
@@ -1,5 +1,7 @@
 2014-11-05  Thomas Schwinge  
 
+   * c-c++-common/goacc/cache-1.c: New file.
+
* gfortran.dg/goacc/data-tree.f95: Remove dg-prune-output directive.
* gfortran.dg/goacc/kernels-tree.f95: Likewise.
* gfortran.dg/goacc/loop-tree-1.f90: Likewise.
diff --git gcc/testsuite/c-c++-common/goacc/cache-1.c 
gcc/testsuite/c-c++-common/goacc/cache-1.c
new file mode 100644
index 000..9503341
--- /dev/null
+++ gcc/testsuite/c-c++-common/goacc/cache-1.c
@@ -0,0 +1,88 @@
+int
+main (int argc, char **argv)
+{
+#define N   2
+int a[N], b[N];
+int i;
+
+for (i = 0; i < N; i++)
+{
+a[i] = 3;
+b[i] = 0;
+}
+
+#pragma acc parallel copyin (a[0:N]) copyout (b[0:N])
+{
+int ii;
+
+for (ii = 0; ii < N; ii++)
+{
+const int idx = ii;
+int n = 1;
+const int len = n;
+
+#pragma acc cache /* { dg-error "expected '\\\(' before end of line" } */
+
+#pragma acc cache a[0:N] /* { dg-error "expected '\\\(' before 'a'" } */
+   /* { dg-bogus "expected end of line before 'a'" "" { xfail c++ } 26 } */
+
+#pragma acc cache (a) /* { dg-error "expected '\\\['" } */
+
+#pragma acc cache ( /* { dg-error "expected (identifier|unqualified-id) before 
end of line" } */
+
+#pragma acc cache () /* { dg-error "expected (identifier|unqualified-id) 
before '\\\)' token" } */
+
+#pragma acc cache (,) /* { dg-error "expected (identifier|unqualified-id) 
before '(,|\\\))' token" } */
+
+#pragma acc cache (a[0:N] /* { dg-error "expected '\\\)' before end of line" } 
*/
+
+#pragma acc cache (a[0:N],) /* { dg-error "expected 
(identifier|unqualified-id) before '(,|\\\))' token" "" { xfail c } } */
+
+#pragma acc cache (a[0:N]) copyin (a[0:N]) /* { dg-error "expected end of line 
before 'copyin'" } */
+
+#pragma acc cache () /* { dg-error "expected (identifier|unqualified-id) 
before '\\\)' token" } */
+
+#pragma acc cache (a[0:N] b[0:N]) /* { dg-error "expected '\\\)' before 'b'" } 
*/
+
+#pragma acc cache (a[0:N] b[0:N}) /* { dg-error "expected '\\\)' before 'b'" } 
*/
+   /* { dg-bogus "expected end of line before '\\\}' token" "" { xfail c++ 
} 47 } */
+
+#pragma acc cache (a[0:N] /* { dg-error "expected '\\\)' before end of line" } 
*/
+
+#pragma acc cache (a[ii]) /* { dg-error "'ii' is not a constant" } */
+
+#pragma acc cache (a[idx:n]) /* { dg-error "'n' is not a constant" } */
+
+#pragma acc cache (a[0:N]) ( /* { dg-error "expected end of line before '\\(' 
token" } */
+
+#pragma acc cache (a[0:N]) ii /* { dg-error "expected end of line before 'ii'" 
} */
+
+#pragma acc cache (a[0:N] ii) /* { dg-error "expected '\\)' before 'ii'" } */
+
+#pragma acc cache (a[0:N])
+
+#pragma acc cache (a[0:N], a[0:N])
+
+#pragma acc cache (a[0:N], b[0:N])
+
+#pragma acc cache (a[0])
+
+#pragma acc cache (a[0], a[1], b[0:N])
+
+#pragma acc cache (a[idx])
+
+#pragma acc cache (a[idx:len])
+
+b[ii] = a[ii];
+}
+}
+
+
+for (i = 0; i < N; i++)
+{
+if (a[i] != b[i])
+__builtin_abort ();
+}
+
+return 0;
+}
diff --git libgomp/ChangeLog.gomp libgomp/ChangeLog.gomp
index 4ac348d..096a2a9 100644
--- libgomp/ChangeLog.gomp
+++ libgomp/ChangeLog.gomp
@@ -1,3 +1,10 @@
+2014-11-05  Thomas Schwinge  
+
+   * testsuite/libgomp.oacc-c/cache-1.c: Remove directives that are
+   expected to fail, and rename the file to...
+   * testsuite/libgomp.oacc-c-c++-common/cache-1.c: ... this.
+   * testsuite/libgomp.oacc

[gomp4] OpenACC update host/self maintenance (was: acc enter/exit data)

[Thomas Schwinge]

Hi!

On Thu, 30 Oct 2014 17:11:04 -0700, Cesar Philippidis  
wrote:
>   gcc/fortran/
>   * gfortran.h (enum OMP_LIST_HOST): Remove.
>   (enum OMP_LIST_DEVICE, OMP_LIST_DEVICE): Remove.
>   * dump-parse-tree.c (show_omp_clauses): Remove OMP_LIST_HOST and
>   OMP_LIST_DEVICE from here also.
>   * openmp.c (OMP_CLAUSE_SELF): New define.
>   (gfc_match_omp_clauses): Update handling of OMP_CLAUSE_HOST and
>   OMP_CLAUSE_DEVICE. Add support for OMP_CLAUSE_SELF.
>   * trans-openmp.c (gfc_trans_omp_clauses): Remove support for
>   OMP_LIST_HOST and OMP_LIST_DEVICE since they are treated as memory
>   maps now.
>   (gfc_trans_oacc_executable_directive): Remove stale EXEC_OACC_WAIT.

Applied to gomp-4_0-branch in r217148:

commit a7bba5ecc7c62a022616f55ff1d8fb48266fcb67
Author: tschwinge 
Date:   Wed Nov 5 16:54:07 2014 +

OpenACC update host/self maintenance.

gcc/c/
* c-parser.c (c_parser_omp_clause_name) <"host">: Return
PRAGMA_OMP_CLAUSE_HOST.

gcc/c/
(c_parser_oacc_data_clause): Group PRAGMA_OMP_CLAUSE_SELF next to
PRAGMA_OMP_CLAUSE_HOST.
gcc/cp/
* parser.c (cp_parser_oacc_data_clause): Group
PRAGMA_OMP_CLAUSE_SELF next to PRAGMA_OMP_CLAUSE_HOST.

gcc/fortran/
* openmp.c (OMP_CLAUSE_HOST, OMP_CLAUSE_SELF): Merge into the new
OMP_CLAUSE_HOST_SELF.  Update all users.

gcc/
* tree-core.h (enum omp_clause_code): Remove OMP_CLAUSE_HOST and
OMP_CLAUSE_OACC_DEVICE.  Update all users.

gcc/testsuite/
* c-c++-common/goacc/update-1.c: Extend.
* gfortran.dg/goacc/assumed.f95: Likewise.
* gfortran.dg/goacc/coarray.f95: Likewise.
* gfortran.dg/goacc/cray.f95: Likewise.
* gfortran.dg/goacc/literal.f95: Likewise.
* gfortran.dg/goacc/parameter.f95: Likewise.
libgomp/
* testsuite/libgomp.oacc-c-c++-common/update-1-2.c: New file.
* testsuite/libgomp.oacc-fortran/data-4-2.f90: Likewise.
* testsuite/libgomp.oacc-fortran/data-4.f90: In one instance, use
the self clause instead of host clause.

git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@217148 
138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog.gomp |   3 +
 gcc/c/ChangeLog.gomp   |   6 +
 gcc/c/c-parser.c   |  10 +-
 gcc/cp/ChangeLog.gomp  |   3 +
 gcc/cp/parser.c|   8 +-
 gcc/fortran/ChangeLog.gomp |   3 +
 gcc/fortran/openmp.c   |  21 +-
 gcc/gimplify.c |   4 -
 gcc/omp-low.c  |   4 -
 gcc/testsuite/ChangeLog.gomp   |   7 +
 gcc/testsuite/c-c++-common/goacc/update-1.c|   5 +
 gcc/testsuite/gfortran.dg/goacc/assumed.f95|   8 +-
 gcc/testsuite/gfortran.dg/goacc/coarray.f95|   3 +-
 gcc/testsuite/gfortran.dg/goacc/cray.f95   |   6 +-
 gcc/testsuite/gfortran.dg/goacc/literal.f95|   5 +-
 gcc/testsuite/gfortran.dg/goacc/parameter.f95  |   3 +-
 gcc/tree-core.h|  10 +-
 gcc/tree-pretty-print.c|   6 -
 gcc/tree.c |   6 -
 libgomp/ChangeLog.gomp |   5 +
 .../libgomp.oacc-c-c++-common/update-1-2.c | 282 +
 .../{data-4.f90 => data-4-2.f90}   |   8 +-
 libgomp/testsuite/libgomp.oacc-fortran/data-4.f90  |   2 +-
 23 files changed, 353 insertions(+), 65 deletions(-)

diff --git gcc/ChangeLog.gomp gcc/ChangeLog.gomp
index 2c2b349..d140a35 100644
--- gcc/ChangeLog.gomp
+++ gcc/ChangeLog.gomp
@@ -1,5 +1,8 @@
 2014-11-05  Thomas Schwinge  
 
+   * tree-core.h (enum omp_clause_code): Remove OMP_CLAUSE_HOST and
+   OMP_CLAUSE_OACC_DEVICE.  Update all users.
+
* gimplify.c (gimplify_oacc_cache): New function.
(gimplify_expr): Use it for OACC_CACHE.
(gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses): Handle
diff --git gcc/c/ChangeLog.gomp gcc/c/ChangeLog.gomp
index 70278b9..a223a17 100644
--- gcc/c/ChangeLog.gomp
+++ gcc/c/ChangeLog.gomp
@@ -1,5 +1,11 @@
 2014-11-05  Thomas Schwinge  
 
+   * c-parser.c (c_parser_omp_clause_name) <"host">: Return
+   PRAGMA_OMP_CLAUSE_HOST.
+
+   * c-parser.c (c_parser_oacc_data_clause): Group
+   PRAGMA_OMP_CLAUSE_SELF next to PRAGMA_OMP_CLAUSE_HOST.
+
* c-parser.c (c_parser_oacc_cache): Generate OACC_CACHE.
* c-typeck.c (c_finish_omp_clauses): Handle OMP_CLAUSE__CACHE_.
 
diff --git gcc/c/c-parser.c gcc/c/c-parser.c
index 40d4314..bd2864f 100644
--- gcc/c/c-parser.c
+++ gcc/c/c-parser.c
@@ -9832,7 +9832,7 @@ c_parser_omp_clause_name (c_parser *parser)
  br

RE: [PATCHv3][MIPS] Implement O32 ABI extensions (GCC)

[Matthew Fortune]

Hi Catherine,

The full patch is attached and the delta from v2 is inline below.

Testing (O32):

MIPS I - FP32, MIPS II - FP32, MIPS II - FPXX
MIPS32 - FP32, MIPS32 - FPXX, MIPS32 - FPXX ODDSPREG,
MIPS32R2 - FP32, MIPS32R2 - FPXX, MIPS32R2 - FPXX ODDSPREG,
MIPS32R2 - FP64, MIPS32R2 - FP64A

One of the new tests fails in the two ODDSPREG configurations but I don't
plan to resolve that as it makes very little sense to pre-configure
GCC to explicitly use -modd-spreg (it only makes sense to use -mno-odd-spreg).
The failure is merely that the expected behaviour does not match up because
of the implicit -modd-spreg option. This same issue affects one of the
pre-existing loongson tests in the same configurations but I am leaving that
for the same reason.

The FP64 configurations enable extra tests which failed before for FP64 and
still fail after so will be cleaned up in later work. In particular there
are some missed vectorization cases which are to do with loongson.

Testing (N64):

The testsuite is still running after a rebase but it had no regressions last
time.

Changes (on top of addressing all comments):

* Handle the fact that -msingle-float, -msoft-float, -mfp32, -mfpxx, -mfp64
  are all part of the FPU ABI selection. As such if one of these is
  given which conflicts with an implicit option then the implicit option is
  not added.  This is necessary to provide a stable and consistent behaviour
  when writing assembly code such that (for example) building a file with
  -msoft-float and using .set hardfloat gets you into the standard hard-float
  ABI which is FP32. This issue was identified as part of resolving kernel
  issues when built with FPXX/.MIPS.abiflags aware tools:
  http://www.linux-mips.org/archives/linux-mips/2014-10/msg00886.html
  There are lots of other strange specs issues to resolve for MIPS but this
  is a step in the right direction.
* Fix a build warning
* Fix a testcase for FP64 configs

OK to commit (pending successful results for the N64 testrun)?

Thanks,
Matthew

---
 gcc/config.in  |  2 +-
 gcc/config/mips/mips.c | 49 +++-
 gcc/config/mips/mips.h | 51 +-
 gcc/config/mips/mips.md|  8 ++---
 gcc/config/mips/mips.opt   |  2 +-
 gcc/config/mips/mti-elf.h  |  6 ++--
 gcc/config/mips/mti-linux.h|  6 ++--
 gcc/configure  | 20 ++--
 gcc/configure.ac   | 12 +++
 gcc/doc/invoke.texi| 17 +-
 gcc/testsuite/gcc.target/mips/oddspreg-3.c |  2 +-
 11 files changed, 93 insertions(+), 82 deletions(-)

diff --git a/gcc/config.in b/gcc/config.in
index 40dd6f5..8fd0e0e 100644
--- a/gcc/config.in
+++ b/gcc/config.in
@@ -474,7 +474,7 @@
 
 /* Define if the assembler understands .module. */
 #ifndef USED_FOR_TARGET
-#undef HAVE_AS_MODULE
+#undef HAVE_AS_DOT_MODULE
 #endif
 
 
diff --git a/gcc/config/mips/mips.c b/gcc/config/mips/mips.c
index 2dc6725..46f3890 100644
--- a/gcc/config/mips/mips.c
+++ b/gcc/config/mips/mips.c
@@ -5466,7 +5466,7 @@ mips_function_arg_boundary (machine_mode mode, const_tree 
type)
 static machine_mode
 mips_get_reg_raw_mode (int regno)
 {
-  if (mips_abi == ABI_32 && !TARGET_FLOAT32 && FP_REG_P (regno))
+  if (TARGET_FLOATXX && FP_REG_P (regno))
 return DFmode;
   return default_get_reg_raw_mode (regno);
 }
@@ -5796,19 +5796,24 @@ mips_libcall_value (machine_mode mode, const_rtx fun 
ATTRIBUTE_UNUSED)
 static bool
 mips_function_value_regno_p (const unsigned int regno)
 {
+  /* Most types only require one GPR or one FPR for return values but for
+ hard-float two FPRs can be used for _Complex types (for all ABIs)
+ and long doubles (for n64).  */
   if (regno == GP_RETURN
   || regno == FP_RETURN
-  || regno == FP_RETURN + 2
-  || (LONG_DOUBLE_TYPE_SIZE == 128
- && FP_RETURN != GP_RETURN
+  || (FP_RETURN != GP_RETURN
  && regno == FP_RETURN + 2))
 return true;
 
-  if ((regno == FP_RETURN + 1
-  || regno == FP_RETURN + 3)
+  /* For o32 FP32, _Complex double will be returned in four 32-bit registers.
+ This does not apply to o32 FPXX as floating-point function argument and
+ return registers are described as 64-bit even though floating-point
+ registers are primarily described as 32-bit internally.
+ See: mips_get_reg_raw_mode.  */
+  if ((mips_abi == ABI_32 && TARGET_FLOAT32)
   && FP_RETURN != GP_RETURN
-  && (mips_abi == ABI_32 && TARGET_FLOAT32)
-  && FP_REG_P (regno))
+  && (regno == FP_RETURN + 1
+ || regno == FP_RETURN + 3))
 return true;
 
   return false;
@@ -8723,14 +8728,13 @@ mips_dwarf_register_span (rtx reg)
   machine_mode mode;
 
   /* TARGET_FLOATXX is implemented as 32-bit floating-point registers but
- ensures that double precision registers are treated as if they were
+ ensure

Re: [PATCH] PR36312

[Manuel López-Ibáñez]

On 4 November 2014 23:40, Jeff Law  wrote:
> On 10/25/14 04:20, Anthony Brandon wrote:
>>
>> Hi,
>>
>> Sorry for the delay. Here are the updated diff and changelog.
>>
>> gcc/testsuite/ChangeLog:
>>
>> 2014-10-25  Anthony Brandon  
>>
>>  PR driver/36312
>>  * gcc.misc-tests/output.exp: New test case for identical input and
>> output files.
>>
>> include/ChangeLog:
>>
>> 2014-10-25  Anthony Brandon  
>>
>>  PR driver/36312
>>  * filenames.h: Add prototype for canonical_filename_eq.
>>
>> gcc/ChangeLog:
>>
>> 2014-10-25  Anthony Brandon  
>>
>>  PR driver/36312
>>  * diagnostic-core.h: Add prototype for fatal_error.
>>  * diagnostic.c (fatal_error): New function fatal_error.
>>  * gcc.c (store_arg): Remove have_o_argbuf_index.
>>  (process_command): Check if input and output files are the same.
>>  * toplev.c (init_asm_output): Check if input and output files are the
>> same.
>>
>> libiberty/ChangeLog:
>>
>> 2014-10-25  Anthony Brandon  
>>
>>  PR driver/36312
>>  * filename_cmp.c (canonical_filename_eq): New function to check if
>> file names are the same.
>>  * functions.texi: Updated with documentation for new function.
>
> This is fine for the trunk.  Please install.
>
> Thanks,
> Jeff
>

I committed this as r217149.

Anthony, Thanks!

Cheers,

Manuel.

Re: libstdc++ new deque failures

[Jonathan Wakely]

On 5 November 2014 14:14, David Edelsohn wrote:
> Jonathan,
>
> I still am seeing new failures in the libstdc++ deque testsuite as of
> last night.  I don't know if you still are working through the fallout
> from the earlier patches, but I wanted to make you aware.

Yes, those tests are meant to fail but I need to adjust the dg-error
line numbers after one of my earlier patches.

I'm working on a patch (I might make other changes to std::deque,
which would require changing the dg-error line numbers yet agan, so
I'm holding off until the other changes are ready ... or I decide not
to make them and just fix the tests.)

Sorry for the noise in the testresults.

>
>
> And these are not related to deque, but appear to be additional issues
> in the libstdc++ implementation:

I hadn't seen these ones, I'll take a look, thanks.

[gomp4] Use GOMP_OFFLOAD_ prefix for (OpenACC) plugin hooks

[Julian Brown]

Hi,

Mirroring changes in Ilya Verbin's libgomp offloading pieces posted to
trunk, this patch adds a prefix of GOMP_OFFLOAD_ to the OpenACC plugin
hooks. Some of these bits will not be needed for a trunk version of the
patch once Ilya's patch is approved (I'm hoping other
incompatibilities haven't crept in other than the renaming!).

I will apply to the gomp4 branch shortly.

Thanks,

Julian

ChangeLog

libgomp/
* oacc-host.c: Add GOMP_OFFLOAD_ prefix for plugin hooks. Rename
device_init to init_device, device_fini to fini_device,
offload_register to register_image and remove extraneous "device_"
from device_alloc, device_free, device_dev2host, device_host2dev and
device_run.
(host_dispatch): Use new names for hooks.
* oacc-init.c: Use new names for hooks, throughout.
* plugin-nvptx.c: Likewise.
* target.c: Likewise.
(gomp_load_plugin_for_device): Likewise. Look for new hook names.
* target.h (gomp_device_descr): Use new hook names.
commit 4e1b71a5e0d15de4c6e89ab5139964e32b563d68
Author: Julian Brown 
Date:   Wed Nov 5 02:34:22 2014 -0800

Use GOMP_OFFLOAD_ prefix for plugin hooks.

diff --git a/libgomp/oacc-host.c b/libgomp/oacc-host.c
index fc3e77c..02794bb 100644
--- a/libgomp/oacc-host.c
+++ b/libgomp/oacc-host.c
@@ -60,7 +60,7 @@ static struct gomp_device_descr host_dispatch;
 #endif
 
 STATIC const char *
-get_name (void)
+GOMP_OFFLOAD_get_name (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -74,7 +74,7 @@ get_name (void)
 }
 
 STATIC int
-get_type (void)
+GOMP_OFFLOAD_get_type (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -88,7 +88,7 @@ get_type (void)
 }
 
 STATIC unsigned int
-get_caps (void)
+GOMP_OFFLOAD_get_caps (void)
 {
   unsigned int caps = TARGET_CAP_OPENACC_200 | TARGET_CAP_OPENMP_400
 		  | TARGET_CAP_NATIVE_EXEC;
@@ -105,7 +105,7 @@ get_caps (void)
 }
 
 STATIC int
-get_num_devices (void)
+GOMP_OFFLOAD_get_num_devices (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -115,7 +115,7 @@ get_num_devices (void)
 }
 
 STATIC void
-offload_register (void *host_table, void *target_data)
+GOMP_OFFLOAD_register_image (void *host_table, void *target_data)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%p, %p)\n", __FILE__, __FUNCTION__, host_table,
@@ -124,17 +124,17 @@ offload_register (void *host_table, void *target_data)
 }
 
 STATIC int
-device_init (void)
+GOMP_OFFLOAD_init_device (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
 #endif
 
-  return get_num_devices ();
+  return GOMP_OFFLOAD_get_num_devices ();
 }
 
 STATIC int
-device_fini (void)
+GOMP_OFFLOAD_fini_device (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -144,7 +144,7 @@ device_fini (void)
 }
 
 STATIC int
-device_get_table (struct mapping_table **table)
+GOMP_OFFLOAD_get_table (struct mapping_table **table)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%p)\n", __FILE__, __FUNCTION__, table);
@@ -154,7 +154,7 @@ device_get_table (struct mapping_table **table)
 }
 
 STATIC bool
-openacc_avail (void)
+GOMP_OFFLOAD_openacc_avail (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -164,7 +164,7 @@ openacc_avail (void)
 }
 
 STATIC void *
-openacc_open_device (int n)
+GOMP_OFFLOAD_openacc_open_device (int n)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%u)\n", __FILE__, __FUNCTION__, n);
@@ -174,7 +174,7 @@ openacc_open_device (int n)
 }
 
 STATIC int
-openacc_close_device (void *hnd)
+GOMP_OFFLOAD_openacc_close_device (void *hnd)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%p)\n", __FILE__, __FUNCTION__, hnd);
@@ -184,7 +184,7 @@ openacc_close_device (void *hnd)
 }
 
 STATIC int
-openacc_get_device_num (void)
+GOMP_OFFLOAD_openacc_get_device_num (void)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s\n", __FILE__, __FUNCTION__);
@@ -194,7 +194,7 @@ openacc_get_device_num (void)
 }
 
 STATIC void
-openacc_set_device_num (int n)
+GOMP_OFFLOAD_openacc_set_device_num (int n)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%u)\n", __FILE__, __FUNCTION__, n);
@@ -205,7 +205,7 @@ openacc_set_device_num (int n)
 }
 
 STATIC void *
-device_alloc (size_t s)
+GOMP_OFFLOAD_alloc (size_t s)
 {
   void *ptr = GOMP(malloc) (s);
 
@@ -217,7 +217,7 @@ device_alloc (size_t s)
 }
 
 STATIC void
-device_free (void *p)
+GOMP_OFFLOAD_free (void *p)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%p)\n", __FILE__, __FUNCTION__, p);
@@ -227,7 +227,7 @@ device_free (void *p)
 }
 
 STATIC void *
-device_host2dev (void *d, const void *h, size_t s)
+GOMP_OFFLOAD_host2dev (void *d, const void *h, size_t s)
 {
 #ifdef DEBUG
   fprintf (stderr, SELF "%s:%s (%p, %p, %zd)\n", __FILE__, __FUNCTION__, d, h,
@@ -242,7 +242,7 @@ device_host2dev (void *d, const void *h, size_t s)
 }
 
 STATIC void *
-device_dev2host (void *h, const void *d, size_t s)
+GOMP_OFFLOAD_

[gomp4] Move libgomp plugins into subdirectory

[Julian Brown]

Hi,

This patch moves plugin-nvptx.c and plugin-host.c (from oacc-host.c)
into a new "plugin" subdirectory, as requested by Jakub, and to match
more closely the layout of the Intel MIC pieces. This also moves the
autotools bits to enable the NVPTX plugin and locate CUDA libraries
into the plugin directory's (new) configury bits.

So far this only changes the location of the source files: the plugins
themselves are still installed to the same place as before (alongside
libgomp itself).

Test results look reasonable with my (patched for PTX support) version
of the gomp4 branch. I'll apply it there shortly.

Thanks,

Julian

ChangeLog

libgomp/
* Makefile.am (SUBDIRS): Add plugin.
(DIST_SUBDIRS): Define.
(libgomp_plugin_nvptx_*): Remove nvptx support from here.
(libgomp_plugin_host_nonshm_*): Likewise.
* Makefile.in: Regenerate.
* configure: Regenerate.
* oacc-host.c: Replace with #include of plugin/plugin-host.c code,
move implementation to the latter.
* plugin/plugin-host.c: New file.
* plugin-nvptx.c: Move to...
* plugin/plugin-nvptx.c: New file.
* plugin/Makefile.am: New.
* plugin/Makefile.in: Regenerate.
* plugin/aclocal.m4: Regenerate.
* plugin/configure: Regenerate.
commit 8994fb8c1b9d52cb9c82a61227a450df29e61806
Author: Julian Brown 
Date:   Wed Nov 5 02:54:30 2014 -0800

Move libgomp plugins into their own directory.

diff --git a/libgomp/Makefile.am b/libgomp/Makefile.am
index e0ab763..f265c5d 100644
--- a/libgomp/Makefile.am
+++ b/libgomp/Makefile.am
@@ -1,7 +1,8 @@
 ## Process this file with automake to produce Makefile.in
 
 ACLOCAL_AMFLAGS = -I .. -I ../config
-SUBDIRS = testsuite
+SUBDIRS = testsuite plugin
+DIST_SUBDIRS = plugin
 
 ## May be used by toolexeclibdir.
 gcc_version := $(shell cat $(top_srcdir)/../gcc/BASE-VER)
@@ -21,27 +22,6 @@ AM_LDFLAGS = $(XLDFLAGS) $(SECTION_LDFLAGS) $(OPT_LDFLAGS)
 toolexeclib_LTLIBRARIES = libgomp.la
 nodist_toolexeclib_HEADERS = libgomp.spec
 
-if PLUGIN_NVPTX
-# Nvidia PTX OpenACC plugin.
-libgomp_plugin_nvptx_version_info = -version-info $(libtool_VERSION)
-toolexeclib_LTLIBRARIES += libgomp-plugin-nvptx.la
-libgomp_plugin_nvptx_la_SOURCES = plugin-nvptx.c
-libgomp_plugin_nvptx_la_CPPFLAGS = $(AM_CPPFLAGS) $(PLUGIN_NVPTX_CPPFLAGS)
-libgomp_plugin_nvptx_la_LDFLAGS = $(libgomp_plugin_nvptx_version_info) \
-	$(lt_host_flags)
-libgomp_plugin_nvptx_la_LDFLAGS += $(PLUGIN_NVPTX_LDFLAGS)
-libgomp_plugin_nvptx_la_LIBADD = $(PLUGIN_NVPTX_LIBS)
-libgomp_plugin_nvptx_la_LIBTOOLFLAGS = --tag=disable-static
-endif
-
-libgomp_plugin_host_nonshm_version_info = -version-info $(libtool_VERSION)
-toolexeclib_LTLIBRARIES += libgomp-plugin-host_nonshm.la
-libgomp_plugin_host_nonshm_la_SOURCES = oacc-host.c
-libgomp_plugin_host_nonshm_la_CPPFLAGS = $(AM_CPPFLAGS) -DHOST_NONSHM_PLUGIN
-libgomp_plugin_host_nonshm_la_LDFLAGS = \
-	$(libgomp_plugin_host_nonshm_version_info) $(lt_host_flags)
-libgomp_plugin_host_nonshm_la_LIBTOOLFLAGS = --tag=disable-static
-
 if LIBGOMP_BUILD_VERSIONED_SHLIB
 # -Wc is only a libtool option.
 comma = ,
diff --git a/libgomp/Makefile.in b/libgomp/Makefile.in
index d12376e..ea3e1ca 100644
diff --git a/libgomp/configure b/libgomp/configure
index 7daccd9..11a7ae0 100755
diff --git a/libgomp/configure.ac b/libgomp/configure.ac
index 89c6b31..e883945 100644
--- a/libgomp/configure.ac
+++ b/libgomp/configure.ac
@@ -30,42 +30,6 @@ LIBGOMP_ENABLE(generated-files-in-srcdir, no, ,
 AC_MSG_RESULT($enable_generated_files_in_srcdir)
 AM_CONDITIONAL(GENINSRC, test "$enable_generated_files_in_srcdir" = yes)
 
-# Look for the CUDA driver package.
-CUDA_DRIVER_INCLUDE=
-CUDA_DRIVER_LIB=
-AC_SUBST(CUDA_DRIVER_INCLUDE)
-AC_SUBST(CUDA_DRIVER_LIB)
-CUDA_DRIVER_CPPFLAGS=
-CUDA_DRIVER_LDFLAGS=
-AC_ARG_WITH(cuda-driver,
-	[AS_HELP_STRING([--with-cuda-driver=PATH],
-		[specify prefix directory for installed CUDA driver package.
-		 Equivalent to --with-cuda-driver-include=PATH/include
-		 plus --with-cuda-driver-lib=PATH/lib])])
-AC_ARG_WITH(cuda-driver-include,
-	[AS_HELP_STRING([--with-cuda-driver-include=PATH],
-		[specify directory for installed CUDA driver include files])])
-AC_ARG_WITH(cuda-driver-lib,
-	[AS_HELP_STRING([--with-cuda-driver-lib=PATH],
-		[specify directory for the installed CUDA driver library])])
-if test "x$with_cuda_driver" != x; then
-  CUDA_DRIVER_INCLUDE=$with_cuda_driver/include
-  CUDA_DRIVER_LIB=$with_cuda_driver/lib
-fi
-if test "x$with_cuda_driver_include" != x; then
-  CUDA_DRIVER_INCLUDE=$with_cuda_driver_include
-fi
-if test "x$with_cuda_driver_lib" != x; then
-  CUDA_DRIVER_LIB=$with_cuda_driver_lib
-fi
-if test "x$CUDA_DRIVER_INCLUDE" != x; then
-  CUDA_DRIVER_CPPFLAGS=-I$CUDA_DRIVER_INCLUDE
-fi
-if test "x$CUDA_DRIVER_LIB" != x; then
-  CUDA_DRIVER_LDFLAGS=-L$CUDA_DRIVER_LIB
-fi
-
-
 # ---
 # ---
 
@@ -241,52 +205,7 @@ elif test "x$enable_accelerator" != xno; then
   AC_MSG_ERROR([Can't have support for accelerators without support for plugins])
 fi
 
-PLUGIN_

[jit] Verify enum values earlier

[David Malcolm]

It wasn't clear to me that all of these enum values were being fully
validated by the internals, and it's better to fail early (so we can
report which function was at fault), so explicitly validate enum values
at the API entrypoints.

The new testcases bring the # of expected passes in jit.sum from
4663 to 4711.

Committed to the dmalcolm/jit branch.

gcc/jit/ChangeLog.jit:
* libgccjit.c (gcc_jit_context_get_type): Verify that "type"
is valid immediately, rather than relying on called code.
(gcc_jit_context_new_function): Likewise for "kind".
(gcc_jit_context_new_unary_op): Likewise for "op".
(valid_binary_op_p): New.
(gcc_jit_context_new_binary_op): Verify that "op" is valid
immediately, rather than relying on called code.
(gcc_jit_context_new_comparison): Likewise.
(gcc_jit_block_add_assignment_op): Likewise.

gcc/testsuite/ChangeLog.jit:
* jit.dg/test-error-get-type-bad-enum.c: New test case.
* jit.dg/test-error-new-binary-op-bad-op.c: Likewise.
* jit.dg/test-error-new-function-bad-kind.c: Likewise.
* jit.dg/test-error-new-unary-op-bad-op.c: Likewise.
---
 gcc/jit/ChangeLog.jit  | 12 ++
 gcc/jit/libgccjit.c| 49 +++---
 gcc/testsuite/ChangeLog.jit|  7 
 .../jit.dg/test-error-get-type-bad-enum.c  | 27 
 .../jit.dg/test-error-new-binary-op-bad-op.c   | 37 
 .../jit.dg/test-error-new-function-bad-kind.c  | 41 ++
 .../jit.dg/test-error-new-unary-op-bad-op.c| 36 
 7 files changed, 204 insertions(+), 5 deletions(-)
 create mode 100644 gcc/testsuite/jit.dg/test-error-get-type-bad-enum.c
 create mode 100644 gcc/testsuite/jit.dg/test-error-new-binary-op-bad-op.c
 create mode 100644 gcc/testsuite/jit.dg/test-error-new-function-bad-kind.c
 create mode 100644 gcc/testsuite/jit.dg/test-error-new-unary-op-bad-op.c

diff --git a/gcc/jit/ChangeLog.jit b/gcc/jit/ChangeLog.jit
index 3d6361c..ce927d6 100644
--- a/gcc/jit/ChangeLog.jit
+++ b/gcc/jit/ChangeLog.jit
@@ -1,5 +1,17 @@
 2014-11-05  David Malcolm  
 
+   * libgccjit.c (gcc_jit_context_get_type): Verify that "type"
+   is valid immediately, rather than relying on called code.
+   (gcc_jit_context_new_function): Likewise for "kind".
+   (gcc_jit_context_new_unary_op): Likewise for "op".
+   (valid_binary_op_p): New.
+   (gcc_jit_context_new_binary_op): Verify that "op" is valid
+   immediately, rather than relying on called code.
+   (gcc_jit_context_new_comparison): Likewise.
+   (gcc_jit_block_add_assignment_op): Likewise.
+
+2014-11-05  David Malcolm  
+
* libgccjit.c: Include safe-ctype.h from libiberty.
(IS_ASCII_ALPHA): Delete.
(IS_ASCII_DIGIT): Delete.
diff --git a/gcc/jit/libgccjit.c b/gcc/jit/libgccjit.c
index d9f259e..c109ba6 100644
--- a/gcc/jit/libgccjit.c
+++ b/gcc/jit/libgccjit.c
@@ -316,7 +316,11 @@ gcc_jit_context_get_type (gcc_jit_context *ctxt,
  enum gcc_jit_types type)
 {
   RETURN_NULL_IF_FAIL (ctxt, NULL, NULL, "NULL context");
-  /* The inner function checks "type" for us.  */
+  RETURN_NULL_IF_FAIL_PRINTF1 (
+(type >= GCC_JIT_TYPE_VOID
+ && type <= GCC_JIT_TYPE_FILE_PTR),
+ctxt, NULL,
+"unrecognized value for enum gcc_jit_types: %i", type);
 
   return (gcc_jit_type *)ctxt->get_type (type);
 }
@@ -574,6 +578,12 @@ gcc_jit_context_new_function (gcc_jit_context *ctxt,
  int is_variadic)
 {
   RETURN_NULL_IF_FAIL (ctxt, NULL, loc, "NULL context");
+  RETURN_NULL_IF_FAIL_PRINTF1 (
+((kind >= GCC_JIT_FUNCTION_EXPORTED)
+ && (kind <= GCC_JIT_FUNCTION_ALWAYS_INLINE)),
+ctxt, loc,
+"unrecognized value for enum gcc_jit_function_kind: %i",
+kind);
   RETURN_NULL_IF_FAIL (return_type, ctxt, loc, "NULL return_type");
   RETURN_NULL_IF_FAIL (name, ctxt, loc, "NULL name");
   /* The assembler can only handle certain names, so for now, enforce
@@ -835,13 +845,29 @@ gcc_jit_context_new_unary_op (gcc_jit_context *ctxt,
  gcc_jit_rvalue *rvalue)
 {
   RETURN_NULL_IF_FAIL (ctxt, NULL, loc, "NULL context");
-  /* op is checked by the inner function.  */
+  RETURN_NULL_IF_FAIL_PRINTF1 (
+(op >= GCC_JIT_UNARY_OP_MINUS
+ && op <= GCC_JIT_UNARY_OP_LOGICAL_NEGATE),
+ctxt, loc,
+"unrecognized value for enum gcc_jit_unary_op: %i",
+op);
   RETURN_NULL_IF_FAIL (result_type, ctxt, loc, "NULL result_type");
   RETURN_NULL_IF_FAIL (rvalue, ctxt, loc, "NULL rvalue");
 
   return (gcc_jit_rvalue *)ctxt->new_unary_op (loc, op, result_type, rvalue);
 }
 
+/* Determine if OP is a valid value for enum gcc_jit_binary_op.
+   For use by both gcc_jit_context_new_binary_op and
+   gcc_jit_block_add_assignment_op.  */
+
+static bool
+valid_binary_op_p (enum gcc_jit_binary_op op)
+{
+  return (op >= GC

Add uniform_inside_sphere_distribution

[Ed Smith-Rowland]

This distribution has come in handy for me.

It relies on uniform_on_sphere_distribution and like it, min and max 
have no real meaning.
Unlike uniform_on_sphere_distribution which really is a random 
multidimensional unit vector users often want to pick the radius of the 
distribution.
Unit radius is a good default and is provided but the user can specify 
radius.


Like the uniform_on_sphere_distribution which is used inside, the 
2-dimensional case uses rejection and higher dimensions use a transform.


Built and tested clean on x86_64-linux.

OK?

Ed

2014-09-05  Edward Smith-Rowland  <3dw...@verizon.net>

* include/ext/random: Add uniform_inside_sphere_distribution.
* include/ext/random.tcc: Out-of-line implementation of
uniform_inside_sphere_distribution.
* testsuite/ext/random/uniform_inside_sphere_distribution/cons/
default.cc: New.
* testsuite/ext/random/uniform_inside_sphere_distribution/cons/
parms.cc: New.
* testsuite/ext/random/uniform_inside_sphere_distribution/operators/
equal.cc: New.
* testsuite/ext/random/uniform_inside_sphere_distribution/operators/
generate.cc: New.
* testsuite/ext/random/uniform_inside_sphere_distribution/operators/
inequal.cc: New.
* testsuite/ext/random/uniform_inside_sphere_distribution/operators/
serialize.cc: New.
Index: include/ext/random
===
--- include/ext/random  (revision 216942)
+++ include/ext/random  (working copy)
@@ -3487,6 +3487,218 @@
   _RealType>& __d2)
 { return !(__d1 == __d2); }
 
+
+  /**
+   * @brief A distribution for random coordinates inside a unit sphere.
+   */
+  template
+class uniform_inside_sphere_distribution
+{
+  static_assert(std::is_floating_point<_RealType>::value,
+   "template argument not a floating point type");
+  static_assert(_Dimen != 0, "dimension is zero");
+
+public:
+  /** The type of the range of the distribution. */
+  using result_type = std::array<_RealType, _Dimen>;
+
+  /** Parameter type. */
+  struct param_type
+  {
+   using distribution_type
+ = uniform_inside_sphere_distribution<_Dimen, _RealType>;
+   friend class uniform_inside_sphere_distribution<_Dimen, _RealType>;
+
+   explicit
+   param_type(_RealType __radius = _RealType(1))
+   : _M_radius(__radius)
+   {
+ _GLIBCXX_DEBUG_ASSERT(_M_radius > _RealType(0));
+   }
+
+   _RealType
+   radius() const
+   { return _M_radius; }
+
+   friend bool
+   operator==(const param_type& __p1, const param_type& __p2)
+   { return __p1._M_radius == __p2._M_radius; }
+
+  private:
+   _RealType _M_radius;
+  };
+
+  /**
+   * @brief Constructors.
+   */
+  explicit
+  uniform_inside_sphere_distribution(_RealType __radius = _RealType(1))
+  : _M_param(__radius), _M_uosd()
+  { }
+
+  explicit
+  uniform_inside_sphere_distribution(const param_type& __p)
+  : _M_param(__p), _M_uosd()
+  { }
+
+  /**
+   * @brief Resets the distribution state.
+   */
+  void
+  reset()
+  { _M_uosd.reset(); }
+
+  /**
+   * @brief Returns the @f$radius@f$ of the distribution.
+   */
+  _RealType
+  radius() const
+  { return _M_param.radius(); }
+
+  /**
+   * @brief Returns the parameter set of the distribution.
+   */
+  param_type
+  param() const
+  { return _M_param; }
+
+  /**
+   * @brief Sets the parameter set of the distribution.
+   * @param __param The new parameter set of the distribution.
+   */
+  void
+  param(const param_type& __param)
+  { _M_param = __param; }
+
+  /**
+   * @brief Returns the greatest lower bound value of the distribution.
+   * This function makes no sense for this distribution.
+   */
+  result_type
+  min() const
+  {
+   result_type __res;
+   __res.fill(0);
+   return __res;
+  }
+
+  /**
+   * @brief Returns the least upper bound value of the distribution.
+   * This function makes no sense for this distribution.
+   */
+  result_type
+  max() const
+  {
+   result_type __res;
+   __res.fill(0);
+   return __res;
+  }
+
+  /**
+   * @brief Generating functions.
+   */
+  template
+   result_type
+   operator()(_UniformRandomNumberGenerator& __urng)
+   { return this->operator()(__urng, _M_param); }
+
+  template
+   result_type
+   operator()(_UniformRandomNumberGenerator& __urng,
+  const param_type& __p);
+
+  template
+   void
+   __generate(_ForwardIterator __f, _ForwardIterator __t,
+  _UniformRandomNumberGenerator& __urng)
+   { this->__generate(__f, __t, __urng, this->param()); }
+
+  templat

Re: [Ada] Changes related to back-end inlining

[Eric Botcazou]

> 2014-10-31  Eric Botcazou  
> 
>   * inline.adb (Has_Excluded_Declaration): With back-end inlining,
>   only return true for nested packages.
>   (Cannot_Inline): Issue errors/warnings whatever the optimization level
>   for back-end inlining and remove assertion.

Here is a follow-up patch for the case of nested subprograms, as well as a 
bunch of testcases for the gnat.dg testsuite.

Tested on x86_64-suse-linux, applied on the mainline.


2014-11-05  Eric Botcazou  

* gcc-interface/utils.c (create_subprog_decl): Move code dealing with
conflicting inlining status of nested subprograms to...
* gcc-interface/trans.c (check_inlining_for_nested_subprog): ...here.
(Attribute_to_gnu) : Call it.
(Call_to_gnu): Likewise.
(Subprogram_Body_to_gnu): Drop the body if it is an inlined external
function that has been marked uninlinable.


2014-11-05  Eric Botcazou  

* gnat.dg/inline1.adb: New test.
* gnat.dg/inline1_pkg.ad[sb]: New helper.
* gnat.dg/inline2.adb: New test.
* gnat.dg/inline2_pkg.ad[sb]: New helper.
* gnat.dg/inline3.adb: New test.
* gnat.dg/inline3_pkg.ad[sb]: New helper.
* gnat.dg/inline4.adb: New test.
* gnat.dg/inline4_pkg.ad[sb]: New helper.
* gnat.dg/inline5.adb: New test.
* gnat.dg/inline5_pkg.ad[sb]: New helper.
* gnat.dg/inline6.adb: New test.
* gnat.dg/inline6_pkg.ad[sb]: New helper.
* gnat.dg/inline7.adb: New test.
* gnat.dg/inline7_pkg1.ad[sb]: New helper.
* gnat.dg/inline7_pkg2.ad[sb]: Likewise.
* gnat.dg/inline8.adb: New test.
* gnat.dg/inline8_pkg1.ad[sb]: New helper.
* gnat.dg/inline8_pkg2.ad[sb]: New helper.
* gnat.dg/inline9.adb: New test.
* gnat.dg/inline9_pkg.ad[sb]: New helper.
* gnat.dg/inline10.adb: New test.
* gnat.dg/inline10_pkg.ad[sb]: New helper.
* gnat.dg/inline11.adb: New test.
* gnat.dg/inline11_pkg.ad[sb]: New helper.


-- 
Eric BotcazouIndex: gcc-interface/utils.c
===
--- gcc-interface/utils.c	(revision 217119)
+++ gcc-interface/utils.c	(working copy)
@@ -3027,18 +3027,6 @@ create_subprog_decl (tree subprog_name,
  TREE_TYPE (subprog_type));
   DECL_ARGUMENTS (subprog_decl) = param_decl_list;
 
-  /* If this is a non-inline function nested inside an inlined external
- function, we cannot honor both requests without cloning the nested
- function in the current unit since it is private to the other unit.
- We could inline the nested function as well but it's probably better
- to err on the side of too little inlining.  */
-  if ((inline_status == is_suppressed || inline_status == is_disabled)
-  && !public_flag
-  && current_function_decl
-  && DECL_DECLARED_INLINE_P (current_function_decl)
-  && DECL_EXTERNAL (current_function_decl))
-DECL_DECLARED_INLINE_P (current_function_decl) = 0;
-
   DECL_ARTIFICIAL (subprog_decl) = artificial_flag;
   DECL_EXTERNAL (subprog_decl) = extern_flag;
 
Index: gcc-interface/trans.c
===
--- gcc-interface/trans.c	(revision 217119)
+++ gcc-interface/trans.c	(working copy)
@@ -1481,6 +1481,49 @@ Pragma_to_gnu (Node_Id gnat_node)
   return gnu_result;
 }
 
+
+/* Check the inlining status of nested function FNDECL in the current context.
+
+   If a non-inline nested function is referenced from an inline external
+   function, we cannot honor both requests at the same time without cloning
+   the nested function in the current unit since it is private to its unit.
+   We could inline it as well but it's probably better to err on the side
+   of too little inlining.
+
+   This must be invoked only on nested functions present in the source code
+   and not on nested functions generated by the compiler, e.g. finalizers,
+   because they are not marked inline and we don't want them to block the
+   inlining of the parent function.  */
+
+static void
+check_inlining_for_nested_subprog (tree fndecl)
+{
+  if (!DECL_DECLARED_INLINE_P (fndecl)
+  && current_function_decl
+  && DECL_EXTERNAL (current_function_decl)
+  && DECL_DECLARED_INLINE_P (current_function_decl))
+{
+  const location_t loc1 = DECL_SOURCE_LOCATION (fndecl);
+  const location_t loc2 = DECL_SOURCE_LOCATION (current_function_decl);
+
+  if (lookup_attribute ("always_inline",
+			DECL_ATTRIBUTES (current_function_decl)))
+	{
+	  error_at (loc1, "subprogram %q+F not marked Inline_Always", fndecl);
+	  error_at (loc2, "parent subprogram cannot be inlined");
+	}
+  else
+	{
+	  warning_at (loc1, OPT_Winline, "subprogram %q+F not marked Inline",
+		  fndecl);
+	  warning_at (loc2, OPT_Winline, "parent subprogram cannot be inlined");
+	}
+
+  DECL_DECLARED_INLINE_P (current_function_decl) = 0;
+  DECL_UNINLINABLE (cur

[Ada] Fix location information of exception block

[Eric Botcazou]

This makes it so that an exception block doesn't inherit a bogus location 
information in SJLJ mode.

Tested on x86_64-suse-linux, applied on the mainline.


2014-11-05  Eric Botcazou  

* gcc-interface/trans.c (Handled_Sequence_Of_Statements_to_gnu): Set
the SLOC of the node on the call to set_jmpbuf_address_soft emitted
on block entry with SJLJ.


-- 
Eric BotcazouIndex: gcc-interface/trans.c
===
--- gcc-interface/trans.c	(revision 217151)
+++ gcc-interface/trans.c	(working copy)
@@ -4629,9 +4629,13 @@ Handled_Sequence_Of_Statements_to_gnu (N
   start_stmt_group ();
 
   if (setjmp_longjmp)
-add_stmt (build_call_n_expr (set_jmpbuf_decl, 1,
- build_unary_op (ADDR_EXPR, NULL_TREE,
-		 gnu_jmpbuf_decl)));
+{
+  gnu_expr = build_call_n_expr (set_jmpbuf_decl, 1,
+build_unary_op (ADDR_EXPR, NULL_TREE,
+		gnu_jmpbuf_decl));
+  set_expr_location_from_node (gnu_expr, gnat_node);
+  add_stmt (gnu_expr);
+}
 
   if (Present (First_Real_Statement (gnat_node)))
 process_decls (Statements (gnat_node), Empty,

Re: [gomp4] Use GOMP_OFFLOAD_ prefix for (OpenACC) plugin hooks

[Ilya Verbin]

Hi,

On 05 Nov 17:56, Julian Brown wrote:
> +GOMP_OFFLOAD_register_image (void *host_table, void *target_data)
> +GOMP_OFFLOAD_get_table (struct mapping_table **table)

FYI, these interfaces may change in the near future.
Currently GOMP_OFFLOAD_get_table returns a joint table for all images, offloaded
to a device.  But this doesn't work properly with offloading from dlopened libs.
Do you plan to support such cases for PTX?
Perhaps it's worth to replace them with a function like GOMP_OFFLOAD_load_image,
which will offload one image, and return a target table for this image.
In this case there is no need to pass host_table to the plugin, and return a
joint table, since libgomp will join host and target tables itself.

Another question is what to do with multiple devices of same type.
Can they have different images?  There are 2 options:
1. GOMP_OFFLOAD_load_image will offload one image to one device and receive a
table from it.
or
2. GOMP_OFFLOAD_register_image will register one image in the plugin for all
devices of same type, and
GOMP_OFFLOAD_get_table will return a table for one image and for one device.

Multiple MICs can't have different images, but for the generality we can use
option #1.

  -- Ilya

[Ada] Fix ICE on type derived from private discriminated type

[Eric Botcazou]

The compiler aborts on a record type derived from a private discriminated 
record type without discriminant contraints, if the private discriminated 
record type is itself derived from another discriminated record type.

Tested on x86_64-suse-linux, applied on the mainline.


2014-11-05  Eric Botcazou  

* gcc-interface/decl.c (gnat_to_gnu_entity) : For a
derived untagged type that renames discriminants, be prepared for
a type derived from a private discriminated type when changing the
type of the stored discriminants.


2014-11-05  Eric Botcazou  

* gnat.dg/specs/private2.ads: New test.
* gnat.dg/specs/private2_pkg.ads: New helper.


-- 
Eric BotcazouIndex: gcc-interface/decl.c
===
--- gcc-interface/decl.c	(revision 217119)
+++ gcc-interface/decl.c	(working copy)
@@ -3056,7 +3056,7 @@ gnat_to_gnu_entity (Entity_Id gnat_entit
 		 gnat_field = Next_Stored_Discriminant (gnat_field))
 		  if (Present (Corresponding_Discriminant (gnat_field)))
 		{
-		  Entity_Id field = Empty;
+		  Entity_Id field;
 		  for (field = First_Stored_Discriminant (gnat_parent);
 			   Present (field);
 			   field = Next_Stored_Discriminant (field))
@@ -3138,8 +3138,30 @@ gnat_to_gnu_entity (Entity_Id gnat_entit
 		&& Ekind (Entity (Node (gnat_constr))) == E_Discriminant)
 	  {
 		Entity_Id gnat_discr = Entity (Node (gnat_constr));
-		tree gnu_discr_type = gnat_to_gnu_type (Etype (gnat_discr));
-		tree gnu_ref
+		tree gnu_discr_type, gnu_ref;
+
+		/* If the scope of the discriminant is not the record type,
+		   this means that we're processing the implicit full view
+		   of a type derived from a private discriminated type: in
+		   this case, the Stored_Constraint list is simply copied
+		   from the partial view, see Build_Derived_Private_Type.
+		   So we need to retrieve the corresponding discriminant
+		   of the implicit full view, otherwise we will abort.  */
+		if (Scope (gnat_discr) != gnat_entity)
+		  {
+		Entity_Id field;
+		for (field = First_Entity (gnat_entity);
+			 Present (field);
+			 field = Next_Entity (field))
+		  if (Ekind (field) == E_Discriminant
+			  && same_discriminant_p (gnat_discr, field))
+			break;
+		gcc_assert (Present (field));
+		gnat_discr = field;
+		  }
+
+		gnu_discr_type = gnat_to_gnu_type (Etype (gnat_discr));
+		gnu_ref
 		  = gnat_to_gnu_entity (Original_Record_Component (gnat_discr),
 	NULL_TREE, 0);
 
-- { dg-do compile }

with Private2_Pkg; use Private2_Pkg;

package Private2 is

   type R is new Rec2;

end Private2;
package Private2_Pkg is

   type Rec2 (D : Natural) is private;

private

   type Rec1 (D : Natural) is null record;

   type Rec2 (D : Natural) is new Rec1 (D);

end Private2_Pkg;

[Ada] Fix crash on function with In-Out param returning discriminated type

[Eric Botcazou]

The compiler crashes on a function with an In-Out parameter which returns a 
discriminated record type with default discriminant.

Tested on x86_64-suse-linux, applied on the mainline.


2014-11-05  Eric Botcazou  

* gcc-interface/trans.c (Subprogram_Body_to_gnu): For a function with
copy-in/copy-out parameters and which returns by invisible reference,
do not create the variable for the return value; instead, manually
generate the indirect copy out statements on exit.
(gnat_to_gnu) : Adjust accordingly and build
a simple indirect assignment for the return value.


2014-11-05  Eric Botcazou  

* gnat.dg/discr42.adb: New test.
* gnat.dg/discr42_pkg.ad[sb]: New helper.


-- 
Eric Botcazou-- { dg-do run }

with Discr42_Pkg; use Discr42_Pkg;

procedure Discr42 is

  R : Rec;
  Pos : Natural := 1;

begin

  R := F (Pos);

  if Pos /= 2 then
raise Program_Error;
  end if;

  if R /= (D => True, N => 4) then
raise Program_Error;
  end if;

end;package body Discr42_Pkg is

   function F (Pos : in out Natural) return Rec is
   begin
  Pos := Pos + 1;
  if Pos > 1 then
return (D => True, N => Pos * 2);
  else
return (D => False);
  end if;
   end;

end Discr42_Pkg;Index: gcc-interface/trans.c
===
--- gcc-interface/trans.c	(revision 217152)
+++ gcc-interface/trans.c	(working copy)
@@ -3547,13 +3547,12 @@ Subprogram_Body_to_gnu (Node_Id gnat_nod
   gnu_result_decl = DECL_RESULT (gnu_subprog_decl);
   gnu_subprog_type = TREE_TYPE (gnu_subprog_decl);
   gnu_cico_list = TYPE_CI_CO_LIST (gnu_subprog_type);
-  if (gnu_cico_list)
-gnu_return_var_elmt = value_member (void_type_node, gnu_cico_list);
+  if (gnu_cico_list && TREE_VALUE (gnu_cico_list) == void_type_node)
+gnu_return_var_elmt = gnu_cico_list;
 
   /* If the function returns by invisible reference, make it explicit in the
- function body.  See gnat_to_gnu_entity, E_Subprogram_Type case.
- Handle the explicit case here and the copy-in/copy-out case below.  */
-  if (TREE_ADDRESSABLE (gnu_subprog_type) && !gnu_return_var_elmt)
+ function body.  See gnat_to_gnu_entity, E_Subprogram_Type case.  */
+  if (TREE_ADDRESSABLE (gnu_subprog_type))
 {
   TREE_TYPE (gnu_result_decl)
 	= build_reference_type (TREE_TYPE (gnu_result_decl));
@@ -3573,9 +3572,10 @@ Subprogram_Body_to_gnu (Node_Id gnat_nod
 
   begin_subprog_body (gnu_subprog_decl);
 
-  /* If there are In Out or Out parameters, we need to ensure that the return
- statement properly copies them out.  We do this by making a new block and
- converting any return into a goto to a label at the end of the block.  */
+  /* If there are copy-in/copy-out parameters, we need to ensure that they are
+ properly copied out by the return statement.  We do this by making a new
+ block and converting any return into a goto to a label at the end of the
+ block.  */
   if (gnu_cico_list)
 {
   tree gnu_return_var = NULL_TREE;
@@ -3586,19 +3586,14 @@ Subprogram_Body_to_gnu (Node_Id gnat_nod
   start_stmt_group ();
   gnat_pushlevel ();
 
-  /* If this is a function with In Out or Out parameters, we also need a
-	 variable for the return value to be placed.  */
-  if (gnu_return_var_elmt)
+  /* If this is a function with copy-in/copy-out parameters and which does
+	 not return by invisible reference, we also need a variable for the
+	 return value to be placed.  */
+  if (gnu_return_var_elmt && !TREE_ADDRESSABLE (gnu_subprog_type))
 	{
 	  tree gnu_return_type
 	= TREE_TYPE (TREE_PURPOSE (gnu_return_var_elmt));
 
-	  /* If the function returns by invisible reference, make it
-	 explicit in the function body.  See gnat_to_gnu_entity,
-	 E_Subprogram_Type case.  */
-	  if (TREE_ADDRESSABLE (gnu_subprog_type))
-	gnu_return_type = build_reference_type (gnu_return_type);
-
 	  gnu_return_var
 	= create_var_decl (get_identifier ("RETVAL"), NULL_TREE,
 			   gnu_return_type, NULL_TREE, false, false,
@@ -3693,7 +3688,8 @@ Subprogram_Body_to_gnu (Node_Id gnat_nod
  the label and copy statement.  */
   if (gnu_cico_list)
 {
-  tree gnu_retval;
+  const Node_Id gnat_end_label
+	= End_Label (Handled_Statement_Sequence (gnat_node));
 
   gnu_return_var_stack->pop ();
 
@@ -3701,14 +3697,45 @@ Subprogram_Body_to_gnu (Node_Id gnat_nod
   add_stmt (build1 (LABEL_EXPR, void_type_node,
 			gnu_return_label_stack->last ()));
 
-  if (list_length (gnu_cico_list) == 1)
-	gnu_retval = TREE_VALUE (gnu_cico_list);
+  /* If this is a function which returns by invisible reference, the
+	 return value has already been dealt with at the return statements,
+	 so we only need to indirectly copy out the parameters.  */
+  if (TREE_ADDRESSABLE (gnu_subprog_type))
+	{
+	  tree gnu_ret_deref
+	= build_unary_op (INDIRECT_REF, NULL_TREE, gnu_result_decl);
+	  tree

[jit] Add comments throughout libgccjit.c, and in libgccjit.h

[David Malcolm]

Committed to branch dmalcolm/jit:

Also, add checking to ensure that gcc_jit_context_new_array_type
fails with an error if given a negative size.

gcc/jit/ChangeLog.jit:
* docs/topics/expressions.rst (Type-coercion): Casts between
pointer types are valid.
* libgccjit.c: Document that gcc_jit_context et al are actually
subclasses of the gcc::jit::recording classes.
(RETURN_VAL_IF_FAIL): Add top-level descriptive comment.
(RETURN_IF_NOT_VALID_BLOCK): Likewise.
(RETURN_NULL_IF_NOT_VALID_BLOCK): Likewise.
(jit_error): Likewise.
(compatible_types): Likewise.
(gcc_jit_context_acquire): Likewise.
(gcc_jit_context_release): Likewise.
(gcc_jit_context_new_child_context): Likewise.
(gcc_jit_context_new_location): Likewise.
(gcc_jit_location_as_object): Likewise.
(gcc_jit_type_as_object): Likewise.
(gcc_jit_context_get_type): Likewise.
(gcc_jit_context_get_int_type): Likewise.
(gcc_jit_type_get_pointer): Likewise.
(gcc_jit_type_get_const): Likewise.
(gcc_jit_type_get_volatile): Likewise.
(gcc_jit_context_new_array_type): Likewise.  Also document that
LOC can be NULL.  Fail with an error on negative size.
(gcc_jit_context_new_field): Add top-level descriptive comment and
document that LOC can be NULL.
(gcc_jit_field_as_object): Add top-level descriptive comment.
(gcc_jit_context_new_struct_type): Likewise.  Also document that
LOC can be NULL.
(gcc_jit_context_new_opaque_struct): Likewise.
(gcc_jit_struct_as_type): Add top-level descriptive comment.
(gcc_jit_struct_set_fields): Likewise.  Also document that LOC can
be NULL.
(gcc_jit_context_new_union_type): Likewise.
(gcc_jit_context_new_function_ptr_type): Likewise.
(gcc_jit_context_new_param): Likewise.
(gcc_jit_param_as_object): Add top-level descriptive comment.
(gcc_jit_param_as_lvalue): Likewise.
(gcc_jit_param_as_rvalue): Likewise.
(gcc_jit_context_new_function): Likewise.  Also document that LOC
can be NULL.
(gcc_jit_context_get_builtin_function): Add top-level descriptive
comment.
(gcc_jit_function_as_object): Likewise.
(gcc_jit_function_get_param): Likewise.
(gcc_jit_function_dump_to_dot): Likewise.
(gcc_jit_function_new_block): Likewise.
(gcc_jit_block_as_object): Likewise.
(gcc_jit_block_get_function): Likewise.
(gcc_jit_context_new_global): Likewise.  Also document that LOC
can be NULL.
(gcc_jit_lvalue_as_object): Add top-level descriptive comment.
(gcc_jit_lvalue_as_rvalue): Likewise.
(gcc_jit_rvalue_as_object): Likewise.
(gcc_jit_rvalue_get_type): Likewise.
(RETURN_NULL_IF_FAIL_NONNULL_NUMERIC_TYPE): Likewise.
(gcc_jit_context_new_rvalue_from_int): Likewise.
(gcc_jit_context_zero): Likewise.
(gcc_jit_context_one): Likewise.
(gcc_jit_context_new_rvalue_from_double): Likewise.
(gcc_jit_context_new_rvalue_from_ptr): Likewise.
(gcc_jit_context_null): Likewise.
(gcc_jit_context_new_string_literal): Likewise.
(gcc_jit_context_new_unary_op): Likewise.  Also document that LOC
can be NULL.
(gcc_jit_context_new_binary_op): Likewise.
(gcc_jit_context_new_comparison): Likewise.
(gcc_jit_context_new_call): Likewise.
(gcc_jit_context_new_call_through_ptr): Likewise.
(is_valid_cast): Add top-level descriptive comment.
(gcc_jit_context_new_cast): Likewise.  Also document that LOC can
be NULL.
(gcc_jit_context_new_array_access): Likewise.
(gcc_jit_object_get_context): Add top-level descriptive comment.
(gcc_jit_object_get_debug_string): Likewise.
(gcc_jit_lvalue_access_field): Likewise.  Also document that LOC can
be NULL.
(gcc_jit_rvalue_access_field): Likewise.
(gcc_jit_rvalue_dereference_field): Likewise.
(gcc_jit_rvalue_dereference): Likewise.
(gcc_jit_lvalue_get_address): Likewise.
(gcc_jit_function_new_local): Likewise.
(gcc_jit_block_add_eval): Likewise.
(gcc_jit_block_add_assignment): Likewise.
(gcc_jit_block_add_assignment_op): Likewise.
(is_bool): Add top-level descriptive comment.
(gcc_jit_block_end_with_conditional): Likewise.  Also document
that LOC can be NULL.
(gcc_jit_block_add_comment): Likewise.
(gcc_jit_block_end_with_jump): Likewise.
(gcc_jit_block_end_with_return): Likewise.
(gcc_jit_block_end_with_void_return): Likewise.
(gcc_jit_context_set_str_option): Add top-level descriptive
comment.
(gcc_jit_context_set_int_option): Likewise.
(gcc_jit_context_set_bool_option): Likewise.
(gcc_jit_context_compile): L

Re: [PATCH 10/27] New file: gcc/jit/libgccjit.c

[David Malcolm]

On Tue, 2014-11-04 at 14:39 -0700, Jeff Law wrote:
> On 11/04/14 09:57, David Malcolm wrote:
> >>> +#define IS_ASCII_DIGIT(CHAR) \
> >>> +  ((CHAR) >= '0' && (CHAR) <='9')
> >>> +
> >>> +#define IS_ASCII_ALNUM(CHAR) \
> >>> +  (IS_ASCII_ALPHA (CHAR) || IS_ASCII_DIGIT (CHAR))
> >> Can't we rely on the C library to give us equivalents?
> >
> > I've been burned in the past by the C library using locales, in
> > particular the two lowercase "i" variants in Turkish.
> >
> > These macros are used by gcc_jit_context_new_function to enforce C's
> > naming restrictions, to avoid errors from the assembler.  The comment I
> > put there was:
> >
> >/* The assembler can only handle certain names, so for now, enforce
> >   C's rules for identifiers upon the name.
> >   Eventually we'll need some way to interact with e.g. C++ name 
> > mangling.  */
> >
> > Am I right in thinking that for the assembler we need to enforce the C
> > naming rules specifically on *ASCII*.
> >
> > (clearly another comment is needed here).
> I guess you've got to do it somewhere.  Presumably there isn't something 
> already in GCC that enforces an input character set?  I guess I just 
> dislike seeing something that feels like it ought to already be available.
> 
> >
> > Presumably by marking it with __attribute__((cold)) ?  (with a suitable
> > macro in case we're not being compiled with a gcc that supports it).
> Yup.  That's precisely what you want since that gives the predictors 
> enough information to mark paths as unlikely without having to mark each 
> path yourself.
> 
> >
> > Sorry.  I'll post a followup with comments added.
> Thanks.  I probably rely more on those for this kind of review than 
> anything, so the lack of them really stood out.
> 
> >
> > Many of the functions are public API entrypoints, where there's a
> > comment in the public header.  Should I simply duplicate the comment
> > from there into the .c file, or put a comment like:
> Good question.  Normally in the past we'd have you duplicate the 
> comment, but with this new usage scenario that may not make a lot of 
> sense since one or the other will likely get out of sync at some point.
> 
> At this point a snarky comment about generating documentation and the 
> interface from a single definition would be appropriate.
> 
> 
> >
> >/* Public entrypoint.  See description in libgccjit.h.  */
> >
> > for each of these?  (perhaps with additional text giving implementation
> > notes?)
> Let's go with this.  If folks want the comment duplicated, they can 
> argue for it after the fact :-)
> 
> >
> >
> > Thanks for all the reviews.  Looks like this and patch 16 are now the
> > only non-approved parts of the kit (I didn't see a review of 16).
> Right.  I didn't get to #16 yesterday.

Thanks.

I've added comments throughout the file.

I didn't bother adding __attribute__((cold)), instead simply dropping
that "TODO".

Attached is the current state of the file gcc/jit/libgccjit.c (on the
branch) for review.

OK for trunk? (conditional on all the rest being approved, and usual
bootstrap®rtesting; I've merely verified a non-bootstrap compile and
successful make check-jit so far).

There were a few other changes relative to what you've approved, which
I'll post for review shortly.

Dave
/* Implementation of the C API; all wrappers into the internal C++ API
   Copyright (C) 2013-2014 Free Software Foundation, Inc.
   Contributed by David Malcolm .

This file is part of GCC.

GCC is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 3, or (at your option)
any later version.

GCC is distributed in the hope that it will be useful, but
WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
General Public License for more details.

You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3.  If not see
.  */

#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "opts.h"
#include "safe-ctype.h"

#include "libgccjit.h"
#include "jit-common.h"
#include "jit-recording.h"

/* The opaque types used by the public API are actually subclasses
   of the gcc::jit::recording classes.  */

struct gcc_jit_context : public gcc::jit::recording::context
{
  gcc_jit_context (gcc_jit_context *parent_ctxt) :
context (parent_ctxt)
  {}
};

struct gcc_jit_result : public gcc::jit::result
{
};

struct gcc_jit_object : public gcc::jit::recording::memento
{
};

struct gcc_jit_location : public gcc::jit::recording::location
{
};

struct gcc_jit_type : public gcc::jit::recording::type
{
};

struct gcc_jit_struct : public gcc::jit::recording::struct_
{
};

struct gcc_jit_field : public gcc::jit::recording::field
{
};

struct gcc_jit_function : public gcc::jit::recording::function
{
};

struct gcc_jit_block : public gc

Re: [jit] Use ISALPHA and ISALNUM rather than writing our own

[Jeff Law]

On 11/05/14 08:48, David Malcolm wrote:

On Tue, 2014-11-04 at 14:39 -0700, Jeff Law wrote:

On 11/04/14 09:57, David Malcolm wrote:

+#define IS_ASCII_DIGIT(CHAR) \
+  ((CHAR) >= '0' && (CHAR) <='9')
+
+#define IS_ASCII_ALNUM(CHAR) \
+  (IS_ASCII_ALPHA (CHAR) || IS_ASCII_DIGIT (CHAR))

Can't we rely on the C library to give us equivalents?


I've been burned in the past by the C library using locales, in
particular the two lowercase "i" variants in Turkish.

These macros are used by gcc_jit_context_new_function to enforce C's
naming restrictions, to avoid errors from the assembler.  The comment I
put there was:

/* The assembler can only handle certain names, so for now, enforce
   C's rules for identifiers upon the name.
   Eventually we'll need some way to interact with e.g. C++ name mangling.  
*/

Am I right in thinking that for the assembler we need to enforce the C
naming rules specifically on *ASCII*.

(clearly another comment is needed here).

I guess you've got to do it somewhere.  Presumably there isn't something
already in GCC that enforces an input character set?  I guess I just
dislike seeing something that feels like it ought to already be available.


It turns out that locale-independent tests for this did already exist in
libiberty, in safe-ctype.h, so I've committed this to the jit branch:

gcc/jit/ChangeLog.jit:
* libgccjit.c: Include safe-ctype.h from libiberty.
(IS_ASCII_ALPHA): Delete.
(IS_ASCII_DIGIT): Delete.
(IS_ASCII_ALNUM): Delete.
(gcc_jit_context_new_function): Replace use of IS_ASCII_ALPHA and
IS_ASCII_ALNUM with ISALPHA and ISALNUM respectively, from
libiberty.

Excellent.  Thanks for the cleanup.

Jeff

Re: The nvptx port [8/11+] Write undefined decls.

[Jeff Law]

On 11/05/14 05:01, Bernd Schmidt wrote:

On 10/22/2014 08:11 PM, Jeff Law wrote:

I'm not going to insist you do this in the same way as the PA.  That was
a different era -- we had significant motivation to make things work in
such a way that everything could be buried in the pa specific files.
That sometimes led to less than optimal approaches to fix certain
problems.


So... is this patch approved?

Yes, sorry for not being explicit.

Jeff