linux-sg2042/fs/fuse/file.c

2130 lines
51 KiB
C
Raw Normal View History

/*
FUSE: Filesystem in Userspace
Copyright (C) 2001-2008 Miklos Szeredi <miklos@szeredi.hu>
This program can be distributed under the terms of the GNU GPL.
See the file COPYING.
*/
#include "fuse_i.h"
#include <linux/pagemap.h>
#include <linux/slab.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/module.h>
#include <linux/compat.h>
#include <linux/swap.h>
static const struct file_operations fuse_direct_io_file_operations;
static int fuse_send_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
int opcode, struct fuse_open_out *outargp)
{
struct fuse_open_in inarg;
struct fuse_req *req;
int err;
req = fuse_get_req(fc);
if (IS_ERR(req))
return PTR_ERR(req);
memset(&inarg, 0, sizeof(inarg));
inarg.flags = file->f_flags & ~(O_CREAT | O_EXCL | O_NOCTTY);
if (!fc->atomic_o_trunc)
inarg.flags &= ~O_TRUNC;
req->in.h.opcode = opcode;
req->in.h.nodeid = nodeid;
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
req->out.numargs = 1;
req->out.args[0].size = sizeof(*outargp);
req->out.args[0].value = outargp;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
return err;
}
struct fuse_file *fuse_file_alloc(struct fuse_conn *fc)
{
struct fuse_file *ff;
ff = kmalloc(sizeof(struct fuse_file), GFP_KERNEL);
if (unlikely(!ff))
return NULL;
ff->fc = fc;
ff->reserved_req = fuse_request_alloc();
if (unlikely(!ff->reserved_req)) {
kfree(ff);
return NULL;
}
INIT_LIST_HEAD(&ff->write_entry);
atomic_set(&ff->count, 0);
RB_CLEAR_NODE(&ff->polled_node);
init_waitqueue_head(&ff->poll_wait);
spin_lock(&fc->lock);
ff->kh = ++fc->khctr;
spin_unlock(&fc->lock);
return ff;
}
void fuse_file_free(struct fuse_file *ff)
{
fuse_request_free(ff->reserved_req);
kfree(ff);
}
struct fuse_file *fuse_file_get(struct fuse_file *ff)
{
atomic_inc(&ff->count);
return ff;
}
static void fuse_release_async(struct work_struct *work)
{
struct fuse_req *req;
struct fuse_conn *fc;
struct path path;
req = container_of(work, struct fuse_req, misc.release.work);
path = req->misc.release.path;
fc = get_fuse_conn(path.dentry->d_inode);
fuse_put_request(fc, req);
path_put(&path);
}
static void fuse_release_end(struct fuse_conn *fc, struct fuse_req *req)
{
if (fc->destroy_req) {
/*
* If this is a fuseblk mount, then it's possible that
* releasing the path will result in releasing the
* super block and sending the DESTROY request. If
* the server is single threaded, this would hang.
* For this reason do the path_put() in a separate
* thread.
*/
atomic_inc(&req->count);
INIT_WORK(&req->misc.release.work, fuse_release_async);
schedule_work(&req->misc.release.work);
} else {
path_put(&req->misc.release.path);
}
}
static void fuse_file_put(struct fuse_file *ff, bool sync)
{
if (atomic_dec_and_test(&ff->count)) {
struct fuse_req *req = ff->reserved_req;
if (sync) {
fuse_request_send(ff->fc, req);
path_put(&req->misc.release.path);
fuse_put_request(ff->fc, req);
} else {
req->end = fuse_release_end;
fuse_request_send_background(ff->fc, req);
}
kfree(ff);
}
}
int fuse_do_open(struct fuse_conn *fc, u64 nodeid, struct file *file,
bool isdir)
{
struct fuse_open_out outarg;
struct fuse_file *ff;
int err;
int opcode = isdir ? FUSE_OPENDIR : FUSE_OPEN;
ff = fuse_file_alloc(fc);
if (!ff)
return -ENOMEM;
err = fuse_send_open(fc, nodeid, file, opcode, &outarg);
if (err) {
fuse_file_free(ff);
return err;
}
if (isdir)
outarg.open_flags &= ~FOPEN_DIRECT_IO;
ff->fh = outarg.fh;
ff->nodeid = nodeid;
ff->open_flags = outarg.open_flags;
file->private_data = fuse_file_get(ff);
return 0;
}
EXPORT_SYMBOL_GPL(fuse_do_open);
void fuse_finish_open(struct inode *inode, struct file *file)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = get_fuse_conn(inode);
if (ff->open_flags & FOPEN_DIRECT_IO)
file->f_op = &fuse_direct_io_file_operations;
if (!(ff->open_flags & FOPEN_KEEP_CACHE))
invalidate_inode_pages2(inode->i_mapping);
if (ff->open_flags & FOPEN_NONSEEKABLE)
nonseekable_open(inode, file);
if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) {
struct fuse_inode *fi = get_fuse_inode(inode);
spin_lock(&fc->lock);
fi->attr_version = ++fc->attr_version;
i_size_write(inode, 0);
spin_unlock(&fc->lock);
fuse_invalidate_attr(inode);
}
}
int fuse_open_common(struct inode *inode, struct file *file, bool isdir)
{
struct fuse_conn *fc = get_fuse_conn(inode);
int err;
/* VFS checks this, but only _after_ ->open() */
if (file->f_flags & O_DIRECT)
return -EINVAL;
err = generic_file_open(inode, file);
if (err)
return err;
err = fuse_do_open(fc, get_node_id(inode), file, isdir);
if (err)
return err;
fuse_finish_open(inode, file);
return 0;
}
static void fuse_prepare_release(struct fuse_file *ff, int flags, int opcode)
{
struct fuse_conn *fc = ff->fc;
struct fuse_req *req = ff->reserved_req;
struct fuse_release_in *inarg = &req->misc.release.in;
spin_lock(&fc->lock);
list_del(&ff->write_entry);
if (!RB_EMPTY_NODE(&ff->polled_node))
rb_erase(&ff->polled_node, &fc->polled_files);
spin_unlock(&fc->lock);
wake_up_interruptible_all(&ff->poll_wait);
inarg->fh = ff->fh;
inarg->flags = flags;
req->in.h.opcode = opcode;
req->in.h.nodeid = ff->nodeid;
req->in.numargs = 1;
req->in.args[0].size = sizeof(struct fuse_release_in);
req->in.args[0].value = inarg;
}
void fuse_release_common(struct file *file, int opcode)
{
struct fuse_file *ff;
struct fuse_req *req;
ff = file->private_data;
if (unlikely(!ff))
return;
req = ff->reserved_req;
fuse_prepare_release(ff, file->f_flags, opcode);
if (ff->flock) {
struct fuse_release_in *inarg = &req->misc.release.in;
inarg->release_flags |= FUSE_RELEASE_FLOCK_UNLOCK;
inarg->lock_owner = fuse_lock_owner_id(ff->fc,
(fl_owner_t) file);
}
/* Hold vfsmount and dentry until release is finished */
path_get(&file->f_path);
req->misc.release.path = file->f_path;
/*
* Normally this will send the RELEASE request, however if
* some asynchronous READ or WRITE requests are outstanding,
* the sending will be delayed.
*
* Make the release synchronous if this is a fuseblk mount,
* synchronous RELEASE is allowed (and desirable) in this case
* because the server can be trusted not to screw up.
*/
fuse_file_put(ff, ff->fc->destroy_req != NULL);
}
static int fuse_open(struct inode *inode, struct file *file)
{
return fuse_open_common(inode, file, false);
}
static int fuse_release(struct inode *inode, struct file *file)
{
fuse_release_common(file, FUSE_RELEASE);
/* return value is ignored by VFS */
return 0;
}
void fuse_sync_release(struct fuse_file *ff, int flags)
{
WARN_ON(atomic_read(&ff->count) > 1);
fuse_prepare_release(ff, flags, FUSE_RELEASE);
ff->reserved_req->force = 1;
fuse_request_send(ff->fc, ff->reserved_req);
fuse_put_request(ff->fc, ff->reserved_req);
kfree(ff);
}
EXPORT_SYMBOL_GPL(fuse_sync_release);
/*
* Scramble the ID space with XTEA, so that the value of the files_struct
* pointer is not exposed to userspace.
*/
u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id)
{
u32 *k = fc->scramble_key;
u64 v = (unsigned long) id;
u32 v0 = v;
u32 v1 = v >> 32;
u32 sum = 0;
int i;
for (i = 0; i < 32; i++) {
v0 += ((v1 << 4 ^ v1 >> 5) + v1) ^ (sum + k[sum & 3]);
sum += 0x9E3779B9;
v1 += ((v0 << 4 ^ v0 >> 5) + v0) ^ (sum + k[sum>>11 & 3]);
}
return (u64) v0 + ((u64) v1 << 32);
}
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* Check if page 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_page_is_writeback(struct inode *inode, pgoff_t index)
{
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
struct fuse_req *req;
bool found = false;
spin_lock(&fc->lock);
list_for_each_entry(req, &fi->writepages, writepages_entry) {
pgoff_t curr_index;
BUG_ON(req->inode != inode);
curr_index = req->misc.write.in.offset >> PAGE_CACHE_SHIFT;
if (curr_index == index) {
found = true;
break;
}
}
spin_unlock(&fc->lock);
return found;
}
/*
* Wait for page writeback to be completed.
*
* Since fuse doesn't rely on the VM writeback tracking, this has to
* use some other means.
*/
static int fuse_wait_on_page_writeback(struct inode *inode, pgoff_t index)
{
struct fuse_inode *fi = get_fuse_inode(inode);
wait_event(fi->page_waitq, !fuse_page_is_writeback(inode, index));
return 0;
}
static int fuse_flush(struct file *file, fl_owner_t id)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_file *ff = file->private_data;
struct fuse_req *req;
struct fuse_flush_in inarg;
int err;
if (is_bad_inode(inode))
return -EIO;
if (fc->no_flush)
return 0;
req = fuse_get_req_nofail(fc, file);
memset(&inarg, 0, sizeof(inarg));
inarg.fh = ff->fh;
inarg.lock_owner = fuse_lock_owner_id(fc, id);
req->in.h.opcode = FUSE_FLUSH;
req->in.h.nodeid = get_node_id(inode);
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
req->force = 1;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
if (err == -ENOSYS) {
fc->no_flush = 1;
err = 0;
}
return err;
}
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* Wait for all pending writepages on the inode to finish.
*
* This is currently done by blocking further writes with FUSE_NOWRITE
* and waiting for all sent writes to complete.
*
* This must be called under i_mutex, otherwise the FUSE_NOWRITE usage
* could conflict with truncation.
*/
static void fuse_sync_writes(struct inode *inode)
{
fuse_set_nowrite(inode);
fuse_release_nowrite(inode);
}
int fuse_fsync_common(struct file *file, loff_t start, loff_t end,
int datasync, int isdir)
{
struct inode *inode = file->f_mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_file *ff = file->private_data;
struct fuse_req *req;
struct fuse_fsync_in inarg;
int err;
if (is_bad_inode(inode))
return -EIO;
err = filemap_write_and_wait_range(inode->i_mapping, start, end);
if (err)
return err;
if ((!isdir && fc->no_fsync) || (isdir && fc->no_fsyncdir))
return 0;
mutex_lock(&inode->i_mutex);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* Start writeback against all dirty pages of the inode, then
* wait for all outstanding writes, before sending the FSYNC
* request.
*/
err = write_inode_now(inode, 0);
if (err)
goto out;
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
fuse_sync_writes(inode);
req = fuse_get_req(fc);
if (IS_ERR(req)) {
err = PTR_ERR(req);
goto out;
}
memset(&inarg, 0, sizeof(inarg));
inarg.fh = ff->fh;
inarg.fsync_flags = datasync ? 1 : 0;
req->in.h.opcode = isdir ? FUSE_FSYNCDIR : FUSE_FSYNC;
req->in.h.nodeid = get_node_id(inode);
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
if (err == -ENOSYS) {
if (isdir)
fc->no_fsyncdir = 1;
else
fc->no_fsync = 1;
err = 0;
}
out:
mutex_unlock(&inode->i_mutex);
return err;
}
static int fuse_fsync(struct file *file, loff_t start, loff_t end,
int datasync)
{
return fuse_fsync_common(file, start, end, datasync, 0);
}
void fuse_read_fill(struct fuse_req *req, struct file *file, loff_t pos,
size_t count, int opcode)
{
struct fuse_read_in *inarg = &req->misc.read.in;
struct fuse_file *ff = file->private_data;
inarg->fh = ff->fh;
inarg->offset = pos;
inarg->size = count;
inarg->flags = file->f_flags;
req->in.h.opcode = opcode;
req->in.h.nodeid = ff->nodeid;
req->in.numargs = 1;
req->in.args[0].size = sizeof(struct fuse_read_in);
req->in.args[0].value = inarg;
req->out.argvar = 1;
req->out.numargs = 1;
req->out.args[0].size = count;
}
static size_t fuse_send_read(struct fuse_req *req, struct file *file,
loff_t pos, size_t count, fl_owner_t owner)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
fuse_read_fill(req, file, pos, count, FUSE_READ);
if (owner != NULL) {
struct fuse_read_in *inarg = &req->misc.read.in;
inarg->read_flags |= FUSE_READ_LOCKOWNER;
inarg->lock_owner = fuse_lock_owner_id(fc, owner);
}
fuse_request_send(fc, req);
return req->out.args[0].size;
}
static void fuse_read_update_size(struct inode *inode, loff_t size,
u64 attr_ver)
{
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
spin_lock(&fc->lock);
if (attr_ver == fi->attr_version && size < inode->i_size) {
fi->attr_version = ++fc->attr_version;
i_size_write(inode, size);
}
spin_unlock(&fc->lock);
}
static int fuse_readpage(struct file *file, struct page *page)
{
struct inode *inode = page->mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_req *req;
size_t num_read;
loff_t pos = page_offset(page);
size_t count = PAGE_CACHE_SIZE;
u64 attr_ver;
int err;
err = -EIO;
if (is_bad_inode(inode))
goto out;
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* Page writeback can extend beyond the lifetime of the
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
* page-cache page, so make sure we read a properly synced
* page.
*/
fuse_wait_on_page_writeback(inode, page->index);
req = fuse_get_req(fc);
err = PTR_ERR(req);
if (IS_ERR(req))
goto out;
attr_ver = fuse_get_attr_version(fc);
req->out.page_zeroing = 1;
req->out.argpages = 1;
req->num_pages = 1;
req->pages[0] = page;
num_read = fuse_send_read(req, file, pos, count, NULL);
err = req->out.h.error;
fuse_put_request(fc, req);
if (!err) {
/*
* Short read means EOF. If file size is larger, truncate it
*/
if (num_read < count)
fuse_read_update_size(inode, pos + num_read, attr_ver);
SetPageUptodate(page);
}
fuse_invalidate_attr(inode); /* atime changed */
out:
unlock_page(page);
return err;
}
static void fuse_readpages_end(struct fuse_conn *fc, struct fuse_req *req)
{
int i;
size_t count = req->misc.read.in.size;
size_t num_read = req->out.args[0].size;
struct address_space *mapping = NULL;
for (i = 0; mapping == NULL && i < req->num_pages; i++)
mapping = req->pages[i]->mapping;
if (mapping) {
struct inode *inode = mapping->host;
/*
* Short read means EOF. If file size is larger, truncate it
*/
if (!req->out.h.error && num_read < count) {
loff_t pos;
pos = page_offset(req->pages[0]) + num_read;
fuse_read_update_size(inode, pos,
req->misc.read.attr_ver);
}
fuse_invalidate_attr(inode); /* atime changed */
}
for (i = 0; i < req->num_pages; i++) {
struct page *page = req->pages[i];
if (!req->out.h.error)
SetPageUptodate(page);
else
SetPageError(page);
unlock_page(page);
page_cache_release(page);
}
if (req->ff)
fuse_file_put(req->ff, false);
}
static void fuse_send_readpages(struct fuse_req *req, struct file *file)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
loff_t pos = page_offset(req->pages[0]);
size_t count = req->num_pages << PAGE_CACHE_SHIFT;
req->out.argpages = 1;
req->out.page_zeroing = 1;
req->out.page_replace = 1;
fuse_read_fill(req, file, pos, count, FUSE_READ);
req->misc.read.attr_ver = fuse_get_attr_version(fc);
if (fc->async_read) {
req->ff = fuse_file_get(ff);
req->end = fuse_readpages_end;
fuse_request_send_background(fc, req);
} else {
fuse_request_send(fc, req);
fuse_readpages_end(fc, req);
fuse_put_request(fc, req);
}
}
struct fuse_fill_data {
struct fuse_req *req;
struct file *file;
struct inode *inode;
};
static int fuse_readpages_fill(void *_data, struct page *page)
{
struct fuse_fill_data *data = _data;
struct fuse_req *req = data->req;
struct inode *inode = data->inode;
struct fuse_conn *fc = get_fuse_conn(inode);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
fuse_wait_on_page_writeback(inode, page->index);
if (req->num_pages &&
(req->num_pages == FUSE_MAX_PAGES_PER_REQ ||
(req->num_pages + 1) * PAGE_CACHE_SIZE > fc->max_read ||
req->pages[req->num_pages - 1]->index + 1 != page->index)) {
fuse_send_readpages(req, data->file);
data->req = req = fuse_get_req(fc);
if (IS_ERR(req)) {
unlock_page(page);
return PTR_ERR(req);
}
}
page_cache_get(page);
req->pages[req->num_pages] = page;
req->num_pages++;
return 0;
}
static int fuse_readpages(struct file *file, struct address_space *mapping,
struct list_head *pages, unsigned nr_pages)
{
struct inode *inode = mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_fill_data data;
int err;
err = -EIO;
if (is_bad_inode(inode))
goto out;
data.file = file;
data.inode = inode;
data.req = fuse_get_req(fc);
err = PTR_ERR(data.req);
if (IS_ERR(data.req))
goto out;
err = read_cache_pages(mapping, pages, fuse_readpages_fill, &data);
if (!err) {
if (data.req->num_pages)
fuse_send_readpages(data.req, file);
else
fuse_put_request(fc, data.req);
}
out:
return err;
}
static ssize_t fuse_file_aio_read(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct inode *inode = iocb->ki_filp->f_mapping->host;
if (pos + iov_length(iov, nr_segs) > i_size_read(inode)) {
int err;
/*
* If trying to read past EOF, make sure the i_size
* attribute is up-to-date.
*/
err = fuse_update_attributes(inode, NULL, iocb->ki_filp, NULL);
if (err)
return err;
}
return generic_file_aio_read(iocb, iov, nr_segs, pos);
}
static void fuse_write_fill(struct fuse_req *req, struct fuse_file *ff,
loff_t pos, size_t count)
{
struct fuse_write_in *inarg = &req->misc.write.in;
struct fuse_write_out *outarg = &req->misc.write.out;
inarg->fh = ff->fh;
inarg->offset = pos;
inarg->size = count;
req->in.h.opcode = FUSE_WRITE;
req->in.h.nodeid = ff->nodeid;
req->in.numargs = 2;
if (ff->fc->minor < 9)
req->in.args[0].size = FUSE_COMPAT_WRITE_IN_SIZE;
else
req->in.args[0].size = sizeof(struct fuse_write_in);
req->in.args[0].value = inarg;
req->in.args[1].size = count;
req->out.numargs = 1;
req->out.args[0].size = sizeof(struct fuse_write_out);
req->out.args[0].value = outarg;
}
static size_t fuse_send_write(struct fuse_req *req, struct file *file,
loff_t pos, size_t count, fl_owner_t owner)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
struct fuse_write_in *inarg = &req->misc.write.in;
fuse_write_fill(req, ff, pos, count);
inarg->flags = file->f_flags;
if (owner != NULL) {
inarg->write_flags |= FUSE_WRITE_LOCKOWNER;
inarg->lock_owner = fuse_lock_owner_id(fc, owner);
}
fuse_request_send(fc, req);
return req->misc.write.out.size;
}
void fuse_write_update_size(struct inode *inode, loff_t pos)
{
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
spin_lock(&fc->lock);
fi->attr_version = ++fc->attr_version;
if (pos > inode->i_size)
i_size_write(inode, pos);
spin_unlock(&fc->lock);
}
static size_t fuse_send_write_pages(struct fuse_req *req, struct file *file,
struct inode *inode, loff_t pos,
size_t count)
{
size_t res;
unsigned offset;
unsigned i;
for (i = 0; i < req->num_pages; i++)
fuse_wait_on_page_writeback(inode, req->pages[i]->index);
res = fuse_send_write(req, file, pos, count, NULL);
offset = req->page_offset;
count = res;
for (i = 0; i < req->num_pages; i++) {
struct page *page = req->pages[i];
if (!req->out.h.error && !offset && count >= PAGE_CACHE_SIZE)
SetPageUptodate(page);
if (count > PAGE_CACHE_SIZE - offset)
count -= PAGE_CACHE_SIZE - offset;
else
count = 0;
offset = 0;
unlock_page(page);
page_cache_release(page);
}
return res;
}
static ssize_t fuse_fill_write_pages(struct fuse_req *req,
struct address_space *mapping,
struct iov_iter *ii, loff_t pos)
{
struct fuse_conn *fc = get_fuse_conn(mapping->host);
unsigned offset = pos & (PAGE_CACHE_SIZE - 1);
size_t count = 0;
int err;
req->in.argpages = 1;
req->page_offset = offset;
do {
size_t tmp;
struct page *page;
pgoff_t index = pos >> PAGE_CACHE_SHIFT;
size_t bytes = min_t(size_t, PAGE_CACHE_SIZE - offset,
iov_iter_count(ii));
bytes = min_t(size_t, bytes, fc->max_write - count);
again:
err = -EFAULT;
if (iov_iter_fault_in_readable(ii, bytes))
break;
err = -ENOMEM;
fs: symlink write_begin allocation context fix With the write_begin/write_end aops, page_symlink was broken because it could no longer pass a GFP_NOFS type mask into the point where the allocations happened. They are done in write_begin, which would always assume that the filesystem can be entered from reclaim. This bug could cause filesystem deadlocks. The funny thing with having a gfp_t mask there is that it doesn't really allow the caller to arbitrarily tinker with the context in which it can be called. It couldn't ever be GFP_ATOMIC, for example, because it needs to take the page lock. The only thing any callers care about is __GFP_FS anyway, so turn that into a single flag. Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on this flag in their write_begin function. Change __grab_cache_page to accept a nofs argument as well, to honour that flag (while we're there, change the name to grab_cache_page_write_begin which is more instructive and does away with random leading underscores). This is really a more flexible way to go in the end anyway -- if a filesystem happens to want any extra allocations aside from the pagecache ones in ints write_begin function, it may now use GFP_KERNEL (rather than GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a random example). [kosaki.motohiro@jp.fujitsu.com: fix ubifs] [kosaki.motohiro@jp.fujitsu.com: fix fuse] Signed-off-by: Nick Piggin <npiggin@suse.de> Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: <stable@kernel.org> [2.6.28.x] Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> [ Cleaned up the calling convention: just pass in the AOP flags untouched to the grab_cache_page_write_begin() function. That just simplifies everybody, and may even allow future expansion of the logic. - Linus ] Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-05 04:00:53 +08:00
page = grab_cache_page_write_begin(mapping, index, 0);
if (!page)
break;
if (mapping_writably_mapped(mapping))
flush_dcache_page(page);
pagefault_disable();
tmp = iov_iter_copy_from_user_atomic(page, ii, offset, bytes);
pagefault_enable();
flush_dcache_page(page);
mark_page_accessed(page);
if (!tmp) {
unlock_page(page);
page_cache_release(page);
bytes = min(bytes, iov_iter_single_seg_count(ii));
goto again;
}
err = 0;
req->pages[req->num_pages] = page;
req->num_pages++;
iov_iter_advance(ii, tmp);
count += tmp;
pos += tmp;
offset += tmp;
if (offset == PAGE_CACHE_SIZE)
offset = 0;
if (!fc->big_writes)
break;
} while (iov_iter_count(ii) && count < fc->max_write &&
req->num_pages < FUSE_MAX_PAGES_PER_REQ && offset == 0);
return count > 0 ? count : err;
}
static ssize_t fuse_perform_write(struct file *file,
struct address_space *mapping,
struct iov_iter *ii, loff_t pos)
{
struct inode *inode = mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
int err = 0;
ssize_t res = 0;
if (is_bad_inode(inode))
return -EIO;
do {
struct fuse_req *req;
ssize_t count;
req = fuse_get_req(fc);
if (IS_ERR(req)) {
err = PTR_ERR(req);
break;
}
count = fuse_fill_write_pages(req, mapping, ii, pos);
if (count <= 0) {
err = count;
} else {
size_t num_written;
num_written = fuse_send_write_pages(req, file, inode,
pos, count);
err = req->out.h.error;
if (!err) {
res += num_written;
pos += num_written;
/* break out of the loop on short write */
if (num_written != count)
err = -EIO;
}
}
fuse_put_request(fc, req);
} while (!err && iov_iter_count(ii));
if (res > 0)
fuse_write_update_size(inode, pos);
fuse_invalidate_attr(inode);
return res > 0 ? res : err;
}
static ssize_t fuse_file_aio_write(struct kiocb *iocb, const struct iovec *iov,
unsigned long nr_segs, loff_t pos)
{
struct file *file = iocb->ki_filp;
struct address_space *mapping = file->f_mapping;
size_t count = 0;
ssize_t written = 0;
struct inode *inode = mapping->host;
ssize_t err;
struct iov_iter i;
WARN_ON(iocb->ki_pos != pos);
err = generic_segment_checks(iov, &nr_segs, &count, VERIFY_READ);
if (err)
return err;
mutex_lock(&inode->i_mutex);
vfs_check_frozen(inode->i_sb, SB_FREEZE_WRITE);
/* We can write back this queue in page reclaim */
current->backing_dev_info = mapping->backing_dev_info;
err = generic_write_checks(file, &pos, &count, S_ISBLK(inode->i_mode));
if (err)
goto out;
if (count == 0)
goto out;
err = file_remove_suid(file);
if (err)
goto out;
file_update_time(file);
iov_iter_init(&i, iov, nr_segs, count, 0);
written = fuse_perform_write(file, mapping, &i, pos);
if (written >= 0)
iocb->ki_pos = pos + written;
out:
current->backing_dev_info = NULL;
mutex_unlock(&inode->i_mutex);
return written ? written : err;
}
static void fuse_release_user_pages(struct fuse_req *req, int write)
{
unsigned i;
for (i = 0; i < req->num_pages; i++) {
struct page *page = req->pages[i];
if (write)
set_page_dirty_lock(page);
put_page(page);
}
}
static int fuse_get_user_pages(struct fuse_req *req, const char __user *buf,
size_t *nbytesp, int write)
{
size_t nbytes = *nbytesp;
unsigned long user_addr = (unsigned long) buf;
unsigned offset = user_addr & ~PAGE_MASK;
int npages;
/* Special case for kernel I/O: can copy directly into the buffer */
if (segment_eq(get_fs(), KERNEL_DS)) {
if (write)
req->in.args[1].value = (void *) user_addr;
else
req->out.args[0].value = (void *) user_addr;
return 0;
}
nbytes = min_t(size_t, nbytes, FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT);
npages = (nbytes + offset + PAGE_SIZE - 1) >> PAGE_SHIFT;
npages = clamp(npages, 1, FUSE_MAX_PAGES_PER_REQ);
npages = get_user_pages_fast(user_addr, npages, !write, req->pages);
if (npages < 0)
return npages;
req->num_pages = npages;
req->page_offset = offset;
if (write)
req->in.argpages = 1;
else
req->out.argpages = 1;
nbytes = (req->num_pages << PAGE_SHIFT) - req->page_offset;
*nbytesp = min(*nbytesp, nbytes);
return 0;
}
ssize_t fuse_direct_io(struct file *file, const char __user *buf,
size_t count, loff_t *ppos, int write)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
size_t nmax = write ? fc->max_write : fc->max_read;
loff_t pos = *ppos;
ssize_t res = 0;
struct fuse_req *req;
req = fuse_get_req(fc);
if (IS_ERR(req))
return PTR_ERR(req);
while (count) {
size_t nres;
fl_owner_t owner = current->files;
size_t nbytes = min(count, nmax);
int err = fuse_get_user_pages(req, buf, &nbytes, write);
if (err) {
res = err;
break;
}
if (write)
nres = fuse_send_write(req, file, pos, nbytes, owner);
else
nres = fuse_send_read(req, file, pos, nbytes, owner);
fuse_release_user_pages(req, !write);
if (req->out.h.error) {
if (!res)
res = req->out.h.error;
break;
} else if (nres > nbytes) {
res = -EIO;
break;
}
count -= nres;
res += nres;
pos += nres;
buf += nres;
if (nres != nbytes)
break;
if (count) {
fuse_put_request(fc, req);
req = fuse_get_req(fc);
if (IS_ERR(req))
break;
}
}
if (!IS_ERR(req))
fuse_put_request(fc, req);
if (res > 0)
*ppos = pos;
return res;
}
EXPORT_SYMBOL_GPL(fuse_direct_io);
static ssize_t fuse_direct_read(struct file *file, char __user *buf,
size_t count, loff_t *ppos)
{
ssize_t res;
struct inode *inode = file->f_path.dentry->d_inode;
if (is_bad_inode(inode))
return -EIO;
res = fuse_direct_io(file, buf, count, ppos, 0);
fuse_invalidate_attr(inode);
return res;
}
static ssize_t fuse_direct_write(struct file *file, const char __user *buf,
size_t count, loff_t *ppos)
{
struct inode *inode = file->f_path.dentry->d_inode;
ssize_t res;
if (is_bad_inode(inode))
return -EIO;
/* Don't allow parallel writes to the same file */
mutex_lock(&inode->i_mutex);
res = generic_write_checks(file, ppos, &count, 0);
if (!res) {
res = fuse_direct_io(file, buf, count, ppos, 1);
if (res > 0)
fuse_write_update_size(inode, *ppos);
}
mutex_unlock(&inode->i_mutex);
fuse_invalidate_attr(inode);
return res;
}
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
static void fuse_writepage_free(struct fuse_conn *fc, struct fuse_req *req)
{
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
__free_page(req->pages[0]);
fuse_file_put(req->ff, false);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
}
static void fuse_writepage_finish(struct fuse_conn *fc, struct fuse_req *req)
{
struct inode *inode = req->inode;
struct fuse_inode *fi = get_fuse_inode(inode);
struct backing_dev_info *bdi = inode->i_mapping->backing_dev_info;
list_del(&req->writepages_entry);
dec_bdi_stat(bdi, BDI_WRITEBACK);
dec_zone_page_state(req->pages[0], NR_WRITEBACK_TEMP);
bdi_writeout_inc(bdi);
wake_up(&fi->page_waitq);
}
/* Called under fc->lock, may release and reacquire it */
static void fuse_send_writepage(struct fuse_conn *fc, struct fuse_req *req)
__releases(fc->lock)
__acquires(fc->lock)
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
{
struct fuse_inode *fi = get_fuse_inode(req->inode);
loff_t size = i_size_read(req->inode);
struct fuse_write_in *inarg = &req->misc.write.in;
if (!fc->connected)
goto out_free;
if (inarg->offset + PAGE_CACHE_SIZE <= size) {
inarg->size = PAGE_CACHE_SIZE;
} else if (inarg->offset < size) {
inarg->size = size & (PAGE_CACHE_SIZE - 1);
} else {
/* Got truncated off completely */
goto out_free;
}
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
req->in.args[1].size = inarg->size;
fi->writectr++;
fuse_request_send_background_locked(fc, req);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
return;
out_free:
fuse_writepage_finish(fc, req);
spin_unlock(&fc->lock);
fuse_writepage_free(fc, req);
fuse_put_request(fc, req);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
spin_lock(&fc->lock);
}
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* If fi->writectr is positive (no truncate or fsync going on) send
* all queued writepage requests.
*
* Called with fc->lock
*/
void fuse_flush_writepages(struct inode *inode)
__releases(fc->lock)
__acquires(fc->lock)
{
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
struct fuse_req *req;
while (fi->writectr >= 0 && !list_empty(&fi->queued_writes)) {
req = list_entry(fi->queued_writes.next, struct fuse_req, list);
list_del_init(&req->list);
fuse_send_writepage(fc, req);
}
}
static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_req *req)
{
struct inode *inode = req->inode;
struct fuse_inode *fi = get_fuse_inode(inode);
mapping_set_error(inode->i_mapping, req->out.h.error);
spin_lock(&fc->lock);
fi->writectr--;
fuse_writepage_finish(fc, req);
spin_unlock(&fc->lock);
fuse_writepage_free(fc, req);
}
static int fuse_writepage_locked(struct page *page)
{
struct address_space *mapping = page->mapping;
struct inode *inode = mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
struct fuse_req *req;
struct fuse_file *ff;
struct page *tmp_page;
set_page_writeback(page);
req = fuse_request_alloc_nofs();
if (!req)
goto err;
tmp_page = alloc_page(GFP_NOFS | __GFP_HIGHMEM);
if (!tmp_page)
goto err_free;
spin_lock(&fc->lock);
BUG_ON(list_empty(&fi->write_files));
ff = list_entry(fi->write_files.next, struct fuse_file, write_entry);
req->ff = fuse_file_get(ff);
spin_unlock(&fc->lock);
fuse_write_fill(req, ff, page_offset(page), 0);
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
copy_highpage(tmp_page, page);
req->misc.write.in.write_flags |= FUSE_WRITE_CACHE;
req->in.argpages = 1;
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
req->num_pages = 1;
req->pages[0] = tmp_page;
req->page_offset = 0;
req->end = fuse_writepage_end;
req->inode = inode;
inc_bdi_stat(mapping->backing_dev_info, BDI_WRITEBACK);
inc_zone_page_state(tmp_page, NR_WRITEBACK_TEMP);
end_page_writeback(page);
spin_lock(&fc->lock);
list_add(&req->writepages_entry, &fi->writepages);
list_add_tail(&req->list, &fi->queued_writes);
fuse_flush_writepages(inode);
spin_unlock(&fc->lock);
return 0;
err_free:
fuse_request_free(req);
err:
end_page_writeback(page);
return -ENOMEM;
}
static int fuse_writepage(struct page *page, struct writeback_control *wbc)
{
int err;
err = fuse_writepage_locked(page);
unlock_page(page);
return err;
}
static int fuse_launder_page(struct page *page)
{
int err = 0;
if (clear_page_dirty_for_io(page)) {
struct inode *inode = page->mapping->host;
err = fuse_writepage_locked(page);
if (!err)
fuse_wait_on_page_writeback(inode, page->index);
}
return err;
}
/*
* Write back dirty pages now, because there may not be any suitable
* open files later
*/
static void fuse_vma_close(struct vm_area_struct *vma)
{
filemap_write_and_wait(vma->vm_file->f_mapping);
}
/*
* Wait for writeback against this page to complete before allowing it
* to be marked dirty again, and hence written back again, possibly
* before the previous writepage completed.
*
* Block here, instead of in ->writepage(), so that the userspace fs
* can only block processes actually operating on the filesystem.
*
* Otherwise unprivileged userspace fs would be able to block
* unrelated:
*
* - page migration
* - sync(2)
* - try_to_free_pages() with order > PAGE_ALLOC_COSTLY_ORDER
*/
static int fuse_page_mkwrite(struct vm_area_struct *vma, struct vm_fault *vmf)
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
{
struct page *page = vmf->page;
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
/*
* Don't use page->mapping as it may become NULL from a
* concurrent truncate.
*/
struct inode *inode = vma->vm_file->f_mapping->host;
fuse_wait_on_page_writeback(inode, page->index);
return 0;
}
static const struct vm_operations_struct fuse_file_vm_ops = {
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
.close = fuse_vma_close,
.fault = filemap_fault,
.page_mkwrite = fuse_page_mkwrite,
};
static int fuse_file_mmap(struct file *file, struct vm_area_struct *vma)
{
if ((vma->vm_flags & VM_SHARED) && (vma->vm_flags & VM_MAYWRITE)) {
struct inode *inode = file->f_dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_inode *fi = get_fuse_inode(inode);
struct fuse_file *ff = file->private_data;
/*
* file may be written through mmap, so chain it onto the
* inodes's write_file list
*/
spin_lock(&fc->lock);
if (list_empty(&ff->write_entry))
list_add(&ff->write_entry, &fi->write_files);
spin_unlock(&fc->lock);
}
file_accessed(file);
vma->vm_ops = &fuse_file_vm_ops;
return 0;
}
static int fuse_direct_mmap(struct file *file, struct vm_area_struct *vma)
{
/* Can't provide the coherency needed for MAP_SHARED */
if (vma->vm_flags & VM_MAYSHARE)
return -ENODEV;
invalidate_inode_pages2(file->f_mapping);
return generic_file_mmap(file, vma);
}
static int convert_fuse_file_lock(const struct fuse_file_lock *ffl,
struct file_lock *fl)
{
switch (ffl->type) {
case F_UNLCK:
break;
case F_RDLCK:
case F_WRLCK:
if (ffl->start > OFFSET_MAX || ffl->end > OFFSET_MAX ||
ffl->end < ffl->start)
return -EIO;
fl->fl_start = ffl->start;
fl->fl_end = ffl->end;
fl->fl_pid = ffl->pid;
break;
default:
return -EIO;
}
fl->fl_type = ffl->type;
return 0;
}
static void fuse_lk_fill(struct fuse_req *req, struct file *file,
const struct file_lock *fl, int opcode, pid_t pid,
int flock)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_file *ff = file->private_data;
struct fuse_lk_in *arg = &req->misc.lk_in;
arg->fh = ff->fh;
arg->owner = fuse_lock_owner_id(fc, fl->fl_owner);
arg->lk.start = fl->fl_start;
arg->lk.end = fl->fl_end;
arg->lk.type = fl->fl_type;
arg->lk.pid = pid;
if (flock)
arg->lk_flags |= FUSE_LK_FLOCK;
req->in.h.opcode = opcode;
req->in.h.nodeid = get_node_id(inode);
req->in.numargs = 1;
req->in.args[0].size = sizeof(*arg);
req->in.args[0].value = arg;
}
static int fuse_getlk(struct file *file, struct file_lock *fl)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_req *req;
struct fuse_lk_out outarg;
int err;
req = fuse_get_req(fc);
if (IS_ERR(req))
return PTR_ERR(req);
fuse_lk_fill(req, file, fl, FUSE_GETLK, 0, 0);
req->out.numargs = 1;
req->out.args[0].size = sizeof(outarg);
req->out.args[0].value = &outarg;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
if (!err)
err = convert_fuse_file_lock(&outarg.lk, fl);
return err;
}
static int fuse_setlk(struct file *file, struct file_lock *fl, int flock)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_req *req;
int opcode = (fl->fl_flags & FL_SLEEP) ? FUSE_SETLKW : FUSE_SETLK;
pid_t pid = fl->fl_type != F_UNLCK ? current->tgid : 0;
int err;
if (fl->fl_lmops && fl->fl_lmops->lm_grant) {
/* NLM needs asynchronous locks, which we don't support yet */
return -ENOLCK;
}
/* Unlock on close is handled by the flush method */
if (fl->fl_flags & FL_CLOSE)
return 0;
req = fuse_get_req(fc);
if (IS_ERR(req))
return PTR_ERR(req);
fuse_lk_fill(req, file, fl, opcode, pid, flock);
fuse_request_send(fc, req);
err = req->out.h.error;
/* locking is restartable */
if (err == -EINTR)
err = -ERESTARTSYS;
fuse_put_request(fc, req);
return err;
}
static int fuse_file_lock(struct file *file, int cmd, struct file_lock *fl)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
int err;
if (cmd == F_CANCELLK) {
err = 0;
} else if (cmd == F_GETLK) {
if (fc->no_lock) {
posix_test_lock(file, fl);
err = 0;
} else
err = fuse_getlk(file, fl);
} else {
if (fc->no_lock)
err = posix_lock_file(file, fl, NULL);
else
err = fuse_setlk(file, fl, 0);
}
return err;
}
static int fuse_file_flock(struct file *file, int cmd, struct file_lock *fl)
{
struct inode *inode = file->f_path.dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
int err;
if (fc->no_flock) {
err = flock_lock_file_wait(file, fl);
} else {
struct fuse_file *ff = file->private_data;
/* emulate flock with POSIX locks */
fl->fl_owner = (fl_owner_t) file;
ff->flock = true;
err = fuse_setlk(file, fl, 1);
}
return err;
}
static sector_t fuse_bmap(struct address_space *mapping, sector_t block)
{
struct inode *inode = mapping->host;
struct fuse_conn *fc = get_fuse_conn(inode);
struct fuse_req *req;
struct fuse_bmap_in inarg;
struct fuse_bmap_out outarg;
int err;
if (!inode->i_sb->s_bdev || fc->no_bmap)
return 0;
req = fuse_get_req(fc);
if (IS_ERR(req))
return 0;
memset(&inarg, 0, sizeof(inarg));
inarg.block = block;
inarg.blocksize = inode->i_sb->s_blocksize;
req->in.h.opcode = FUSE_BMAP;
req->in.h.nodeid = get_node_id(inode);
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
req->out.numargs = 1;
req->out.args[0].size = sizeof(outarg);
req->out.args[0].value = &outarg;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
if (err == -ENOSYS)
fc->no_bmap = 1;
return err ? 0 : outarg.block;
}
static loff_t fuse_file_llseek(struct file *file, loff_t offset, int origin)
{
loff_t retval;
struct inode *inode = file->f_path.dentry->d_inode;
/* No i_mutex protection necessary for SEEK_CUR and SEEK_SET */
if (origin == SEEK_CUR || origin == SEEK_SET)
return generic_file_llseek(file, offset, origin);
mutex_lock(&inode->i_mutex);
retval = fuse_update_attributes(inode, NULL, file, NULL);
if (!retval)
retval = generic_file_llseek(file, offset, origin);
mutex_unlock(&inode->i_mutex);
return retval;
}
static int fuse_ioctl_copy_user(struct page **pages, struct iovec *iov,
unsigned int nr_segs, size_t bytes, bool to_user)
{
struct iov_iter ii;
int page_idx = 0;
if (!bytes)
return 0;
iov_iter_init(&ii, iov, nr_segs, bytes, 0);
while (iov_iter_count(&ii)) {
struct page *page = pages[page_idx++];
size_t todo = min_t(size_t, PAGE_SIZE, iov_iter_count(&ii));
void *kaddr;
kaddr = kmap(page);
while (todo) {
char __user *uaddr = ii.iov->iov_base + ii.iov_offset;
size_t iov_len = ii.iov->iov_len - ii.iov_offset;
size_t copy = min(todo, iov_len);
size_t left;
if (!to_user)
left = copy_from_user(kaddr, uaddr, copy);
else
left = copy_to_user(uaddr, kaddr, copy);
if (unlikely(left))
return -EFAULT;
iov_iter_advance(&ii, copy);
todo -= copy;
kaddr += copy;
}
kunmap(page);
}
return 0;
}
/*
* CUSE servers compiled on 32bit broke on 64bit kernels because the
* ABI was defined to be 'struct iovec' which is different on 32bit
* and 64bit. Fortunately we can determine which structure the server
* used from the size of the reply.
*/
static int fuse_copy_ioctl_iovec_old(struct iovec *dst, void *src,
size_t transferred, unsigned count,
bool is_compat)
{
#ifdef CONFIG_COMPAT
if (count * sizeof(struct compat_iovec) == transferred) {
struct compat_iovec *ciov = src;
unsigned i;
/*
* With this interface a 32bit server cannot support
* non-compat (i.e. ones coming from 64bit apps) ioctl
* requests
*/
if (!is_compat)
return -EINVAL;
for (i = 0; i < count; i++) {
dst[i].iov_base = compat_ptr(ciov[i].iov_base);
dst[i].iov_len = ciov[i].iov_len;
}
return 0;
}
#endif
if (count * sizeof(struct iovec) != transferred)
return -EIO;
memcpy(dst, src, transferred);
return 0;
}
/* Make sure iov_length() won't overflow */
static int fuse_verify_ioctl_iov(struct iovec *iov, size_t count)
{
size_t n;
u32 max = FUSE_MAX_PAGES_PER_REQ << PAGE_SHIFT;
for (n = 0; n < count; n++) {
if (iov->iov_len > (size_t) max)
return -ENOMEM;
max -= iov->iov_len;
}
return 0;
}
static int fuse_copy_ioctl_iovec(struct fuse_conn *fc, struct iovec *dst,
void *src, size_t transferred, unsigned count,
bool is_compat)
{
unsigned i;
struct fuse_ioctl_iovec *fiov = src;
if (fc->minor < 16) {
return fuse_copy_ioctl_iovec_old(dst, src, transferred,
count, is_compat);
}
if (count * sizeof(struct fuse_ioctl_iovec) != transferred)
return -EIO;
for (i = 0; i < count; i++) {
/* Did the server supply an inappropriate value? */
if (fiov[i].base != (unsigned long) fiov[i].base ||
fiov[i].len != (unsigned long) fiov[i].len)
return -EIO;
dst[i].iov_base = (void __user *) (unsigned long) fiov[i].base;
dst[i].iov_len = (size_t) fiov[i].len;
#ifdef CONFIG_COMPAT
if (is_compat &&
(ptr_to_compat(dst[i].iov_base) != fiov[i].base ||
(compat_size_t) dst[i].iov_len != fiov[i].len))
return -EIO;
#endif
}
return 0;
}
/*
* For ioctls, there is no generic way to determine how much memory
* needs to be read and/or written. Furthermore, ioctls are allowed
* to dereference the passed pointer, so the parameter requires deep
* copying but FUSE has no idea whatsoever about what to copy in or
* out.
*
* This is solved by allowing FUSE server to retry ioctl with
* necessary in/out iovecs. Let's assume the ioctl implementation
* needs to read in the following structure.
*
* struct a {
* char *buf;
* size_t buflen;
* }
*
* On the first callout to FUSE server, inarg->in_size and
* inarg->out_size will be NULL; then, the server completes the ioctl
* with FUSE_IOCTL_RETRY set in out->flags, out->in_iovs set to 1 and
* the actual iov array to
*
* { { .iov_base = inarg.arg, .iov_len = sizeof(struct a) } }
*
* which tells FUSE to copy in the requested area and retry the ioctl.
* On the second round, the server has access to the structure and
* from that it can tell what to look for next, so on the invocation,
* it sets FUSE_IOCTL_RETRY, out->in_iovs to 2 and iov array to
*
* { { .iov_base = inarg.arg, .iov_len = sizeof(struct a) },
* { .iov_base = a.buf, .iov_len = a.buflen } }
*
* FUSE will copy both struct a and the pointed buffer from the
* process doing the ioctl and retry ioctl with both struct a and the
* buffer.
*
* This time, FUSE server has everything it needs and completes ioctl
* without FUSE_IOCTL_RETRY which finishes the ioctl call.
*
* Copying data out works the same way.
*
* Note that if FUSE_IOCTL_UNRESTRICTED is clear, the kernel
* automatically initializes in and out iovs by decoding @cmd with
* _IOC_* macros and the server is not allowed to request RETRY. This
* limits ioctl data transfers to well-formed ioctls and is the forced
* behavior for all FUSE servers.
*/
long fuse_do_ioctl(struct file *file, unsigned int cmd, unsigned long arg,
unsigned int flags)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
struct fuse_ioctl_in inarg = {
.fh = ff->fh,
.cmd = cmd,
.arg = arg,
.flags = flags
};
struct fuse_ioctl_out outarg;
struct fuse_req *req = NULL;
struct page **pages = NULL;
struct iovec *iov_page = NULL;
struct iovec *in_iov = NULL, *out_iov = NULL;
unsigned int in_iovs = 0, out_iovs = 0, num_pages = 0, max_pages;
size_t in_size, out_size, transferred;
int err;
#if BITS_PER_LONG == 32
inarg.flags |= FUSE_IOCTL_32BIT;
#else
if (flags & FUSE_IOCTL_COMPAT)
inarg.flags |= FUSE_IOCTL_32BIT;
#endif
/* assume all the iovs returned by client always fits in a page */
BUILD_BUG_ON(sizeof(struct fuse_ioctl_iovec) * FUSE_IOCTL_MAX_IOV > PAGE_SIZE);
err = -ENOMEM;
pages = kcalloc(FUSE_MAX_PAGES_PER_REQ, sizeof(pages[0]), GFP_KERNEL);
iov_page = (struct iovec *) __get_free_page(GFP_KERNEL);
if (!pages || !iov_page)
goto out;
/*
* If restricted, initialize IO parameters as encoded in @cmd.
* RETRY from server is not allowed.
*/
if (!(flags & FUSE_IOCTL_UNRESTRICTED)) {
struct iovec *iov = iov_page;
iov->iov_base = (void __user *)arg;
iov->iov_len = _IOC_SIZE(cmd);
if (_IOC_DIR(cmd) & _IOC_WRITE) {
in_iov = iov;
in_iovs = 1;
}
if (_IOC_DIR(cmd) & _IOC_READ) {
out_iov = iov;
out_iovs = 1;
}
}
retry:
inarg.in_size = in_size = iov_length(in_iov, in_iovs);
inarg.out_size = out_size = iov_length(out_iov, out_iovs);
/*
* Out data can be used either for actual out data or iovs,
* make sure there always is at least one page.
*/
out_size = max_t(size_t, out_size, PAGE_SIZE);
max_pages = DIV_ROUND_UP(max(in_size, out_size), PAGE_SIZE);
/* make sure there are enough buffer pages and init request with them */
err = -ENOMEM;
if (max_pages > FUSE_MAX_PAGES_PER_REQ)
goto out;
while (num_pages < max_pages) {
pages[num_pages] = alloc_page(GFP_KERNEL | __GFP_HIGHMEM);
if (!pages[num_pages])
goto out;
num_pages++;
}
req = fuse_get_req(fc);
if (IS_ERR(req)) {
err = PTR_ERR(req);
req = NULL;
goto out;
}
memcpy(req->pages, pages, sizeof(req->pages[0]) * num_pages);
req->num_pages = num_pages;
/* okay, let's send it to the client */
req->in.h.opcode = FUSE_IOCTL;
req->in.h.nodeid = ff->nodeid;
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
if (in_size) {
req->in.numargs++;
req->in.args[1].size = in_size;
req->in.argpages = 1;
err = fuse_ioctl_copy_user(pages, in_iov, in_iovs, in_size,
false);
if (err)
goto out;
}
req->out.numargs = 2;
req->out.args[0].size = sizeof(outarg);
req->out.args[0].value = &outarg;
req->out.args[1].size = out_size;
req->out.argpages = 1;
req->out.argvar = 1;
fuse_request_send(fc, req);
err = req->out.h.error;
transferred = req->out.args[1].size;
fuse_put_request(fc, req);
req = NULL;
if (err)
goto out;
/* did it ask for retry? */
if (outarg.flags & FUSE_IOCTL_RETRY) {
void *vaddr;
/* no retry if in restricted mode */
err = -EIO;
if (!(flags & FUSE_IOCTL_UNRESTRICTED))
goto out;
in_iovs = outarg.in_iovs;
out_iovs = outarg.out_iovs;
/*
* Make sure things are in boundary, separate checks
* are to protect against overflow.
*/
err = -ENOMEM;
if (in_iovs > FUSE_IOCTL_MAX_IOV ||
out_iovs > FUSE_IOCTL_MAX_IOV ||
in_iovs + out_iovs > FUSE_IOCTL_MAX_IOV)
goto out;
vaddr = kmap_atomic(pages[0], KM_USER0);
err = fuse_copy_ioctl_iovec(fc, iov_page, vaddr,
transferred, in_iovs + out_iovs,
(flags & FUSE_IOCTL_COMPAT) != 0);
kunmap_atomic(vaddr, KM_USER0);
if (err)
goto out;
in_iov = iov_page;
out_iov = in_iov + in_iovs;
err = fuse_verify_ioctl_iov(in_iov, in_iovs);
if (err)
goto out;
err = fuse_verify_ioctl_iov(out_iov, out_iovs);
if (err)
goto out;
goto retry;
}
err = -EIO;
if (transferred > inarg.out_size)
goto out;
err = fuse_ioctl_copy_user(pages, out_iov, out_iovs, transferred, true);
out:
if (req)
fuse_put_request(fc, req);
free_page((unsigned long) iov_page);
while (num_pages)
__free_page(pages[--num_pages]);
kfree(pages);
return err ? err : outarg.result;
}
EXPORT_SYMBOL_GPL(fuse_do_ioctl);
long fuse_ioctl_common(struct file *file, unsigned int cmd,
unsigned long arg, unsigned int flags)
{
struct inode *inode = file->f_dentry->d_inode;
struct fuse_conn *fc = get_fuse_conn(inode);
if (!fuse_allow_task(fc, current))
return -EACCES;
if (is_bad_inode(inode))
return -EIO;
return fuse_do_ioctl(file, cmd, arg, flags);
}
static long fuse_file_ioctl(struct file *file, unsigned int cmd,
unsigned long arg)
{
return fuse_ioctl_common(file, cmd, arg, 0);
}
static long fuse_file_compat_ioctl(struct file *file, unsigned int cmd,
unsigned long arg)
{
return fuse_ioctl_common(file, cmd, arg, FUSE_IOCTL_COMPAT);
}
/*
* All files which have been polled are linked to RB tree
* fuse_conn->polled_files which is indexed by kh. Walk the tree and
* find the matching one.
*/
static struct rb_node **fuse_find_polled_node(struct fuse_conn *fc, u64 kh,
struct rb_node **parent_out)
{
struct rb_node **link = &fc->polled_files.rb_node;
struct rb_node *last = NULL;
while (*link) {
struct fuse_file *ff;
last = *link;
ff = rb_entry(last, struct fuse_file, polled_node);
if (kh < ff->kh)
link = &last->rb_left;
else if (kh > ff->kh)
link = &last->rb_right;
else
return link;
}
if (parent_out)
*parent_out = last;
return link;
}
/*
* The file is about to be polled. Make sure it's on the polled_files
* RB tree. Note that files once added to the polled_files tree are
* not removed before the file is released. This is because a file
* polled once is likely to be polled again.
*/
static void fuse_register_polled_file(struct fuse_conn *fc,
struct fuse_file *ff)
{
spin_lock(&fc->lock);
if (RB_EMPTY_NODE(&ff->polled_node)) {
struct rb_node **link, *parent;
link = fuse_find_polled_node(fc, ff->kh, &parent);
BUG_ON(*link);
rb_link_node(&ff->polled_node, parent, link);
rb_insert_color(&ff->polled_node, &fc->polled_files);
}
spin_unlock(&fc->lock);
}
unsigned fuse_file_poll(struct file *file, poll_table *wait)
{
struct fuse_file *ff = file->private_data;
struct fuse_conn *fc = ff->fc;
struct fuse_poll_in inarg = { .fh = ff->fh, .kh = ff->kh };
struct fuse_poll_out outarg;
struct fuse_req *req;
int err;
if (fc->no_poll)
return DEFAULT_POLLMASK;
poll_wait(file, &ff->poll_wait, wait);
/*
* Ask for notification iff there's someone waiting for it.
* The client may ignore the flag and always notify.
*/
if (waitqueue_active(&ff->poll_wait)) {
inarg.flags |= FUSE_POLL_SCHEDULE_NOTIFY;
fuse_register_polled_file(fc, ff);
}
req = fuse_get_req(fc);
if (IS_ERR(req))
return POLLERR;
req->in.h.opcode = FUSE_POLL;
req->in.h.nodeid = ff->nodeid;
req->in.numargs = 1;
req->in.args[0].size = sizeof(inarg);
req->in.args[0].value = &inarg;
req->out.numargs = 1;
req->out.args[0].size = sizeof(outarg);
req->out.args[0].value = &outarg;
fuse_request_send(fc, req);
err = req->out.h.error;
fuse_put_request(fc, req);
if (!err)
return outarg.revents;
if (err == -ENOSYS) {
fc->no_poll = 1;
return DEFAULT_POLLMASK;
}
return POLLERR;
}
EXPORT_SYMBOL_GPL(fuse_file_poll);
/*
* This is called from fuse_handle_notify() on FUSE_NOTIFY_POLL and
* wakes up the poll waiters.
*/
int fuse_notify_poll_wakeup(struct fuse_conn *fc,
struct fuse_notify_poll_wakeup_out *outarg)
{
u64 kh = outarg->kh;
struct rb_node **link;
spin_lock(&fc->lock);
link = fuse_find_polled_node(fc, kh, NULL);
if (*link) {
struct fuse_file *ff;
ff = rb_entry(*link, struct fuse_file, polled_node);
wake_up_interruptible_sync(&ff->poll_wait);
}
spin_unlock(&fc->lock);
return 0;
}
static const struct file_operations fuse_file_operations = {
.llseek = fuse_file_llseek,
.read = do_sync_read,
.aio_read = fuse_file_aio_read,
.write = do_sync_write,
.aio_write = fuse_file_aio_write,
.mmap = fuse_file_mmap,
.open = fuse_open,
.flush = fuse_flush,
.release = fuse_release,
.fsync = fuse_fsync,
.lock = fuse_file_lock,
.flock = fuse_file_flock,
.splice_read = generic_file_splice_read,
.unlocked_ioctl = fuse_file_ioctl,
.compat_ioctl = fuse_file_compat_ioctl,
.poll = fuse_file_poll,
};
static const struct file_operations fuse_direct_io_file_operations = {
.llseek = fuse_file_llseek,
.read = fuse_direct_read,
.write = fuse_direct_write,
.mmap = fuse_direct_mmap,
.open = fuse_open,
.flush = fuse_flush,
.release = fuse_release,
.fsync = fuse_fsync,
.lock = fuse_file_lock,
.flock = fuse_file_flock,
.unlocked_ioctl = fuse_file_ioctl,
.compat_ioctl = fuse_file_compat_ioctl,
.poll = fuse_file_poll,
/* no splice_read */
};
static const struct address_space_operations fuse_file_aops = {
.readpage = fuse_readpage,
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
.writepage = fuse_writepage,
.launder_page = fuse_launder_page,
.readpages = fuse_readpages,
fuse: support writable mmap Quoting Linus (3 years ago, FUSE inclusion discussions): "User-space filesystems are hard to get right. I'd claim that they are almost impossible, unless you limit them somehow (shared writable mappings are the nastiest part - if you don't have those, you can reasonably limit your problems by limiting the number of dirty pages you accept through normal "write()" calls)." Instead of attempting the impossible, I've just waited for the dirty page accounting infrastructure to materialize (thanks to Peter Zijlstra and others). This nicely solved the biggest problem: limiting the number of pages used for write caching. Some small details remained, however, which this largish patch attempts to address. It provides a page writeback implementation for fuse, which is completely safe against VM related deadlocks. Performance may not be very good for certain usage patterns, but generally it should be acceptable. It has been tested extensively with fsx-linux and bash-shared-mapping. Fuse page writeback design -------------------------- fuse_writepage() allocates a new temporary page with GFP_NOFS|__GFP_HIGHMEM. It copies the contents of the original page, and queues a WRITE request to the userspace filesystem using this temp page. The writeback is finished instantly from the MM's point of view: the page is removed from the radix trees, and the PageDirty and PageWriteback flags are cleared. For the duration of the actual write, the NR_WRITEBACK_TEMP counter is incremented. The per-bdi writeback count is not decremented until the actual write completes. On dirtying the page, fuse waits for a previous write to finish before proceeding. This makes sure, there can only be one temporary page used at a time for one cached page. This approach is wasteful in both memory and CPU bandwidth, so why is this complication needed? The basic problem is that there can be no guarantee about the time in which the userspace filesystem will complete a write. It may be buggy or even malicious, and fail to complete WRITE requests. We don't want unrelated parts of the system to grind to a halt in such cases. Also a filesystem may need additional resources (particularly memory) to complete a WRITE request. There's a great danger of a deadlock if that allocation may wait for the writepage to finish. Currently there are several cases where the kernel can block on page writeback: - allocation order is larger than PAGE_ALLOC_COSTLY_ORDER - page migration - throttle_vm_writeout (through NR_WRITEBACK) - sync(2) Of course in some cases (fsync, msync) we explicitly want to allow blocking. So for these cases new code has to be added to fuse, since the VM is not tracking writeback pages for us any more. As an extra safetly measure, the maximum dirty ratio allocated to a single fuse filesystem is set to 1% by default. This way one (or several) buggy or malicious fuse filesystems cannot slow down the rest of the system by hogging dirty memory. With appropriate privileges, this limit can be raised through '/sys/class/bdi/<bdi>/max_ratio'. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-30 15:54:41 +08:00
.set_page_dirty = __set_page_dirty_nobuffers,
.bmap = fuse_bmap,
};
void fuse_init_file_inode(struct inode *inode)
{
inode->i_fop = &fuse_file_operations;
inode->i_data.a_ops = &fuse_file_aops;
}