[Patch 0/2] Checkpoint and restart of file locks

Sukadev Bhattiprolu sukadev at linux.vnet.ibm.com
Tue Jan 11 16:17:00 PST 2011

Oren Laadan [orenl at cs.columbia.edu] wrote:
| The following two patches add c/r support for POSIX file locks. They
| are based on the series posted by Suka some time ago:
| https://lists.linux-foundation.org/pipermail/containers/2010-October/025855.html
| Suka:
| When I read your post I thought that boht posix and flock were   
| supported, but looking at the code I saw only posix. Am I missing
| anything ?

Hmm. I have been working with only the posix locks for starters since
it checkpoints/restarts the more general, 'struct file_lock' object.

But I should make that explicit. If a patch description implies that
it covers flocks can be checkpointed, please let me know and I will
fix it.

C/R of leases/file-owner info is blocked on C/R of 'struct pids'.

| Also, I'm not sure that all the input is well sanitized during
| restart - for example, the possible values for the lock flags.
| Any thoughts ?

I have this check in restore_file_locks():

+               ret = -EBADF;
+               if (h->fl_flags & FL_POSIX)
+                       ret = restore_one_posix_lock(ctx, file, fd, 

The other fields (lock type for instance) will be validated by flock64_set()
as for a new fcntl() call - no ?

| Lastly, do you have test cases to run against this patch and test
| it ?

Yes, I added some tests to cr-tests a while ago and then updated them.
Will find the updates and post them.


More information about the Containers mailing list