| Message ID | 20220331153130.41287-2-jlayton@kernel.org (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | ceph+fscrypt: fully-working prototype | expand |
On Thu, Mar 31, 2022 at 11:30:37AM -0400, Jeff Layton wrote: > Ceph needs to be able to allocate inodes ahead of a create that might > involve a fscrypt-encrypted inode. new_inode() almost fits the bill, > but it puts the inode on the sb->s_inodes list and when we go to hash > it, that might be done again. > > We could work around that by setting I_CREATING on the new inode, but > that causes ilookup5 to return -ESTALE if something tries to find it > before I_NEW is cleared. This is desirable behavior for most > filesystems, but doesn't work for ceph. > > To work around all of this, just use new_inode_pseudo which doesn't add > it to the sb->s_inodes list. Umm... I can live with that, but... why not just leave the hash insertion until the thing is fully set up and you are ready to clear I_NEW?
On Thu, 2022-03-31 at 19:50 +0000, Al Viro wrote: > On Thu, Mar 31, 2022 at 11:30:37AM -0400, Jeff Layton wrote: > > Ceph needs to be able to allocate inodes ahead of a create that might > > involve a fscrypt-encrypted inode. new_inode() almost fits the bill, > > but it puts the inode on the sb->s_inodes list and when we go to hash > > it, that might be done again. > > > > We could work around that by setting I_CREATING on the new inode, but > > that causes ilookup5 to return -ESTALE if something tries to find it > > before I_NEW is cleared. This is desirable behavior for most > > filesystems, but doesn't work for ceph. > > > > To work around all of this, just use new_inode_pseudo which doesn't add > > it to the sb->s_inodes list. > > Umm... I can live with that, but... why not just leave the hash insertion > until the thing is fully set up and you are ready to clear I_NEW? If the thing is already in the hash at the end then we have to go back and redo the inode update with the correct inode. That can be messy too -- in some cases we hand off strings and such. On IRC, Al suggested that we instead change the test in inode_insert5 so we can avoid the double list_add. I'm testing a patch now that seems to be working, so I'll plan to drop this one in favor of that approach. Thanks for the help!
diff --git a/fs/inode.c b/fs/inode.c index 63324df6fa27..9ddf7d1a7359 100644 --- a/fs/inode.c +++ b/fs/inode.c @@ -1025,6 +1025,7 @@ struct inode *new_inode_pseudo(struct super_block *sb) } return inode; } +EXPORT_SYMBOL(new_inode_pseudo); /** * new_inode - obtain an inode
Ceph needs to be able to allocate inodes ahead of a create that might involve a fscrypt-encrypted inode. new_inode() almost fits the bill, but it puts the inode on the sb->s_inodes list and when we go to hash it, that might be done again. We could work around that by setting I_CREATING on the new inode, but that causes ilookup5 to return -ESTALE if something tries to find it before I_NEW is cleared. This is desirable behavior for most filesystems, but doesn't work for ceph. To work around all of this, just use new_inode_pseudo which doesn't add it to the sb->s_inodes list. Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Jeff Layton <jlayton@kernel.org> --- fs/inode.c | 1 + 1 file changed, 1 insertion(+)