2006-07-12 23:44:04 +02:00
|
|
|
/*
|
2010-02-24 18:08:18 +01:00
|
|
|
* Copyright (C) 2006-2010 Red Hat, Inc. All rights reserved.
|
2006-07-12 23:44:04 +02:00
|
|
|
*
|
|
|
|
* This copyrighted material is made available to anyone wishing to use,
|
|
|
|
* modify, copy, or redistribute it subject to the terms and conditions
|
|
|
|
* of the GNU General Public License v.2.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/miscdevice.h>
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/wait.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/file.h>
|
|
|
|
#include <linux/fs.h>
|
|
|
|
#include <linux/poll.h>
|
|
|
|
#include <linux/signal.h>
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/dlm.h>
|
|
|
|
#include <linux/dlm_device.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 09:04:11 +01:00
|
|
|
#include <linux/slab.h>
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
#include "dlm_internal.h"
|
|
|
|
#include "lockspace.h"
|
|
|
|
#include "lock.h"
|
|
|
|
#include "lvb_table.h"
|
2007-02-26 00:18:42 +01:00
|
|
|
#include "user.h"
|
2011-02-21 21:58:21 +01:00
|
|
|
#include "ast.h"
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2008-01-29 06:50:16 +01:00
|
|
|
static const char name_prefix[] = "dlm";
|
2007-02-12 09:55:34 +01:00
|
|
|
static const struct file_operations device_fops;
|
2008-08-18 18:43:30 +02:00
|
|
|
static atomic_t dlm_monitor_opened;
|
|
|
|
static int dlm_monitor_unused = 1;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
|
|
|
|
struct dlm_lock_params32 {
|
|
|
|
__u8 mode;
|
|
|
|
__u8 namelen;
|
2007-05-18 16:00:32 +02:00
|
|
|
__u16 unused;
|
|
|
|
__u32 flags;
|
2006-07-12 23:44:04 +02:00
|
|
|
__u32 lkid;
|
|
|
|
__u32 parent;
|
2007-05-18 16:00:32 +02:00
|
|
|
__u64 xid;
|
|
|
|
__u64 timeout;
|
2006-07-12 23:44:04 +02:00
|
|
|
__u32 castparam;
|
|
|
|
__u32 castaddr;
|
|
|
|
__u32 bastparam;
|
|
|
|
__u32 bastaddr;
|
|
|
|
__u32 lksb;
|
|
|
|
char lvb[DLM_USER_LVB_LEN];
|
|
|
|
char name[0];
|
|
|
|
};
|
|
|
|
|
|
|
|
struct dlm_write_request32 {
|
|
|
|
__u32 version[3];
|
|
|
|
__u8 cmd;
|
|
|
|
__u8 is64bit;
|
|
|
|
__u8 unused[2];
|
|
|
|
|
|
|
|
union {
|
|
|
|
struct dlm_lock_params32 lock;
|
|
|
|
struct dlm_lspace_params lspace;
|
2007-03-30 22:06:16 +02:00
|
|
|
struct dlm_purge_params purge;
|
2006-07-12 23:44:04 +02:00
|
|
|
} i;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct dlm_lksb32 {
|
|
|
|
__u32 sb_status;
|
|
|
|
__u32 sb_lkid;
|
|
|
|
__u8 sb_flags;
|
|
|
|
__u32 sb_lvbptr;
|
|
|
|
};
|
|
|
|
|
|
|
|
struct dlm_lock_result32 {
|
2007-05-18 16:00:32 +02:00
|
|
|
__u32 version[3];
|
2006-07-12 23:44:04 +02:00
|
|
|
__u32 length;
|
|
|
|
__u32 user_astaddr;
|
|
|
|
__u32 user_astparam;
|
|
|
|
__u32 user_lksb;
|
|
|
|
struct dlm_lksb32 lksb;
|
|
|
|
__u8 bast_mode;
|
|
|
|
__u8 unused[3];
|
|
|
|
/* Offsets may be zero if no data is present */
|
|
|
|
__u32 lvb_offset;
|
|
|
|
};
|
|
|
|
|
|
|
|
static void compat_input(struct dlm_write_request *kb,
|
2008-01-17 11:25:28 +01:00
|
|
|
struct dlm_write_request32 *kb32,
|
2009-03-04 18:17:23 +01:00
|
|
|
int namelen)
|
2006-07-12 23:44:04 +02:00
|
|
|
{
|
|
|
|
kb->version[0] = kb32->version[0];
|
|
|
|
kb->version[1] = kb32->version[1];
|
|
|
|
kb->version[2] = kb32->version[2];
|
|
|
|
|
|
|
|
kb->cmd = kb32->cmd;
|
|
|
|
kb->is64bit = kb32->is64bit;
|
|
|
|
if (kb->cmd == DLM_USER_CREATE_LOCKSPACE ||
|
|
|
|
kb->cmd == DLM_USER_REMOVE_LOCKSPACE) {
|
|
|
|
kb->i.lspace.flags = kb32->i.lspace.flags;
|
|
|
|
kb->i.lspace.minor = kb32->i.lspace.minor;
|
2009-03-04 18:17:23 +01:00
|
|
|
memcpy(kb->i.lspace.name, kb32->i.lspace.name, namelen);
|
2007-03-30 22:06:16 +02:00
|
|
|
} else if (kb->cmd == DLM_USER_PURGE) {
|
|
|
|
kb->i.purge.nodeid = kb32->i.purge.nodeid;
|
|
|
|
kb->i.purge.pid = kb32->i.purge.pid;
|
2006-07-12 23:44:04 +02:00
|
|
|
} else {
|
|
|
|
kb->i.lock.mode = kb32->i.lock.mode;
|
|
|
|
kb->i.lock.namelen = kb32->i.lock.namelen;
|
|
|
|
kb->i.lock.flags = kb32->i.lock.flags;
|
|
|
|
kb->i.lock.lkid = kb32->i.lock.lkid;
|
|
|
|
kb->i.lock.parent = kb32->i.lock.parent;
|
2007-05-18 16:00:32 +02:00
|
|
|
kb->i.lock.xid = kb32->i.lock.xid;
|
|
|
|
kb->i.lock.timeout = kb32->i.lock.timeout;
|
2006-07-12 23:44:04 +02:00
|
|
|
kb->i.lock.castparam = (void *)(long)kb32->i.lock.castparam;
|
|
|
|
kb->i.lock.castaddr = (void *)(long)kb32->i.lock.castaddr;
|
|
|
|
kb->i.lock.bastparam = (void *)(long)kb32->i.lock.bastparam;
|
|
|
|
kb->i.lock.bastaddr = (void *)(long)kb32->i.lock.bastaddr;
|
|
|
|
kb->i.lock.lksb = (void *)(long)kb32->i.lock.lksb;
|
|
|
|
memcpy(kb->i.lock.lvb, kb32->i.lock.lvb, DLM_USER_LVB_LEN);
|
2009-03-04 18:17:23 +01:00
|
|
|
memcpy(kb->i.lock.name, kb32->i.lock.name, namelen);
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static void compat_output(struct dlm_lock_result *res,
|
|
|
|
struct dlm_lock_result32 *res32)
|
|
|
|
{
|
2007-05-18 16:00:32 +02:00
|
|
|
res32->version[0] = res->version[0];
|
|
|
|
res32->version[1] = res->version[1];
|
|
|
|
res32->version[2] = res->version[2];
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
res32->user_astaddr = (__u32)(long)res->user_astaddr;
|
|
|
|
res32->user_astparam = (__u32)(long)res->user_astparam;
|
|
|
|
res32->user_lksb = (__u32)(long)res->user_lksb;
|
|
|
|
res32->bast_mode = res->bast_mode;
|
|
|
|
|
|
|
|
res32->lvb_offset = res->lvb_offset;
|
|
|
|
res32->length = res->length;
|
|
|
|
|
|
|
|
res32->lksb.sb_status = res->lksb.sb_status;
|
|
|
|
res32->lksb.sb_flags = res->lksb.sb_flags;
|
|
|
|
res32->lksb.sb_lkid = res->lksb.sb_lkid;
|
|
|
|
res32->lksb.sb_lvbptr = (__u32)(long)res->lksb.sb_lvbptr;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
2007-05-29 15:44:23 +02:00
|
|
|
/* Figure out if this lock is at the end of its life and no longer
|
|
|
|
available for the application to use. The lkb still exists until
|
|
|
|
the final ast is read. A lock becomes EOL in three situations:
|
|
|
|
1. a noqueue request fails with EAGAIN
|
|
|
|
2. an unlock completes with EUNLOCK
|
|
|
|
3. a cancel of a waiting request completes with ECANCEL/EDEADLK
|
|
|
|
An EOL lock needs to be removed from the process's list of locks.
|
|
|
|
And we can't allow any new operation on an EOL lock. This is
|
|
|
|
not related to the lifetime of the lkb struct which is managed
|
|
|
|
entirely by refcount. */
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
static int lkb_is_endoflife(int mode, int status)
|
2007-05-29 15:44:23 +02:00
|
|
|
{
|
2011-02-21 21:58:21 +01:00
|
|
|
switch (status) {
|
2007-05-29 15:44:23 +02:00
|
|
|
case -DLM_EUNLOCK:
|
|
|
|
return 1;
|
|
|
|
case -DLM_ECANCEL:
|
|
|
|
case -ETIMEDOUT:
|
2007-05-29 15:46:00 +02:00
|
|
|
case -EDEADLK:
|
2007-05-29 15:44:23 +02:00
|
|
|
case -EAGAIN:
|
2011-02-21 21:58:21 +01:00
|
|
|
if (mode == DLM_LOCK_IV)
|
2007-05-29 15:44:23 +02:00
|
|
|
return 1;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-03-28 16:56:46 +02:00
|
|
|
/* we could possibly check if the cancel of an orphan has resulted in the lkb
|
|
|
|
being removed and then remove that lkb from the orphans list and free it */
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
void dlm_user_add_ast(struct dlm_lkb *lkb, uint32_t flags, int mode,
|
|
|
|
int status, uint32_t sbflags, uint64_t seq)
|
2006-07-12 23:44:04 +02:00
|
|
|
{
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
struct dlm_user_args *ua;
|
|
|
|
struct dlm_user_proc *proc;
|
2011-02-21 21:58:21 +01:00
|
|
|
int rv;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2007-03-28 16:56:46 +02:00
|
|
|
if (lkb->lkb_flags & (DLM_IFL_ORPHAN | DLM_IFL_DEAD))
|
2006-07-12 23:44:04 +02:00
|
|
|
return;
|
|
|
|
|
|
|
|
ls = lkb->lkb_resource->res_ls;
|
|
|
|
mutex_lock(&ls->ls_clear_proc_locks);
|
|
|
|
|
|
|
|
/* If ORPHAN/DEAD flag is set, it means the process is dead so an ast
|
|
|
|
can't be delivered. For ORPHAN's, dlm_clear_proc_locks() freed
|
2007-03-28 16:56:46 +02:00
|
|
|
lkb->ua so we can't try to use it. This second check is necessary
|
|
|
|
for cases where a completion ast is received for an operation that
|
|
|
|
began before clear_proc_locks did its cancel/unlock. */
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2007-03-28 16:56:46 +02:00
|
|
|
if (lkb->lkb_flags & (DLM_IFL_ORPHAN | DLM_IFL_DEAD))
|
2006-07-12 23:44:04 +02:00
|
|
|
goto out;
|
|
|
|
|
2008-02-07 06:27:04 +01:00
|
|
|
DLM_ASSERT(lkb->lkb_ua, dlm_print_lkb(lkb););
|
|
|
|
ua = lkb->lkb_ua;
|
2006-07-12 23:44:04 +02:00
|
|
|
proc = ua->proc;
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if ((flags & DLM_CB_BAST) && ua->bastaddr == NULL)
|
2006-07-12 23:44:04 +02:00
|
|
|
goto out;
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if ((flags & DLM_CB_CAST) && lkb_is_endoflife(mode, status))
|
|
|
|
lkb->lkb_flags |= DLM_IFL_ENDOFLIFE;
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
spin_lock(&proc->asts_spin);
|
2007-03-28 16:56:46 +02:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
rv = dlm_add_lkb_callback(lkb, flags, mode, status, sbflags, seq);
|
|
|
|
if (rv < 0) {
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
goto out;
|
|
|
|
}
|
2007-03-28 16:56:46 +02:00
|
|
|
|
2011-04-05 20:16:24 +02:00
|
|
|
if (list_empty(&lkb->lkb_cb_list)) {
|
2006-07-12 23:44:04 +02:00
|
|
|
kref_get(&lkb->lkb_ref);
|
2011-04-05 20:16:24 +02:00
|
|
|
list_add_tail(&lkb->lkb_cb_list, &proc->asts);
|
2006-07-12 23:44:04 +02:00
|
|
|
wake_up_interruptible(&proc->wait);
|
|
|
|
}
|
|
|
|
spin_unlock(&proc->asts_spin);
|
2006-07-18 18:24:04 +02:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (lkb->lkb_flags & DLM_IFL_ENDOFLIFE) {
|
|
|
|
/* N.B. spin_lock locks_spin, not asts_spin */
|
2008-01-14 22:48:58 +01:00
|
|
|
spin_lock(&proc->locks_spin);
|
2007-03-28 16:56:46 +02:00
|
|
|
if (!list_empty(&lkb->lkb_ownqueue)) {
|
|
|
|
list_del_init(&lkb->lkb_ownqueue);
|
|
|
|
dlm_put_lkb(lkb);
|
|
|
|
}
|
2008-01-14 22:48:58 +01:00
|
|
|
spin_unlock(&proc->locks_spin);
|
2006-07-18 18:24:04 +02:00
|
|
|
}
|
2006-07-12 23:44:04 +02:00
|
|
|
out:
|
|
|
|
mutex_unlock(&ls->ls_clear_proc_locks);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int device_user_lock(struct dlm_user_proc *proc,
|
|
|
|
struct dlm_lock_params *params)
|
|
|
|
{
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
struct dlm_user_args *ua;
|
2014-10-17 18:05:50 +02:00
|
|
|
uint32_t lkid;
|
2006-07-12 23:44:04 +02:00
|
|
|
int error = -ENOMEM;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(proc->lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
if (!params->castaddr || !params->lksb) {
|
|
|
|
error = -EINVAL;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
2009-11-30 23:34:43 +01:00
|
|
|
ua = kzalloc(sizeof(struct dlm_user_args), GFP_NOFS);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (!ua)
|
|
|
|
goto out;
|
|
|
|
ua->proc = proc;
|
|
|
|
ua->user_lksb = params->lksb;
|
|
|
|
ua->castparam = params->castparam;
|
|
|
|
ua->castaddr = params->castaddr;
|
|
|
|
ua->bastparam = params->bastparam;
|
|
|
|
ua->bastaddr = params->bastaddr;
|
2007-05-18 16:00:32 +02:00
|
|
|
ua->xid = params->xid;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2014-10-17 18:05:50 +02:00
|
|
|
if (params->flags & DLM_LKF_CONVERT) {
|
2006-07-12 23:44:04 +02:00
|
|
|
error = dlm_user_convert(ls, ua,
|
|
|
|
params->mode, params->flags,
|
2007-05-18 16:00:32 +02:00
|
|
|
params->lkid, params->lvb,
|
|
|
|
(unsigned long) params->timeout);
|
2014-10-17 18:05:50 +02:00
|
|
|
} else if (params->flags & DLM_LKF_ORPHAN) {
|
|
|
|
error = dlm_user_adopt_orphan(ls, ua,
|
|
|
|
params->mode, params->flags,
|
|
|
|
params->name, params->namelen,
|
|
|
|
(unsigned long) params->timeout,
|
|
|
|
&lkid);
|
|
|
|
if (!error)
|
|
|
|
error = lkid;
|
|
|
|
} else {
|
2006-07-12 23:44:04 +02:00
|
|
|
error = dlm_user_request(ls, ua,
|
|
|
|
params->mode, params->flags,
|
|
|
|
params->name, params->namelen,
|
2007-05-18 16:00:32 +02:00
|
|
|
(unsigned long) params->timeout);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (!error)
|
|
|
|
error = ua->lksb.sb_lkid;
|
|
|
|
}
|
|
|
|
out:
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int device_user_unlock(struct dlm_user_proc *proc,
|
|
|
|
struct dlm_lock_params *params)
|
|
|
|
{
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
struct dlm_user_args *ua;
|
|
|
|
int error = -ENOMEM;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(proc->lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
2009-11-30 23:34:43 +01:00
|
|
|
ua = kzalloc(sizeof(struct dlm_user_args), GFP_NOFS);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (!ua)
|
|
|
|
goto out;
|
|
|
|
ua->proc = proc;
|
|
|
|
ua->user_lksb = params->lksb;
|
|
|
|
ua->castparam = params->castparam;
|
|
|
|
ua->castaddr = params->castaddr;
|
|
|
|
|
|
|
|
if (params->flags & DLM_LKF_CANCEL)
|
|
|
|
error = dlm_user_cancel(ls, ua, params->flags, params->lkid);
|
|
|
|
else
|
|
|
|
error = dlm_user_unlock(ls, ua, params->flags, params->lkid,
|
|
|
|
params->lvb);
|
|
|
|
out:
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2007-05-29 15:46:00 +02:00
|
|
|
static int device_user_deadlock(struct dlm_user_proc *proc,
|
|
|
|
struct dlm_lock_params *params)
|
|
|
|
{
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(proc->lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
error = dlm_user_deadlock(ls, params->flags, params->lkid);
|
|
|
|
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2008-08-06 20:30:24 +02:00
|
|
|
static int dlm_device_register(struct dlm_ls *ls, char *name)
|
2007-03-21 10:23:53 +01:00
|
|
|
{
|
|
|
|
int error, len;
|
|
|
|
|
2008-08-06 20:30:24 +02:00
|
|
|
/* The device is already registered. This happens when the
|
|
|
|
lockspace is created multiple times from userspace. */
|
|
|
|
if (ls->ls_device.name)
|
|
|
|
return 0;
|
|
|
|
|
2007-03-21 10:23:53 +01:00
|
|
|
error = -ENOMEM;
|
|
|
|
len = strlen(name) + strlen(name_prefix) + 2;
|
2009-11-30 23:34:43 +01:00
|
|
|
ls->ls_device.name = kzalloc(len, GFP_NOFS);
|
2007-03-21 10:23:53 +01:00
|
|
|
if (!ls->ls_device.name)
|
|
|
|
goto fail;
|
|
|
|
|
|
|
|
snprintf((char *)ls->ls_device.name, len, "%s_%s", name_prefix,
|
|
|
|
name);
|
|
|
|
ls->ls_device.fops = &device_fops;
|
|
|
|
ls->ls_device.minor = MISC_DYNAMIC_MINOR;
|
|
|
|
|
|
|
|
error = misc_register(&ls->ls_device);
|
|
|
|
if (error) {
|
|
|
|
kfree(ls->ls_device.name);
|
|
|
|
}
|
|
|
|
fail:
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2008-08-06 20:30:24 +02:00
|
|
|
int dlm_device_deregister(struct dlm_ls *ls)
|
|
|
|
{
|
|
|
|
/* The device is not registered. This happens when the lockspace
|
|
|
|
was never used from userspace, or when device_create_lockspace()
|
|
|
|
calls dlm_release_lockspace() after the register fails. */
|
|
|
|
if (!ls->ls_device.name)
|
|
|
|
return 0;
|
|
|
|
|
2015-07-31 00:59:57 +02:00
|
|
|
misc_deregister(&ls->ls_device);
|
|
|
|
kfree(ls->ls_device.name);
|
|
|
|
return 0;
|
2008-08-06 20:30:24 +02:00
|
|
|
}
|
|
|
|
|
2007-03-30 22:06:16 +02:00
|
|
|
static int device_user_purge(struct dlm_user_proc *proc,
|
|
|
|
struct dlm_purge_params *params)
|
|
|
|
{
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(proc->lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
error = dlm_user_purge(ls, proc, params->nodeid, params->pid);
|
|
|
|
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
static int device_create_lockspace(struct dlm_lspace_params *params)
|
|
|
|
{
|
|
|
|
dlm_lockspace_t *lockspace;
|
|
|
|
struct dlm_ls *ls;
|
2007-03-21 10:23:53 +01:00
|
|
|
int error;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
|
2011-11-02 20:30:58 +01:00
|
|
|
error = dlm_new_lockspace(params->name, NULL, params->flags,
|
|
|
|
DLM_USER_LVB_LEN, NULL, NULL, NULL,
|
|
|
|
&lockspace);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (error)
|
|
|
|
return error;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
2008-08-06 20:30:24 +02:00
|
|
|
error = dlm_device_register(ls, params->name);
|
2006-07-12 23:44:04 +02:00
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
|
2007-03-21 10:23:53 +01:00
|
|
|
if (error)
|
|
|
|
dlm_release_lockspace(lockspace, 0);
|
|
|
|
else
|
|
|
|
error = ls->ls_device.minor;
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int device_remove_lockspace(struct dlm_lspace_params *params)
|
|
|
|
{
|
|
|
|
dlm_lockspace_t *lockspace;
|
|
|
|
struct dlm_ls *ls;
|
2006-08-30 17:50:18 +02:00
|
|
|
int error, force = 0;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
if (!capable(CAP_SYS_ADMIN))
|
|
|
|
return -EPERM;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_device(params->minor);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
2006-08-30 17:50:18 +02:00
|
|
|
if (params->flags & DLM_USER_LSFLG_FORCEFREE)
|
|
|
|
force = 2;
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
lockspace = ls->ls_local_handle;
|
2008-08-06 20:30:24 +02:00
|
|
|
dlm_put_lockspace(ls);
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2008-08-06 20:30:24 +02:00
|
|
|
/* The final dlm_release_lockspace waits for references to go to
|
|
|
|
zero, so all processes will need to close their device for the
|
|
|
|
ls before the release will proceed. release also calls the
|
|
|
|
device_deregister above. Converting a positive return value
|
|
|
|
from release to zero means that userspace won't know when its
|
|
|
|
release was the final one, but it shouldn't need to know. */
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2006-08-30 17:50:18 +02:00
|
|
|
error = dlm_release_lockspace(lockspace, force);
|
2008-08-06 20:30:24 +02:00
|
|
|
if (error > 0)
|
|
|
|
error = 0;
|
2006-07-12 23:44:04 +02:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Check the user's version matches ours */
|
|
|
|
static int check_version(struct dlm_write_request *req)
|
|
|
|
{
|
|
|
|
if (req->version[0] != DLM_DEVICE_VERSION_MAJOR ||
|
|
|
|
(req->version[0] == DLM_DEVICE_VERSION_MAJOR &&
|
|
|
|
req->version[1] > DLM_DEVICE_VERSION_MINOR)) {
|
|
|
|
|
|
|
|
printk(KERN_DEBUG "dlm: process %s (%d) version mismatch "
|
|
|
|
"user (%d.%d.%d) kernel (%d.%d.%d)\n",
|
|
|
|
current->comm,
|
2007-10-19 08:40:40 +02:00
|
|
|
task_pid_nr(current),
|
2006-07-12 23:44:04 +02:00
|
|
|
req->version[0],
|
|
|
|
req->version[1],
|
|
|
|
req->version[2],
|
|
|
|
DLM_DEVICE_VERSION_MAJOR,
|
|
|
|
DLM_DEVICE_VERSION_MINOR,
|
|
|
|
DLM_DEVICE_VERSION_PATCH);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* device_write
|
|
|
|
*
|
|
|
|
* device_user_lock
|
|
|
|
* dlm_user_request -> request_lock
|
|
|
|
* dlm_user_convert -> convert_lock
|
|
|
|
*
|
|
|
|
* device_user_unlock
|
|
|
|
* dlm_user_unlock -> unlock_lock
|
|
|
|
* dlm_user_cancel -> cancel_lock
|
|
|
|
*
|
|
|
|
* device_create_lockspace
|
|
|
|
* dlm_new_lockspace
|
|
|
|
*
|
|
|
|
* device_remove_lockspace
|
|
|
|
* dlm_release_lockspace
|
|
|
|
*/
|
|
|
|
|
|
|
|
/* a write to a lockspace device is a lock or unlock request, a write
|
|
|
|
to the control device is to create/remove a lockspace */
|
|
|
|
|
|
|
|
static ssize_t device_write(struct file *file, const char __user *buf,
|
|
|
|
size_t count, loff_t *ppos)
|
|
|
|
{
|
|
|
|
struct dlm_user_proc *proc = file->private_data;
|
|
|
|
struct dlm_write_request *kbuf;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
if (count < sizeof(struct dlm_write_request32))
|
|
|
|
#else
|
|
|
|
if (count < sizeof(struct dlm_write_request))
|
|
|
|
#endif
|
|
|
|
return -EINVAL;
|
|
|
|
|
2013-02-04 22:31:22 +01:00
|
|
|
/*
|
|
|
|
* can't compare against COMPAT/dlm_write_request32 because
|
|
|
|
* we don't yet know if is64bit is zero
|
|
|
|
*/
|
2012-09-09 16:16:58 +02:00
|
|
|
if (count > sizeof(struct dlm_write_request) + DLM_RESNAME_MAXLEN)
|
|
|
|
return -EINVAL;
|
|
|
|
|
2009-11-30 23:34:43 +01:00
|
|
|
kbuf = kzalloc(count + 1, GFP_NOFS);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (!kbuf)
|
|
|
|
return -ENOMEM;
|
|
|
|
|
|
|
|
if (copy_from_user(kbuf, buf, count)) {
|
|
|
|
error = -EFAULT;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (check_version(kbuf)) {
|
|
|
|
error = -EBADE;
|
|
|
|
goto out_free;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
if (!kbuf->is64bit) {
|
|
|
|
struct dlm_write_request32 *k32buf;
|
2009-03-04 18:17:23 +01:00
|
|
|
int namelen = 0;
|
|
|
|
|
|
|
|
if (count > sizeof(struct dlm_write_request32))
|
|
|
|
namelen = count - sizeof(struct dlm_write_request32);
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
k32buf = (struct dlm_write_request32 *)kbuf;
|
2009-03-04 18:17:23 +01:00
|
|
|
|
|
|
|
/* add 1 after namelen so that the name string is terminated */
|
|
|
|
kbuf = kzalloc(sizeof(struct dlm_write_request) + namelen + 1,
|
2009-11-30 23:34:43 +01:00
|
|
|
GFP_NOFS);
|
2008-07-29 22:21:19 +02:00
|
|
|
if (!kbuf) {
|
|
|
|
kfree(k32buf);
|
2006-07-12 23:44:04 +02:00
|
|
|
return -ENOMEM;
|
2008-07-29 22:21:19 +02:00
|
|
|
}
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
if (proc)
|
|
|
|
set_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags);
|
2009-03-04 18:17:23 +01:00
|
|
|
|
|
|
|
compat_input(kbuf, k32buf, namelen);
|
2006-07-12 23:44:04 +02:00
|
|
|
kfree(k32buf);
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* do we really need this? can a write happen after a close? */
|
|
|
|
if ((kbuf->cmd == DLM_USER_LOCK || kbuf->cmd == DLM_USER_UNLOCK) &&
|
2008-07-29 22:21:19 +02:00
|
|
|
(proc && test_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags))) {
|
|
|
|
error = -EINVAL;
|
|
|
|
goto out_free;
|
|
|
|
}
|
2006-07-12 23:44:04 +02:00
|
|
|
|
|
|
|
error = -EINVAL;
|
|
|
|
|
|
|
|
switch (kbuf->cmd)
|
|
|
|
{
|
|
|
|
case DLM_USER_LOCK:
|
|
|
|
if (!proc) {
|
|
|
|
log_print("no locking on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
error = device_user_lock(proc, &kbuf->i.lock);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case DLM_USER_UNLOCK:
|
|
|
|
if (!proc) {
|
|
|
|
log_print("no locking on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
error = device_user_unlock(proc, &kbuf->i.lock);
|
|
|
|
break;
|
|
|
|
|
2007-05-29 15:46:00 +02:00
|
|
|
case DLM_USER_DEADLOCK:
|
|
|
|
if (!proc) {
|
|
|
|
log_print("no locking on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2007-05-29 15:46:00 +02:00
|
|
|
}
|
|
|
|
error = device_user_deadlock(proc, &kbuf->i.lock);
|
|
|
|
break;
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
case DLM_USER_CREATE_LOCKSPACE:
|
|
|
|
if (proc) {
|
|
|
|
log_print("create/remove only on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
error = device_create_lockspace(&kbuf->i.lspace);
|
|
|
|
break;
|
|
|
|
|
|
|
|
case DLM_USER_REMOVE_LOCKSPACE:
|
|
|
|
if (proc) {
|
|
|
|
log_print("create/remove only on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
error = device_remove_lockspace(&kbuf->i.lspace);
|
|
|
|
break;
|
|
|
|
|
2007-03-30 22:06:16 +02:00
|
|
|
case DLM_USER_PURGE:
|
|
|
|
if (!proc) {
|
|
|
|
log_print("no locking on control device");
|
2013-08-12 22:22:43 +02:00
|
|
|
goto out_free;
|
2007-03-30 22:06:16 +02:00
|
|
|
}
|
|
|
|
error = device_user_purge(proc, &kbuf->i.purge);
|
|
|
|
break;
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
default:
|
|
|
|
log_print("Unknown command passed to DLM device : %d\n",
|
|
|
|
kbuf->cmd);
|
|
|
|
}
|
|
|
|
|
|
|
|
out_free:
|
|
|
|
kfree(kbuf);
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Every process that opens the lockspace device has its own "proc" structure
|
|
|
|
hanging off the open file that's used to keep track of locks owned by the
|
|
|
|
process and asts that need to be delivered to the process. */
|
|
|
|
|
|
|
|
static int device_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct dlm_user_proc *proc;
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_device(iminor(inode));
|
2008-09-04 19:51:20 +02:00
|
|
|
if (!ls)
|
2006-07-12 23:44:04 +02:00
|
|
|
return -ENOENT;
|
|
|
|
|
2009-11-30 23:34:43 +01:00
|
|
|
proc = kzalloc(sizeof(struct dlm_user_proc), GFP_NOFS);
|
2006-07-12 23:44:04 +02:00
|
|
|
if (!proc) {
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
|
|
|
proc->lockspace = ls->ls_local_handle;
|
|
|
|
INIT_LIST_HEAD(&proc->asts);
|
|
|
|
INIT_LIST_HEAD(&proc->locks);
|
2007-01-15 17:34:52 +01:00
|
|
|
INIT_LIST_HEAD(&proc->unlocking);
|
2006-07-12 23:44:04 +02:00
|
|
|
spin_lock_init(&proc->asts_spin);
|
|
|
|
spin_lock_init(&proc->locks_spin);
|
|
|
|
init_waitqueue_head(&proc->wait);
|
|
|
|
file->private_data = proc;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int device_close(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
struct dlm_user_proc *proc = file->private_data;
|
|
|
|
struct dlm_ls *ls;
|
|
|
|
|
|
|
|
ls = dlm_find_lockspace_local(proc->lockspace);
|
|
|
|
if (!ls)
|
|
|
|
return -ENOENT;
|
|
|
|
|
|
|
|
set_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags);
|
|
|
|
|
|
|
|
dlm_clear_proc_locks(ls, proc);
|
|
|
|
|
|
|
|
/* at this point no more lkb's should exist for this lockspace,
|
|
|
|
so there's no chance of dlm_user_add_ast() being called and
|
|
|
|
looking for lkb->ua->proc */
|
|
|
|
|
|
|
|
kfree(proc);
|
|
|
|
file->private_data = NULL;
|
|
|
|
|
|
|
|
dlm_put_lockspace(ls);
|
|
|
|
dlm_put_lockspace(ls); /* for the find in device_open() */
|
|
|
|
|
|
|
|
/* FIXME: AUTOFREE: if this ls is no longer used do
|
|
|
|
device_remove_lockspace() */
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
static int copy_result_to_user(struct dlm_user_args *ua, int compat,
|
|
|
|
uint32_t flags, int mode, int copy_lvb,
|
|
|
|
char __user *buf, size_t count)
|
2006-07-12 23:44:04 +02:00
|
|
|
{
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
struct dlm_lock_result32 result32;
|
|
|
|
#endif
|
|
|
|
struct dlm_lock_result result;
|
|
|
|
void *resultptr;
|
|
|
|
int error=0;
|
|
|
|
int len;
|
|
|
|
int struct_len;
|
|
|
|
|
|
|
|
memset(&result, 0, sizeof(struct dlm_lock_result));
|
2007-05-18 16:00:32 +02:00
|
|
|
result.version[0] = DLM_DEVICE_VERSION_MAJOR;
|
|
|
|
result.version[1] = DLM_DEVICE_VERSION_MINOR;
|
|
|
|
result.version[2] = DLM_DEVICE_VERSION_PATCH;
|
2006-07-12 23:44:04 +02:00
|
|
|
memcpy(&result.lksb, &ua->lksb, sizeof(struct dlm_lksb));
|
|
|
|
result.user_lksb = ua->user_lksb;
|
|
|
|
|
|
|
|
/* FIXME: dlm1 provides for the user's bastparam/addr to not be updated
|
|
|
|
in a conversion unless the conversion is successful. See code
|
|
|
|
in dlm_user_convert() for updating ua from ua_tmp. OpenVMS, though,
|
|
|
|
notes that a new blocking AST address and parameter are set even if
|
|
|
|
the conversion fails, so maybe we should just do that. */
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (flags & DLM_CB_BAST) {
|
2006-07-12 23:44:04 +02:00
|
|
|
result.user_astaddr = ua->bastaddr;
|
|
|
|
result.user_astparam = ua->bastparam;
|
2010-03-25 20:23:13 +01:00
|
|
|
result.bast_mode = mode;
|
2006-07-12 23:44:04 +02:00
|
|
|
} else {
|
|
|
|
result.user_astaddr = ua->castaddr;
|
|
|
|
result.user_astparam = ua->castparam;
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
if (compat)
|
|
|
|
len = sizeof(struct dlm_lock_result32);
|
|
|
|
else
|
|
|
|
#endif
|
|
|
|
len = sizeof(struct dlm_lock_result);
|
|
|
|
struct_len = len;
|
|
|
|
|
|
|
|
/* copy lvb to userspace if there is one, it's been updated, and
|
|
|
|
the user buffer has space for it */
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (copy_lvb && ua->lksb.sb_lvbptr && count >= len + DLM_USER_LVB_LEN) {
|
2006-07-12 23:44:04 +02:00
|
|
|
if (copy_to_user(buf+len, ua->lksb.sb_lvbptr,
|
|
|
|
DLM_USER_LVB_LEN)) {
|
|
|
|
error = -EFAULT;
|
|
|
|
goto out;
|
|
|
|
}
|
|
|
|
|
|
|
|
result.lvb_offset = len;
|
|
|
|
len += DLM_USER_LVB_LEN;
|
|
|
|
}
|
|
|
|
|
|
|
|
result.length = len;
|
|
|
|
resultptr = &result;
|
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
if (compat) {
|
|
|
|
compat_output(&result, &result32);
|
|
|
|
resultptr = &result32;
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
if (copy_to_user(buf, resultptr, struct_len))
|
|
|
|
error = -EFAULT;
|
|
|
|
else
|
|
|
|
error = len;
|
|
|
|
out:
|
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
2007-05-18 16:00:32 +02:00
|
|
|
static int copy_version_to_user(char __user *buf, size_t count)
|
|
|
|
{
|
|
|
|
struct dlm_device_version ver;
|
|
|
|
|
|
|
|
memset(&ver, 0, sizeof(struct dlm_device_version));
|
|
|
|
ver.version[0] = DLM_DEVICE_VERSION_MAJOR;
|
|
|
|
ver.version[1] = DLM_DEVICE_VERSION_MINOR;
|
|
|
|
ver.version[2] = DLM_DEVICE_VERSION_PATCH;
|
|
|
|
|
|
|
|
if (copy_to_user(buf, &ver, sizeof(struct dlm_device_version)))
|
|
|
|
return -EFAULT;
|
|
|
|
return sizeof(struct dlm_device_version);
|
|
|
|
}
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
/* a read returns a single ast described in a struct dlm_lock_result */
|
|
|
|
|
|
|
|
static ssize_t device_read(struct file *file, char __user *buf, size_t count,
|
|
|
|
loff_t *ppos)
|
|
|
|
{
|
|
|
|
struct dlm_user_proc *proc = file->private_data;
|
|
|
|
struct dlm_lkb *lkb;
|
|
|
|
DECLARE_WAITQUEUE(wait, current);
|
2011-02-21 21:58:21 +01:00
|
|
|
struct dlm_callback cb;
|
|
|
|
int rv, resid, copy_lvb = 0;
|
2015-08-25 19:51:44 +02:00
|
|
|
int old_mode, new_mode;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2007-05-18 16:00:32 +02:00
|
|
|
if (count == sizeof(struct dlm_device_version)) {
|
2011-02-21 21:58:21 +01:00
|
|
|
rv = copy_version_to_user(buf, count);
|
|
|
|
return rv;
|
2007-05-18 16:00:32 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
if (!proc) {
|
|
|
|
log_print("non-version read from control device %zu", count);
|
|
|
|
return -EINVAL;
|
|
|
|
}
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
#ifdef CONFIG_COMPAT
|
|
|
|
if (count < sizeof(struct dlm_lock_result32))
|
|
|
|
#else
|
|
|
|
if (count < sizeof(struct dlm_lock_result))
|
|
|
|
#endif
|
|
|
|
return -EINVAL;
|
|
|
|
|
2010-03-25 20:23:13 +01:00
|
|
|
try_another:
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
/* do we really need this? can a read happen after a close? */
|
|
|
|
if (test_bit(DLM_PROC_FLAGS_CLOSING, &proc->flags))
|
|
|
|
return -EINVAL;
|
|
|
|
|
|
|
|
spin_lock(&proc->asts_spin);
|
|
|
|
if (list_empty(&proc->asts)) {
|
|
|
|
if (file->f_flags & O_NONBLOCK) {
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
return -EAGAIN;
|
|
|
|
}
|
|
|
|
|
|
|
|
add_wait_queue(&proc->wait, &wait);
|
|
|
|
|
|
|
|
repeat:
|
|
|
|
set_current_state(TASK_INTERRUPTIBLE);
|
|
|
|
if (list_empty(&proc->asts) && !signal_pending(current)) {
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
schedule();
|
|
|
|
spin_lock(&proc->asts_spin);
|
|
|
|
goto repeat;
|
|
|
|
}
|
|
|
|
set_current_state(TASK_RUNNING);
|
|
|
|
remove_wait_queue(&proc->wait, &wait);
|
|
|
|
|
|
|
|
if (signal_pending(current)) {
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
return -ERESTARTSYS;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
/* if we empty lkb_callbacks, we don't want to unlock the spinlock
|
2011-04-05 20:16:24 +02:00
|
|
|
without removing lkb_cb_list; so empty lkb_cb_list is always
|
2011-02-21 21:58:21 +01:00
|
|
|
consistent with empty lkb_callbacks */
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2011-04-05 20:16:24 +02:00
|
|
|
lkb = list_entry(proc->asts.next, struct dlm_lkb, lkb_cb_list);
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2015-08-25 19:51:44 +02:00
|
|
|
/* rem_lkb_callback sets a new lkb_last_cast */
|
|
|
|
old_mode = lkb->lkb_last_cast.mode;
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
rv = dlm_rem_lkb_callback(lkb->lkb_resource->res_ls, lkb, &cb, &resid);
|
|
|
|
if (rv < 0) {
|
|
|
|
/* this shouldn't happen; lkb should have been removed from
|
|
|
|
list when resid was zero */
|
|
|
|
log_print("dlm_rem_lkb_callback empty %x", lkb->lkb_id);
|
2011-04-05 20:16:24 +02:00
|
|
|
list_del_init(&lkb->lkb_cb_list);
|
2011-02-21 21:58:21 +01:00
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
/* removes ref for proc->asts, may cause lkb to be freed */
|
|
|
|
dlm_put_lkb(lkb);
|
|
|
|
goto try_another;
|
2010-03-25 20:23:13 +01:00
|
|
|
}
|
2011-02-21 21:58:21 +01:00
|
|
|
if (!resid)
|
2011-04-05 20:16:24 +02:00
|
|
|
list_del_init(&lkb->lkb_cb_list);
|
2011-02-21 21:58:21 +01:00
|
|
|
spin_unlock(&proc->asts_spin);
|
2010-03-25 20:23:13 +01:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (cb.flags & DLM_CB_SKIP) {
|
|
|
|
/* removes ref for proc->asts, may cause lkb to be freed */
|
|
|
|
if (!resid)
|
|
|
|
dlm_put_lkb(lkb);
|
|
|
|
goto try_another;
|
2010-03-25 20:23:13 +01:00
|
|
|
}
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (cb.flags & DLM_CB_CAST) {
|
|
|
|
new_mode = cb.mode;
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
if (!cb.sb_status && lkb->lkb_lksb->sb_lvbptr &&
|
|
|
|
dlm_lvb_operations[old_mode + 1][new_mode + 1])
|
|
|
|
copy_lvb = 1;
|
2010-03-25 20:23:13 +01:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
lkb->lkb_lksb->sb_status = cb.sb_status;
|
|
|
|
lkb->lkb_lksb->sb_flags = cb.sb_flags;
|
2010-03-25 20:23:13 +01:00
|
|
|
}
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
rv = copy_result_to_user(lkb->lkb_ua,
|
|
|
|
test_bit(DLM_PROC_FLAGS_COMPAT, &proc->flags),
|
|
|
|
cb.flags, cb.mode, copy_lvb, buf, count);
|
2010-03-25 20:23:13 +01:00
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
/* removes ref for proc->asts, may cause lkb to be freed */
|
|
|
|
if (!resid)
|
2006-07-12 23:44:04 +02:00
|
|
|
dlm_put_lkb(lkb);
|
|
|
|
|
2011-02-21 21:58:21 +01:00
|
|
|
return rv;
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
|
|
|
|
static unsigned int device_poll(struct file *file, poll_table *wait)
|
|
|
|
{
|
|
|
|
struct dlm_user_proc *proc = file->private_data;
|
|
|
|
|
|
|
|
poll_wait(file, &proc->wait, wait);
|
|
|
|
|
|
|
|
spin_lock(&proc->asts_spin);
|
|
|
|
if (!list_empty(&proc->asts)) {
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
return POLLIN | POLLRDNORM;
|
|
|
|
}
|
|
|
|
spin_unlock(&proc->asts_spin);
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-08-18 18:43:30 +02:00
|
|
|
int dlm_user_daemon_available(void)
|
|
|
|
{
|
|
|
|
/* dlm_controld hasn't started (or, has started, but not
|
|
|
|
properly populated configfs) */
|
|
|
|
|
|
|
|
if (!dlm_our_nodeid())
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
/* This is to deal with versions of dlm_controld that don't
|
|
|
|
know about the monitor device. We assume that if the
|
|
|
|
dlm_controld was started (above), but the monitor device
|
|
|
|
was never opened, that it's an old version. dlm_controld
|
|
|
|
should open the monitor device before populating configfs. */
|
|
|
|
|
|
|
|
if (dlm_monitor_unused)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
return atomic_read(&dlm_monitor_opened) ? 1 : 0;
|
|
|
|
}
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
static int ctl_device_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
file->private_data = NULL;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int ctl_device_close(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2008-08-18 18:43:30 +02:00
|
|
|
static int monitor_device_open(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
atomic_inc(&dlm_monitor_opened);
|
|
|
|
dlm_monitor_unused = 0;
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int monitor_device_close(struct inode *inode, struct file *file)
|
|
|
|
{
|
|
|
|
if (atomic_dec_and_test(&dlm_monitor_opened))
|
|
|
|
dlm_stop_lockspaces();
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2007-02-12 09:55:34 +01:00
|
|
|
static const struct file_operations device_fops = {
|
2006-07-12 23:44:04 +02:00
|
|
|
.open = device_open,
|
|
|
|
.release = device_close,
|
|
|
|
.read = device_read,
|
|
|
|
.write = device_write,
|
|
|
|
.poll = device_poll,
|
|
|
|
.owner = THIS_MODULE,
|
llseek: automatically add .llseek fop
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>
2010-08-15 18:52:59 +02:00
|
|
|
.llseek = noop_llseek,
|
2006-07-12 23:44:04 +02:00
|
|
|
};
|
|
|
|
|
2007-02-12 09:55:34 +01:00
|
|
|
static const struct file_operations ctl_device_fops = {
|
2006-07-12 23:44:04 +02:00
|
|
|
.open = ctl_device_open,
|
|
|
|
.release = ctl_device_close,
|
2007-05-18 16:00:32 +02:00
|
|
|
.read = device_read,
|
2006-07-12 23:44:04 +02:00
|
|
|
.write = device_write,
|
|
|
|
.owner = THIS_MODULE,
|
llseek: automatically add .llseek fop
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>
2010-08-15 18:52:59 +02:00
|
|
|
.llseek = noop_llseek,
|
2006-07-12 23:44:04 +02:00
|
|
|
};
|
|
|
|
|
2008-01-29 06:50:16 +01:00
|
|
|
static struct miscdevice ctl_device = {
|
|
|
|
.name = "dlm-control",
|
|
|
|
.fops = &ctl_device_fops,
|
|
|
|
.minor = MISC_DYNAMIC_MINOR,
|
|
|
|
};
|
|
|
|
|
2008-08-18 18:43:30 +02:00
|
|
|
static const struct file_operations monitor_device_fops = {
|
|
|
|
.open = monitor_device_open,
|
|
|
|
.release = monitor_device_close,
|
|
|
|
.owner = THIS_MODULE,
|
llseek: automatically add .llseek fop
All file_operations should get a .llseek operation so we can make
nonseekable_open the default for future file operations without a
.llseek pointer.
The three cases that we can automatically detect are no_llseek, seq_lseek
and default_llseek. For cases where we can we can automatically prove that
the file offset is always ignored, we use noop_llseek, which maintains
the current behavior of not returning an error from a seek.
New drivers should normally not use noop_llseek but instead use no_llseek
and call nonseekable_open at open time. Existing drivers can be converted
to do the same when the maintainer knows for certain that no user code
relies on calling seek on the device file.
The generated code is often incorrectly indented and right now contains
comments that clarify for each added line why a specific variant was
chosen. In the version that gets submitted upstream, the comments will
be gone and I will manually fix the indentation, because there does not
seem to be a way to do that using coccinelle.
Some amount of new code is currently sitting in linux-next that should get
the same modifications, which I will do at the end of the merge window.
Many thanks to Julia Lawall for helping me learn to write a semantic
patch that does all this.
===== begin semantic patch =====
// This adds an llseek= method to all file operations,
// as a preparation for making no_llseek the default.
//
// The rules are
// - use no_llseek explicitly if we do nonseekable_open
// - use seq_lseek for sequential files
// - use default_llseek if we know we access f_pos
// - use noop_llseek if we know we don't access f_pos,
// but we still want to allow users to call lseek
//
@ open1 exists @
identifier nested_open;
@@
nested_open(...)
{
<+...
nonseekable_open(...)
...+>
}
@ open exists@
identifier open_f;
identifier i, f;
identifier open1.nested_open;
@@
int open_f(struct inode *i, struct file *f)
{
<+...
(
nonseekable_open(...)
|
nested_open(...)
)
...+>
}
@ read disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ read_no_fpos disable optional_qualifier exists @
identifier read_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off)
{
... when != off
}
@ write @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
expression E;
identifier func;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
<+...
(
*off = E
|
*off += E
|
func(..., off, ...)
|
E = *off
)
...+>
}
@ write_no_fpos @
identifier write_f;
identifier f, p, s, off;
type ssize_t, size_t, loff_t;
@@
ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off)
{
... when != off
}
@ fops0 @
identifier fops;
@@
struct file_operations fops = {
...
};
@ has_llseek depends on fops0 @
identifier fops0.fops;
identifier llseek_f;
@@
struct file_operations fops = {
...
.llseek = llseek_f,
...
};
@ has_read depends on fops0 @
identifier fops0.fops;
identifier read_f;
@@
struct file_operations fops = {
...
.read = read_f,
...
};
@ has_write depends on fops0 @
identifier fops0.fops;
identifier write_f;
@@
struct file_operations fops = {
...
.write = write_f,
...
};
@ has_open depends on fops0 @
identifier fops0.fops;
identifier open_f;
@@
struct file_operations fops = {
...
.open = open_f,
...
};
// use no_llseek if we call nonseekable_open
////////////////////////////////////////////
@ nonseekable1 depends on !has_llseek && has_open @
identifier fops0.fops;
identifier nso ~= "nonseekable_open";
@@
struct file_operations fops = {
... .open = nso, ...
+.llseek = no_llseek, /* nonseekable */
};
@ nonseekable2 depends on !has_llseek @
identifier fops0.fops;
identifier open.open_f;
@@
struct file_operations fops = {
... .open = open_f, ...
+.llseek = no_llseek, /* open uses nonseekable */
};
// use seq_lseek for sequential files
/////////////////////////////////////
@ seq depends on !has_llseek @
identifier fops0.fops;
identifier sr ~= "seq_read";
@@
struct file_operations fops = {
... .read = sr, ...
+.llseek = seq_lseek, /* we have seq_read */
};
// use default_llseek if there is a readdir
///////////////////////////////////////////
@ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier readdir_e;
@@
// any other fop is used that changes pos
struct file_operations fops = {
... .readdir = readdir_e, ...
+.llseek = default_llseek, /* readdir is present */
};
// use default_llseek if at least one of read/write touches f_pos
/////////////////////////////////////////////////////////////////
@ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read.read_f;
@@
// read fops use offset
struct file_operations fops = {
... .read = read_f, ...
+.llseek = default_llseek, /* read accesses f_pos */
};
@ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write.write_f;
@@
// write fops use offset
struct file_operations fops = {
... .write = write_f, ...
+ .llseek = default_llseek, /* write accesses f_pos */
};
// Use noop_llseek if neither read nor write accesses f_pos
///////////////////////////////////////////////////////////
@ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
identifier write_no_fpos.write_f;
@@
// write fops use offset
struct file_operations fops = {
...
.write = write_f,
.read = read_f,
...
+.llseek = noop_llseek, /* read and write both use no f_pos */
};
@ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier write_no_fpos.write_f;
@@
struct file_operations fops = {
... .write = write_f, ...
+.llseek = noop_llseek, /* write uses no f_pos */
};
@ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
identifier read_no_fpos.read_f;
@@
struct file_operations fops = {
... .read = read_f, ...
+.llseek = noop_llseek, /* read uses no f_pos */
};
@ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @
identifier fops0.fops;
@@
struct file_operations fops = {
...
+.llseek = noop_llseek, /* no read or write fn */
};
===== End semantic patch =====
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Julia Lawall <julia@diku.dk>
Cc: Christoph Hellwig <hch@infradead.org>
2010-08-15 18:52:59 +02:00
|
|
|
.llseek = noop_llseek,
|
2008-08-18 18:43:30 +02:00
|
|
|
};
|
|
|
|
|
|
|
|
static struct miscdevice monitor_device = {
|
|
|
|
.name = "dlm-monitor",
|
|
|
|
.fops = &monitor_device_fops,
|
|
|
|
.minor = MISC_DYNAMIC_MINOR,
|
|
|
|
};
|
|
|
|
|
2008-02-01 18:53:46 +01:00
|
|
|
int __init dlm_user_init(void)
|
2006-07-12 23:44:04 +02:00
|
|
|
{
|
|
|
|
int error;
|
|
|
|
|
2008-08-18 18:43:30 +02:00
|
|
|
atomic_set(&dlm_monitor_opened, 0);
|
|
|
|
|
2006-07-12 23:44:04 +02:00
|
|
|
error = misc_register(&ctl_device);
|
2008-08-18 18:43:30 +02:00
|
|
|
if (error) {
|
2006-07-12 23:44:04 +02:00
|
|
|
log_print("misc_register failed for control device");
|
2008-08-18 18:43:30 +02:00
|
|
|
goto out;
|
|
|
|
}
|
2006-07-12 23:44:04 +02:00
|
|
|
|
2008-08-18 18:43:30 +02:00
|
|
|
error = misc_register(&monitor_device);
|
|
|
|
if (error) {
|
|
|
|
log_print("misc_register failed for monitor device");
|
|
|
|
misc_deregister(&ctl_device);
|
|
|
|
}
|
|
|
|
out:
|
2006-07-12 23:44:04 +02:00
|
|
|
return error;
|
|
|
|
}
|
|
|
|
|
|
|
|
void dlm_user_exit(void)
|
|
|
|
{
|
|
|
|
misc_deregister(&ctl_device);
|
2008-08-18 18:43:30 +02:00
|
|
|
misc_deregister(&monitor_device);
|
2006-07-12 23:44:04 +02:00
|
|
|
}
|
|
|
|
|