linux/fs/ceph/locks.c

382 lines
9.9 KiB
C
Raw Normal View History

#include <linux/ceph/ceph_debug.h>
#include <linux/file.h>
#include <linux/namei.h>
#include <linux/random.h>
#include "super.h"
#include "mds_client.h"
#include <linux/ceph/pagelist.h>
static u64 lock_secret;
static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc,
struct ceph_mds_request *req);
static inline u64 secure_addr(void *addr)
{
u64 v = lock_secret ^ (u64)(unsigned long)addr;
/*
* Set the most significant bit, so that MDS knows the 'owner'
* is sufficient to identify the owner of lock. (old code uses
* both 'owner' and 'pid')
*/
v |= (1ULL << 63);
return v;
}
void __init ceph_flock_init(void)
{
get_random_bytes(&lock_secret, sizeof(lock_secret));
}
/**
* Implement fcntl and flock locking functions.
*/
static int ceph_lock_message(u8 lock_type, u16 operation, struct file *file,
int cmd, u8 wait, struct file_lock *fl)
{
struct inode *inode = file_inode(file);
struct ceph_mds_client *mdsc = ceph_sb_to_client(inode->i_sb)->mdsc;
struct ceph_mds_request *req;
int err;
u64 length = 0;
u64 owner;
if (operation != CEPH_MDS_OP_SETFILELOCK || cmd == CEPH_LOCK_UNLOCK)
wait = 0;
req = ceph_mdsc_create_request(mdsc, operation, USE_AUTH_MDS);
if (IS_ERR(req))
return PTR_ERR(req);
req->r_inode = inode;
ihold(inode);
req->r_num_caps = 1;
/* mds requires start and length rather than start and end */
if (LLONG_MAX == fl->fl_end)
length = 0;
else
length = fl->fl_end - fl->fl_start + 1;
owner = secure_addr(fl->fl_owner);
dout("ceph_lock_message: rule: %d, op: %d, owner: %llx, pid: %llu, "
"start: %llu, length: %llu, wait: %d, type: %d", (int)lock_type,
(int)operation, owner, (u64)fl->fl_pid, fl->fl_start, length,
wait, fl->fl_type);
req->r_args.filelock_change.rule = lock_type;
req->r_args.filelock_change.type = cmd;
req->r_args.filelock_change.owner = cpu_to_le64(owner);
req->r_args.filelock_change.pid = cpu_to_le64((u64)fl->fl_pid);
req->r_args.filelock_change.start = cpu_to_le64(fl->fl_start);
req->r_args.filelock_change.length = cpu_to_le64(length);
req->r_args.filelock_change.wait = wait;
if (wait)
req->r_wait_for_completion = ceph_lock_wait_for_completion;
err = ceph_mdsc_do_request(mdsc, inode, req);
if (operation == CEPH_MDS_OP_GETFILELOCK) {
fl->fl_pid = le64_to_cpu(req->r_reply_info.filelock_reply->pid);
if (CEPH_LOCK_SHARED == req->r_reply_info.filelock_reply->type)
fl->fl_type = F_RDLCK;
else if (CEPH_LOCK_EXCL == req->r_reply_info.filelock_reply->type)
fl->fl_type = F_WRLCK;
else
fl->fl_type = F_UNLCK;
fl->fl_start = le64_to_cpu(req->r_reply_info.filelock_reply->start);
length = le64_to_cpu(req->r_reply_info.filelock_reply->start) +
le64_to_cpu(req->r_reply_info.filelock_reply->length);
if (length >= 1)
fl->fl_end = length -1;
else
fl->fl_end = 0;
}
ceph_mdsc_put_request(req);
dout("ceph_lock_message: rule: %d, op: %d, pid: %llu, start: %llu, "
"length: %llu, wait: %d, type: %d, err code %d", (int)lock_type,
(int)operation, (u64)fl->fl_pid, fl->fl_start,
length, wait, fl->fl_type, err);
return err;
}
static int ceph_lock_wait_for_completion(struct ceph_mds_client *mdsc,
struct ceph_mds_request *req)
{
struct ceph_mds_request *intr_req;
struct inode *inode = req->r_inode;
int err, lock_type;
BUG_ON(req->r_op != CEPH_MDS_OP_SETFILELOCK);
if (req->r_args.filelock_change.rule == CEPH_LOCK_FCNTL)
lock_type = CEPH_LOCK_FCNTL_INTR;
else if (req->r_args.filelock_change.rule == CEPH_LOCK_FLOCK)
lock_type = CEPH_LOCK_FLOCK_INTR;
else
BUG_ON(1);
BUG_ON(req->r_args.filelock_change.type == CEPH_LOCK_UNLOCK);
err = wait_for_completion_interruptible(&req->r_completion);
if (!err)
return 0;
dout("ceph_lock_wait_for_completion: request %llu was interrupted\n",
req->r_tid);
intr_req = ceph_mdsc_create_request(mdsc, CEPH_MDS_OP_SETFILELOCK,
USE_AUTH_MDS);
if (IS_ERR(intr_req))
return PTR_ERR(intr_req);
intr_req->r_inode = inode;
ihold(inode);
intr_req->r_num_caps = 1;
intr_req->r_args.filelock_change = req->r_args.filelock_change;
intr_req->r_args.filelock_change.rule = lock_type;
intr_req->r_args.filelock_change.type = CEPH_LOCK_UNLOCK;
err = ceph_mdsc_do_request(mdsc, inode, intr_req);
ceph_mdsc_put_request(intr_req);
if (err && err != -ERESTARTSYS)
return err;
wait_for_completion(&req->r_completion);
return 0;
}
/**
* Attempt to set an fcntl lock.
* For now, this just goes away to the server. Later it may be more awesome.
*/
int ceph_lock(struct file *file, int cmd, struct file_lock *fl)
{
u8 lock_cmd;
int err;
u8 wait = 0;
u16 op = CEPH_MDS_OP_SETFILELOCK;
if (!(fl->fl_flags & FL_POSIX))
return -ENOLCK;
/* No mandatory locks */
if (__mandatory_lock(file->f_mapping->host) && fl->fl_type != F_UNLCK)
return -ENOLCK;
dout("ceph_lock, fl_owner: %p", fl->fl_owner);
/* set wait bit as appropriate, then make command as Ceph expects it*/
if (IS_GETLK(cmd))
op = CEPH_MDS_OP_GETFILELOCK;
else if (IS_SETLKW(cmd))
wait = 1;
if (F_RDLCK == fl->fl_type)
lock_cmd = CEPH_LOCK_SHARED;
else if (F_WRLCK == fl->fl_type)
lock_cmd = CEPH_LOCK_EXCL;
else
lock_cmd = CEPH_LOCK_UNLOCK;
err = ceph_lock_message(CEPH_LOCK_FCNTL, op, file, lock_cmd, wait, fl);
if (!err) {
if (op != CEPH_MDS_OP_GETFILELOCK) {
dout("mds locked, locking locally");
err = posix_lock_file(file, fl, NULL);
if (err && (CEPH_MDS_OP_SETFILELOCK == op)) {
/* undo! This should only happen if
* the kernel detects local
* deadlock. */
ceph_lock_message(CEPH_LOCK_FCNTL, op, file,
CEPH_LOCK_UNLOCK, 0, fl);
dout("got %d on posix_lock_file, undid lock",
err);
}
}
}
return err;
}
int ceph_flock(struct file *file, int cmd, struct file_lock *fl)
{
u8 lock_cmd;
int err;
u8 wait = 0;
if (!(fl->fl_flags & FL_FLOCK))
return -ENOLCK;
/* No mandatory locks */
if (__mandatory_lock(file->f_mapping->host) && fl->fl_type != F_UNLCK)
return -ENOLCK;
dout("ceph_flock, fl_file: %p", fl->fl_file);
if (IS_SETLKW(cmd))
wait = 1;
if (F_RDLCK == fl->fl_type)
lock_cmd = CEPH_LOCK_SHARED;
else if (F_WRLCK == fl->fl_type)
lock_cmd = CEPH_LOCK_EXCL;
else
lock_cmd = CEPH_LOCK_UNLOCK;
err = ceph_lock_message(CEPH_LOCK_FLOCK, CEPH_MDS_OP_SETFILELOCK,
file, lock_cmd, wait, fl);
if (!err) {
err = locks_lock_file_wait(file, fl);
if (err) {
ceph_lock_message(CEPH_LOCK_FLOCK,
CEPH_MDS_OP_SETFILELOCK,
file, CEPH_LOCK_UNLOCK, 0, fl);
dout("got %d on locks_lock_file_wait, undid lock", err);
}
}
return err;
}
/*
* Fills in the passed counter variables, so you can prepare pagelist metadata
* before calling ceph_encode_locks.
*/
void ceph_count_locks(struct inode *inode, int *fcntl_count, int *flock_count)
{
struct file_lock *lock;
struct file_lock_context *ctx;
*fcntl_count = 0;
*flock_count = 0;
ctx = inode->i_flctx;
if (ctx) {
spin_lock(&ctx->flc_lock);
list_for_each_entry(lock, &ctx->flc_posix, fl_list)
++(*fcntl_count);
list_for_each_entry(lock, &ctx->flc_flock, fl_list)
++(*flock_count);
spin_unlock(&ctx->flc_lock);
}
dout("counted %d flock locks and %d fcntl locks",
*flock_count, *fcntl_count);
}
/**
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
* Encode the flock and fcntl locks for the given inode into the ceph_filelock
* array. Must be called with inode->i_lock already held.
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
* If we encounter more of a specific lock type than expected, return -ENOSPC.
*/
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
int ceph_encode_locks_to_buffer(struct inode *inode,
struct ceph_filelock *flocks,
int num_fcntl_locks, int num_flock_locks)
{
struct file_lock *lock;
struct file_lock_context *ctx = inode->i_flctx;
int err = 0;
int seen_fcntl = 0;
int seen_flock = 0;
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
int l = 0;
dout("encoding %d flock and %d fcntl locks", num_flock_locks,
num_fcntl_locks);
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
if (!ctx)
return 0;
spin_lock(&ctx->flc_lock);
list_for_each_entry(lock, &ctx->flc_posix, fl_list) {
++seen_fcntl;
if (seen_fcntl > num_fcntl_locks) {
err = -ENOSPC;
goto fail;
}
err = lock_to_ceph_filelock(lock, &flocks[l]);
if (err)
goto fail;
++l;
}
list_for_each_entry(lock, &ctx->flc_flock, fl_list) {
++seen_flock;
if (seen_flock > num_flock_locks) {
err = -ENOSPC;
goto fail;
}
err = lock_to_ceph_filelock(lock, &flocks[l]);
if (err)
goto fail;
++l;
}
fail:
spin_unlock(&ctx->flc_lock);
return err;
}
ceph: ceph_pagelist_append might sleep while atomic Ceph's encode_caps_cb() worked hard to not call __page_cache_alloc() while holding a lock, but it's spoiled because ceph_pagelist_addpage() always calls kmap(), which might sleep. Here's the result: [13439.295457] ceph: mds0 reconnect start [13439.300572] BUG: sleeping function called from invalid context at include/linux/highmem.h:58 [13439.309243] in_atomic(): 1, irqs_disabled(): 0, pid: 12059, name: kworker/1:1 . . . [13439.376225] Call Trace: [13439.378757] [<ffffffff81076f4c>] __might_sleep+0xfc/0x110 [13439.384353] [<ffffffffa03f4ce0>] ceph_pagelist_append+0x120/0x1b0 [libceph] [13439.391491] [<ffffffffa0448fe9>] ceph_encode_locks+0x89/0x190 [ceph] [13439.398035] [<ffffffff814ee849>] ? _raw_spin_lock+0x49/0x50 [13439.403775] [<ffffffff811cadf5>] ? lock_flocks+0x15/0x20 [13439.409277] [<ffffffffa045e2af>] encode_caps_cb+0x41f/0x4a0 [ceph] [13439.415622] [<ffffffff81196748>] ? igrab+0x28/0x70 [13439.420610] [<ffffffffa045e9f8>] ? iterate_session_caps+0xe8/0x250 [ceph] [13439.427584] [<ffffffffa045ea25>] iterate_session_caps+0x115/0x250 [ceph] [13439.434499] [<ffffffffa045de90>] ? set_request_path_attr+0x2d0/0x2d0 [ceph] [13439.441646] [<ffffffffa0462888>] send_mds_reconnect+0x238/0x450 [ceph] [13439.448363] [<ffffffffa0464542>] ? ceph_mdsmap_decode+0x5e2/0x770 [ceph] [13439.455250] [<ffffffffa0462e42>] check_new_map+0x352/0x500 [ceph] [13439.461534] [<ffffffffa04631ad>] ceph_mdsc_handle_map+0x1bd/0x260 [ceph] [13439.468432] [<ffffffff814ebc7e>] ? mutex_unlock+0xe/0x10 [13439.473934] [<ffffffffa043c612>] extra_mon_dispatch+0x22/0x30 [ceph] [13439.480464] [<ffffffffa03f6c2c>] dispatch+0xbc/0x110 [libceph] [13439.486492] [<ffffffffa03eec3d>] process_message+0x1ad/0x1d0 [libceph] [13439.493190] [<ffffffffa03f1498>] ? read_partial_message+0x3e8/0x520 [libceph] . . . [13439.587132] ceph: mds0 reconnect success [13490.720032] ceph: mds0 caps stale [13501.235257] ceph: mds0 recovery completed [13501.300419] ceph: mds0 caps renewed Fix it up by encoding locks into a buffer first, and when the number of encoded locks is stable, copy that into a ceph_pagelist. [elder@inktank.com: abbreviated the stack info a bit.] Cc: stable@vger.kernel.org # 3.4+ Signed-off-by: Jim Schutt <jaschut@sandia.gov> Reviewed-by: Alex Elder <elder@inktank.com>
2013-05-15 18:03:35 +00:00
/**
* Copy the encoded flock and fcntl locks into the pagelist.
* Format is: #fcntl locks, sequential fcntl locks, #flock locks,
* sequential flock locks.
* Returns zero on success.
*/
int ceph_locks_to_pagelist(struct ceph_filelock *flocks,
struct ceph_pagelist *pagelist,
int num_fcntl_locks, int num_flock_locks)
{
int err = 0;
__le32 nlocks;
nlocks = cpu_to_le32(num_fcntl_locks);
err = ceph_pagelist_append(pagelist, &nlocks, sizeof(nlocks));
if (err)
goto out_fail;
err = ceph_pagelist_append(pagelist, flocks,
num_fcntl_locks * sizeof(*flocks));
if (err)
goto out_fail;
nlocks = cpu_to_le32(num_flock_locks);
err = ceph_pagelist_append(pagelist, &nlocks, sizeof(nlocks));
if (err)
goto out_fail;
err = ceph_pagelist_append(pagelist,
&flocks[num_fcntl_locks],
num_flock_locks * sizeof(*flocks));
out_fail:
return err;
}
/*
* Given a pointer to a lock, convert it to a ceph filelock
*/
int lock_to_ceph_filelock(struct file_lock *lock,
struct ceph_filelock *cephlock)
{
int err = 0;
cephlock->start = cpu_to_le64(lock->fl_start);
cephlock->length = cpu_to_le64(lock->fl_end - lock->fl_start + 1);
cephlock->client = cpu_to_le64(0);
cephlock->pid = cpu_to_le64((u64)lock->fl_pid);
cephlock->owner = cpu_to_le64(secure_addr(lock->fl_owner));
switch (lock->fl_type) {
case F_RDLCK:
cephlock->type = CEPH_LOCK_SHARED;
break;
case F_WRLCK:
cephlock->type = CEPH_LOCK_EXCL;
break;
case F_UNLCK:
cephlock->type = CEPH_LOCK_UNLOCK;
break;
default:
dout("Have unknown lock type %d", lock->fl_type);
err = -EINVAL;
}
return err;
}