Re: Unordered container insertion hints
On 19 June 2013 20:56, François Dumont wrote: > Still no chance to have a look ? I'll try to finish reviewing it today, thanks for the reminder.
[ping] [patch 4/5] fix bugs with -fstrict-volatile-bitfields and packed structures
On 06/16/2013 01:08 PM, Sandra Loosemore wrote: This part of the patch series fixes problems with bad code being emitted for unaligned bitfield accesses, as reported in PRs 48784, 56341, and 56997. A secondary goal of this patch was making the bitfield store and extract code follow similar logic, at least for the parts relating to -fstrict-volatile-bitfield handling. Is it possible to get this part of the patch series reviewed? Except for the documentation change, it is independent of the controversy surrounding part 3 regarding whether the target ABI or C/C++ standard should take precedence when they conflict, and is independent of any further patches to change the default -fstrict-volatile-bitfields setting. If the rest of the patch is approved, I'll take care to fix up invoke.texi to accurately reflect the behavior of the approved patches before checking anything in. http://gcc.gnu.org/ml/gcc-patches/2013-06/msg00911.html -Sandra
[C++ Patch] PR 57682
Hi,
in this C++11 reject valid parsing issue, in a function-try-block we reject:
try
: f { func } { } (1)
and we normally accept the C++98:
try
: f ( func ) { } (2)
Considering cp_parser_save_member_function_body, I think that a neat
solution would be simply consuming first any 'try' keyword and then
proceeding as usual. What is happening now is that, later on, the closed
curly brace in 'f{ func }' is confused with the closed curly brace of
(2) and things wrong when the former is not followed by 'catch'.
Tested x86_64-linux.
Thanks,
Paolo.
/
/cp
2013-06-23 Paolo Carlini
PR c++/57682
* parser.c (cp_parser_save_member_function_body): Handle correctly
curly braces in function-try-block mem-initializers.
/testsuite
2013-06-23 Paolo Carlini
PR c++/57682
* g++.dg/cpp0x/initlist72.C: New.
Index: cp/parser.c
===
--- cp/parser.c (revision 200350)
+++ cp/parser.c (working copy)
@@ -22797,6 +22797,9 @@ cp_parser_save_member_function_body (cp_parser* pa
/* Save away the tokens that make up the body of the
function. */
first = parser->lexer->next_token;
+ /* Handle function try blocks. */
+ if (cp_lexer_next_token_is_keyword (parser->lexer, RID_TRY))
+cp_lexer_consume_token (parser->lexer);
/* We can have braced-init-list mem-initializers before the fn body. */
if (cp_lexer_next_token_is (parser->lexer, CPP_COLON))
{
Index: testsuite/g++.dg/cpp0x/initlist72.C
===
--- testsuite/g++.dg/cpp0x/initlist72.C (revision 0)
+++ testsuite/g++.dg/cpp0x/initlist72.C (working copy)
@@ -0,0 +1,13 @@
+// PR c++/57682
+// { dg-do compile { target c++11 } }
+
+struct Class
+{
+ Class (int func)
+ try
+ : f { func } { }
+ catch ( ... ) { }
+
+private:
+ int f;
+};
[committed] Fix -mlzcnt (PR target/57688)
On Fri, May 17, 2013 at 09:00:21PM -0700, Sriraman Tallam wrote:
> --- common/config/i386/i386-common.c (revision 198950)
> +++ common/config/i386/i386-common.c (working copy)
> @@ -438,6 +440,18 @@ ix86_handle_option (struct gcc_options *opts,
> }
>return true;
>
> +case OPT_mlzcnt:
> + if (value)
> + {
> + opts->x_ix86_isa_flags |= OPTION_MASK_ISA_LZCNT_SET;
> + opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_LZCNT_SET;
> + }
> + else
> + {
> + opts->x_ix86_isa_flags &= ~OPTION_MASK_ISA_LZCNT_UNSET;
> + opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_LZCNT_UNSET;
> + }
> +
> case OPT_mtbm:
>if (value)
> {
This change broke PR target/57688, because the -mlzcnt case falls thru
into the -mtbm case, so if you have -mlzcnt, but not -mtbm, it will enable
-mtbm anyway.
Fixed thusly, committed as obvious.
2013-06-23 Jakub Jelinek
PR target/57688
* common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
add missing return true.
--- gcc/common/config/i386/i386-common.c.jj 2013-06-23 20:43:34.0
+0200
+++ gcc/common/config/i386/i386-common.c2013-06-23 21:00:17.056390879
+0200
@@ -451,6 +451,7 @@ ix86_handle_option (struct gcc_options *
opts->x_ix86_isa_flags &= ~OPTION_MASK_ISA_LZCNT_UNSET;
opts->x_ix86_isa_flags_explicit |= OPTION_MASK_ISA_LZCNT_UNSET;
}
+ return true;
case OPT_mtbm:
if (value)
Jakub
Re: [PATCH] Basic support for MIPS r5900
Hello Richard,
> >> Does it still work with those changes, as below? If so, I'll check it in.
> >
> > I tested it. It is still working. So the patch is OK, please check it in.
>
> OK, I've applied this and the config.gcc patch.
Thanks.
In the native PS2SDK (i.e. no Linux) I detected that there are undefined
references to `__fixtfsi', `__floatsitf', `__subtf3', `__multf3',
`__extenddftf2', `__lttf2' and `__gttf2'. These are the functions for handling
"long double".
There is a fix needed for this in libgcc. I attached 2 patches to this e-mail.
There should be only one of the patches used to fix this.
The first one fixes the problem for all mips systems. The second one fixes this
only for r5900. I think it is a general problem. I assume the undefined
references could appear on all mips64 systems which are not Linux. So the first
patch should be applied.
Checking for "ac_cv_sizeof_long_double" should work on all mips systems, but I
don't know why the check was only in the mips*-linux part of the file
libgcc/config.host.
Best regards
JürgenIndex: libgcc/config.host
===
--- libgcc/config.host (Revision 200155)
+++ libgcc/config.host (Arbeitskopie)
@@ -136,6 +136,9 @@
mips*-*-*)
cpu_type=mips
tmake_file=mips/t-mips
+ if test "${ac_cv_sizeof_long_double}" = 16; then
+ tmake_file="${tmake_file} mips/t-tpbit"
+ fi
;;
powerpc*-*-*)
cpu_type=rs6000
@@ -751,9 +754,6 @@
;;
esac
md_unwind_header=mips/linux-unwind.h
- if test "${ac_cv_sizeof_long_double}" = 16; then
- tmake_file="${tmake_file} mips/t-tpbit"
- fi
;;
mips*-sde-elf*)
tmake_file="$tmake_file mips/t-crtstuff mips/t-mips16"
Index: libgcc/config.host
===
--- libgcc/config.host (Revision 200155)
+++ libgcc/config.host (Arbeitskopie)
@@ -798,6 +798,9 @@
mips64r5900-*-elf* | mips64r5900el-*-elf*)
tmake_file="$tmake_file mips/t-elf mips/t-crtstuff"
extra_parts="$extra_parts crti.o crtn.o"
+ if test "${ac_cv_sizeof_long_double}" = 16; then
+ tmake_file="${tmake_file} mips/t-tpbit"
+ fi
;;
mips64vr-*-elf* | mips64vrel-*-elf*)
tmake_file="$tmake_file mips/t-elf mips/t-vr mips/t-crtstuff"
Re: [PATCH] Basic support for MIPS r5900
"Jürgen Urban" writes: > Hello Richard, > >> >> Does it still work with those changes, as below? If so, I'll check it in. >> > >> > I tested it. It is still working. So the patch is OK, please check it in. >> >> OK, I've applied this and the config.gcc patch. > > Thanks. > > In the native PS2SDK (i.e. no Linux) I detected that there are undefined > references to `__fixtfsi', `__floatsitf', `__subtf3', `__multf3', > __extenddftf2', `__lttf2' and `__gttf2'. These are the functions for > handling "long double". > There is a fix needed for this in libgcc. I attached 2 patches to this > e-mail. There should be only one of the patches used to fix this. > The first one fixes the problem for all mips systems. The second one > fixes this only for r5900. I think it is a general problem. I assume the > undefined references could appear on all mips64 systems which are not > Linux. So the first patch should be applied. The second patch is right. Not all targets have 128-bit long doubles, so t-tpbit is intentionally an opt-in rather than opt-out thing. E.g. mips*-sde-elf and mips*-mti-elf both use: #undef LONG_DOUBLE_TYPE_SIZE #define LONG_DOUBLE_TYPE_SIZE 64 to get a form of n32 in which long doubles are the same as doubles. Do you want to stick with 128-bit long doubles for PS2, or would you prefer to do what SDE does? Thanks, Richard
[Patch, Fortran] PR52413 - Incorrect behavior of FRACTION when applied to a constant
This patch packages the Steven G. Kargl's patch when MPFR 3.1.0 or newer
is available and fixes the original code for older versions using the
FX Coudert's machinery (see PR for details).
Tested on powerpc-apple-darwin9 and x86_64-apple-darwin10 without
regression.
OK for the trunk? If yes, could someone commit it for me?
Since this is a wrong-code issue, I think the light version of the patch
should be applied to 4.7 and 4.8. If this is accepted I'll do the testing
and the packaging.
Dominique
2013-06-23 Steven G. Kargl
Francois-Xavier Coudert
Dominique d'Humieres
PR fortran/52413
* simplify.c: Fix the sign of negative values in gfc_simplify_fraction.
2013-06-23 Francois-Xavier Coudert
Dominique d'Humieres
PR fortran/52413
* gfortran.dg/fraction.f90: New.
--- ../_clean/gcc/fortran/simplify.c2013-06-08 21:50:33.0 +0200
+++ gcc/fortran/simplify.c 2013-06-23 17:19:55.0 +0200
@@ -2342,16 +2342,26 @@ gfc_expr *
gfc_simplify_fraction (gfc_expr *x)
{
gfc_expr *result;
+
+#if MPFR_VERSION < MPFR_VERSION_NUM(3,1,0)
mpfr_t absv, exp, pow2;
+#else
+ mpfr_exp_t e;
+#endif
if (x->expr_type != EXPR_CONSTANT)
return NULL;
result = gfc_get_constant_expr (BT_REAL, x->ts.kind, &x->where);
+#if MPFR_VERSION < MPFR_VERSION_NUM(3,1,0)
+
+ /* MPFR versions before 3.1.0 do not include mpfr_frexp.
+ TODO: remove the kludge when MPFR 3.1.0 or newer will be required */
+
if (mpfr_sgn (x->value.real) == 0)
{
- mpfr_set_ui (result->value.real, 0, GFC_RND_MODE);
+ mpfr_set (result->value.real, x->value.real, GFC_RND_MODE);
return result;
}
@@ -2368,10 +2378,16 @@ gfc_simplify_fraction (gfc_expr *x)
mpfr_ui_pow (pow2, 2, exp, GFC_RND_MODE);
- mpfr_div (result->value.real, absv, pow2, GFC_RND_MODE);
+ mpfr_div (result->value.real, x->value.real, pow2, GFC_RND_MODE);
mpfr_clears (exp, absv, pow2, NULL);
+#else
+
+ mpfr_frexp (&e, result->value.real, x->value.real, GFC_RND_MODE);
+
+#endif
+
return range_check (result, "FRACTION");
}
--- ../_clean/gcc/testsuite/gfortran.dg/fraction.f901970-01-01
01:00:00.0 +0100
+++ gcc/testsuite/gfortran.dg/fraction.f90 2013-06-23 22:06:00.0
+0200
@@ -0,0 +1,15 @@
+! { dg-do run }
+!
+! Test for pr52413
+!
+
+program test_frac
+
+ real :: y
+ y=fraction (-2.0)
+ if (fraction (-2.0) /= -0.5) call abort ()
+ if (fraction (-0.0) /= 0.0) call abort ()
+ if (sign(1.0, fraction(-0.0)) /= -1.0) call abort ()
+ if (fraction (-2.0_8) /= -0.5) call abort ()
+
+end program test_frac
[C++ testcase, committed] Add SFINAE testcase
Hi,
I added to mainline the testcase in Comment 16 of PR51213.
Thanks,
Paolo.
/
2013-06-23 Paolo Carlini
* g++.dg/cpp0x/sfinae47.C: New.
Index: g++.dg/cpp0x/sfinae47.C
===
--- g++.dg/cpp0x/sfinae47.C (revision 0)
+++ g++.dg/cpp0x/sfinae47.C (working copy)
@@ -0,0 +1,25 @@
+// Source: Comment 16 of PR51213
+// { dg-do compile { target c++11 } }
+
+template
+T && declval();
+
+template
+constexpr auto hasSize(int) -> decltype(declval().size(), bool())
+{ return true; }
+
+template
+constexpr bool hasSize(...)
+{ return false; }
+
+struct A
+{
+ int size();
+};
+
+struct B : private A
+{
+};
+
+static_assert(hasSize(0), "A");
+static_assert(!hasSize(0), "B");
Re: [Patch, Fortran] Print floating-point exception status after STOP/ERROR STOP
On Sun, 23 Jun 2013, Tobias Burnus wrote: > Gerald and Eric: Would the following work for you? (Sorry, no > config.h.in/configure patch; I haven't build in maintainer's mode, yet.) Yes, that looks like what I had in mind. Thanks, Tobias! Gerald
Re: [Patch wwwdocs] gcc-4.9 changes: mention support of the Intel Silvermont microarchitecture
On Mon, 10 Jun 2013, Igor Zamyatin wrote: > This patch mentions support of Silvermont architecture in the > gcc-4.9/changes.html page. Looks good to me; just say "the new Intel...". Thanks, Gerald
Re: [RS6000] libffi little-endian
On Fri, Jun 07, 2013 at 12:12:17AM -0400, David Edelsohn wrote:
> On Thu, Jun 6, 2013 at 9:34 PM, Alan Modra wrote:
> > Bootstrapped and regression tested powerpc64-linux. OK to apply?
> >
> > * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Support
> > little-endian.
> > * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Likewise.
>
> This patch needs to be applied upstream in the libffi repository.
>
> All of the handling of structs in ffi.c and ffi_darwin.c doesn't need
> any changes? Cool. I thought there might be a padding issue.
You were right, of course. When we finally got around to running an
all languages bootstrap on powerpc64le, we discovered some missing
pieces in libffi. The following adds some ffi.c changes to the
previous patch (closure.S patches are unchanged). I haven't tackled
ffi_darwin.c.
Bootstrapped and regression tested powerpc64-linux. This one passes
the libffi testsuite on powerpc64le-linux. OK mainline and 4.8?
* src/powerpc/ffi.c (ffi_prep_args_SYSV): Move var declaration
before statements.
(ffi_prep_args64): Support little-endian.
(ffi_closure_helper_SYSV, ffi_closure_helper_LINUX64): Likewise.
* src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Likewise.
* src/powerpc/ppc_closure.S (ffi_closure_SYSV): Likewise.
Index: libffi/src/powerpc/ffi.c
===
--- libffi/src/powerpc/ffi.c(revision 200159)
+++ libffi/src/powerpc/ffi.c(working copy)
@@ -127,6 +127,9 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *
int i;
ffi_type **ptr;
+#ifndef __NO_FPRS__
+ double double_tmp;
+#endif
union {
void **v;
char **c;
@@ -146,7 +149,6 @@ ffi_prep_args_SYSV (extended_cif *ecif, unsigned *
gpr_base.u = stacktop.u - ASM_NEEDS_REGISTERS - NUM_GPR_ARG_REGISTERS;
intarg_count = 0;
#ifndef __NO_FPRS__
- double double_tmp;
fpr_base.d = gpr_base.d - NUM_FPR_ARG_REGISTERS;
fparg_count = 0;
copy_space.c = ((flags & FLAG_FP_ARGUMENTS) ? fpr_base.c : gpr_base.c);
@@ -542,11 +544,12 @@ ffi_prep_args64 (extended_cif *ecif, unsigned long
{
char *where = next_arg.c;
+#ifndef __LITTLE_ENDIAN__
/* Structures with size less than eight bytes are passed
left-padded. */
if ((*ptr)->size < 8)
where += 8 - (*ptr)->size;
-
+#endif
memcpy (where, *p_argv.c, (*ptr)->size);
next_arg.ul += words;
if (next_arg.ul == gpr_end.ul)
@@ -1208,6 +1211,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, voi
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
+#ifndef __LITTLE_ENDIAN__
/* there are 8 gpr registers used to pass values */
if (ng < 8)
{
@@ -1221,9 +1225,10 @@ ffi_closure_helper_SYSV (ffi_closure *closure, voi
pst++;
}
break;
-
+#endif
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
+#ifndef __LITTLE_ENDIAN__
/* there are 8 gpr registers used to pass values */
if (ng < 8)
{
@@ -1237,7 +1242,7 @@ ffi_closure_helper_SYSV (ffi_closure *closure, voi
pst++;
}
break;
-
+#endif
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
case FFI_TYPE_POINTER:
@@ -1367,22 +1372,25 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure,
{
case FFI_TYPE_SINT8:
case FFI_TYPE_UINT8:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 7;
pst++;
break;
-
+#endif
case FFI_TYPE_SINT16:
case FFI_TYPE_UINT16:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 6;
pst++;
break;
-
+#endif
case FFI_TYPE_SINT32:
case FFI_TYPE_UINT32:
+#ifndef __LITTLE_ENDIAN__
avalue[i] = (char *) pst + 4;
pst++;
break;
-
+#endif
case FFI_TYPE_SINT64:
case FFI_TYPE_UINT64:
case FFI_TYPE_POINTER:
@@ -1391,11 +1399,13 @@ ffi_closure_helper_LINUX64 (ffi_closure *closure,
break;
case FFI_TYPE_STRUCT:
+#ifndef __LITTLE_ENDIAN__
/* Structures with size less than eight bytes are passed
left-padded. */
if (arg_types[i]->size < 8)
avalue[i] = (char *) pst + 8 - arg_types[i]->size;
else
+#endif
avalue[i] = pst;
pst += (arg_types[i]->size + 7) / 8;
break;
Index: libffi/src/powerpc/linux64_closure.S
===
--- libffi/src/powerpc/linux64_closure.S(revision 200159)
+++ libffi/src/powerpc/linux64_closure.S(working copy)
@@ -132,7 +132,11 @@ ffi_closure_LINUX64:
blr
nop
# case FFI_TYPE_INT
+#ifdef __LITTLE_ENDIAN__
+ lwa %r3, 112+0(%r1)
+#else
lwa %r3, 112+4(%r1)
+#endif
mtlr %r0
addi %
Re: [RS6000] libffi little-endian
On Sun, Jun 23, 2013 at 8:26 PM, Alan Modra wrote: > On Fri, Jun 07, 2013 at 12:12:17AM -0400, David Edelsohn wrote: >> On Thu, Jun 6, 2013 at 9:34 PM, Alan Modra wrote: >> > Bootstrapped and regression tested powerpc64-linux. OK to apply? >> > >> > * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Support >> > little-endian. >> > * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Likewise. >> >> This patch needs to be applied upstream in the libffi repository. >> >> All of the handling of structs in ffi.c and ffi_darwin.c doesn't need >> any changes? Cool. I thought there might be a padding issue. > > You were right, of course. When we finally got around to running an > all languages bootstrap on powerpc64le, we discovered some missing > pieces in libffi. The following adds some ffi.c changes to the > previous patch (closure.S patches are unchanged). I haven't tackled > ffi_darwin.c. > > Bootstrapped and regression tested powerpc64-linux. This one passes > the libffi testsuite on powerpc64le-linux. OK mainline and 4.8? > > * src/powerpc/ffi.c (ffi_prep_args_SYSV): Move var declaration > before statements. > (ffi_prep_args64): Support little-endian. > (ffi_closure_helper_SYSV, ffi_closure_helper_LINUX64): Likewise. > * src/powerpc/linux64_closure.S (ffi_closure_LINUX64): Likewise. > * src/powerpc/ppc_closure.S (ffi_closure_SYSV): Likewise. This looks okay to me and more like the changes I expected. Assuming Anthony accepts it upstream. Thanks, David
