This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] fix target/44606, reload bug on SPE
- From: Bernd Schmidt <bernds at codesourcery dot com>
- To: Eric Botcazou <ebotcazou at adacore dot com>
- Cc: Nathan Froyd <froydnj at codesourcery dot com>, gcc-patches at gcc dot gnu dot org
- Date: Sat, 02 Oct 2010 01:16:35 +0200
- Subject: Re: [PATCH] fix target/44606, reload bug on SPE
- References: <20100930182927.GL32503@codesourcery.com> <201010010016.50677.ebotcazou@adacore.com> <20101001163350.GP32503@codesourcery.com> <201010012338.39855.ebotcazou@adacore.com>
On 10/01/2010 11:38 PM, Eric Botcazou wrote:
>> I don't see a good way to tighten the above condition. Deleting the
>> cleverness is an option, though. ;)
>
> How about adding a check that the modes have the same size when this is for an
> output reload?
That's just papering over the real problem. I think the intent is
clearly that if we find an equivalence, the move insn being reloaded is
removed (as it serves no purpose if we already have the value in a
register) and only the output reload remains.
For the release branches it's conceivable to disable this optimization,
it shouldn't trigger that often anyway.
Bernd