Messages in this thread | ![/](/images/icornerl.gif) | | Date | Thu, 12 Mar 2015 08:24:09 -0400 | From | Josef Bacik <> | Subject | Re: [PATCH 2/9] inode: add IOP_NOTHASHED to avoid inode hash lock in evict |
| |
On 03/12/2015 05:52 AM, Al Viro wrote: > On Tue, Mar 10, 2015 at 03:45:17PM -0400, Josef Bacik wrote: >> From: Dave Chinner <dchinner@redhat.com> >> >> Some filesystems don't use the VFS inode hash and fake the fact they >> are hashed so that all the writeback code works correctly. However, >> this means the evict() path still tries to remove the inode from the >> hash, meaning that the inode_hash_lock() needs to be taken >> unnecessarily. Hence under certain workloads the inode_hash_lock can >> be contended even if the inode is never actually hashed. >> >> To avoid this, add an inode opflag to allow inode_hash_remove() to >> avoid taking the hash lock on inodes have never actually been >> hashed. > > Why bother with flags, etc. when we could just do > > static inline bool hlist_fake(struct hlist_node *h) > { > return h->pprev == &h->next; > } > >> - if (!inode_unhashed(inode)) >> + if (!((inode->i_opflags & IOP_NOTHASHED) || inode_unhashed(inode))) > > and turn this check into > if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash)) > > instead? >
I made the change and sent the new patch, hopefully it worked, sometimes my laptop screws up git-send-email. If everything looks good to you Al you can pull from my tree if you don't want to scrape from the list
git://git.kernel.org/pub/scm/linux/kernel/git/josef/btrfs-next.git superblock-scaling
Thanks,
Josef
| ![\](/images/icornerr.gif) |