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: Nathan Froyd <froydnj at codesourcery dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 30 Sep 2010 22:49:47 +0200
- Subject: Re: [PATCH] fix target/44606, reload bug on SPE
- References: <20100930182927.GL32503@codesourcery.com>
On 09/30/2010 08:29 PM, Nathan Froyd wrote:
> The solution Bernd came up with is this: if we find an equivalent
> register for an output reload, we remember it in reload_override_in
> rather than reg_rtx.
[...]
> Once we have the, the fix becomes straightforward: emit_reload_insns can
> check to see if we inherited an output reload. If we did, then there's
> no point in keeping that insn around, so we might as well delete it,
> which is what reload_as_needed has bee modified to do.
>
> Bootstrapping in progress on x86-64. OK to commit and backport to 4.5
> and 4.4?
Thanks for debugging and submitting the patch. I think ideally we'd get
a second opinion on this, but if no one comments within a week you can
check it in.
Bernd