forked from WebAssembly/binaryen
-
Notifications
You must be signed in to change notification settings - Fork 0
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Allow DWARF and multivalue together (WebAssembly#6570)
This allows writing of binaries with DWARF info when multivalue is enabled. Currently we just crash when both are enabled together. This just assumes, unless we have run DWARF-invalidating passes, all locals added for tuples or scratch locals would have been added at the end of the local list, so just printing all locals in order would preserve the DWARF info. Tuple locals are expanded in place and scratch locals are added at the end.
- Loading branch information
Showing
3 changed files
with
135 additions
and
9 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,106 @@ | ||
;; Test that we handle multivalue + DWARF correctly. When we need to preserve | ||
;; DWARF info, we don't do any local reordering and all newly added locals | ||
;; during parsing/writing are added at the end of the local list. | ||
|
||
;; Generated from this c file with the following command: | ||
;; $ emcc -g -Xclang -target-abi -Xclang experimental-mv dwarf-multivalue.c -o dwarf-multivalue.wasm | ||
;; | ||
;; struct MyStruct { | ||
;; int a; | ||
;; float b; | ||
;; }; | ||
;; | ||
;; struct MyStruct foo() { | ||
;; struct MyStruct ret = {.a = 3, .b = 3.5}; | ||
;; return ret; | ||
;; } | ||
;; | ||
;; void test() { | ||
;; struct MyStruct s = foo(); | ||
;; } | ||
;; | ||
;; int main() { | ||
;; test(); | ||
;; return 0; | ||
;; } | ||
|
||
;; The original wasm file's $test function's locals are as follows: | ||
;; (func $test | ||
;; (local $0 i32) | ||
;; (local $1 i32) | ||
;; (local $2 i32) | ||
;; (local $3 i32) | ||
;; (local $4 f32) | ||
;; (local $5 i32) | ||
;; (local $6 i32) | ||
;; (local $7 i32) | ||
;; (local $8 f32) | ||
;; (local $9 i32) | ||
;; (local $10 f32) | ||
|
||
;; If we parse this wasm file into Binaryen IR, two locals are added in the | ||
;; process. Here $11 is added for tuple parsing and $12 is added for stacky IR | ||
;; resolving during binary reading process. | ||
;; RUN: wasm-dis %s.wasm -o - | filecheck %s --check-prefix=ORIG | ||
;; ORIG: (func $test | ||
;; ORIG-NEXT: (local $0 i32) | ||
;; ORIG-NEXT: (local $1 i32) | ||
;; ORIG-NEXT: (local $2 i32) | ||
;; ORIG-NEXT: (local $3 i32) | ||
;; ORIG-NEXT: (local $4 f32) | ||
;; ORIG-NEXT: (local $5 i32) | ||
;; ORIG-NEXT: (local $6 i32) | ||
;; ORIG-NEXT: (local $7 i32) | ||
;; ORIG-NEXT: (local $8 f32) | ||
;; ORIG-NEXT: (local $9 i32) | ||
;; ORIG-NEXT: (local $10 f32) | ||
;; ORIG-NEXT: (local $11 (tuple i32 f32)) | ||
;; ORIG-NEXT: (local $12 i32) | ||
|
||
;; If we write this IR into binary, even if this cannot be displayed in the wast | ||
;; format, the local order of $test will look like this, because we don't | ||
;; reorder locals: | ||
;; (func $test | ||
;; (local $0 i32) | ||
;; (local $1 i32) | ||
;; (local $2 i32) | ||
;; (local $3 i32) | ||
;; (local $4 f32) | ||
;; (local $5 i32) | ||
;; (local $6 i32) | ||
;; (local $7 i32) | ||
;; (local $8 f32) | ||
;; (local $9 i32) | ||
;; (local $10 f32) | ||
;; (local $11 i32) ;; Previous (local $11 (tuple i32 f32))'s first element | ||
;; (local $12 f32) ;; Previous (local $11 (tuple i32 f32))'s second element | ||
;; (local $13 i32) ;; Previous (local $12 i32) | ||
;; (local $14 f32) ;; scratch local for f32 | ||
|
||
;; We parse this binary again into Binaryen IR, roundtripping the original | ||
;; binary. Here until (local $14) is the same with the previous list, and $15 is | ||
;; is added for tuple parsing and $16 is added for stacky IR resolving during | ||
;; binary reading process. | ||
;; RUN: wasm-opt -all -g --roundtrip %s.wasm -S -o - | filecheck %s --check-prefix=ROUNDTRIP | ||
;; ROUNDTRIP: (func $test | ||
;; ROUNDTRIP-NEXT: (local $0 i32) | ||
;; ROUNDTRIP-NEXT: (local $1 i32) | ||
;; ROUNDTRIP-NEXT: (local $2 i32) | ||
;; ROUNDTRIP-NEXT: (local $3 i32) | ||
;; ROUNDTRIP-NEXT: (local $4 f32) | ||
;; ROUNDTRIP-NEXT: (local $5 i32) | ||
;; ROUNDTRIP-NEXT: (local $6 i32) | ||
;; ROUNDTRIP-NEXT: (local $7 i32) | ||
;; ROUNDTRIP-NEXT: (local $8 f32) | ||
;; ROUNDTRIP-NEXT: (local $9 i32) | ||
;; ROUNDTRIP-NEXT: (local $10 f32) | ||
;; ROUNDTRIP-NEXT: (local $11 i32) | ||
;; ROUNDTRIP-NEXT: (local $12 f32) | ||
;; ROUNDTRIP-NEXT: (local $13 i32) | ||
;; ROUNDTRIP-NEXT: (local $14 f32) | ||
;; ROUNDTRIP-NEXT: (local $15 (tuple i32 f32)) | ||
;; ROUNDTRIP-NEXT: (local $16 i32) | ||
|
||
;; We can see that we don't reorder the locals during the process and the | ||
;; original list of locals, local $0~$10, is untouched, to NOT invalidate DWARF | ||
;; info. |
Binary file not shown.