Bug 2300 - reexec.sh and login-timeout.sh fail due to a race condition
Summary: reexec.sh and login-timeout.sh fail due to a race condition
Status: CLOSED FIXED
Alias: None
Product: Portable OpenSSH
Classification: Unclassified
Component: Regression tests (show other bugs)
Version: 6.7p1
Hardware: Other Linux
: P5 enhancement
Assignee: Assigned to nobody
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-10-25 09:08 AEDT by Petr Lautrbach
Modified: 2021-04-23 15:03 AEST (History)
1 user (show)

See Also:

Attachments
fix race condition between kill sshd and start_sshd (1.97 KB, patch)
2014-10-25 09:08 AEDT, Petr Lautrbach 		no flags Details | Diff
View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Petr Lautrbach 2014-10-25 09:08:42 AEDT 
Created attachment 2499 [details]
fix race condition between kill sshd and start_sshd

run test reexec.sh ...
test config passing
FATAL: no sshd running on port 4242
Makefile:168: recipe for target 't-exec' failed
make[1]: *** [t-exec] Error 1
make[1]: Leaving directory '/home/plautrba/devel/src/bachradsusi.github/openssh-portable.git/regress'
Makefile:483: recipe for target 'tests' failed
make: *** [tests] Error 2

The problem is in the race between

$SUDO kill `$SUDO cat $PIDFILE`

and following

start_sshd

While sshd is still runnning and closing opened socket after it got a signal, the test tries to run a new sshd instance which cannot bind to not-yet-released socket. The attached patch tries to fix it.
Comment 1 Damien Miller 2020-01-25 23:56:42 AEDT 
the tests have had a stop_sshd function that waits for sshd to actually exit since 2016.
Comment 2 Damien Miller 2021-04-23 15:03:37 AEST 
closing resolved bugs as of 8.6p1 release