Messages in this thread | | | From | Fox Chen <> | Date | Thu, 3 Dec 2020 14:35:54 +0800 | Subject | Re: [PATCH 2/2] kernfs: remove mutex in kernfs_dop_revalidate |
| |
On Thu, Dec 3, 2020 at 2:26 AM Greg KH <gregkh@linuxfoundation.org> wrote: > > On Wed, Dec 02, 2020 at 10:58:37PM +0800, Fox Chen wrote: > > There is a big mutex in kernfs_dop_revalidate which slows down the > > concurrent performance of kernfs. > > > > Since kernfs_dop_revalidate only does some checks, the lock is > > largely unnecessary. Also, according to kernel filesystem locking > > document: > > https://www.kernel.org/doc/html/latest/filesystems/locking.html > > locking is not in the protocal for d_revalidate operation. > > > > This patch remove this mutex from > > kernfs_dop_revalidate, so kernfs_dop_revalidate > > can run concurrently. > > > > Signed-off-by: Fox Chen <foxhlchen@gmail.com> > > --- > > fs/kernfs/dir.c | 9 +++------ > > 1 file changed, 3 insertions(+), 6 deletions(-) > > > > diff --git a/fs/kernfs/dir.c b/fs/kernfs/dir.c > > index 9aec80b9d7c6..c2267c93f546 100644 > > --- a/fs/kernfs/dir.c > > +++ b/fs/kernfs/dir.c > > @@ -26,7 +26,6 @@ static DEFINE_SPINLOCK(kernfs_idr_lock); /* root->ino_idr */ > > > > static bool kernfs_active(struct kernfs_node *kn) > > { > > - lockdep_assert_held(&kernfs_mutex); > > return atomic_read(&kn->active) >= 0; > > } > > > > @@ -557,10 +556,9 @@ static int kernfs_dop_revalidate(struct dentry *dentry, unsigned int flags) > > > > /* Always perform fresh lookup for negatives */ > > if (d_really_is_negative(dentry)) > > - goto out_bad_unlocked; > > + goto out_bad; > > > > kn = kernfs_dentry_node(dentry); > > - mutex_lock(&kernfs_mutex); > > > > /* The kernfs node has been deactivated */ > > if (!kernfs_active(kn)) > > @@ -579,11 +577,8 @@ static int kernfs_dop_revalidate(struct dentry *dentry, unsigned int flags) > > kernfs_info(dentry->d_sb)->ns != kn->ns) > > goto out_bad; > > > > - mutex_unlock(&kernfs_mutex); > > return 1; > > out_bad: > > - mutex_unlock(&kernfs_mutex); > > -out_bad_unlocked: > > return 0; > > } > > > > @@ -650,6 +645,8 @@ static struct kernfs_node *__kernfs_new_node(struct kernfs_root *root, > > kn->mode = mode; > > kn->flags = flags; > > > > + rwlock_init(&kn->iattr_rwlock); > > Ah, now you initialize this, it should go into patch 1, right? :) > Yes, it's my fault. It should be in patch 1. Sorry.
| |