https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #46 from Jakub Jelinek ---
Author: jakub
Date: Tue Apr 25 16:47:32 2017
New Revision: 247259
URL: https://gcc.gnu.org/viewcvs?rev=247259=gcc=rev
Log:
PR target/77728
* config/arm/arm.c: Include gimple.h.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #45 from Jakub Jelinek ---
Author: jakub
Date: Tue Apr 25 16:46:34 2017
New Revision: 247258
URL: https://gcc.gnu.org/viewcvs?rev=247258=gcc=rev
Log:
PR target/77728
* config/arm/arm.c: Include gimple.h.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #44 from Jakub Jelinek ---
Works for me. That would mean roughly applying the two patches, but instead of
doing else if (res < 0) do if (res) (and something similar for aarch64).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #43 from Richard Earnshaw ---
Hmm, so how about just inserting the warning in the broken compilers?
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #42 from Jakub Jelinek ---
Oh, and we surely need to document this in gcc-7/changes.html, and I'd think we
should make sure to also document the earlier ARM ABI change in
gcc-5/changes.html for GCC 5.2 when it has changed and
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #41 from Jakub Jelinek ---
That is true, but we've got only a single bugreport about this since the 5.2
release (21 months), so it doesn't trigger that often. For static data members
the ABI is self-consistent, for some template
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #40 from Richard Earnshaw ---
(In reply to Jakub Jelinek from comment #39)
> It is an ABI change, so I think it is highly undesirable to backport. It is
> enough that people will have to rebuild many packages built by GCC 7
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #39 from Jakub Jelinek ---
It is an ABI change, so I think it is highly undesirable to backport. It is
enough that people will have to rebuild many packages built by GCC 7
prereleases, if it is backported, they would have to rebuild
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Christophe Lyon changed:
What|Removed |Added
CC||clyon at gcc dot gnu.org
--- Comment
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #37 from Jakub Jelinek ---
Author: jakub
Date: Tue Apr 25 13:56:10 2017
New Revision: 247241
URL: https://gcc.gnu.org/viewcvs?rev=247241=gcc=rev
Log:
PR target/77728
* config/aarch64/aarch64.c (struct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #36 from Jakub Jelinek ---
Author: jakub
Date: Tue Apr 25 13:52:33 2017
New Revision: 247239
URL: https://gcc.gnu.org/viewcvs?rev=247239=gcc=rev
Log:
PR target/77728
* config/aarch64/aarch64.c (struct
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #35 from Jakub Jelinek ---
Created attachment 41264
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41264=edit
gcc7-pr77728-aarch64.patch
Similarly adjusted AArch64 patch.
In the earlier AArch64 patch, warning_alignment didn't
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Jakub Jelinek changed:
What|Removed |Added
Attachment #41262|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Jakub Jelinek changed:
What|Removed |Added
Attachment #41261|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Jakub Jelinek changed:
What|Removed |Added
Attachment #41259|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #31 from Jakub Jelinek ---
(In reply to Jakub Jelinek from comment #30)
> Created attachment 41259 [details]
> gcc7-pr77728.patch
The above attached version successfully bootstrapped with
--enable-languages=c,c++
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #30 from Jakub Jelinek ---
Created attachment 41259
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41259=edit
gcc7-pr77728.patch
This is how I'd have written the aarch32 patch. It has smaller amount of
changes and calls the
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #29 from Jakub Jelinek ---
Also, the aarch64_function_arg_alignment and arm_needs_doubleword_align
function comments need updating.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #28 from Jakub Jelinek ---
The
+ else
+ warn_align = MAX (warn_align, DECL_ALIGN (field));
last line here is misindented, should be below se.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #27 from Ramana Radhakrishnan ---
(In reply to Ramana Radhakrishnan from comment #26)
> Created attachment 41257 [details]
> AArch32 wip patch.
>
> I think there's one line > 80 chars here but this is what I'm testing
> currently.
>
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Ramana Radhakrishnan changed:
What|Removed |Added
Attachment #41255|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #25 from Ramana Radhakrishnan ---
Created attachment 41255
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41255=edit
AArch32 wip patch.
- There are some line > 80 characters, cosmetic issues in this patch .
- However this
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #24 from Jakub Jelinek ---
std::min/max was used in the backend previously too, so it is your decision and
doesn't need to be changed to resolve this regression.
Do you have arm32 patch too, or would it be helpful if I write one
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
Ramana Radhakrishnan changed:
What|Removed |Added
Attachment #41237|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #22 from Ramana Radhakrishnan ---
(In reply to Jakub Jelinek from comment #20)
> BTW, the wording e.g. i386 backend has is:
> inform (input_location,
> "the ABI of passing structure with complex
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #21 from Jakub Jelinek ---
inform, beyond being consistent with what i386 and rs6000 backends do here, has
the advantage that it doesn't break the build even with -Werror, generally
there is nothing wrong on the code we want to
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #20 from Jakub Jelinek ---
BTW, the wording e.g. i386 backend has is:
inform (input_location,
"the ABI of passing structure with complex float"
" member has changed in GCC
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #19 from Jakub Jelinek ---
Indeed, with the above plus:
#include
int
fn1 (int a, int b, int c, int d, int e, int f, int g, int h, int i, int j, int
k, int l, int m, B<0> n, ...)
{
va_list ap;
va_start (ap, n);
int x = va_arg
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #18 from Jakub Jelinek ---
If you want a testcase for aarch64, e.g.
template
struct alignas (16) A { char p[16]; };
A<0> v;
template
struct B
{
typedef A T;
int i, j, k, l;
};
int
foo (int a, B<0> b)
{
return a + b.i;
}
int
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #17 from Ramana Radhakrishnan ---
Created attachment 41237
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=41237=edit
AArch64 prototype.
Patch.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77728
--- Comment #16 from Ramana Radhakrishnan ---
(In reply to Jakub Jelinek from comment #10)
> aarch64_function_arg_alignment
> has:
> for (tree field = TYPE_FIELDS (type); field; field = DECL_CHAIN (field))
> alignment = std::max
31 matches
Mail list logo
