public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
* [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler
@ 2020-11-14 11:02 doko at debian dot org
2020-11-16 7:17 ` [Bug target/97827] " rguenth at gcc dot gnu.org
` (18 more replies)
0 siblings, 19 replies; 20+ messages in thread
From: doko at debian dot org @ 2020-11-14 11:02 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Bug ID: 97827
Summary: [11 Regression] bootstrap error building the
amdgcn-amdhsa offload compiler
Product: gcc
Version: 10.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: doko at debian dot org
Target Milestone: ---
builds fine with trunk 20201112, fails with trunk 20201113. newlib 3.3.0 is
used.
[...]
/tmp/ccpzv56q.s:1129:2: error: changed section flags for .rodata.cst8,
expected: 0x12
.section .rodata.cst8
^
/tmp/ccpzv56q.s:1129:2: error: changed section entsize for .rodata.cst8,
expected: 8
.section .rodata.cst8
^
make[10]: *** [Makefile:980: lib_a-wcstod.o] Error 1
make[10]: *** Waiting for unfinished jobs....
/tmp/ccDmKs3n.s:4450:2: error: changed section flags for .rodata.cst8,
expected: 0x12
.section .rodata.cst8
^
/tmp/ccDmKs3n.s:4450:2: error: changed section entsize for .rodata.cst8,
expected: 8
.section .rodata.cst8
^
/tmp/ccDmKs3n.s:4643:2: error: changed section flags for .rodata.cst8,
expected: 0x12
.section .rodata.cst8
^
/tmp/ccDmKs3n.s:4643:2: error: changed section entsize for .rodata.cst8,
expected: 8
.section .rodata.cst8
^
make[10]: *** [Makefile:944: lib_a-strtod.o] Error 1
make[10]: Leaving directory
'/packages/gcc/11/gcc-11-11-20201113/build-gcn/amdgcn-amdhsa/gfx900/newlib/libc/stdlib'
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
@ 2020-11-16 7:17 ` rguenth at gcc dot gnu.org
2020-11-16 8:06 ` doko at debian dot org
` (17 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: rguenth at gcc dot gnu.org @ 2020-11-16 7:17 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target| |gcn
Version|10.2.1 |11.0
Target Milestone|--- |11.0
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
2020-11-16 7:17 ` [Bug target/97827] " rguenth at gcc dot gnu.org
@ 2020-11-16 8:06 ` doko at debian dot org
2020-11-17 11:44 ` [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11 burnus at gcc dot gnu.org
` (16 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: doko at debian dot org @ 2020-11-16 8:06 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #1 from Matthias Klose <doko at debian dot org> ---
my bad, this is not a regression. Seen when building the offload compiler with
LLVM 11 instead of LLVM 10 or 9
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
2020-11-16 7:17 ` [Bug target/97827] " rguenth at gcc dot gnu.org
2020-11-16 8:06 ` doko at debian dot org
@ 2020-11-17 11:44 ` burnus at gcc dot gnu.org
2020-11-17 12:45 ` burnus at gcc dot gnu.org
` (15 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-11-17 11:44 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |ams at gcc dot gnu.org,
| |burnus at gcc dot gnu.org
--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Changed in LLVM in https://reviews.llvm.org/D73999
cf. https://lists.llvm.org/pipermail/llvm-dev/2020-February/139390.html
and related Binutils discussion:
https://sourceware.org/legacy-ml/binutils/2020-02/msg00093.html
LLVM commit 75af9da755721123e62b45cd0bc0c5e688a9722a:
" GNU as started to emit warnings for changed sh_type or sh_flags in 2000.
GNU as>=2.35 will emit errors for most sh_type/sh_flags change, and error
for entsize change."
Cf. also http://web.mit.edu/rhel-doc/3/rhel-as-en-3/section.html
A grep shows:
.section .rodata.cst8
.section .rodata.cst8
.section .rodata.cst8,"aM",@progbits,8
Codewise, we have:
mergeable_constant_section (mode=E_DFmode, align=64, flags=0)
which calls:
884 flags |= (align / 8) | SECTION_MERGE;
885 return get_section (name, flags, NULL);
with flags == 32776.
In the latter:
flags |= SECTION_NAMED;
→ 2129928
and as "slot = section_htab->find_slot_with_hash" → *slot == NULL:
sect = ggc_alloc<section> ();
sect->named.common.flags = flags;
sect->named.name = ggc_strdup (name);
sect->named.decl = decl;
However, once the same function is called, we have:
sect->common.flags == 3178504
while still/again:
flags == 2129928
Next comes the call to:
switch_to_section
which ends with
new_section->common.flags |= SECTION_DECLARED;
which is
3178504
If we look at default_elf_asm_named_section, we find:
/* If we have already declared this section, we can use an
abbreviated form to switch back to it -- unless this section is
part of a COMDAT groups, in which case GAS requires the full
declaration every time. */
if (!(HAVE_COMDAT_GROUP && (flags & SECTION_LINKONCE))
&& (flags & SECTION_DECLARED))
{
fprintf (asm_out_file, "\t.section\t%s\n", name);
return;
}
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (2 preceding siblings ...)
2020-11-17 11:44 ` [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11 burnus at gcc dot gnu.org
@ 2020-11-17 12:45 ` burnus at gcc dot gnu.org
2020-11-17 16:31 ` burnus at gcc dot gnu.org
` (14 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-11-17 12:45 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #3 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Now filled https://bugs.llvm.org/show_bug.cgi?id=48201 for this
LLVM and GNU as ("gas") tried to align, cf. discussion
https://sourceware.org/legacy-ml/binutils/2020-02/msg00091.html
and the GAS patch to turn a warning to an error:
https://sourceware.org/legacy-ml/binutils/2020-02/msg00129.html
However, while llvm/lib/MC/MCParser/ELFAsmParser.cpp always issues an error:
+ if (Section->getType() != Type)
+ Error(loc, "changed section type for " + SectionName + ", expected: 0x" +
+ utohexstr(Section->getType()));
etc.
The GNU assembler only does so if flags have been specified,
gas/config/obj-elf.c:
if (attr != 0)
{
/* If section attributes are specified the second time we see a
particular section, then check that they are the same as we
saw the first time. */
if (((old_sec->flags ^ flags)
...
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (3 preceding siblings ...)
2020-11-17 12:45 ` burnus at gcc dot gnu.org
@ 2020-11-17 16:31 ` burnus at gcc dot gnu.org
2020-11-18 12:32 ` burnus at gcc dot gnu.org
` (13 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-11-17 16:31 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #4 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Draft patch (untested):
diff --git a/gcc/varasm.c b/gcc/varasm.c
index ada99940f65..51a507393a8 100644
--- a/gcc/varasm.c
+++ b/gcc/varasm.c
@@ -6738,9 +6738,11 @@ default_elf_asm_named_section (const char *name,
unsigned int flags,
/* If we have already declared this section, we can use an
abbreviated form to switch back to it -- unless this section is
part of a COMDAT groups, in which case GAS requires the full
- declaration every time. */
+ declaration every time. LLVM's MC linker requires that the
+ flags are identical, thus avoid the abbreviated form with MERGE. */
if (!(HAVE_COMDAT_GROUP && (flags & SECTION_LINKONCE))
- && (flags & SECTION_DECLARED))
+ && (flags & SECTION_DECLARED)
+ && !(flags & SECTION_MERGE))
{
fprintf (asm_out_file, "\t.section\t%s\n", name);
return;
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (4 preceding siblings ...)
2020-11-17 16:31 ` burnus at gcc dot gnu.org
@ 2020-11-18 12:32 ` burnus at gcc dot gnu.org
2020-11-18 12:46 ` jakub at gcc dot gnu.org
` (12 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-11-18 12:32 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #5 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Patch: https://gcc.gnu.org/pipermail/gcc-patches/2020-November/559487.html
First review regards this as LLVM bug.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (5 preceding siblings ...)
2020-11-18 12:32 ` burnus at gcc dot gnu.org
@ 2020-11-18 12:46 ` jakub at gcc dot gnu.org
2020-11-24 20:01 ` burnus at gcc dot gnu.org
` (11 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: jakub at gcc dot gnu.org @ 2020-11-18 12:46 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Might not be a LLVM bug if llvm-as doesn't pretend to be assembler compatible
with other assemblers. But in that case we should treat it as a target with
specific assembler requirements (e.g. like Darwin).
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (6 preceding siblings ...)
2020-11-18 12:46 ` jakub at gcc dot gnu.org
@ 2020-11-24 20:01 ` burnus at gcc dot gnu.org
2020-12-11 17:02 ` burnus at gcc dot gnu.org
` (10 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-11-24 20:01 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #7 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Submitted LLVM patch at https://reviews.llvm.org/D92052
If it gets accepted for LLVM + backported to 11, we are done.
Otherwise, we have to proceed as suggested in the email thread.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (7 preceding siblings ...)
2020-11-24 20:01 ` burnus at gcc dot gnu.org
@ 2020-12-11 17:02 ` burnus at gcc dot gnu.org
2020-12-14 18:31 ` i at maskray dot me
` (9 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-12-11 17:02 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|UNCONFIRMED |RESOLVED
Resolution|--- |WONTFIX
--- Comment #8 from Tobias Burnus <burnus at gcc dot gnu.org> ---
FIXED in LLVM 'main' in commit 1deff4009e0ae661b03682901bf6932297ce7ea1
→ https://reviews.llvm.org/D92052 + https://bugs.llvm.org/show_bug.cgi?id=48201
Close as WONTFIX on the GCC side.
I try to get it backported to LLVM 11.0.1 or, failing that, LLVM 11.0.2.
Thanks Matthias for pointing out the issue.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (8 preceding siblings ...)
2020-12-11 17:02 ` burnus at gcc dot gnu.org
@ 2020-12-14 18:31 ` i at maskray dot me
2020-12-14 18:38 ` i at maskray dot me
` (8 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: i at maskray dot me @ 2020-12-14 18:31 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Fangrui Song <i at maskray dot me> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |i at maskray dot me
--- Comment #9 from Fangrui Song <i at maskray dot me> ---
I want to know whether this is really a wontfix on GCC's side.
Richard Sandiford on
https://gcc.gnu.org/pipermail/gcc-patches/2020-November/559572.html
"I'm not saying we should bend over backwards to support difficult
quirks. But here we're talking about a choice between (a) doing
something that works “everywhere” unconditionally (and keeping things
simple) vs. (b) having both code that takes a shortcut and code that
doesn't take a shortcut and trying to predict which one we should do."
This makes a lot of sense to me. For the LLVM "fix", we had not known this PR
before https://reviews.llvm.org/D92052#2452577
To me personally, I might have a different opinion if I knew this is not an
entire dead end on gcc -S output.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (9 preceding siblings ...)
2020-12-14 18:31 ` i at maskray dot me
@ 2020-12-14 18:38 ` i at maskray dot me
2020-12-15 10:26 ` burnus at gcc dot gnu.org
` (7 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: i at maskray dot me @ 2020-12-14 18:38 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #10 from Fangrui Song <i at maskray dot me> ---
Note: the section key is not just (name, group name "G"). It is a quadruple:
(name, group name "G", linked-to "o", unique ID)
Keeping just name works for the simplest case.
If GCC decides to support PR95095 -fno-unique-section-names, unique ID can be
common. https://sourceware.org/bugzilla/show_bug.cgi?id=25490#c3 added the
support for `.section ....,unique` to GNU as.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (10 preceding siblings ...)
2020-12-14 18:38 ` i at maskray dot me
@ 2020-12-15 10:26 ` burnus at gcc dot gnu.org
2020-12-22 16:17 ` doko at debian dot org
` (6 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-12-15 10:26 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #11 from Tobias Burnus <burnus at gcc dot gnu.org> ---
For completeness, the LLVM 'main' patch was backported/cherry-picked for
LLVM 11.0.1 with commit 700baa009dc685a0adc5f94d258be4ae24742471
Regarding the .section discussion, see also last few comments in
https://reviews.llvm.org/D92052
Thanks again for the bug report report, Matthias – and to those providing
background information and helping to reach a good solution :-)
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (11 preceding siblings ...)
2020-12-15 10:26 ` burnus at gcc dot gnu.org
@ 2020-12-22 16:17 ` doko at debian dot org
2020-12-27 20:26 ` burnus at gcc dot gnu.org
` (5 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: doko at debian dot org @ 2020-12-22 16:17 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #12 from Matthias Klose <doko at debian dot org> ---
Fyi, this is still seen with LLVM 11.0.1 rc2
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (12 preceding siblings ...)
2020-12-22 16:17 ` doko at debian dot org
@ 2020-12-27 20:26 ` burnus at gcc dot gnu.org
2020-12-28 10:12 ` doko at debian dot org
` (4 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2020-12-27 20:26 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
Ever confirmed|0 |1
Resolution|WONTFIX |---
Last reconfirmed| |2020-12-27
--- Comment #13 from Tobias Burnus <burnus at gcc dot gnu.org> ---
(In reply to Matthias Klose from comment #12)
> Fyi, this is still seen with LLVM 11.0.1 rc2
But the fix (commit 700baa009dc685a0adc5f94d258be4ae24742471) is included in
llvmorg-11.0.1-rc2.
If the issue is still present, something else goes wrong. What's the error
message? Is it really identical to the one of comment 0?
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (13 preceding siblings ...)
2020-12-27 20:26 ` burnus at gcc dot gnu.org
@ 2020-12-28 10:12 ` doko at debian dot org
2021-01-05 10:31 ` burnus at gcc dot gnu.org
` (3 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: doko at debian dot org @ 2020-12-28 10:12 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #14 from Matthias Klose <doko at debian dot org> ---
[...]
/tmp/ccQFzQVM.s:1082:2: error: changed section flags for .rodata.cst8,
expected:
0x12
.section .rodata.cst8
^
/tmp/ccQFzQVM.s:1082:2: error: changed section entsize for .rodata.cst8,
expected: 8
.section .rodata.cst8
^
make[10]: *** [Makefile:980: lib_a-wcstod.o] Error 1
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (14 preceding siblings ...)
2020-12-28 10:12 ` doko at debian dot org
@ 2021-01-05 10:31 ` burnus at gcc dot gnu.org
2021-01-28 15:21 ` burnus at gcc dot gnu.org
` (2 subsequent siblings)
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2021-01-05 10:31 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #15 from Tobias Burnus <burnus at gcc dot gnu.org> ---
I unfortunately missed in my the LLVM patch that '.rodata' implies flags and
messed up the check. Should by fixed by: https://reviews.llvm.org/D94072
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (15 preceding siblings ...)
2021-01-05 10:31 ` burnus at gcc dot gnu.org
@ 2021-01-28 15:21 ` burnus at gcc dot gnu.org
2021-07-02 15:04 ` ams at gcc dot gnu.org
2021-12-17 9:57 ` burnus at gcc dot gnu.org
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2021-01-28 15:21 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Tobias Burnus <burnus at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
Status|REOPENED |RESOLVED
--- Comment #16 from Tobias Burnus <burnus at gcc dot gnu.org> ---
(In reply to Tobias Burnus from comment #15)
> I unfortunately missed in my the LLVM patch that '.rodata' implies flags and
> messed up the check. Should by fixed by: https://reviews.llvm.org/D94072
Now merged to LLVM 12/trunk (see ↑ or next link); for LLVM 11 backporting, see:
https://bugs.llvm.org/show_bug.cgi?id=48922
For related Debian bug, see
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=975692
Close as FIXED – hopefully, now for real.
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (16 preceding siblings ...)
2021-01-28 15:21 ` burnus at gcc dot gnu.org
@ 2021-07-02 15:04 ` ams at gcc dot gnu.org
2021-12-17 9:57 ` burnus at gcc dot gnu.org
18 siblings, 0 replies; 20+ messages in thread
From: ams at gcc dot gnu.org @ 2021-07-02 15:04 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
Andrew Stubbs <ams at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |xw111luoye at gmail dot com
--- Comment #17 from Andrew Stubbs <ams at gcc dot gnu.org> ---
*** Bug 95023 has been marked as a duplicate of this bug. ***
^ permalink raw reply [flat|nested] 20+ messages in thread
* [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
` (17 preceding siblings ...)
2021-07-02 15:04 ` ams at gcc dot gnu.org
@ 2021-12-17 9:57 ` burnus at gcc dot gnu.org
18 siblings, 0 replies; 20+ messages in thread
From: burnus at gcc dot gnu.org @ 2021-12-17 9:57 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97827
--- Comment #18 from Tobias Burnus <burnus at gcc dot gnu.org> ---
POST SCRIPT / STATUS UPDATE (This PR remains FIXED)
(In reply to Tobias Burnus from comment #16)
> > I unfortunately missed in my the LLVM patch that '.rodata' implies flags and
> > messed up the check. Should by fixed by: https://reviews.llvm.org/D94072
>
> Now merged to LLVM 12/trunk (see ↑ or next link); for LLVM 11 backporting,
> see: https://bugs.llvm.org/show_bug.cgi?id=48922
For completeness:
It turned out that it missed the LLVM 12 branch point - such that is is only in
LLVM 13+ and not LLVM 12. (The LLVM 11 request was also ignored.)
* * *
Additionally, a very hard to hit corner case was missed that became easier to
trigger due to an LLVM 13 change. - That one is still rare to hit (and does not
prevent bootstrapping).
It was fixed in LLVM 'main' and backported to LLVM releases/13.x (just after
13.0.1-RC1); thus, it will be in RC2 (if any) and 13.0.1.
Cf. https://reviews.llvm.org/D114717
main:
https://github.com/llvm/llvm-project/commit/c01c62c76c60a5a5da0496e41faae907944c92dd
LLVM 13.x backport (for 13.0.1):
https://github.com/llvm/llvm-project/issues/52607
https://github.com/llvm/llvm-project/commit/5932c004778cf251302db0d46d1dfb247325ed3f
Thus: FIXED (only) in LLVM 13 - but best to use at least LLVM release/13.x or
13.0.1(-rc2).
* * *
GCC 12 also has several changes (mostly: configure checks to detect and act on
the LLVM version) - to be able to use LLVM 13; several were backported to GCC
11 (releases/gcc-11).
* * *
Crossref: PR 100181 (-> PR 100181 comment 10) is another GCN-related issue,
which was solved in LLVM 13
^ permalink raw reply [flat|nested] 20+ messages in thread
end of thread, other threads:[~2021-12-17 9:57 UTC | newest]
Thread overview: 20+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-11-14 11:02 [Bug target/97827] New: [11 Regression] bootstrap error building the amdgcn-amdhsa offload compiler doko at debian dot org
2020-11-16 7:17 ` [Bug target/97827] " rguenth at gcc dot gnu.org
2020-11-16 8:06 ` doko at debian dot org
2020-11-17 11:44 ` [Bug target/97827] bootstrap error building the amdgcn-amdhsa offload compiler with LLVM 11 burnus at gcc dot gnu.org
2020-11-17 12:45 ` burnus at gcc dot gnu.org
2020-11-17 16:31 ` burnus at gcc dot gnu.org
2020-11-18 12:32 ` burnus at gcc dot gnu.org
2020-11-18 12:46 ` jakub at gcc dot gnu.org
2020-11-24 20:01 ` burnus at gcc dot gnu.org
2020-12-11 17:02 ` burnus at gcc dot gnu.org
2020-12-14 18:31 ` i at maskray dot me
2020-12-14 18:38 ` i at maskray dot me
2020-12-15 10:26 ` burnus at gcc dot gnu.org
2020-12-22 16:17 ` doko at debian dot org
2020-12-27 20:26 ` burnus at gcc dot gnu.org
2020-12-28 10:12 ` doko at debian dot org
2021-01-05 10:31 ` burnus at gcc dot gnu.org
2021-01-28 15:21 ` burnus at gcc dot gnu.org
2021-07-02 15:04 ` ams at gcc dot gnu.org
2021-12-17 9:57 ` burnus 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).