Re: [PATCH] fuse: add fast path for fuse_range_is_writeback

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Sorry for the late reply.

On Wed, Aug 14, 2024 at 05:56:06PM +0800, Jingbo Xu wrote:
> 
> 
> On 8/14/24 5:36 PM, yangyun wrote:
> > In some cases, the fi->writepages may be empty. And there is no need
> > to check fi->writepages with spin_lock, which may have an impact on
> > performance due to lock contention. For example, in scenarios where
> > multiple readers read the same file without any writers, or where
> > the page cache is not enabled.
> > 
> > Also remove the outdated comment since commit 6b2fb79963fb ("fuse:
> > optimize writepages search") has optimize the situation by replacing
> > list with rb-tree.
> > 
> > Signed-off-by: yangyun <yangyun50@xxxxxxxxxx>
> > ---
> >  fs/fuse/file.c | 6 +++---
> >  1 file changed, 3 insertions(+), 3 deletions(-)
> > 
> > diff --git a/fs/fuse/file.c b/fs/fuse/file.c
> > index f39456c65ed7..59c911b61000 100644
> > --- a/fs/fuse/file.c
> > +++ b/fs/fuse/file.c
> > @@ -448,9 +448,6 @@ static struct fuse_writepage_args *fuse_find_writeback(struct fuse_inode *fi,
> >  
> >  /*
> >   * Check if any page in a range is under writeback
> > - *
> > - * This is currently done by walking the list of writepage requests
> > - * for the inode, which can be pretty inefficient.
> >   */
> >  static bool fuse_range_is_writeback(struct inode *inode, pgoff_t idx_from,
> >  				   pgoff_t idx_to)
> > @@ -458,6 +455,9 @@ static bool fuse_range_is_writeback(struct inode *inode, pgoff_t idx_from,
> >  	struct fuse_inode *fi = get_fuse_inode(inode);
> >  	bool found;
> >  
> > +	if (RB_EMPTY_ROOT(&fi->writepages))
> > +		return false;
> 
> fi->lock is held when inserting wpa into fi->writepages rbtree (see
> fuse_writepage_add()).  I doubt if there is race condition when checking
> if fi->writepages rbtree is empty without fi->lock held.

The code can make sure that there are no race conditions because:
1. For O_DIRECT and FOPEN_DIRECT_IO with fc->direct_io_allow_mmap, the `filemap_write_and_wait_range` before can make the insert operation to be happend before the check operation.
2. For other cases, there are no pagecache operaions so the fi->writepages is always empty.

In my usercase, the fi->writepages is usually empty but the spin_lock associated with it contributes a great impact on the performace of my filesystem due to lock contention. So optimize it.
> 
> -- 
> Thanks,
> Jingbo




[Index of Archives]     [Linux Ext4 Filesystem]     [Union Filesystem]     [Filesystem Testing]     [Ceph Users]     [Ecryptfs]     [NTFS 3]     [AutoFS]     [Kernel Newbies]     [Share Photos]     [Security]     [Netfilter]     [Bugtraq]     [Yosemite News]     [MIPS Linux]     [ARM Linux]     [Linux Security]     [Linux Cachefs]     [Reiser Filesystem]     [Linux RAID]     [NTFS 3]     [Samba]     [Device Mapper]     [CEPH Development]

  Powered by Linux