fuse: separate queue for FORGET requests

Terje Malmedal reports that a fuse filesystem with 32 million inodes
on a machine with lots of memory can go unresponsive for up to 30
minutes when all those inodes are evicted from the icache.

The reason is that FORGET messages, sent when the inode is evicted,
are queued up together with regular filesystem requests, and while the
huge queue of FORGET messages are processed no other filesystem
operation can proceed.

Since a full fuse request structure is allocated for each inode, these
take up quite a bit of memory as well.

To solve these issues, create a slim 'fuse_forget_link' structure
containing just the minimum of information required to send the FORGET
request and chain these on a separate queue.

When userspace is asking for a request make sure that FORGET and
non-FORGET requests are selected fairly: for each 8 non-FORGET allow
16 FORGET requests.  This will make sure FORGETs do not pile up, yet
other requests are also allowed to proceed while the queued FORGETs
are processed.

Reported-by: Terje Malmedal <terje.malmedal@usit.uio.no>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>

1 files changed, 26 insertions, 27 deletions

diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index c9627c95482d..6ea42e98cb17 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -10,9 +10,9 @@
 
 #include <linux/pagemap.h>
 #include <linux/file.h>
-#include <linux/gfp.h>
 #include <linux/sched.h>
 #include <linux/namei.h>
+#include <linux/slab.h>
 
 #if BITS_PER_LONG >= 64
 static inline void fuse_dentry_settime(struct dentry *entry, u64 time)
@@ -165,7 +165,7 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
                 struct fuse_entry_out outarg;
                 struct fuse_conn *fc;
                 struct fuse_req *req;
-                struct fuse_req *forget_req;
+                struct fuse_forget_link *forget;
                 struct dentry *parent;
                 u64 attr_version;
 
@@ -178,8 +178,8 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
                 if (IS_ERR(req))
                         return 0;
 
-                forget_req = fuse_get_req(fc);
-                if (IS_ERR(forget_req)) {
+                forget = fuse_alloc_forget();
+                if (!forget) {
                         fuse_put_request(fc, req);
                         return 0;
                 }
@@ -199,15 +199,14 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
                 if (!err) {
                         struct fuse_inode *fi = get_fuse_inode(inode);
                         if (outarg.nodeid != get_node_id(inode)) {
-                                fuse_send_forget(fc, forget_req,
-                                                 outarg.nodeid, 1);
+                                fuse_queue_forget(fc, forget, outarg.nodeid, 1);
                                 return 0;
                         }
                         spin_lock(&fc->lock);
                         fi->nlookup++;
                         spin_unlock(&fc->lock);
                 }
-                fuse_put_request(fc, forget_req);
+                kfree(forget);
                 if (err || (outarg.attr.mode ^ inode->i_mode) & S_IFMT)
                         return 0;
 
@@ -259,7 +258,7 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
 {
         struct fuse_conn *fc = get_fuse_conn_super(sb);
         struct fuse_req *req;
-        struct fuse_req *forget_req;
+        struct fuse_forget_link *forget;
         u64 attr_version;
         int err;
 
@@ -273,9 +272,9 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
         if (IS_ERR(req))
                 goto out;
 
-        forget_req = fuse_get_req(fc);
-        err = PTR_ERR(forget_req);
-        if (IS_ERR(forget_req)) {
+        forget = fuse_alloc_forget();
+        err = -ENOMEM;
+        if (!forget) {
                 fuse_put_request(fc, req);
                 goto out;
         }
@@ -301,13 +300,13 @@ int fuse_lookup_name(struct super_block *sb, u64 nodeid, struct qstr *name,
                            attr_version);
         err = -ENOMEM;
         if (!*inode) {
-                fuse_send_forget(fc, forget_req, outarg->nodeid, 1);
+                fuse_queue_forget(fc, forget, outarg->nodeid, 1);
                 goto out;
         }
         err = 0;
 
  out_put_forget:
-        fuse_put_request(fc, forget_req);
+        kfree(forget);
  out:
         return err;
 }
@@ -374,7 +373,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
         struct inode *inode;
         struct fuse_conn *fc = get_fuse_conn(dir);
         struct fuse_req *req;
-        struct fuse_req *forget_req;
+        struct fuse_forget_link *forget;
         struct fuse_create_in inarg;
         struct fuse_open_out outopen;
         struct fuse_entry_out outentry;
@@ -388,9 +387,9 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
         if (flags & O_DIRECT)
                 return -EINVAL;
 
-        forget_req = fuse_get_req(fc);
-        if (IS_ERR(forget_req))
-                return PTR_ERR(forget_req);
+        forget = fuse_alloc_forget();
+        if (!forget)
+                return -ENOMEM;
 
         req = fuse_get_req(fc);
         err = PTR_ERR(req);
@@ -448,10 +447,10 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
         if (!inode) {
                 flags &= ~(O_CREAT | O_EXCL | O_TRUNC);
                 fuse_sync_release(ff, flags);
-                fuse_send_forget(fc, forget_req, outentry.nodeid, 1);
+                fuse_queue_forget(fc, forget, outentry.nodeid, 1);
                 return -ENOMEM;
         }
-        fuse_put_request(fc, forget_req);
+        kfree(forget);
         d_instantiate(entry, inode);
         fuse_change_entry_timeout(entry, &outentry);
         fuse_invalidate_attr(dir);
@@ -469,7 +468,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, int mode,
  out_put_request:
         fuse_put_request(fc, req);
  out_put_forget_req:
-        fuse_put_request(fc, forget_req);
+        kfree(forget);
         return err;
 }
 
@@ -483,12 +482,12 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
         struct fuse_entry_out outarg;
         struct inode *inode;
         int err;
-        struct fuse_req *forget_req;
+        struct fuse_forget_link *forget;
 
-        forget_req = fuse_get_req(fc);
-        if (IS_ERR(forget_req)) {
+        forget = fuse_alloc_forget();
+        if (!forget) {
                 fuse_put_request(fc, req);
-                return PTR_ERR(forget_req);
+                return -ENOMEM;
         }
 
         memset(&outarg, 0, sizeof(outarg));
@@ -515,10 +514,10 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
         inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation,
                           &outarg.attr, entry_attr_timeout(&outarg), 0);
         if (!inode) {
-                fuse_send_forget(fc, forget_req, outarg.nodeid, 1);
+                fuse_queue_forget(fc, forget, outarg.nodeid, 1);
                 return -ENOMEM;
         }
-        fuse_put_request(fc, forget_req);
+        kfree(forget);
 
         if (S_ISDIR(inode->i_mode)) {
                 struct dentry *alias;
@@ -541,7 +540,7 @@ static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
         return 0;
 
  out_put_forget_req:
-        fuse_put_request(fc, forget_req);
+        kfree(forget);
         return err;
 }