[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-30 Thread akrl at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

akrl at gcc dot gnu.org changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 Resolution|--- |FIXED

--- Comment #10 from akrl at gcc dot gnu.org ---
Thanks, closing.

[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-29 Thread iains at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #9 from Iain Sandoe  ---
(In reply to Andrea Corallo from comment #8)
> "iains at gcc dot gnu.org via Gcc-bugs"  writes:
> [...]
> > unfortunately, I've not been able to test since you applied this - currently
> > bootstrap is broken on aarch64-darwin for reasons outside our control (new
> > security provisions stopping the gcc/build/gen* programs from running).  If 
> > it
> > could stay open for a few more days, while we try to find a fix for that - 
> > (or
> > close it and we can reopen if needed).
> 
> Sure no rush, thanks for testing it!

I managed to do a manual build on arm64-darwin, and the ICE is gone.  Not sure
if the flags are supported by the hardware but that's a different issue.  Fine
to close this from my pov.

We get this for the simple test-case (mach-o/darwin format object):
.arch armv8-a
.text
.align  2
.globl _main
_main:
LFB0:
.cfi_startproc
sub sp, sp, #16
.cfi_def_cfa_offset 16
mrs x0, fpcr
str w0, [sp, 12]
mov w0, 0
add sp, sp, 16
.cfi_def_cfa_offset 0
ret
.cfi_endproc
LFE0:
.ident  "GCC: (GNU) 11.0.0 20200929 (experimental)"
.subsections_via_symbols

Re: [Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-24 Thread Andrea Corallo
"iains at gcc dot gnu.org via Gcc-bugs"  writes:
[...]
> unfortunately, I've not been able to test since you applied this - currently
> bootstrap is broken on aarch64-darwin for reasons outside our control (new
> security provisions stopping the gcc/build/gen* programs from running).  If it
> could stay open for a few more days, while we try to find a fix for that - (or
> close it and we can reopen if needed).

Sure no rush, thanks for testing it!


[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-24 Thread andrea.corallo at arm dot com via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #8 from Andrea Corallo  ---
"iains at gcc dot gnu.org via Gcc-bugs"  writes:
[...]
> unfortunately, I've not been able to test since you applied this - currently
> bootstrap is broken on aarch64-darwin for reasons outside our control (new
> security provisions stopping the gcc/build/gen* programs from running).  If it
> could stay open for a few more days, while we try to find a fix for that - (or
> close it and we can reopen if needed).

Sure no rush, thanks for testing it!

[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-24 Thread iains at gcc dot gnu.org via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #7 from Iain Sandoe  ---
(In reply to Andrea Corallo from comment #6)
> I believe f5e73de00e9c853ce65333efada7409b0d00f758 should have fixed
> this.
> 
> Okay to close?

unfortunately, I've not been able to test since you applied this - currently
bootstrap is broken on aarch64-darwin for reasons outside our control (new
security provisions stopping the gcc/build/gen* programs from running).  If it
could stay open for a few more days, while we try to find a fix for that - (or
close it and we can reopen if needed).

[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-24 Thread andrea.corallo at arm dot com via Gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #6 from Andrea Corallo  ---
I believe f5e73de00e9c853ce65333efada7409b0d00f758 should have fixed
this.

Okay to close?

Thanks

  Andrea

Re: [Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-24 Thread Andrea Corallo
I believe f5e73de00e9c853ce65333efada7409b0d00f758 should have fixed
this.

Okay to close?

Thanks

  Andrea


[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-21 Thread cvs-commit at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #5 from CVS Commits  ---
The master branch has been updated by Andrea Corallo :

https://gcc.gnu.org/g:f5e73de00e9c853ce65333efada7409b0d00f758

commit r11-3317-gf5e73de00e9c853ce65333efada7409b0d00f758
Author: Andrea Corallo 
Date:   Mon Sep 14 14:47:24 2020 +0100

aarch64: Fix ICE on fpsr fpcr getters [PR96968]

gcc/ChangeLog

2020-09-14  Andrea Corallo  

PR target/96968
* config/aarch64/aarch64-builtins.c
(aarch64_expand_fpsr_fpcr_setter): Fix comment nit.
(aarch64_expand_fpsr_fpcr_getter): New function, expand these
getters using expand_insn machinery.
(aarch64_general_expand_builtin): Make use of.

gcc/testsuite/ChangeLog

2020-09-14  Andrea Corallo  

PR target/96968
* gcc.target/aarch64/pr96968.c: New test.

Re: [Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-14 Thread Andrea Corallo
Patch posted at:

https://gcc.gnu.org/pipermail/gcc-patches/2020-September/553882.html


[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-14 Thread andrea.corallo at arm dot com
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #4 from Andrea Corallo  ---
Patch posted at:

https://gcc.gnu.org/pipermail/gcc-patches/2020-September/553882.html

[Bug target/96968] aarch64 : ICE in vregs or expand pass, lowering __builtin_aarch64_get_{fpcr,fpsr,fpcr64,fpsr64}

2020-09-08 Thread akrl at gcc dot gnu.org
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96968

--- Comment #3 from akrl at gcc dot gnu.org ---
Confirm this was introduced by:

0d7e5fa655e59c99035bf94a46c912e369bb9fa0
"aarch64: Add 64 bit setter getter fpsr fpcr"

I'll have a look as soon as I have a slot of time.