This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [Patch ARM] Fix PR40482.
- From: Richard Earnshaw <rearnsha at arm dot com>
- To: Ramana Radhakrishnan <ramana dot radhakrishnan at arm dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Thu, 18 Jun 2009 11:43:30 +0100
- Subject: Re: [Patch ARM] Fix PR40482.
- References: <001301c9efff$29e67d80$7db37880$@radhakrishnan@arm.com>
On Thu, 2009-06-18 at 11:25 +0100, Ramana Radhakrishnan wrote:
> Hi,
>
> This fixes PR40482. In configurations with HOST_WIDE_INT being set to 64
> bits we now need to fix thumb_shiftable_const to only work on the lower 32
> bits. This allows us to generate more constants inline rather than loading
> them from memory.
>
> Ok to commit if no regressions on a cross arm-none-eabi testsuite ?
>
>
> cheers
> Ramana
>
> 2009-06-18 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
>
> PR target/40482
> * config/arm/arm.c (thumb_shiftable_const): Truncate val to
> 32 bits.
> * config/arm/arm.md: Likewise.
>
> 2009-06-18 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
>
> PR target/40482
> * gcc.target/arm/pr40482.c: New test.
>
> ---
>
> Ramana Radhakrishnan
> GNU Tools Engineer
> ARM Ltd. (www.arm.com)
The test will probably fail if the underlying CPU supports thumb2 (since
we then won't get an lsl instruction. For this test it should be
sufficient to check that LDR was not used. It would be nice if for
certain tests like this we could control the entire environment rather
than having to deal with the dejagnu site file messing with the options
used.
Ok with that change.