yamt wrote:
> That said, the description of the use case is ~3 short sentences.
ok.
here is a concrete example and a bit longer explanation:
https://github.com/yamt/toywasm/blob/9ee6ec86f56723819fd8411866094f72247dba78/lib/insn.c#L515-L527
https://github.com/llvm/llvm-project/pull/92499
yamt wrote:
> If we're going to do this, it should probably also work for constants.
for completeness, maybe. i myself have no use cases though.
> Also, I think I'd prefer to sort out the situation with the C++ standard's
> rules for constant merging before we start extending those rules. See
yamt wrote:
> Hmm... I'm not sure this meets our requirements for inclusion as an
> attribute. The semantics of this are pretty opaque, no obvious significant
> motivation/applicability in the base languages, etc. There doesn't seem to be
> any reasonable use case that I can see.
do you mean
@@ -1944,6 +1944,13 @@ def ReturnsTwice : InheritableAttr {
let SimpleHandler = 1;
}
+def UnnamedAddr : InheritableAttr {
+ let Spellings = [Clang<"unnamed_addr">];
+ let Subjects = SubjectList<[Function]>;
+ let Documentation = [Undocumented];
yamt
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/92499
>From 52b744c91bdba1cf8cda9d6164ec8fc130d75fab Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 17 May 2024 14:47:06 +0900
Subject: [PATCH 1/3] [clang] add unnamed_addr function attribute
It simply
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/92499
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/92499
>From 52b744c91bdba1cf8cda9d6164ec8fc130d75fab Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 17 May 2024 14:47:06 +0900
Subject: [PATCH] [clang] add unnamed_addr function attribute
It simply applies
https://github.com/yamt created https://github.com/llvm/llvm-project/pull/92499
It simply applies the LLVM attribute with the same name to the function.
Sometimes, a programmer knows that function pointer uniqueness doesn't really
matter for some of their functions. In that case, this
yamt wrote:
> I'm not sure about the common practice in this repo, but Github in general
> supports merge workflows much better. Especially in a long-running large PRs,
> if you force-push, it's hard for the reviewers to know what has changed since
> their last review, because all the history
yamt wrote:
> @yamt Given that this PR is not large and we already went through it anyway,
> nevermind what I said about rebasing, and please use whatever method you are
> convenient with. But just FYI you can do `git merge` to avoid rebasing (I'm
> not asking you to do it here; just in case
yamt wrote:
@aheejin may i rebase this PR as it now has conflicts?
https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
yamt wrote:
@aheejin i updated the rest of comments
https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
yamt wrote:
> i tried to run the relevant emscripten tests. but i have not succeeded yet
> even w/o these changes. (i'm an emscripten newbie.)
they passed with the latest version of this PR and the emscripten PR.
https://github.com/llvm/llvm-project/pull/84137
@@ -198,9 +198,18 @@
///
/// If there are calls to setjmp()
///
-/// 2) and 3): The same as 2) and 3) in Emscripten SjLj.
-/// (setjmpTable/setjmpTableSize initialization + setjmp callsite
-/// transformation)
+/// 2) In the function entry that calls setjmp, initialize
+///
@@ -198,9 +198,18 @@
///
/// If there are calls to setjmp()
///
-/// 2) and 3): The same as 2) and 3) in Emscripten SjLj.
-/// (setjmpTable/setjmpTableSize initialization + setjmp callsite
-/// transformation)
+/// 2) In the function entry that calls setjmp, initialize
+///
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 6c1bfa0aa292d81102a9aeb86ca4264516beb634 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 01/26] [WebAssembly] Implement an alternative translation for
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
yamt wrote:
> If possible, can you not force-push? It's hard to track changes between
> commits with force-pushes. Thank you!
i had to rebase for some unrelated reasons. i will avoid it next time.
https://github.com/llvm/llvm-project/pull/84137
___
yamt wrote:
@aheejin @sbc100
i think i resolved most of your feedback in this PR and the emscripten PR.
i tried to run the relevant emscripten tests. but i have not succeeded yet even
w/o these changes. (i'm an emscripten newbie.)
https://github.com/llvm/llvm-project/pull/84137
https://github.com/yamt ready_for_review
https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 6c1bfa0aa292d81102a9aeb86ca4264516beb634 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 01/20] [WebAssembly] Implement an alternative translation for
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 6c1bfa0aa292d81102a9aeb86ca4264516beb634 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 01/19] [WebAssembly] Implement an alternative translation for
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 6c1bfa0aa292d81102a9aeb86ca4264516beb634 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 01/17] [WebAssembly] Implement an alternative translation for
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt edited https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt converted_to_draft
https://github.com/llvm/llvm-project/pull/84137
___
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 6c1bfa0aa292d81102a9aeb86ca4264516beb634 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 01/13] [WebAssembly] Implement an alternative translation for
yamt wrote:
@aheejin @sbc100
let me confirm the plan on this PR.
i can remove the option `-mllvm -experimental-wasm-enable-alt-sjlj` by making
it unconditionally true, and update tests, right?
https://github.com/llvm/llvm-project/pull/84137
___
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 1283ae6b5536810f8fbe183eda80997aa9f5cdc3 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 1/7] [WebAssembly] Implement an alternative translation for
yamt wrote:
> > Given that we don't need `setjmpTableSize` anymore and `setjmpTable`
> > doesn't change, we don't need the whole block here from line 1463 ~ line
> > 1503 doing SSA updates anymore:
>
> i get errors like the following if i simply put the SSA update things in
>
yamt wrote:
> Given that we don't need `setjmpTableSize` anymore and `setjmpTable` doesn't
> change, we don't need the whole block here from line 1463 ~ line 1503 doing
> SSA updates anymore:
i get errors like the following if i simply put the SSA update things in
`!EnableWasmAltSjLj` block.
yamt wrote:
> @yamt, do you want to submit a PR to emscripten adding new library functions?
> I can do that too, but given that this is your code, if you want to do it
> it'd be good. About where to add them... to make the transition smooth,
>
yamt wrote:
> > In terms of getting this landed and tested, I wonder which path we should
> > take:
> >
> > 1. Land this now, without tests, then update emscripten then come back and
> > flip the default, at which point the existing tests will get updated.
> > 2. Duplicate/update the the
@@ -1738,10 +1792,16 @@ void
WebAssemblyLowerEmscriptenEHSjLj::handleLongjmpableCallsForWasmSjLj(
BasicBlock *ThenBB = BasicBlock::Create(C, "if.then", );
BasicBlock *EndBB = BasicBlock::Create(C, "if.end", );
Value *EnvP = IRB.CreateBitCast(Env, getAddrPtrType(),
@@ -999,25 +1017,43 @@ bool
WebAssemblyLowerEmscriptenEHSjLj::runOnModule(Module ) {
// Register __wasm_longjmp function, which calls __builtin_wasm_longjmp.
FunctionType *FTy = FunctionType::get(
IRB.getVoidTy(), {Int8PtrTy, IRB.getInt32Ty()}, false);
-
@@ -1291,19 +1327,29 @@ bool
WebAssemblyLowerEmscriptenEHSjLj::runSjLjOnFunction(Function ) {
Type *IntPtrTy = getAddrIntType();
Constant *size = ConstantInt::get(IntPtrTy, 40);
IRB.SetInsertPoint(SetjmpTableSize);
- auto *SetjmpTable = IRB.CreateMalloc(IntPtrTy,
@@ -1291,19 +1327,29 @@ bool
WebAssemblyLowerEmscriptenEHSjLj::runSjLjOnFunction(Function ) {
Type *IntPtrTy = getAddrIntType();
Constant *size = ConstantInt::get(IntPtrTy, 40);
IRB.SetInsertPoint(SetjmpTableSize);
- auto *SetjmpTable = IRB.CreateMalloc(IntPtrTy,
@@ -1291,19 +1327,29 @@ bool
WebAssemblyLowerEmscriptenEHSjLj::runSjLjOnFunction(Function ) {
Type *IntPtrTy = getAddrIntType();
Constant *size = ConstantInt::get(IntPtrTy, 40);
IRB.SetInsertPoint(SetjmpTableSize);
- auto *SetjmpTable = IRB.CreateMalloc(IntPtrTy,
@@ -999,25 +1002,42 @@ bool
WebAssemblyLowerEmscriptenEHSjLj::runOnModule(Module ) {
// Register __wasm_longjmp function, which calls __builtin_wasm_longjmp.
FunctionType *FTy = FunctionType::get(
IRB.getVoidTy(), {Int8PtrTy, IRB.getInt32Ty()}, false);
-
https://github.com/yamt updated https://github.com/llvm/llvm-project/pull/84137
>From 1283ae6b5536810f8fbe183eda80997aa9f5cdc3 Mon Sep 17 00:00:00 2001
From: YAMAMOTO Takashi
Date: Fri, 9 Feb 2024 15:49:55 +0900
Subject: [PATCH 1/3] [WebAssembly] Implement an alternative translation for
https://github.com/yamt created https://github.com/llvm/llvm-project/pull/84137
Instead of maintaining per-function-invocation malloc()'ed tables to track
which functions each label belongs to, store the equivalent info in jump
buffers (jmp_buf) themselves.
Also, use a less emscripten-looking
43 matches
Mail list logo
