@@ -490,6 +490,17 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
else
A.renderAsInput(Args, CmdArgs);
}
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+if (A->getNumValues() == 1) {
+ const llvm::Triple &Tr
https://github.com/pawosm-arm edited
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
pawosm-arm wrote:
> This LGTM but I think the other reviewers need to take a look too. Can you
> change the title and description of the PR/commit to represent that what
> we're doing is different now?
The commit title and message has been changed already, the PR description
comment too, I've
pawosm-arm wrote:
> This looks good to me but no doc updates?
Would a paragraph in clang/docs/UsersManual.rst "Configuration files" section
do? If so, I'd write one.
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-co
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/117573
>From c450f66e28facca1c394e1b5db4e38c4e505f45b Mon Sep 17 00:00:00 2001
From: Pawel Osmialowski
Date: Mon, 25 Nov 2024 14:46:55 +
Subject: [PATCH] [clang][driver] Special care for -l and -Wl, flags in con
pawosm-arm wrote:
> This looks good to me but no doc updates?
Added.
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/117573
>From 65958806ff828cb0f41a2587e50abe0d221ae88c Mon Sep 17 00:00:00 2001
From: Pawel Osmialowski
Date: Mon, 25 Nov 2024 14:46:55 +
Subject: [PATCH] [clang][driver] Special care for linker flags in config f
@@ -82,3 +82,29 @@
// CHECK-TWO-CONFIGS: -isysroot
// CHECK-TWO-CONFIGS-SAME: /opt/data
// CHECK-TWO-CONFIGS-SAME: -Wall
+
+//--- The linker input flags should be moved to the end of input list and
appear only when linking.
+// RUN: %clang --target=aarch64-unknown-linux-gnu --
@@ -490,6 +490,35 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
else
A.renderAsInput(Args, CmdArgs);
}
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+const llvm::Triple &Triple = TC.getTriple();
+StringRef
@@ -490,6 +490,35 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
else
A.renderAsInput(Args, CmdArgs);
}
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+const llvm::Triple &Triple = TC.getTriple();
+StringRef
pawosm-arm wrote:
> > with -fveclib=ArmPL -lamath put into a config file, it will always be
> > linked with libamath, even if the user decides to use -fveclib=none
>
> I think there might be a related issue with the approach in this patch too
> though, in that `-fveclib=ArmPL` will silently fo
https://github.com/pawosm-arm approved this pull request.
Gosh, sorry for the confusion it caused. LGTM.
https://github.com/llvm/llvm-project/pull/119249
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/li
https://github.com/pawosm-arm closed
https://github.com/llvm/llvm-project/pull/116432
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/116432
>From d19c8ea5e02ad92c1f89b129cd8fe2fd3cf1d4e3 Mon Sep 17 00:00:00 2001
From: Paul Osmialowski
Date: Tue, 3 Dec 2024 12:41:15 +
Subject: [PATCH] [clang][driver] When -fveclib=ArmPL flag is in use, always
@@ -490,6 +490,35 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
else
A.renderAsInput(Args, CmdArgs);
}
+ if (const Arg *A = Args.getLastArg(options::OPT_fveclib)) {
+const llvm::Triple &Triple = TC.getTriple();
+StringRef
https://github.com/pawosm-arm edited
https://github.com/llvm/llvm-project/pull/116432
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
pawosm-arm wrote:
I suppose this is the reason check-clang fails today.
https://github.com/llvm/llvm-project/pull/119873
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
pawosm-arm wrote:
> This should fix https://lab.llvm.org/buildbot/#/builders/190/builds/11300
>
> I am compiling clang locally to verify that there would be no other problems
> with the test on macOS.
>
> > I suppose this is the reason check-clang fails today.
>
> Are there other test failure
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/117573
>From 537a1c6529410d2a8411b39f0072ef5c60ee865d Mon Sep 17 00:00:00 2001
From: Pawel Osmialowski
Date: Mon, 25 Nov 2024 14:46:55 +
Subject: [PATCH] [clang][driver] Use $ prefix with config file options to
@@ -1243,13 +1268,14 @@ Compilation *Driver::BuildCompilation(ArrayRef ArgList) {
// Try parsing configuration file.
if (!ContainsError)
ContainsError = loadConfigFiles();
- bool HasConfigFile = !ContainsError && (CfgOptions.get() != nullptr);
+ bool HasConfigFileHead
https://github.com/pawosm-arm edited
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
pawosm-arm wrote:
> The subject (linker flags) is no longer accurate.
>
> It's better to mention that Add $ for
Reworded
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/117573
>From c599850a039d407061fb6b20d7365184a68c9938 Mon Sep 17 00:00:00 2001
From: Pawel Osmialowski
Date: Mon, 25 Nov 2024 14:46:55 +
Subject: [PATCH] [clang][driver] Use $ prefix with config file options to
@@ -297,6 +297,9 @@ class Driver {
/// Object that stores strings read from configuration file.
llvm::StringSaver Saver;
+ /// Linker inputs originated from configuration file.
+ std::unique_ptr CfgLinkerInputs;
pawosm-arm wrote:
changed
https://github
pawosm-arm wrote:
> It's worth spending more time discussing the metacharacter.
>
> `^` can be interpreted as `^` in regex, which means the beginning. `$`, on
> the other side, suggests the end. While I agree that `@` for response files
> is not a good choice, I personally would not rule out `
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/117573
>From 7c7649d9e18bdf18671531065e48bb6ec49e0eaf Mon Sep 17 00:00:00 2001
From: Pawel Osmialowski
Date: Mon, 25 Nov 2024 14:46:55 +
Subject: [PATCH] [clang][driver] Special care for linker flags in config f
https://github.com/pawosm-arm edited
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/pawosm-arm edited
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/pawosm-arm closed
https://github.com/llvm/llvm-project/pull/117573
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
pawosm-arm wrote:
> I guessing we pass `-lm` twice here to work around `ld.bfd` not handling
> static libraries in a pleasant way?
This is all due to the way libamath incremetnally emerged into its present
existence. And yes, static linking added another layer of problems here.
https://githu
https://github.com/pawosm-arm closed
https://github.com/llvm/llvm-project/pull/133578
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
@@ -116,11 +116,17 @@
/// Verify that vectorized routines library is being linked in.
// RUN: %clang -### --target=aarch64-pc-windows-msvc -fveclib=ArmPL %s 2>&1 |
FileCheck --check-prefix=CHECK-LINKING-ARMPL-MSVC %s
// RUN: %clang -### --target=aarch64-linux-gnu -fveclib=ArmP
https://github.com/pawosm-arm updated
https://github.com/llvm/llvm-project/pull/133578
>From 77dfec4939b3bba8d76cb7c43d2c309aa4c16588 Mon Sep 17 00:00:00 2001
From: Paul Osmialowski
Date: Sat, 29 Mar 2025 07:51:26 +
Subject: [PATCH] [clang][driver] Fix -fveclib=ArmPL issue: with -nostdlib d
@@ -515,17 +515,19 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
//
// 1. On Linux, link only when actually needed.
//
- // 2. Prefer libm functions over libamath.
+ // 2. Prefer libm functions over libamath (when
pawosm-arm wrote:
Seriously, this is something CI people signaled to me. Our CI builds everything
with `-fveclib=ArmPL` and one of the packages failed on unexpected linker
behavior when libm was artificially added by `-fveclib=ArmPL` despite the
`-nostdlib` flag being used. My suggestion to us
@@ -116,11 +116,17 @@
/// Verify that vectorized routines library is being linked in.
// RUN: %clang -### --target=aarch64-pc-windows-msvc -fveclib=ArmPL %s 2>&1 |
FileCheck --check-prefix=CHECK-LINKING-ARMPL-MSVC %s
// RUN: %clang -### --target=aarch64-linux-gnu -fveclib=ArmP
@@ -515,17 +515,19 @@ void tools::AddLinkerInputs(const ToolChain &TC, const
InputInfoList &Inputs,
//
// 1. On Linux, link only when actually needed.
//
- // 2. Prefer libm functions over libamath.
+ // 2. Prefer libm functions over libamath (when
https://github.com/pawosm-arm created
https://github.com/llvm/llvm-project/pull/133578
Although combining -fveclib=ArmPL with -nostdlib is a rare situation, it should
still be supported correctly and should effect in avoidance of linking against
libm.
>From c033be3cf15b3320b24d52f2414acf88ba
pawosm-arm wrote:
We're facing a regression after this one. We can't build
https://github.com/dslarm/Financial-Services-Workload-Samples/tree/main/MonteCarloEuropeanOptions
with `-g` for AArch64 with -fopenmp now:
```
Stack dump:
0. Program arguments: /usr/bin/clang-21 -cc1 -triple
aarch
pawosm-arm wrote:
One thing that worries me is that before this commit, there was no ICE caused
by compiling this snippet of code (regardless its validity or usefulness), and
now there is. It would be great if it could print a normal error message (which
would encourage my resistant colleagues
101 - 140 of 140 matches
Mail list logo
