> > Since Catherine asked for further info then I will leave her to say if she
> is
> > happy to accept on this basis.
> >
>
> I withdraw my request for a testcase.
>
> Catherine
Committed as r220200.
Regards,
Robert
> -Original Message-
> From: Matthew Fortune [mailto:matthew.fort...@imgtec.com]
> Sent: Tuesday, January 27, 2015 7:19 AM
> To: Richard Sandiford
> Cc: Robert Suchanek; gcc-patches@gcc.gnu.org; Moore, Catherine
> Subject: RE: [PATCH RFA MIPS] Prohibit vector mod
Richard Sandiford writes:
> Matthew Fortune writes:
> >> 2015-01-23 Robert Suchanek
> >>
> >>* config/mips/mips.c (mips_hard_regno_mode_ok_p): Prohibit
> >> accumulators
> >>for all vector modes.
> >
> > This seems like a genuine bug and although it can only be triggered by
> > loongso
Matthew Fortune writes:
>> 2015-01-23 Robert Suchanek
>>
>> * config/mips/mips.c (mips_hard_regno_mode_ok_p): Prohibit
>> accumulators
>> for all vector modes.
>
> This seems like a genuine bug and although it can only be triggered by
> loongson or paired-single support it probably q
Hi Catherine,
> The patch looks reasonable, but I'd like to see a test case that fails before
> we agree to include for GCC 5.0.
It's possible to reproduce ICEs with SVN revision 212354 on mipsel-linux-gnu
target. The concerned tests are
loongson-simd.c and loongson-shift-count-truncated-1.c in
> -Original Message-
> From: Matthew Fortune [mailto:matthew.fort...@imgtec.com]
> Sent: Friday, January 23, 2015 2:51 PM
> To: Robert Suchanek; gcc-patches@gcc.gnu.org
> Cc: Moore, Catherine
> Subject: RE: [PATCH RFA MIPS] Prohibit vector modes in accumulators
>
&
> 2015-01-23 Robert Suchanek
>
> * config/mips/mips.c (mips_hard_regno_mode_ok_p): Prohibit
> accumulators
> for all vector modes.
This seems like a genuine bug and although it can only be triggered by
loongson or paired-single support it probably qualifies for fixing.
My suspicion
Hi,
A bug was exposed by LRA for loongson-shift-count-truncated-1.c and
loongson-simd.c with -O0 optimization. These testcases were ICEing
with 'Max. number of generated reloads insns per insn is achieved (90)'
error.
The problem appears to be with vector modes where contents of memory
is meant
