* [PATCH] testsuite: Fix up sse2-v1ti-shift-3.c test [PR102986]
@ 2022-03-23 8:45 Jakub Jelinek
2022-03-23 9:03 ` Richard Biener
0 siblings, 1 reply; 2+ messages in thread
From: Jakub Jelinek @ 2022-03-23 8:45 UTC (permalink / raw)
To: Richard Biener, Roger Sayle; +Cc: gcc-patches
Hi!
This test is dg-do run and invokes UB when these rotate functions
are called with 0 as second argument. There are some other tests
that do this but they are dg-do compile only and not even call those
functions at all, so it IMHO doesn't matter that they are only well
defined for [1,127] and not [0,127].
The following patch fixes it, we pattern recognize both forms as rotates
and we emit identical assembly.
Tested on x86_64-linux -m32/-m64, ok for trunk?
2022-03-23 Jakub Jelinek <jakub@redhat.com>
PR target/102986
* gcc.target/i386/sse2-v1ti-shift-3.c (rotr_v1ti, rotl_v1ti, rotr_ti,
rotl_ti): Use -i&127 instead of 128-i to avoid UB on i == 0.
--- gcc/testsuite/gcc.target/i386/sse2-v1ti-shift-3.c.jj 2021-12-30 15:12:43.709143657 +0100
+++ gcc/testsuite/gcc.target/i386/sse2-v1ti-shift-3.c 2022-03-23 09:36:41.622181842 +0100
@@ -14,14 +14,14 @@ typedef __int128 ti;
uv1ti ashl_v1ti(uv1ti x, unsigned int i) { return x << i; }
uv1ti lshr_v1ti(uv1ti x, unsigned int i) { return x >> i; }
sv1ti ashr_v1ti(sv1ti x, unsigned int i) { return x >> i; }
-uv1ti rotr_v1ti(uv1ti x, unsigned int i) { return (x >> i) | (x << (128-i)); }
-uv1ti rotl_v1ti(uv1ti x, unsigned int i) { return (x << i) | (x >> (128-i)); }
+uv1ti rotr_v1ti(uv1ti x, unsigned int i) { return (x >> i) | (x << (-i&127)); }
+uv1ti rotl_v1ti(uv1ti x, unsigned int i) { return (x << i) | (x >> (-i&127)); }
uti ashl_ti(uti x, unsigned int i) { return x << i; }
uti lshr_ti(uti x, unsigned int i) { return x >> i; }
sti ashr_ti(sti x, unsigned int i) { return x >> i; }
-uti rotr_ti(uti x, unsigned int i) { return (x >> i) | (x << (128-i)); }
-uti rotl_ti(uti x, unsigned int i) { return (x << i) | (x >> (128-i)); }
+uti rotr_ti(uti x, unsigned int i) { return (x >> i) | (x << (-i&127)); }
+uti rotl_ti(uti x, unsigned int i) { return (x << i) | (x >> (-i&127)); }
void test(ti x)
{
Jakub
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [PATCH] testsuite: Fix up sse2-v1ti-shift-3.c test [PR102986]
2022-03-23 8:45 [PATCH] testsuite: Fix up sse2-v1ti-shift-3.c test [PR102986] Jakub Jelinek
@ 2022-03-23 9:03 ` Richard Biener
0 siblings, 0 replies; 2+ messages in thread
From: Richard Biener @ 2022-03-23 9:03 UTC (permalink / raw)
To: Jakub Jelinek; +Cc: Roger Sayle, gcc-patches
On Wed, 23 Mar 2022, Jakub Jelinek wrote:
> Hi!
>
> This test is dg-do run and invokes UB when these rotate functions
> are called with 0 as second argument. There are some other tests
> that do this but they are dg-do compile only and not even call those
> functions at all, so it IMHO doesn't matter that they are only well
> defined for [1,127] and not [0,127].
>
> The following patch fixes it, we pattern recognize both forms as rotates
> and we emit identical assembly.
>
> Tested on x86_64-linux -m32/-m64, ok for trunk?
OK.
> 2022-03-23 Jakub Jelinek <jakub@redhat.com>
>
> PR target/102986
> * gcc.target/i386/sse2-v1ti-shift-3.c (rotr_v1ti, rotl_v1ti, rotr_ti,
> rotl_ti): Use -i&127 instead of 128-i to avoid UB on i == 0.
>
> --- gcc/testsuite/gcc.target/i386/sse2-v1ti-shift-3.c.jj 2021-12-30 15:12:43.709143657 +0100
> +++ gcc/testsuite/gcc.target/i386/sse2-v1ti-shift-3.c 2022-03-23 09:36:41.622181842 +0100
> @@ -14,14 +14,14 @@ typedef __int128 ti;
> uv1ti ashl_v1ti(uv1ti x, unsigned int i) { return x << i; }
> uv1ti lshr_v1ti(uv1ti x, unsigned int i) { return x >> i; }
> sv1ti ashr_v1ti(sv1ti x, unsigned int i) { return x >> i; }
> -uv1ti rotr_v1ti(uv1ti x, unsigned int i) { return (x >> i) | (x << (128-i)); }
> -uv1ti rotl_v1ti(uv1ti x, unsigned int i) { return (x << i) | (x >> (128-i)); }
> +uv1ti rotr_v1ti(uv1ti x, unsigned int i) { return (x >> i) | (x << (-i&127)); }
> +uv1ti rotl_v1ti(uv1ti x, unsigned int i) { return (x << i) | (x >> (-i&127)); }
>
> uti ashl_ti(uti x, unsigned int i) { return x << i; }
> uti lshr_ti(uti x, unsigned int i) { return x >> i; }
> sti ashr_ti(sti x, unsigned int i) { return x >> i; }
> -uti rotr_ti(uti x, unsigned int i) { return (x >> i) | (x << (128-i)); }
> -uti rotl_ti(uti x, unsigned int i) { return (x << i) | (x >> (128-i)); }
> +uti rotr_ti(uti x, unsigned int i) { return (x >> i) | (x << (-i&127)); }
> +uti rotl_ti(uti x, unsigned int i) { return (x << i) | (x >> (-i&127)); }
>
> void test(ti x)
> {
>
> Jakub
>
>
--
Richard Biener <rguenther@suse.de>
SUSE Software Solutions Germany GmbH, Maxfeldstrasse 5, 90409 Nuernberg,
Germany; GF: Ivo Totev; HRB 36809 (AG Nuernberg)
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2022-03-23 9:03 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-03-23 8:45 [PATCH] testsuite: Fix up sse2-v1ti-shift-3.c test [PR102986] Jakub Jelinek
2022-03-23 9:03 ` Richard Biener
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).