mirror/runc
1
0
Fork 0
mirror of https://github.com/opencontainers/runc.git synced 2026-07-10 21:53:57 +08:00
Code Issues Packages Projects Releases Wiki Activity

Fix SIGCHLD race in signal handler setup

Browse Source 
When signal installation was moved to a goroutine for performance,
containers that exited quickly could complete before SIGCHLD was
registered, causing runc to hang waiting for the signal.

This fix ensures SIGCHLD is registered immediately in the main thread
before other signals are handled in the goroutine, maintaining performance
while guaranteeing no missed SIGCHLD notifications for fast-exiting
containers.

Reported-by: Ayato Tokubi <atokubi@redhat.com>
Signed-off-by: lifubang <lifubang@acmcoder.com>
This commit is contained in:
lifubang
2026-04-01 01:17:16 +00:00
parent 348d92f497
commit 404181e4cc
1 changed files with 7 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
signals.go
+7 -4
View File
@@ -26,14 +26,17 @@ func newSignalHandler(enableSubreaper bool, notifySocket *notifySocket) chan *si
}
}
handler := make(chan *signalHandler)
// Ensure that we have a large buffer size so that we do not miss any
// signals in case we are not processing them fast enough.
s := make(chan os.Signal, signalBufferSize)
// signal.Notify is actually quite expensive, as it has to configure the
// signal mask and add signal handlers for all signals (all ~65 of them).
// So, defer this to a background thread while doing the rest of the io/tty
// setup.
// setup, except for SIGCHLD which is very important (see #5208).
signal.Notify(s, unix.SIGCHLD)
go func() {
// ensure that we have a large buffer size so that we do not miss any
// signals in case we are not processing them fast enough.
s := make(chan os.Signal, signalBufferSize)
// handle all signals for the process.
signal.Notify(s)
handler <- &signalHandler{