https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #22 from sandra at gcc dot gnu.org ---
My nios2-elf test results look good now with this patch. Thanks!
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #21 from CVS Commits ---
The master branch has been updated by Richard Biener :
https://gcc.gnu.org/g:dd9ca9d770a18ce4b16d867f49fef3293b483ff5
commit r10-7635-gdd9ca9d770a18ce4b16d867f49fef3293b483ff5
Author: Richard Biener
Date:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #20 from Richard Biener ---
So for the CSE issue we go through the equivalence chain and find
(gdb) p debug_rtx (p->exp)
(mem/j:SI (reg/v/f:SI 48 [ ptr ]) [1 MEM[(struct aa *)ptr_1(D)].a.u.i+0 S4
A32])
5076 if (GET_CODE
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #19 from Richard Biener ---
(In reply to sandra from comment #18)
> Hmmm, it looks to me like things are going wrong in the tree fre1 pass too.
> I commented out the redundant zero store in the test case to see what would
> happen, l
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #18 from sandra at gcc dot gnu.org ---
Hmmm, it looks to me like things are going wrong in the tree fre1 pass too. I
commented out the redundant zero store in the test case to see what would
happen, like
long __attribute__((noipa))
f
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #17 from Richard Biener ---
x86 splits large stores only after reload it seems. Also on x86 GIMPLE
store-merging triggers, merging the two = 0 stores so I need -fno-store-merging
to
even get two stores there. Still the issue in CSE
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #16 from Richard Biener ---
OK, now looking myself. RTL expansion creates
(insn 8 7 9 2 (set (mem/j:SI (reg/v/f:SI 47 [ bv ]) [1 bv_3(D)->b.u.f+0 S4
A32])
(reg:SI 49)) "t.c":12:13 -1
(nil))
(insn 9 8 10 2 (set (mem/j:SI
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #15 from sandra at gcc dot gnu.org ---
Hmmm. I've gone over this code 2 or 3 times now, and I'm still convinced the
problem is in the alias analysis, not the scheduler.
I've stepped deeper into the code in the debugger, and here is t
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #14 from Richard Biener ---
(In reply to sandra from comment #13)
> Well, no. The problem is that the scheduler is moving
>
> ldw r2, 0(r4)
>
> ahead of
>
> stw zero, 0(r5)
>
> which is incorrect because
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #13 from sandra at gcc dot gnu.org ---
Well, no. The problem is that the scheduler is moving
ldw r2, 0(r4)
ahead of
stw zero, 0(r5)
which is incorrect because the pointers in r4 and r5 are aliases.
So at
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #12 from rguenther at suse dot de ---
On March 27, 2020 9:19:33 PM GMT+01:00, "sandra at gcc dot gnu.org"
wrote:
>https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
>
>--- Comment #11 from sandra at gcc dot gnu.org ---
>RTL before sc
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #11 from sandra at gcc dot gnu.org ---
RTL before sched2 does look sane and similar to that generated for x86 with
-m32.
I've been trying to step through sched2. I think that where things get
interesting is the call to true_dependenc
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #10 from Richard Biener ---
(In reply to sandra from comment #9)
> Both the new test cases are failing on nios2 at -Os, -O2, and -O3. I've
> done some analysis, but I'm not sure exactly where the problem lies, and
> whether this is a
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
sandra at gcc dot gnu.org changed:
What|Removed |Added
CC||sandra at gcc dot gnu.org
---
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
Richard Biener changed:
What|Removed |Added
Keywords||wrong-code
Status|ASSIGNED
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #7 from CVS Commits ---
The master branch has been updated by Richard Guenther :
https://gcc.gnu.org/g:3d6fd7ce6dc4b6baa11920387d62dc001980aa70
commit r10-6993-g3d6fd7ce6dc4b6baa11920387d62dc001980aa70
Author: Richard Biener
Date:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #6 from Richard Biener ---
Created attachment 47948
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=47948&action=edit
DSE part
I'm now testing the combined (and will squash if that succeeds).
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
Richard Biener changed:
What|Removed |Added
Attachment #47922|0 |1
is obsolete|
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #4 from Richard Biener ---
OK, so the fix works but on x86_64 with -m32 RTL opts (sched2) end up wrecking
things...
foo:
.LFB0:
.cfi_startproc
movl4(%esp), %eax
movl8(%esp), %edx
movl$1, (%
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
--- Comment #3 from Richard Biener ---
Created attachment 47922
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=47922&action=edit
patch I am testing
This is what I have now. Got to factor out that alias set. In VN consider
passing ao_ref
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
Richard Biener changed:
What|Removed |Added
CC||law at gcc dot gnu.org
--- Comment #2 f
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93946
Richard Biener changed:
What|Removed |Added
Status|UNCONFIRMED |ASSIGNED
Last reconfirmed|
22 matches
Mail list logo
