public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/106279] New: reload problem on arm iwmmxt
@ 2022-07-13 8:01 dcb314 at hotmail dot com
2022-07-13 9:10 ` [Bug target/106279] " wilco at gcc dot gnu.org
` (4 more replies)
0 siblings, 5 replies; 6+ messages in thread
From: dcb314 at hotmail dot com @ 2022-07-13 8:01 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
Bug ID: 106279
Summary: reload problem on arm iwmmxt
Product: gcc
Version: 12.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: dcb314 at hotmail dot com
Target Milestone: ---
For the C testsuite file ./gcc.target/aarch64/vect_unary_1.c, compiled
by an ARM cross compiler, with target arch set to iwmmxt, does this:
$ /home/dcb/raspberrypi/results/bin/arm-linux-gnueabihf-gcc -c -march=iwmmxt
bug827.c
during RTL pass: reload
./gcc.target/aarch64/vect_unary_1.c: In function ‘test2_int_clz_int’:
./gcc.target/aarch64/vect_unary_1.c:15:1: internal compiler error: maximum
number of generated reload insns per insn achieved (90)
15 | }
\
| ^
./gcc.target/aarch64/vect_unary_1.c:161:1: note: in expansion of macro ‘TEST2’
161 | TEST2 (int, clz, int)
| ^~~~~
0xe06010 lra_constraints(bool)
/home/dcb/gcc/trunk.git/gcc/lra-constraints.cc:5126
Similar things happen with target arch set to iwmmxt2.
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug target/106279] reload problem on arm iwmmxt
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
@ 2022-07-13 9:10 ` wilco at gcc dot gnu.org
2022-07-13 9:15 ` dcb314 at hotmail dot com
` (3 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: wilco at gcc dot gnu.org @ 2022-07-13 9:10 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
Wilco <wilco at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |wilco at gcc dot gnu.org
--- Comment #1 from Wilco <wilco at gcc dot gnu.org> ---
iwmmxt has been dead for 2 decades now - it's support has most likely
bitrotted, so I'm surprised anyone is trying to use it...
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug target/106279] reload problem on arm iwmmxt
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
2022-07-13 9:10 ` [Bug target/106279] " wilco at gcc dot gnu.org
@ 2022-07-13 9:15 ` dcb314 at hotmail dot com
2022-07-13 9:17 ` wilco at gcc dot gnu.org
` (2 subsequent siblings)
4 siblings, 0 replies; 6+ messages in thread
From: dcb314 at hotmail dot com @ 2022-07-13 9:15 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
--- Comment #2 from David Binderman <dcb314 at hotmail dot com> ---
(In reply to Wilco from comment #1)
> iwmmxt has been dead for 2 decades now - it's support has most likely
> bitrotted, so I'm surprised anyone is trying to use it...
Time to remove support for it, or just mark this bug report as WONTFIX ?
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug target/106279] reload problem on arm iwmmxt
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
2022-07-13 9:10 ` [Bug target/106279] " wilco at gcc dot gnu.org
2022-07-13 9:15 ` dcb314 at hotmail dot com
@ 2022-07-13 9:17 ` wilco at gcc dot gnu.org
2022-09-01 13:01 ` rearnsha at gcc dot gnu.org
2022-09-01 13:06 ` rearnsha at gcc dot gnu.org
4 siblings, 0 replies; 6+ messages in thread
From: wilco at gcc dot gnu.org @ 2022-07-13 9:17 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
--- Comment #3 from Wilco <wilco at gcc dot gnu.org> ---
(In reply to David Binderman from comment #2)
> (In reply to Wilco from comment #1)
> > iwmmxt has been dead for 2 decades now - it's support has most likely
> > bitrotted, so I'm surprised anyone is trying to use it...
>
> Time to remove support for it, or just mark this bug report as WONTFIX ?
Both!
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug target/106279] reload problem on arm iwmmxt
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
` (2 preceding siblings ...)
2022-07-13 9:17 ` wilco at gcc dot gnu.org
@ 2022-09-01 13:01 ` rearnsha at gcc dot gnu.org
2022-09-01 13:06 ` rearnsha at gcc dot gnu.org
4 siblings, 0 replies; 6+ messages in thread
From: rearnsha at gcc dot gnu.org @ 2022-09-01 13:01 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2022-09-01
Ever confirmed|0 |1
Status|UNCONFIRMED |NEW
--- Comment #4 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
In a general cross compiler you need: -march=iwmmxt -mfpu=vfp -mfloat-abi=hard
to reproduce.
This seems to need a combination of VFP, iWMMX and the hard-float ABI. I don't
think we've ever claimed to support this. Whilst there may be a few devices
with both iWMMX and VFP present, I think that's mostly for backwards
compatibility.
So I think we should resolve this as WONTFIX, but it would be nice if the
compiler could reject this upfront rather than ICEing later on.
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Bug target/106279] reload problem on arm iwmmxt
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
` (3 preceding siblings ...)
2022-09-01 13:01 ` rearnsha at gcc dot gnu.org
@ 2022-09-01 13:06 ` rearnsha at gcc dot gnu.org
4 siblings, 0 replies; 6+ messages in thread
From: rearnsha at gcc dot gnu.org @ 2022-09-01 13:06 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106279
--- Comment #5 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
Technical note: the compiler is missing pattern alternatives to move iwmmxt
register values to VFP registers. There's no way to do this directly, so we'd
need to allow copying via core registers.
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2022-09-01 13:06 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-13 8:01 [Bug target/106279] New: reload problem on arm iwmmxt dcb314 at hotmail dot com
2022-07-13 9:10 ` [Bug target/106279] " wilco at gcc dot gnu.org
2022-07-13 9:15 ` dcb314 at hotmail dot com
2022-07-13 9:17 ` wilco at gcc dot gnu.org
2022-09-01 13:01 ` rearnsha at gcc dot gnu.org
2022-09-01 13:06 ` rearnsha at gcc dot gnu.org
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).