[PATCH] D69825: [Clang][Driver] Re-use the calling process instead of creating a new process for the cc1 invocation

[Puyan Lotfi via Phabricator via cfe-commits]

plotfi added a comment.

In D69825#1831611 <https://reviews.llvm.org/D69825#1831611>, @smeenai wrote:

> CC @plotfi for the InterfaceStubs failures


Thanks

In D69825#1831611 <https://reviews.llvm.org/D69825#1831611>, @smeenai wrote:

> CC @plotfi for the InterfaceStubs failures


@smeenai thanks for the CC. I have been looking into this asan failure and it isn't actually caused by the interface stubs AST consumer at all. What is triggering it is the pipeline setup. Here is what I gather:

- Two standard cc1 CompileJobs followed by two Interface Stubs cc1 CompileJobs (or any other Job type for that matter): asan failure
- if I clear the jobs list in Driver.cpp prior to appending the Interface Stubs jobs: No asan failure.
- One standard cc1 CompileJob followed by one Interface Stubs/Alternate cc1 Compile job: No asan failure.

TL;DR
The InterfaceStubs ASTConsumer CompileJob is not causing the ASAN failure here, it's the setup of having a standard compilation pipeline followed by a secondary side-care pipeline thats triggering it in  -fintegrated-cc1  mode. I am still looking into this and any help would be much appreciated. In the meantime I could  -fno-integrated-cc1 the driver.c InterfaceStubs tests to quite the bots if that is what people want (and file a bug on bugzilla).


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D69825/new/

https://reviews.llvm.org/D69825