This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, middle-end]: Fix also PR46292 (was: Fix PR 44569)
- From: Uros Bizjak <ubizjak at gmail dot com>
- To: Eric Botcazou <ebotcazou at adacore dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Paolo Bonzini <bonzini at gnu dot org>, Richard Guenther <richard dot guenther at gmail dot com>, richard dot sandiford at linaro dot org, John David Anglin <dave at hiauly1 dot hia dot nrc dot ca>
- Date: Wed, 3 Nov 2010 22:46:16 +0100
- Subject: Re: [PATCH, middle-end]: Fix also PR46292 (was: Fix PR 44569)
- References: <AANLkTikhFGbPbtsCRCxqTHqkU7Wjgpr_MB4QRfjrho=o@mail.gmail.com> <201011032238.51735.ebotcazou@adacore.com>
On Wed, Nov 3, 2010 at 10:38 PM, Eric Botcazou <ebotcazou@adacore.com> wrote:
>> This RTX breaks GET_MODE_INNER, but we can win the game with another
>> trick in the sleeve:
>
> Playing tricks on the 4.3/4.4 branches doesn't seem very appropriate to me at
> this stage. ?PR middle-end/44569 wasn't a regression so I wonder why the
> patch was even installed on the 4.5 branch; and there is no testcase AFAICS.
It was "obvious" in the sense that avoided ICE by correcting VOIDmode
for vector mode operands. So, it was either ICE or to dig a little
deeper in the operand itself for a correct mode. Due to this logic,
the patch can't cause any regression.
Uros.