This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] gdb passes and returns incorrect values when dealing with small array on Sparc
- From: Vladimir Mezentsev <vladimir dot mezentsev at oracle dot com>
- To: Pedro Alves <palves at redhat dot com>, binutils at sourceware dot org, gdb-patches at sourceware dot org
- Date: Wed, 29 Mar 2017 09:38:10 -0700
- Subject: Re: [PATCH] gdb passes and returns incorrect values when dealing with small array on Sparc
- Authentication-results: sourceware.org; auth=none
- References: <1490721669-12560-1-git-send-email-vladimir.mezentsev@oracle.com> <5a3397ae-e986-8a36-13ba-ed2cfeb68a12@redhat.com>
- Reply-to: vladimir dot mezentsev at oracle dot com
On 03/29/2017 02:56 AM, Pedro Alves wrote:
Hi Vladimir,
gdb development happens on a separate list - gdb-patches@. Please resend
the patch there. TIA!
Done.
A question I'll have is whether this had any observable effect in testsuite
results, other than "no regressions". I.e., whether there were progressions,
or whether we're missing testsuite coverage.
6 tests ( from gdb/testsuite/gdb.base/gnu_vector.exp) failed on
sparc64-Linux and on sparc-Solaris in 32- and 64-bit mode.
Now all these tests passed.
gdb/testsuite/gdb.base/gnu_vector.exp has 117 different cases for small
(and not small) arrays and structures.
I think we don't need an additional test in gdb/testsuite.
-Vladimir
On 03/28/2017 06:21 PM, vladimir.mezentsev@oracle.com wrote:
From: Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
gdb has a special type (TYPE_CODE_ARRAY) to support the gcc extension
(https://gcc.gnu.org/onlinedocs/gcc/Vector-Extensions.html).
TYPE_CODE_ARRAY is handled incorrectly for both (32- and 64-bit) modes on Sparc machines.
Tested on sparc64-linux-gnu and sparc-solaris (32- and 64-bit mode).
No regressions.
2017-03-27 Vladimir Mezentsev <vladimir.mezentsev@oracle.com>
* gdb/sparc-tdep.c (sparc_structure_return_p): New function.
Also, please drop "gdb/" here. Filename entries in ChangeLogs
are relative to the ChangeLog file path.
Thanks,
Pedro Alves