Repository navigation
pam: back off when auth fails without a conversation - #1085
Open
martintrojer wants to merge 1 commit into
Open
martintrojer wants to merge 1 commit into
martintrojer wants to merge 1 commit into
Conversation
martintrojer
marked this pull request as ready for review
October 28, 2026 10:32
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
If PAM fails without ever calling the conversation (pam_deny in the stack, pam_start failure, empty username) the auth thread retries in a tight loop. It spins a core and the fail counter/log grow by thousands per second. With no username it eventually segfaults.
Now it waits up to 1 s on the existing
inputSubmittedConditionbefore retrying, soterminate()still wakes it (the #1059 deadlock fix is untouched). Empty username is a permanent failure: enqueue it once and stop.pam_deny, ~3 s locked: 11990 failed attempts before, 4 after. SIGUSR1 to exit goes from 25 ms to ~117 ms on that stack. 340 lock+terminate runs, no hang. Normal prompting stack unchanged. Real build in an Arch container.
Related: #1069 waits on the same condvar in
init(), probably a textual conflict.