glider added a comment.
> Should we have in AddressSanitizer.cpp the following for consistency with
> other sanitizers?
Good catch, without that disable_sanitizer_instrumentation could not override
sanitize_address.
I fixed this and added a test.
================
Comment at: clang/test/CodeGen/asan-globals.cpp:62
// CHECK: ![[ATTR_GLOBAL]] = !{{{.*}}, null, null, i1 false, i1 true}
+// CHECK: ![[DISABLE_INSTR_GLOBAL]] = !{{{.*}}, null, null, i1 false, i1 true}
// CHECK: ![[BLACKLISTED_GLOBAL]] = !{{{.*}}, null, null, i1 false, i1 true}
----------------
eugenis wrote:
> Does this test rely on the metadata being in the same order as the global
> declarations in the source? Feels brittle and hard to understand - could you
> check if the regexps could be made more precise?
`llvm.asan.globals` above actually look as follows: `!llvm.asan.globals = !{!0,
!2, !4, !6, !7, !8, !9, !11, !13, !15}`
, so the placeholder variables (e.g. `ATTR_GLOBAL` and `DISABLE_INSTR_GLOBAL`)
actually match just the metadata node numbers.
I replaced the `{{.*}}` with variable/file names where applicable, but we'll
still depend on the order of globals.
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D114421/new/
https://reviews.llvm.org/D114421
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
