Subrata Modak [[email protected]] wrote:
|
| > > pidns21:
| > > The pidns21.c testcase verifies that container-init is terminated
| > > by SIGUSR1 when:
| > > - a handler is specified for SIGUSR1,
| > > - container-init blocks SIGUSR1,
| > > - parent queues SIGUSR1 and
| > > - handler for SIGUSR1 is set to system default before SIGUSR1 is
| > > unblocked.
I know I had acked this test before, but back then the actual implementation
of the signal semantics in the kernel were not complete.
To simplify the implementation of the semantics, it was decided that
SIGKILL/SIGSTOP would be the only reliable signals from a parent
container. IOW, container-init would ignore SIGUSR1 or SIGINT, SIGQUIT
etc even if sent from a parent container.
See patchset/discussion:
http://lkml.org/lkml/2009/1/17/131
(which is not yet merged, but appears to be close to consensus)
The rationale for this simplification is that any serious
'container-init' would explicitly SIG_IGN all signals that it is
not interested in. So the only signals that would be in SIG_DFL
state would be SIGKILL/SIGSTOP.
Effectively, testcase pidns21 will fail if/when the above patchset
(specifically, patch 5/6) is merged.
------------------------------------------------------------------------------
Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
-OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
-Strategies to boost innovation and cut costs with open source participation
-Receive a $600 discount off the registration fee with the source code: SFAD
http://p.sf.net/sfu/XcvMzF8H
_______________________________________________
Ltp-list mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/ltp-list
