Re: [PATCH] vfs: dont chain pipe/anon/socket on superblock s_inodes list
flat view
From: Eric Dumazet <hidden>
Date: 2011-07-26 09:36:34
Also in:
linux-fsdevel, lkml
Subsystem:
filesystems (vfs and infrastructure), networking [general], networking [sockets], the rest · Maintainers:
Alexander Viro, Christian Brauner, "David S. Miller", Eric Dumazet, Jakub Kicinski, Paolo Abeni, Kuniyuki Iwashima, Willem de Bruijn, Linus Torvalds
Le mardi 26 juillet 2011 à 05:03 -0400, Christoph Hellwig a écrit :
On Tue, Jul 26, 2011 at 10:21:06AM +0200, Eric Dumazet wrote:quoted
Well, not 'last' contention point, as we still hit remove_inode_hash(),There should be no ned to put pipe or anon inodes on the inode hash. Probably sockets don't need it either, but I'd need to look at it in detail.quoted
inode_wb_list_del()The should never be on the wb list either, doing an unlocked check for actually beeing on the list before taking the lock should help you.
Yes, it might even help regular inodes ;)
quoted
inode_lru_list_del(),No real need to keep inodes in the LRU if we only allocate them using new_inode but never look them up either. You might want to try setting .drop_inode to generic_delete_inode for these.
Yes, I'll take a look, thanks.
quoted
+struct inode *__new_inode(struct super_block *sb) +{ + struct inode *inode = alloc_inode(sb); + + if (inode) { + spin_lock(&inode->i_lock); + inode->i_state = 0; + spin_unlock(&inode->i_lock); + INIT_LIST_HEAD(&inode->i_sb_list); + } + return inode; +}This needs a much better name like new_inode_pseudo, and a kerneldoc comment explaining when it is safe to use, and the consequences, which appear to me: - fs may never be unmount - quotas can't work on the filesystem - writeback can't work on the filesystem
Thanks for reviewing, here is v2 of the patch, addressing your comments. [PATCH v2] vfs: dont chain pipe/anon/socket on superblock s_inodes list Workloads using pipes and sockets hit inode_sb_list_lock contention. superblock s_inodes list is needed for quota, dirty, pagecache and fsnotify management. pipe/anon/socket fs are clearly not candidates for these. Signed-off-by: Eric Dumazet <redacted> --- v2: address Christoph comments fs/anon_inodes.c | 2 +- fs/inode.c | 39 ++++++++++++++++++++++++++++++--------- fs/pipe.c | 2 +- include/linux/fs.h | 3 ++- net/socket.c | 2 +- 5 files changed, 35 insertions(+), 13 deletions(-)
diff --git a/fs/anon_inodes.c b/fs/anon_inodes.c
index 4d433d3..f11e43e 100644
--- a/fs/anon_inodes.c
+++ b/fs/anon_inodes.c@@ -187,7 +187,7 @@ EXPORT_SYMBOL_GPL(anon_inode_getfd); */ static struct inode *anon_inode_mkinode(void) { - struct inode *inode = new_inode(anon_inode_mnt->mnt_sb); + struct inode *inode = new_inode_pseudo(anon_inode_mnt->mnt_sb); if (!inode) return ERR_PTR(-ENOMEM);
diff --git a/fs/inode.c b/fs/inode.c
index 96c77b8..319b93b 100644
--- a/fs/inode.c
+++ b/fs/inode.c@@ -362,9 +362,11 @@ EXPORT_SYMBOL_GPL(inode_sb_list_add); static inline void inode_sb_list_del(struct inode *inode) { - spin_lock(&inode_sb_list_lock); - list_del_init(&inode->i_sb_list); - spin_unlock(&inode_sb_list_lock); + if (!list_empty(&inode->i_sb_list)) { + spin_lock(&inode_sb_list_lock); + list_del_init(&inode->i_sb_list); + spin_unlock(&inode_sb_list_lock); + } } static unsigned long hash(struct super_block *sb, unsigned long hashval)
@@ -797,6 +799,29 @@ unsigned int get_next_ino(void) EXPORT_SYMBOL(get_next_ino); /** + * new_inode_pseudo - obtain an inode + * @sb: superblock + * + * Allocates a new inode for given superblock. + * Inode wont be chained in superblock s_inodes list + * This means : + * - fs can't be unmount + * - quotas, fsnotify, writeback can't work + */ +struct inode *new_inode_pseudo(struct super_block *sb) +{ + struct inode *inode = alloc_inode(sb); + + if (inode) { + spin_lock(&inode->i_lock); + inode->i_state = 0; + spin_unlock(&inode->i_lock); + INIT_LIST_HEAD(&inode->i_sb_list); + } + return inode; +} + +/** * new_inode - obtain an inode * @sb: superblock *
@@ -814,13 +839,9 @@ struct inode *new_inode(struct super_block *sb) spin_lock_prefetch(&inode_sb_list_lock); - inode = alloc_inode(sb); - if (inode) { - spin_lock(&inode->i_lock); - inode->i_state = 0; - spin_unlock(&inode->i_lock); + inode = new_inode_pseudo(sb); + if (inode) inode_sb_list_add(inode); - } return inode; } EXPORT_SYMBOL(new_inode);
diff --git a/fs/pipe.c b/fs/pipe.c
index 1b7f9af..0e0be1d 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c@@ -948,7 +948,7 @@ static const struct dentry_operations pipefs_dentry_operations = { static struct inode * get_pipe_inode(void) { - struct inode *inode = new_inode(pipe_mnt->mnt_sb); + struct inode *inode = new_inode_pseudo(pipe_mnt->mnt_sb); struct pipe_inode_info *pipe; if (!inode)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index a665804..cc363fa 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h@@ -2310,7 +2310,8 @@ extern void __iget(struct inode * inode); extern void iget_failed(struct inode *); extern void end_writeback(struct inode *); extern void __destroy_inode(struct inode *); -extern struct inode *new_inode(struct super_block *); +extern struct inode *new_inode_pseudo(struct super_block *sb); +extern struct inode *new_inode(struct super_block *sb); extern void free_inode_nonrcu(struct inode *inode); extern int should_remove_suid(struct dentry *); extern int file_remove_suid(struct file *);
diff --git a/net/socket.c b/net/socket.c
index 02dc82d..26ed35c 100644
--- a/net/socket.c
+++ b/net/socket.c@@ -467,7 +467,7 @@ static struct socket *sock_alloc(void) struct inode *inode; struct socket *sock; - inode = new_inode(sock_mnt->mnt_sb); + inode = new_inode_pseudo(sock_mnt->mnt_sb); if (!inode) return NULL;