Refactor clone(2) child stack creation. #167
Merged
Conversation
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
1. Allocate the child stack using mmap, instead of on the stack. The LLVM may optimize the memory in the wrong way to create a race condition where to memory is still in use but free'ed. 2. Instead of having fix size 4K, we use the ulimit -s set on the system. This is consistant with what fork and pthread_create uses.
utam0k
reviewed
Jul 29, 2021
utam0k
reviewed
Jul 29, 2021
utam0k
reviewed
Jul 29, 2021
utam0k
reviewed
Jul 29, 2021
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Fix #163
After looking into more about how pthread_create and fork sets up child stack, refactored/fix how the child stack was created. Before this PR, as #163 indicates, LLVM with high optimization level will make memory behave not as we originally imagined. Therefore, we delved into lower APIs using
mmapto create the child stack memory. This also gives us the benefit of reserving the max process/thread stack size on the system, without committing to actual physical memory. As a result, we don't have to hardcode a child's stack size or guess how much we need for the child process.