Fix PMI non-determinism and a potential deadlock. #252
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PMI in DRIVEALL mode invokes subprocesses in PREPALL mode
in order to be able to recover from errors. We had COMPlus environment
variables (including COMPlus_JitStdOutFile) set for both the parent and
child processes so the jit emitted disassembly, etc. to the same file.
The parent process was still active when the child was running (e.g.,
it called p.StandardOutput.ReadToEnd()). That resulted in duplicate
methods (especially from System.Private.Corelib) in the dasm file and
non-deterministic output interleaving.
The fix is not to set the COMPlus environment variables in jit-dasm-pmi
when invoking pmi but instead pass them in PMIENV environment variable
as semicolon-separated name=value pairs and set the actual COMPlus
environment variables only on child processes.
I also fixed a potential deadlock in output/error redirection: at least
one of the redirected steams has to be read asynchronously.
See https://docs.microsoft.com/en-us/dotnet/api/system.diagnostics.process.standardoutput
for details.