https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
Iain Sandoe changed:
What|Removed |Added
Status|REOPENED|RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
Jakub Jelinek changed:
What|Removed |Added
Target Milestone|9.0 |9.2
--- Comment #36 from Jakub Jelinek
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #35 from Alan Modra ---
Author: amodra
Date: Sat Feb 9 12:44:02 2019
New Revision: 268722
URL: https://gcc.gnu.org/viewcvs?rev=268722&root=gcc&view=rev
Log:
[RS6000] Correct save_reg_p
PR target/88343
* config/rs600
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #34 from Iain Sandoe ---
(In reply to Alan Modra from comment #33)
> It looks to me like that hunk is just removing some dead code, so it doesn't
> matter whether it stays or goes.
yes, just a tidy-up, should not affect function (but
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #33 from Alan Modra ---
It looks to me like that hunk is just removing some dead code, so it doesn't
matter whether it stays or goes.
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #32 from Iain Sandoe ---
Thanks for the patches Alan.
Note that 7 and 8 back ports might need this hunk which was applied to trunk
but removed from 7 and 8 when we backed out the change because of breakage
there.
@@ -23970,13 +23974
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #31 from Alan Modra ---
Author: amodra
Date: Sat Feb 9 08:39:58 2019
New Revision: 268715
URL: https://gcc.gnu.org/viewcvs?rev=268715&root=gcc&view=rev
Log:
[RS6000] Correct save_reg_p
PR target/88343
* config/rs600
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #30 from Alan Modra ---
Author: amodra
Date: Fri Feb 8 22:20:58 2019
New Revision: 268708
URL: https://gcc.gnu.org/viewcvs?rev=268708&root=gcc&view=rev
Log:
[RS6000] Correct save_reg_p
Fixes lack of r30 save/restore on
// -m32 -fp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #29 from Segher Boessenkool ---
This:
===
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index 401e719..f0adef7 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6000.c
@@ -37988,7 +37988,10 @@ r
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #28 from Segher Boessenkool ---
(In reply to Alan Modra from comment #27)
> And possibly -msecure-plt
Yeah that does the trick, thanks :-)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #27 from Alan Modra ---
And possibly -msecure-plt
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #26 from Alan Modra ---
> I don't see that; I get
You need -fpic
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #25 from Segher Boessenkool ---
I don't see that; I get
f1:
.LFB0:
.cfi_startproc
b foo@plt
.cfi_endproc
.LFE0:
What extra options do I need? Or, what other target (you didn't say, I used
powerpc-linux).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #24 from Iain Sandoe ---
so, it seems that there are more cases where the RS6000_PIC_OFFSET_TABLE_REGNUM
is used without setting the uses_pic_offset_table.
We can easily back the change out to "fix" master - but that seems to be
pape
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
Alan Modra changed:
What|Removed |Added
CC||amodra at gmail dot com
--- Comment #23 fro
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #22 from joseph at codesourcery dot com ---
A large proportion of the glibc libm tests still segfault for
powerpc-linux-gnu soft-float with that patch applied to GCC trunk.
(Although I don't see nextafter tests among those falling
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #21 from Segher Boessenkool ---
Before the holidays I did this patch:
===
diff --git a/gcc/config/rs6000/rs6000.c b/gcc/config/rs6000/rs6000.c
index fa5f032..2ffe7d9 100644
--- a/gcc/config/rs6000/rs6000.c
+++ b/gcc/config/rs6000/rs6
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #20 from Iain Sandoe ---
(In reply to Joseph S. Myers from comment #19)
> Created attachment 45330 [details]
> Bad assembly (from trunk r267560 with the patch still present)
Thanks, Joseph, that's very helpful.
.. so the code plainl
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #19 from Joseph S. Myers ---
Created attachment 45330
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45330&action=edit
Bad assembly (from trunk r267560 with the patch still present)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #18 from Joseph S. Myers ---
Created attachment 45329
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45329&action=edit
Good assembly (with the GCC patch reverted)
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #17 from Joseph S. Myers ---
Created attachment 45328
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=45328&action=edit
Preprocessed source
Replacing s_nextafterf.os built with trunk with that patch reverted, by the
same file as
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #16 from joseph at codesourcery dot com ---
On Thu, 3 Jan 2019, iains at gcc dot gnu.org wrote:
> Unfortunately, there doesn't appear to be anything in the GCC test suite that
> catches this (all languages reg-strap was clean, on gcc
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #15 from Iain Sandoe ---
Author: iains
Date: Thu Jan 3 08:45:35 2019
New Revision: 267543
URL: https://gcc.gnu.org/viewcvs?rev=267543&root=gcc&view=rev
Log:
revert fix for pr88343
This causes problems for soft-f on GLIBC, see pr co
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #14 from Iain Sandoe ---
Author: iains
Date: Thu Jan 3 08:34:41 2019
New Revision: 267542
URL: https://gcc.gnu.org/viewcvs?rev=267542&root=gcc&view=rev
Log:
revert fix for pr88343
causes problems with soft-fp in GLIBC, see pr comme
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #13 from Iain Sandoe ---
(In reply to Joseph S. Myers from comment #11)
> This change results in miscompilation of glibc for 32-bit soft-float powerpc
> (symptoms: many libm tests as run by "make regen-ulps" either segfault, or
> prod
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
Segher Boessenkool changed:
What|Removed |Added
Status|RESOLVED|REOPENED
Resolution|FIXED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #11 from Joseph S. Myers ---
This change results in miscompilation of glibc for 32-bit soft-float powerpc
(symptoms: many libm tests as run by "make regen-ulps" either segfault, or
produce spurious errors that don't occur with compile
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
Iain Sandoe changed:
What|Removed |Added
Status|NEW |RESOLVED
Resolution|---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #9 from Iain Sandoe ---
Author: iains
Date: Sun Dec 30 13:20:19 2018
New Revision: 267477
URL: https://gcc.gnu.org/viewcvs?rev=267477&root=gcc&view=rev
Log:
fix PR target/88343 for 32b powerpc.
2018-12-30 Iain Sandoe
bac
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88343
--- Comment #8 from Iain Sandoe ---
Author: iains
Date: Sun Dec 23 21:17:46 2018
New Revision: 267387
URL: https://gcc.gnu.org/viewcvs?rev=267387&root=gcc&view=rev
Log:
fix PR target/88343 by backporting r267049
The PR is about unnecessary save
30 matches
Mail list logo
