[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

rsandifo at gcc dot gnu.org  changed:

   What|Removed |Added

 Status|ASSIGNED|RESOLVED
 Resolution|--- |FIXED

--- Comment #11 from rsandifo at gcc dot gnu.org  
---
Fixed on master.

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

--- Comment #10 from CVS Commits  ---
The master branch has been updated by Richard Sandiford :

https://gcc.gnu.org/g:a044dfb0934c3caea0886dbb497181e3aca660df

commit r10-6261-ga044dfb0934c3caea0886dbb497181e3aca660df
Author: Richard Sandiford 
Date:   Sun Jan 26 12:47:23 2020 +

aarch64: Fix pr71727.c failure

This test started failing after the switch to -fno-common because we can
now force the array to be aligned to 16 bytes, which in turn lets us use
SIMD accesses.  Locally restoring -fcommon seems the most faithful to
the original PR.

2020-01-27  Richard Sandiford  

gcc/testsuite/
PR testsuite/71727
* gcc.target/aarch64/pr71727.c: Add -fcommon.

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[rsandifo at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

rsandifo at gcc dot gnu.org  changed:

   What|Removed |Added

 Status|NEW |ASSIGNED
 CC||rsandifo at gcc dot gnu.org
   Assignee|unassigned at gcc dot gnu.org  |rsandifo at gcc dot 
gnu.org

--- Comment #9 from rsandifo at gcc dot gnu.org  
---
(In reply to Wilco from comment #8)
> (In reply to Martin Liška from comment #7)
> > Christophe: Can the bug be marked as resolved?
> 
> See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91927 - it's still failing
> since the underlying issues haven't been resolved.

I think we just need to change the test.  Now that -fno-common is the
default (thanks), the definition of the array is local and so its
alignment can be changed.  And since the test is hard-coded to access
the first element, we realise that we can use SIMD accesses if we
increase the alignment of the array to 16.

I think we should force -fcommon here.

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[wilco at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

Wilco  changed:

   What|Removed |Added

 CC||wilco at gcc dot gnu.org

--- Comment #8 from Wilco  ---
(In reply to Martin Liška from comment #7)
> Christophe: Can the bug be marked as resolved?

See https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91927 - it's still failing
since the underlying issues haven't been resolved.

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

Martin Liška  changed:

   What|Removed |Added

 CC||marxin at gcc dot gnu.org

--- Comment #7 from Martin Liška  ---
Christophe: Can the bug be marked as resolved?

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[sje at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

Steve Ellcey  changed:

   What|Removed |Added

 CC||sje at gcc dot gnu.org

--- Comment #6 from Steve Ellcey  ---
Is there any reason this defect cannot be closed out?

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

--- Comment #5 from Christophe Lyon  ---
Author: clyon
Date: Fri Oct  6 07:38:51 2017
New Revision: 253477

URL: https://gcc.gnu.org/viewcvs?rev=253477=gcc=rev
Log:
[AArch64] Backport to gcc-7 PR71727 fix -mstrict-align

2017-10-06  Christophe Lyon  

gcc/
Backport from mainline r253242.
2017-09-27  Christophe Lyon  

PR target/71727
* config/aarch64/aarch64.c
(aarch64_builtin_support_vector_misalignment): Always return false
when misalignment is unknown.

gcc/testsuite/
Backport from mainline r253251.
2017-09-27  Christophe Lyon  

PR target/71727
* gcc.target/aarch64/pr71727-2.c: New test.



Added:
branches/gcc-7-branch/gcc/testsuite/gcc.target/aarch64/pr71727-2.c
Modified:
branches/gcc-7-branch/gcc/ChangeLog
branches/gcc-7-branch/gcc/config/aarch64/aarch64.c
branches/gcc-7-branch/gcc/testsuite/ChangeLog

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

--- Comment #4 from Christophe Lyon  ---
Author: clyon
Date: Wed Sep 27 23:52:58 2017
New Revision: 253242

URL: https://gcc.gnu.org/viewcvs?rev=253242=gcc=rev
Log:
[AArch64] PR71727 fix -mstrict-align

2017-09-27  Christophe Lyon  

PR target/71727
gcc/
* config/aarch64/aarch64.c
(aarch64_builtin_support_vector_misalignment): Always return false
when misalignment is unknown.

gcc/testsuite/
* gcc.target/aarch64/pr71727-2.c: New test


Modified:
trunk/gcc/ChangeLog
trunk/gcc/config/aarch64/aarch64.c
trunk/gcc/testsuite/ChangeLog

[Bug target/71727] -O3 -mstrict-align produces code which assumes unaligned vector accesses work

[hs.naveen2u at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71727

hs.naveen2u at gmail dot com changed:

   What|Removed |Added

 CC||hs.naveen2u at gmail dot com

--- Comment #3 from hs.naveen2u at gmail dot com ---
Fixed and Committed as:-
https://gcc.gnu.org/ml/gcc-cvs/2016-12/msg00244.html