jyknight updated this revision to Diff 39644.
jyknight marked 2 inline comments as done.
jyknight added a comment.
Respond to comments.
http://reviews.llvm.org/D14385
Files:
lib/CodeGen/CGAtomic.cpp
test/CodeGen/atomic-ops-libcall.c
Index: test/CodeGen/atomic-ops-libcall.c
jyknight added inline comments.
Comment at: lib/CodeGen/CGAtomic.cpp:901
@@ -897,1 +900,3 @@
+ PostOp = llvm::Instruction::Add;
+// Fall through.
case AtomicExpr::AO__c11_atomic_fetch_add:
compnerd wrote:
> I think we should use `[[clang::fallthrough
jyknight marked 2 inline comments as done.
Comment at: lib/AST/Decl.cpp:3122
@@ +3121,3 @@
+ void *Buffer = Context.Allocate(
+ totalSizeToAlloc(
+ TArgs.size(), Ts.size()));
rsmith wrote:
> OK, I guessed that might be the reason. Seems like a reaso
jyknight created this revision.
jyknight added reviewers: rnk, echristo.
jyknight added a subscriber: cfe-commits.
Herald added subscribers: srhines, danalbert, tberghammer.
On a few architectures, parsing of assembly files annoyingly depends
on whether PIC is enabled or not. This was handled for
Author: jyknight
Date: Fri Aug 7 17:52:34 2015
New Revision: 244370
URL: http://llvm.org/viewvc/llvm-project?rev=244370&view=rev
Log:
[Sparc] XFAIL CodeGen/atomic_ops test.
Modified:
cfe/trunk/test/CodeGen/atomic_ops.c
Modified: cfe/trunk/test/CodeGen/atomic_ops.c
URL:
http://llvm.org/view
jyknight added a subscriber: jyknight.
Comment at: docs/Proposals/GitHub.rst:209
@@ +208,3 @@
+ well as a webhook to update the umbrella project (see below).
+3. Make sure we have an llvm-project (with submodules) setup in the official
+ account, with all necessary hooks (hist
jyknight created this revision.
jyknight added a reviewer: echristo.
jyknight added a subscriber: cfe-commits.
Use it to calculate UserLabelPrefix, instead of specifying it (often
incorrectly).
Note that the *actual* user label prefix has always come from the
DataLayout, and is handled within LLV
jyknight added a comment.
In http://reviews.llvm.org/D17183#351065, @yaron.keren wrote:
> We have tried to keep one copy of DataLayout around
> http://reviews.llvm.org/D11103
> Can it share the Module->getDataLayout() ?
We're at the beginning of clang here, before preprocessing. There's no
l
Those changes were necessary due to a name (IIRC it was "Module") existing
in both the clang and llvm namespaces, and after this change, becoming
ambiguous in those files.
On Feb 12, 2016 3:16 AM, "Paul Robinson"
wrote:
> probinson added a subscriber: probinson.
> probinson added a comment.
>
> D
jyknight added a comment.
Ping. I'm pretty sure you said you were going to look at this. :)
http://reviews.llvm.org/D17183
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Author: jyknight
Date: Tue Feb 23 20:59:33 2016
New Revision: 261717
URL: http://llvm.org/viewvc/llvm-project?rev=261717&view=rev
Log:
Default vaarg lowering should support indirect struct types.
Fixes PR11517 for SPARC.
On most targets, clang lowers va_arg itself, eschewing the use of the
llvm
This revision was automatically updated to reflect the committed changes.
Closed by commit rL261717: Default vaarg lowering should support indirect
struct types. (authored by jyknight).
Changed prior to commit:
http://reviews.llvm.org/D16154?vs=44785&id=48881#toc
Repository:
rL LLVM
http://
jyknight added a comment.
ping again.
http://reviews.llvm.org/D17933
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight added a comment.
Ping.
http://reviews.llvm.org/D17933
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight marked an inline comment as done.
Comment at: test/CodeGen/atomic-ops.c:1
@@ -1,2 +1,2 @@
-// RUN: %clang_cc1 %s -emit-llvm -o - -ffreestanding -ffake-address-space-map
-triple=i686-apple-darwin9 | FileCheck %s
+// RUN: %clang_cc1 %s -emit-llvm -o - -ffreestanding -ffake
jyknight updated this revision to Diff 52430.
jyknight added a comment.
Review tweaks.
http://reviews.llvm.org/D17933
Files:
lib/Basic/Targets.cpp
lib/Frontend/InitPreprocessor.cpp
test/CodeGen/atomic-ops.c
test/CodeGen/ms-volatile.c
test/CodeGenCXX/atomicinit.cpp
test/OpenMP/atomic
jyknight added a comment.
Done and done.
http://reviews.llvm.org/D17933
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight updated this revision to Diff 52587.
jyknight marked 2 inline comments as done.
jyknight added a comment.
review fixes
http://reviews.llvm.org/D17933
Files:
lib/Basic/Targets.cpp
lib/Frontend/InitPreprocessor.cpp
test/CodeGen/atomic-ops.c
test/CodeGen/ms-volatile.c
test/CodeG
Author: jyknight
Date: Mon Apr 4 17:35:56 2016
New Revision: 265355
URL: http://llvm.org/viewvc/llvm-project?rev=265355&view=rev
Log:
Fix "suggest parentheses" warning.
Modified:
cfe/trunk/lib/CodeGen/SwiftCallingConv.cpp
Modified: cfe/trunk/lib/CodeGen/SwiftCallingConv.cpp
URL:
http://llv
jyknight accepted this revision.
jyknight added a comment.
BTW, this change doesn't actually depend on the LLVM change; it could be
committed first, as a "normal" enum puts its members both inside the enum
name's scope and outside.
http://reviews.llvm.org/D18776
jyknight added a comment.
ping.
http://reviews.llvm.org/D17933
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight added a comment.
The large amount of casting to/from integers for AtomicOrderingCABI makes me
think that it probably ought not actually be converted to an enum class after
all.
http://reviews.llvm.org/D18876
___
cfe-commits mailing list
c
I'm confused: why does it break targets if __float128 wasn't supported
before, and still isn't supported?
Surely enabling float128 support on some targets, and leaving it disabled
on linux/x86 shouldn't actually *break* linux/x86?
On Fri, Apr 15, 2016 at 11:53 AM, Nemanja Ivanovic via cfe-commits
jyknight added a comment.
In http://reviews.llvm.org/D18876#402855, @jfb wrote:
> In http://reviews.llvm.org/D18876#399934, @jyknight wrote:
>
> > The large amount of casting to/from integers for AtomicOrderingCABI makes
> > me think that it probably ought not actually be converted to an enum cl
;
> (I'll still need to remove the pointer type in favor of a number of bytes
> for byval (byval to copy (or moving the type from the argument's pointer
> type, into a parameter to byval, eg: %ptr byval(struct_foo) %arg) but
> knowing the alignment's totally handled would b
jyknight added a comment.
This looks okay. The only downside: it will increase sizeof(Decl) on 32bit
platforms, since the layout is:
vtable ptr (4)
NextInContextAndBits ptr (4)
DeclCtx ptr (4)
Loc int (4)
bitfield int (4)
Totaling 20 bytes, which will now be rounded up to 24 bytes. I don't know
jyknight accepted this revision.
jyknight added a comment.
This revision is now accepted and ready to land.
Fine with me.
Repository:
rL LLVM
http://reviews.llvm.org/D20732
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
http://lists.llvm
Author: jyknight
Date: Sat Jun 25 11:40:53 2016
New Revision: 273783
URL: http://llvm.org/viewvc/llvm-project?rev=273783&view=rev
Log:
Fix unused variable warning after r273754.
Modified:
cfe/trunk/lib/Sema/Sema.cpp
Modified: cfe/trunk/lib/Sema/Sema.cpp
URL:
http://llvm.org/viewvc/llvm-proj
jyknight added a subscriber: jyknight.
jyknight commandeered this revision.
jyknight added a reviewer: stbuehler.
jyknight added a comment.
I believe this was fully superseded by
http://reviews.llvm.org/D17567: [GCC] PR23529 Sema part of attrbute abi_tag
support
http://reviews.llvm.org/D18035: [G
jyknight added a subscriber: jyknight.
jyknight added a comment.
This seems to trigger even for the implicitly generated copier of a packed
struct. E.g.
#include
void copyit(epoll_event&out, const epoll_event &in) {
out = in;
}
Is that as intended?
Repository:
rL LLVM
https:/
jyknight added a comment.
Regardless, I think this should not have added a new un-disableable error, but
instead only a default-on warning.
The ""binding reference to packed member" error is firing on some of our code,
and even if it's not a false-positive, it should be possible to disable it
Sorry for not following up here; this will actually be solved in an
entirely different way by http://reviews.llvm.org/D17183
On Thu, Mar 3, 2016 at 5:53 PM, Rafael Espíndola
wrote:
> As a first patch I think you can just leave every
>
> UserLabelPrefix = ""
>
> in place.
>
> Cheers,
> Rafael
___
It'd be nice to have a comment here that mentions that the clang
behavior which is being preserved for Darwin, FreeBSD, and PS4 is a
*bug* which is being intentionally left unfixed. The previous clang
behavior directly contradicts the x86_64 ABI document, which I believe
all of these platforms clai
This revision was automatically updated to reflect the committed changes.
jyknight marked an inline comment as done.
Closed by commit rL262737: Make TargetInfo store an actual DataLayout instead
of a string. (authored by jyknight).
Changed prior to commit:
http://reviews.llvm.org/D17183?vs=4776
Author: jyknight
Date: Fri Mar 4 13:00:41 2016
New Revision: 262737
URL: http://llvm.org/viewvc/llvm-project?rev=262737&view=rev
Log:
Make TargetInfo store an actual DataLayout instead of a string.
Use it to calculate UserLabelPrefix, instead of specifying it (often
incorrectly).
Note that the
Ah, great! I always love it when people document their ABIs.
Is your ABI document public? If so, could you add it to
docs/CompilerWriterInfo.rst?
On Fri, Mar 4, 2016 at 11:54 AM, Robinson, Paul
wrote:
>> It'd be nice to have a comment here that mentions that the clang
>> behavior which is being
jyknight created this revision.
jyknight added a reviewer: t.p.northover.
jyknight added a subscriber: cfe-commits.
Herald added subscribers: dschuff, jfb, aemerson.
Also, remove all the manual definition of
__GCC_HAVE_SYNC_COMPARE_AND_SWAP_* macros. Instead, define it based on
MaxAtomicInlineWidt
jyknight accepted this commit.
jyknight added a comment.
This issue was resolved; hopefully this will inform Phab of that.
Users:
jyknight (Author, Auditor)
3.7-release (Auditor)
cfe-commits (Auditor)
tstellarAMD (Auditor)
compnerd (Auditor)
majnemer (Auditor)
rsmith (Auditor)
htt
> A 4 byte cmpxchg could be lock free, while a 1 byte cmpxchg may not be
lock free.
That's not possible: if you have a 4-byte cmpxchg instruction, you can use
it to implement a 1-byte cmpxchg, too. Many targets do this already. (It
would be better if that was available generically so that code did
On Thu, Mar 17, 2016 at 12:55 PM, Craig, Ben
wrote:
> I know that MIPS does that, and an out-of-tree implementation of hexagon
> implements 1-byte cmpxchg in terms of the 4-byte version. The emulation
> code isn't particularly small, and it seems reasonable to make it a libcall.
That's fine. Ei
jyknight added a subscriber: jyknight.
jyknight added a comment.
This conflicts with http://reviews.llvm.org/D17933. Most of this change also
seems unnecessary.
- I think the `is_always_lock_free` function should be defined based on the
existing `__atomic_always_lock_free` builtin, not on defin
jyknight added inline comments.
Comment at: lib/Frontend/InitPreprocessor.cpp:305
@@ +304,3 @@
+if (TypeWidth == TypeAlign && (TypeWidth & (TypeWidth - 1)) == 0 &&
+TypeWidth <= InlineWidth)
+ return Always;
jfb wrote:
> bcraig wrote:
> > On some
jyknight added a comment.
> Changed to what you suggested. Much nicer. I don't remember why I thought it
> was a bad idea.
Thanks, great! I don't have any opinion on what remains in this patch; someone
else should review now.
http://reviews.llvm.org/D17950
Author: jyknight
Date: Mon Dec 28 22:34:11 2015
New Revision: 256534
URL: http://llvm.org/viewvc/llvm-project?rev=256534&view=rev
Log:
[TrailingObjects] Use a different technique to determine if a getDecl
member function exists on a class.
The previous trick depended on inheriting from the class
Author: jyknight
Date: Mon Dec 28 22:46:43 2015
New Revision: 256535
URL: http://llvm.org/viewvc/llvm-project?rev=256535&view=rev
Log:
Revert "[TrailingObjects] Use a different technique to determine if a getDecl"
This reverts commit r256534.
Failed to build on MSVC with error:
clang/ASTMatchers
Author: jyknight
Date: Tue Dec 29 10:44:11 2015
New Revision: 256564
URL: http://llvm.org/viewvc/llvm-project?rev=256564&view=rev
Log:
[TrailingObjects] Use a different technique to determine if a getDecl
member function exists on a class.
The previous trick depended on inheriting from the class
Author: jyknight
Date: Tue Dec 29 12:15:14 2015
New Revision: 256570
URL: http://llvm.org/viewvc/llvm-project?rev=256570&view=rev
Log:
[TrailingObjects] Convert AST classes that had a ASTTemplateKWAndArgsInfo.
So, also:
- Moved the TemplateArgumentLoc array out of the
ASTTemplateKWAndArgsInfo
Author: jyknight
Date: Tue Dec 29 16:13:13 2015
New Revision: 256585
URL: http://llvm.org/viewvc/llvm-project?rev=256585&view=rev
Log:
[TrailingObjects] Convert Decl* classes.
Also remove now-redundant explicit alignment specification on some of
the classes converted prior to TrailingObjects auto
Author: jyknight
Date: Tue Dec 29 16:31:18 2015
New Revision: 256590
URL: http://llvm.org/viewvc/llvm-project?rev=256590&view=rev
Log:
[TrailingObjects] Convert OffsetOfExpr.
That necessitated moving the OffsetOfNode class out of OffsetOfExpr.
Modified:
cfe/trunk/include/clang/AST/Expr.h
Author: jyknight
Date: Tue Dec 29 20:27:28 2015
New Revision: 256608
URL: http://llvm.org/viewvc/llvm-project?rev=256608&view=rev
Log:
[TrailingObjects] Convert CastExpr and subclasses.
Modified:
cfe/trunk/include/clang/AST/Expr.h
cfe/trunk/include/clang/AST/ExprCXX.h
cfe/trunk/includ
Author: jyknight
Date: Tue Dec 29 21:58:33 2015
New Revision: 256619
URL: http://llvm.org/viewvc/llvm-project?rev=256619&view=rev
Log:
Fix alignment issue in CodeGenFunction::PopCleanupBlock.
It was copying an EHCleanupStack::Cleanup object into a
SmallVector, with a comment saying that SmallVect
Author: jyknight
Date: Wed Dec 30 22:18:25 2015
New Revision: 256658
URL: http://llvm.org/viewvc/llvm-project?rev=256658&view=rev
Log:
[TrailingObjects] Convert remaining classes in Expr.h and ExprCXX.h
Modified:
cfe/trunk/include/clang/AST/Expr.h
cfe/trunk/include/clang/AST/ExprCXX.h
Author: jyknight
Date: Wed Dec 30 22:43:19 2015
New Revision: 256659
URL: http://llvm.org/viewvc/llvm-project?rev=256659&view=rev
Log:
[TrailingObjects] Convert classes in ExprObjC.h
Modified:
cfe/trunk/include/clang/AST/ExprObjC.h
cfe/trunk/lib/AST/ExprObjC.cpp
cfe/trunk/lib/Sema/Sem
Author: jyknight
Date: Thu Dec 31 00:01:19 2015
New Revision: 256665
URL: http://llvm.org/viewvc/llvm-project?rev=256665&view=rev
Log:
[TrailingObjects] Fix bug in "Convert classes in ExprObjC.h"
(Detected by asan)
Modified:
cfe/trunk/lib/AST/ExprObjC.cpp
Modified: cfe/trunk/lib/AST/ExprObj
Author: jyknight
Date: Thu Dec 31 18:38:24 2015
New Revision: 256683
URL: http://llvm.org/viewvc/llvm-project?rev=256683&view=rev
Log:
[TrailingObjects] Convert classes in OpenMPClause.h
Modified:
cfe/trunk/include/clang/AST/OpenMPClause.h
cfe/trunk/lib/AST/OpenMPClause.cpp
Modified: cfe
Author: jyknight
Date: Thu Dec 31 19:12:48 2015
New Revision: 256684
URL: http://llvm.org/viewvc/llvm-project?rev=256684&view=rev
Log:
[TrailingObjects] Fix "Convert classes in OpenMPClause.h" for MSVC
"friend class OMPVarListClause" -> "friend OMPVarListClause". It's a
template, not a class.
M
Author: jyknight
Date: Mon Jan 11 16:00:22 2016
New Revision: 257392
URL: http://llvm.org/viewvc/llvm-project?rev=257392&view=rev
Log:
PR26087: Use nonstandard MSVC extension for VS2015 as well.
In r256564, I had conditioned the workaround in has_getDecl to only be
used for MSVC before the 2015 r
jyknight created this revision.
jyknight added a subscriber: cfe-commits.
Herald added subscribers: jyknight, dschuff, jfb.
Fixes PR11517 for SPARC.
On most targets, clang lowers va_arg itself, eschewing the use of the
llvm vaarg instruction. This is necessary (at least for now) as the type
argum
Author: jyknight
Date: Thu Jan 14 23:57:41 2016
New Revision: 257871
URL: http://llvm.org/viewvc/llvm-project?rev=257871&view=rev
Log:
Fix a -Wparentheses warning in ASTDiagnostic.cpp.
Modified:
cfe/trunk/lib/AST/ASTDiagnostic.cpp
Modified: cfe/trunk/lib/AST/ASTDiagnostic.cpp
URL:
http://ll
Author: jyknight
Date: Tue Jan 26 19:04:51 2016
New Revision: 258894
URL: http://llvm.org/viewvc/llvm-project?rev=258894&view=rev
Log:
Revert "Change of UserLabelPrefix default value from "_" to """
This reverts commit r258504.
This commit breaks (at least) sparc-rtems -- the OS (RTEMS) used to
I reverted this change with r258894, as it breaks (at least) sparc-rtems.
Clearly this area of the code was not sufficiently covered by the testsuite.
On Fri, Jan 22, 2016 at 10:24 AM, Andrey Bokhanko via cfe-commits <
cfe-commits@lists.llvm.org> wrote:
> Author: asbokhan
> Date: Fri Jan 22 09:24
d quite
> extensively in test/Preprocessor/init.c. There is a line for generic
> SPARC as well:
>
> // SPARC:#define __USER_LABEL_PREFIX__ _
>
> and I added initialization of UserLabelPrefix to "_" in
> SparcTargetInfo constructor specifically to accommodate this test.
Author: jyknight
Date: Thu Nov 12 12:37:29 2015
New Revision: 252920
URL: http://llvm.org/viewvc/llvm-project?rev=252920&view=rev
Log:
Correct atomic libcall support for __atomic_*_fetch builtins.
In r244063, I had caused these builtins to call the same-named library
functions, __atomic_*_fetch_S
This revision was automatically updated to reflect the committed changes.
Closed by commit rL252920: Correct atomic libcall support for __atomic_*_fetch
builtins. (authored by jyknight).
Changed prior to commit:
http://reviews.llvm.org/D14385?vs=39644&id=40068#toc
Repository:
rL LLVM
http:/
jyknight added a comment.
The concerns should be fixed by the followup commit
http://reviews.llvm.org/rL252920. If merging to 3.7 branch, both revisions will
need to be included.
Users:
jyknight (Author, Auditor)
3.7-release (Auditor)
cfe-commits (Auditor)
tstellarAMD (Auditor)
compn
jyknight added inline comments.
Comment at: lib/Driver/ToolChains.cpp:4319
@@ +4318,3 @@
+ // This contains libc, libg (a superset of libc), libm, libstdc++, libssp.
+ SmallString<128> LibDir(GCCInstallation.getParentLibPath());
+ if (Triple.getArch() == llvm::Triple::sparcel)
jyknight added a comment.
The commit message could use some work: please summarize what you're actually
changing, and why. (Here you mention only vaguely why, and not what.)
Comment at: test/Driver/nodefaultlib_stdc.c:1
@@ +1,2 @@
+// RUN: %clang -stdlib=libc++ -nodefaultlibs -
jyknight added a comment.
What's the status of this? It's showing bright red on my review dashboard, and
I'd like to clear it off. :) Am I expected to do anything else here, or is it
just waiting to be merged for the next release now (or, is it not going to be
after all)?
Users:
jyknight (A
jyknight added a comment.
In http://reviews.llvm.org/D15455#308532, @joerg wrote:
> As before, I don't see why a linker flag should change code generation. That
> seems to be a gross violation of layering and quite surprising.
I agree.
http://reviews.llvm.org/D15455
__
jyknight added a comment.
> #define CC1_SPEC "%{!mkernel:%{!static:%{!mdynamic-no-pic:-fPIC}}}
That's sad.
> We can change this to be Darwin-specific if you prefer, but we should
> maintain compatibility with GCC and previous Clang releases in this behavior.
Yes, the broken behavior should b
Author: jyknight
Date: Mon Dec 21 14:30:49 2015
New Revision: 256191
URL: http://llvm.org/viewvc/llvm-project?rev=256191&view=rev
Log:
Tweak myriad-toolchain test.
The test failed when run on a SPARC host, since it was finding the
native gcc installation by accident.
Modified:
cfe/trunk/test
Author: jyknight
Date: Wed Dec 23 20:59:37 2015
New Revision: 256359
URL: http://llvm.org/viewvc/llvm-project?rev=256359&view=rev
Log:
[TrailingObjects] Convert ASTTemplateKWAndArgsInfo and
ASTTemplateArgumentListInfo.
Doing so required separating them so that the former doesn't inherit
from the
guard. It's also very rarely
included by anything.
Addresses
https://github.com/llvm/llvm-project/pull/96246#issuecomment-2322453809
>From 861906e432651966238ea764f5a3ac61f0d55c0a Mon Sep 17 00:00:00 2001
From: James Y Knight
Date: Tue, 3 Sep 2024 17:03:49 -0400
Subject: [PATCH] Adjust
jyknight wrote:
The only impact is that any modular header which includes this file will end up
with a definition of the header guard, instead of importing it, right? It's not
clear to me why it's useful to care about that?
(I totally do understand worrying about cases which result in e.g. hav
jyknight wrote:
If we don't keep the include-guard, then we'll report the same warning once per
`#include`, instead of once per TU, which is probably not ideal.
https://github.com/llvm/llvm-project/pull/107155
___
cfe-commits mailing list
cfe-commits@
jyknight wrote:
I think the bug this change was attempting to fix is actually the same as
#38554 and related bugs -- which, it appears, was not really fixed.
The underlying problem here is that `#pragma once` should should work
identically to ifndef guards, as far as what macros/decls are made
jyknight wrote:
Oh -- I'd also note that even if this is reverted, ilya-biryukov may want to
continue to investigate the source-location issue -- it's entirely possible
that the correct fix will trigger that same problem!
https://github.com/llvm/llvm-project/pull/83660
jyknight wrote:
> The problem I'm trying to fix is that nobody knows when it's appropriate to
> use #import vs #include
But you haven't really (and I think cannot) fixed that.
> using header guards or #pragma once is very "un-Objective-C".
Yes, this is quite unfortunate. The best answer woul
jyknight wrote:
>> The end result should be that #imported and #pragma once-guarded files are
>> treated the same way as #ifndef-guarded files.
> While I don't necessarily disagree with that goal in principle, it wasn't
> true before this change either.
Well, yes, I know it isn't true yet --
jyknight wrote:
To get back to my previous point about the semantics implemented by this patch
being unfortunate -- the upshot is that, now:
```
#include
#define NDEBUG
#import
```
will include assert.h twice (even though the latter is an "import") _only_ if
modules are enabled. If modules ar
@@ -1825,32 +1818,32 @@ def : ProcModel;
-def : Proc<"k6-2", [FeatureX87, FeatureCX8, Feature3DNow],
+def : Proc<"k6-2", [FeatureX87, FeatureCX8],
jyknight wrote:
Oops! I had fixed that but in a not-yet-committed diff in my checkout. (Tests
fail, otherwise!)
ht
@@ -7,151 +7,16 @@
*===---===
*/
+// 3dNow intrinsics are no longer supported, and this header remains only as a
+// stub for users who were including it to get to _m_prefetch or
+// _m_prefetchw. Such uses
https://github.com/jyknight updated
https://github.com/llvm/llvm-project/pull/98273
>From 546963316017935bfbbc05d7095d645344fbd5f5 Mon Sep 17 00:00:00 2001
From: James Y Knight
Date: Wed, 10 Jul 2024 00:49:25 -0400
Subject: [PATCH 1/3] Clang: don't unnecessarily convert inline-asm ope
@@ -49,12 +49,10 @@ typedef __bf16 __m128bh __attribute__((__vector_size__(16),
__aligned__(16)));
#endif
/* Define the default attributes for the functions in this file. */
-#define __DEFAULT_FN_ATTRS
\
- __attribute__((_
https://github.com/jyknight closed
https://github.com/llvm/llvm-project/pull/96540
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
Author: James Y Knight
Date: 2024-07-24T18:39:16-04:00
New Revision: b79568654e38a14ef921af932ed96abd8961b1ed
URL:
https://github.com/llvm/llvm-project/commit/b79568654e38a14ef921af932ed96abd8961b1ed
DIFF:
https://github.com/llvm/llvm-project/commit
@@ -159,6 +159,20 @@ AMDGPU Support
X86 Support
^^^
+- The MMX vector intrinsic functions from ``*mmintrin.h`` which
+ operate on `__m64` vectors, such as ``_mm_add_pi8``, have been
+ reimplemented to use the SSE2 instruction-set and XMM registers
+ unconditionally.
@@ -94,14 +94,15 @@ entry:
;
; MMX Store
+; Note: doesn't actually emit a non-temporal store here.
;
define void @test_mmx(ptr nocapture %a0, ptr nocapture %a1) {
; ALL-LABEL: test_mmx:
; ALL: # %bb.0: # %entry
; ALL-NEXT:movq (%rdi), %mm0
; ALL-NEXT:psrl
jyknight wrote:
@RKSimon has fixed in abacc5220e228012b0b86ca2ed540ef1a715f1ee, thanks. Sorry
for missing that!
https://github.com/llvm/llvm-project/pull/98505
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mail
jyknight wrote:
It seems a bit confusing to me to create a situation where `-ffast-math` is
equivalent to `-ffp-model=aggressive`, not `-ffp-model=fast`.
https://github.com/llvm/llvm-project/pull/100453
___
cfe-commits mailing list
cfe-commits@lists.l
https://github.com/jyknight approved this pull request.
LGTM, thanks.
https://github.com/llvm/llvm-project/pull/101005
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight wrote:
We aren't compliant for assignment either with x87 float/double -- e.g. enable
`-O2` on your example.
I think we need to say that we conform except on 32-bit x86 with SSE2 disabled,
and link to an appropriate bug. Maybe we'll fix that backend someday. Or maybe
not.
https://gi
https://github.com/jyknight approved this pull request.
Sounds good enough to me, thank you!
https://github.com/llvm/llvm-project/pull/98736
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-com
https://github.com/jyknight closed
https://github.com/llvm/llvm-project/pull/96246
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
jyknight wrote:
Thanks for the report! Looks like I missed some 3dnow mentions in
llvm/lib/TargetParser/X86TargetParser.cpp, so `-mno-mmx` is still emitting this
feature.
I don't believe it should have any impact other than the warning spam, but I
will fix it.
https://github.com/llvm/llvm-pr
#75081.
>From 2cb327b0377d4196982842455e2aaf4e6a21f667 Mon Sep 17 00:00:00 2001
From: James Y Knight
Date: Wed, 17 Jul 2024 10:43:45 -0400
Subject: [PATCH] Handle constant "pointers" for
`__atomic_always_lock_free`/`__atomic_is_lock_free`.
The second argument passed to these builtins is used to validate
wh
6, but was missed.
>From 70160e033ac94de92fc2cd6289548094cd893498 Mon Sep 17 00:00:00 2001
From: James Y Knight
Date: Wed, 17 Jul 2024 11:40:38 -0400
Subject: [PATCH] Remove 3dnow from X86TargetParser.
This should've been part of PR #96246, but was missed.
This addresses the spurious inclusion of (now unsuppor
@@ -171,14 +171,14 @@ constexpr FeatureBitset FeaturesClearwaterforest =
// Geode Processor.
constexpr FeatureBitset FeaturesGeode =
-FeatureX87 | FeatureCMPXCHG8B | FeatureMMX | Feature3DNOW | Feature3DNOWA;
jyknight wrote:
Possibly not?
We didn't enabl
@@ -171,14 +171,14 @@ constexpr FeatureBitset FeaturesClearwaterforest =
// Geode Processor.
constexpr FeatureBitset FeaturesGeode =
-FeatureX87 | FeatureCMPXCHG8B | FeatureMMX | Feature3DNOW | Feature3DNOWA;
jyknight wrote:
(I somewhat lean towards submi
https://github.com/jyknight approved this pull request.
Thanks, LGTM. I like this version.
Did this change also fix any other bugs? Since it looks like we were previously
not skipping ProcessDeclAttribute for ``[[]]`` attributes in C23 mode, when
IncludeCXX11Attributes == false, and we were su
201 - 300 of 886 matches
Mail list logo
