05.03.2012 19:16, Jeff Layton пишет:
On Mon, 05 Mar 2012 18:53:27 +0400
Stanislav Kinsbursky<skinsbursky@xxxxxxxxxxxxx> wrote:
05.03.2012 18:39, Jeff Layton пишет:
Ok, I've started looking at the existing examples to try and ease the
eventual containerization of this. Am I correct that the existing
blocklayout pipe containerization is not yet complete?
AIUI, you have a rpc_pipe per nfs_net struct, and those structs are
allocated on a per-net basis. bl_pipe_downcall does this however:
if (copy_from_user(&bl_mount_reply, src, mlen) != 0)
return -EFAULT;
...but bl_mount_reply is global:
static struct bl_dev_msg bl_mount_reply;
So, if you're getting downcalls in two different network namespaces at
the same time this could end up corrupt. It seems like bl_mount_reply
needs to be part of nfs_net as well.
Thanks for this catch, Jeff!
I need to fix this.
When you get a downcall, you have a filp pointer. Is there any way to
determine the "correct" nfs_net object for a particular filp?
Yes, sure.
This nfs_net object will be, obviously, in the same network namespace, as PipeFS
superblock. Network namespace is stored on PipeFS sb->s_fs_info.
IOW, code below should works:
struct nfs_net *nn = net_generic(file->f_dentry->d_sb->s_fs_info, nfs_net_id);
Ok, I'm going to have to do something similar for the cld upcall
eventually. We'll need a per-ns list of upcalls in flight, xid counter
and spinlock.
What may be best is to go ahead and introduce a nfsd_net struct as a
per-net object start putting stuff inside it.
Agreed.
Would be cool, if you'll introduce this with your patch set.
--
Best regards,
Stanislav Kinsbursky
--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
