Add FILE* structure manipulation exploit#35
Open
nataliakokoromyti wants to merge 2 commits intogpu-mode:masterfrom
Open
Add FILE* structure manipulation exploit#35nataliakokoromyti wants to merge 2 commits intogpu-mode:masterfrom
nataliakokoromyti wants to merge 2 commits intogpu-mode:masterfrom
Conversation
Walks glibc's _IO_list_all linked list to find the FILE* for the output pipe, then modifies its _fileno field to redirect fprintf writes through a filtering thread that rewrites timing values. Unlike pipe fd interposition, this modifies glibc-internal data structures rather than OS-level file descriptors.
Updated comments to clarify the exploit's mechanism and defense strategies.
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
C's fprintf() (which we use to write timing results) writes through glibc's internal FILE* struct so the exploit hijacks that struct to reroute the harness's timing output through a filter thread. the filter sees every line before the parent process does, so it can rewrite the reported times to whatever it wants while the kernel just computes the correct result normally.
here for example, the kernel itself is the same correct kernel as submission_correct.py but every measurement is rewritten to 1.0 µs before forwarding it to the real output.