linux_dsm_epyc7002/fs/fscache/object-list.c
David Howells 1362729b16 FS-Cache: Simplify cookie retention for fscache_objects, fixing oops
Simplify the way fscache cache objects retain their cookie.  The way I
implemented the cookie storage handling made synchronisation a pain (ie. the
object state machine can't rely on the cookie actually still being there).

Instead of the the object being detached from the cookie and the cookie being
freed in __fscache_relinquish_cookie(), we defer both operations:

 (*) The detachment of the object from the list in the cookie now takes place
     in fscache_drop_object() and is thus governed by the object state machine
     (fscache_detach_from_cookie() has been removed).

 (*) The release of the cookie is now in fscache_object_destroy() - which is
     called by the cache backend just before it frees the object.

This means that the fscache_cookie struct is now available to the cache all the
way through from ->alloc_object() to ->drop_object() and ->put_object() -
meaning that it's no longer necessary to take object->lock to guarantee access.

However, __fscache_relinquish_cookie() doesn't wait for the object to go all
the way through to destruction before letting the netfs proceed.  That would
massively slow down the netfs.  Since __fscache_relinquish_cookie() leaves the
cookie around, in must therefore break all attachments to the netfs - which
includes ->def, ->netfs_data and any outstanding page read/writes.

To handle this, struct fscache_cookie now has an n_active counter:

 (1) This starts off initialised to 1.

 (2) Any time the cache needs to get at the netfs data, it calls
     fscache_use_cookie() to increment it - if it is not zero.  If it was zero,
     then access is not permitted.

 (3) When the cache has finished with the data, it calls fscache_unuse_cookie()
     to decrement it.  This does a wake-up on it if it reaches 0.

 (4) __fscache_relinquish_cookie() decrements n_active and then waits for it to
     reach 0.  The initialisation to 1 in step (1) ensures that we only get
     wake ups when we're trying to get rid of the cookie.

This leaves __fscache_relinquish_cookie() a lot simpler.


***
This fixes a problem in the current code whereby if fscache_invalidate() is
followed sufficiently quickly by fscache_relinquish_cookie() then it is
possible for __fscache_relinquish_cookie() to have detached the cookie from the
object and cleared the pointer before a thread is dispatched to process the
invalidation state in the object state machine.

Since the pending write clearance was deferred to the invalidation state to
make it asynchronous, we need to either wait in relinquishment for the stores
tree to be cleared in the invalidation state or we need to handle the clearance
in relinquishment.

Further, if the relinquishment code does clear the tree, then the invalidation
state need to make the clearance contingent on still having the cookie to hand
(since that's where the tree is rooted) and we have to prevent the cookie from
disappearing for the duration.

This can lead to an oops like the following:

BUG: unable to handle kernel NULL pointer dereference at 000000000000000c
...
RIP: 0010:[<ffffffff8151023e>] _spin_lock+0xe/0x30
...
CR2: 000000000000000c ...
...
Process kslowd002 (...)
....
Call Trace:
 [<ffffffffa01c3278>] fscache_invalidate_writes+0x38/0xd0 [fscache]
 [<ffffffff810096f0>] ? __switch_to+0xd0/0x320
 [<ffffffff8105e759>] ? find_busiest_queue+0x69/0x150
 [<ffffffff8110ddd4>] ? slow_work_enqueue+0x104/0x180
 [<ffffffffa01c1303>] fscache_object_slow_work_execute+0x5e3/0x9d0 [fscache]
 [<ffffffff81096b67>] ? bit_waitqueue+0x17/0xd0
 [<ffffffff8110e233>] slow_work_execute+0x233/0x310
 [<ffffffff8110e515>] slow_work_thread+0x205/0x360
 [<ffffffff81096ca0>] ? autoremove_wake_function+0x0/0x40
 [<ffffffff8110e310>] ? slow_work_thread+0x0/0x360
 [<ffffffff81096936>] kthread+0x96/0xa0
 [<ffffffff8100c0ca>] child_rip+0xa/0x20
 [<ffffffff810968a0>] ? kthread+0x0/0xa0
 [<ffffffff8100c0c0>] ? child_rip+0x0/0x20

The parameter to fscache_invalidate_writes() was object->cookie which is NULL.

Signed-off-by: David Howells <dhowells@redhat.com>
Tested-By: Milosz Tanski <milosz@adfin.com>
Acked-by: Jeff Layton <jlayton@redhat.com>
2013-06-19 14:16:47 +01:00

420 lines
12 KiB
C

/* Global fscache object list maintainer and viewer
*
* Copyright (C) 2009 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public Licence
* as published by the Free Software Foundation; either version
* 2 of the Licence, or (at your option) any later version.
*/
#define FSCACHE_DEBUG_LEVEL COOKIE
#include <linux/module.h>
#include <linux/seq_file.h>
#include <linux/slab.h>
#include <linux/key.h>
#include <keys/user-type.h>
#include "internal.h"
static struct rb_root fscache_object_list;
static DEFINE_RWLOCK(fscache_object_list_lock);
struct fscache_objlist_data {
unsigned long config; /* display configuration */
#define FSCACHE_OBJLIST_CONFIG_KEY 0x00000001 /* show object keys */
#define FSCACHE_OBJLIST_CONFIG_AUX 0x00000002 /* show object auxdata */
#define FSCACHE_OBJLIST_CONFIG_COOKIE 0x00000004 /* show objects with cookies */
#define FSCACHE_OBJLIST_CONFIG_NOCOOKIE 0x00000008 /* show objects without cookies */
#define FSCACHE_OBJLIST_CONFIG_BUSY 0x00000010 /* show busy objects */
#define FSCACHE_OBJLIST_CONFIG_IDLE 0x00000020 /* show idle objects */
#define FSCACHE_OBJLIST_CONFIG_PENDWR 0x00000040 /* show objects with pending writes */
#define FSCACHE_OBJLIST_CONFIG_NOPENDWR 0x00000080 /* show objects without pending writes */
#define FSCACHE_OBJLIST_CONFIG_READS 0x00000100 /* show objects with active reads */
#define FSCACHE_OBJLIST_CONFIG_NOREADS 0x00000200 /* show objects without active reads */
#define FSCACHE_OBJLIST_CONFIG_EVENTS 0x00000400 /* show objects with events */
#define FSCACHE_OBJLIST_CONFIG_NOEVENTS 0x00000800 /* show objects without no events */
#define FSCACHE_OBJLIST_CONFIG_WORK 0x00001000 /* show objects with work */
#define FSCACHE_OBJLIST_CONFIG_NOWORK 0x00002000 /* show objects without work */
u8 buf[512]; /* key and aux data buffer */
};
/*
* Add an object to the object list
* - we use the address of the fscache_object structure as the key into the
* tree
*/
void fscache_objlist_add(struct fscache_object *obj)
{
struct fscache_object *xobj;
struct rb_node **p = &fscache_object_list.rb_node, *parent = NULL;
write_lock(&fscache_object_list_lock);
while (*p) {
parent = *p;
xobj = rb_entry(parent, struct fscache_object, objlist_link);
if (obj < xobj)
p = &(*p)->rb_left;
else if (obj > xobj)
p = &(*p)->rb_right;
else
BUG();
}
rb_link_node(&obj->objlist_link, parent, p);
rb_insert_color(&obj->objlist_link, &fscache_object_list);
write_unlock(&fscache_object_list_lock);
}
/*
* Remove an object from the object list.
*/
void fscache_objlist_remove(struct fscache_object *obj)
{
write_lock(&fscache_object_list_lock);
BUG_ON(RB_EMPTY_ROOT(&fscache_object_list));
rb_erase(&obj->objlist_link, &fscache_object_list);
write_unlock(&fscache_object_list_lock);
}
/*
* find the object in the tree on or after the specified index
*/
static struct fscache_object *fscache_objlist_lookup(loff_t *_pos)
{
struct fscache_object *pobj, *obj = NULL, *minobj = NULL;
struct rb_node *p;
unsigned long pos;
if (*_pos >= (unsigned long) ERR_PTR(-ENOENT))
return NULL;
pos = *_pos;
/* banners (can't represent line 0 by pos 0 as that would involve
* returning a NULL pointer) */
if (pos == 0)
return (struct fscache_object *)(long)++(*_pos);
if (pos < 3)
return (struct fscache_object *)pos;
pobj = (struct fscache_object *)pos;
p = fscache_object_list.rb_node;
while (p) {
obj = rb_entry(p, struct fscache_object, objlist_link);
if (pobj < obj) {
if (!minobj || minobj > obj)
minobj = obj;
p = p->rb_left;
} else if (pobj > obj) {
p = p->rb_right;
} else {
minobj = obj;
break;
}
obj = NULL;
}
if (!minobj)
*_pos = (unsigned long) ERR_PTR(-ENOENT);
else if (minobj != obj)
*_pos = (unsigned long) minobj;
return minobj;
}
/*
* set up the iterator to start reading from the first line
*/
static void *fscache_objlist_start(struct seq_file *m, loff_t *_pos)
__acquires(&fscache_object_list_lock)
{
read_lock(&fscache_object_list_lock);
return fscache_objlist_lookup(_pos);
}
/*
* move to the next line
*/
static void *fscache_objlist_next(struct seq_file *m, void *v, loff_t *_pos)
{
(*_pos)++;
return fscache_objlist_lookup(_pos);
}
/*
* clean up after reading
*/
static void fscache_objlist_stop(struct seq_file *m, void *v)
__releases(&fscache_object_list_lock)
{
read_unlock(&fscache_object_list_lock);
}
/*
* display an object
*/
static int fscache_objlist_show(struct seq_file *m, void *v)
{
struct fscache_objlist_data *data = m->private;
struct fscache_object *obj = v;
struct fscache_cookie *cookie;
unsigned long config = data->config;
char _type[3], *type;
u8 *buf = data->buf, *p;
if ((unsigned long) v == 1) {
seq_puts(m, "OBJECT PARENT STAT CHLDN OPS OOP IPR EX READS"
" EM EV FL S"
" | NETFS_COOKIE_DEF TY FL NETFS_DATA");
if (config & (FSCACHE_OBJLIST_CONFIG_KEY |
FSCACHE_OBJLIST_CONFIG_AUX))
seq_puts(m, " ");
if (config & FSCACHE_OBJLIST_CONFIG_KEY)
seq_puts(m, "OBJECT_KEY");
if ((config & (FSCACHE_OBJLIST_CONFIG_KEY |
FSCACHE_OBJLIST_CONFIG_AUX)) ==
(FSCACHE_OBJLIST_CONFIG_KEY | FSCACHE_OBJLIST_CONFIG_AUX))
seq_puts(m, ", ");
if (config & FSCACHE_OBJLIST_CONFIG_AUX)
seq_puts(m, "AUX_DATA");
seq_puts(m, "\n");
return 0;
}
if ((unsigned long) v == 2) {
seq_puts(m, "======== ======== ==== ===== === === === == ====="
" == == == ="
" | ================ == == ================");
if (config & (FSCACHE_OBJLIST_CONFIG_KEY |
FSCACHE_OBJLIST_CONFIG_AUX))
seq_puts(m, " ================");
seq_puts(m, "\n");
return 0;
}
/* filter out any unwanted objects */
#define FILTER(criterion, _yes, _no) \
do { \
unsigned long yes = FSCACHE_OBJLIST_CONFIG_##_yes; \
unsigned long no = FSCACHE_OBJLIST_CONFIG_##_no; \
if (criterion) { \
if (!(config & yes)) \
return 0; \
} else { \
if (!(config & no)) \
return 0; \
} \
} while(0)
cookie = obj->cookie;
if (~config) {
FILTER(cookie->def,
COOKIE, NOCOOKIE);
FILTER(fscache_object_is_active(obj) ||
obj->n_ops != 0 ||
obj->n_obj_ops != 0 ||
obj->flags ||
!list_empty(&obj->dependents),
BUSY, IDLE);
FILTER(test_bit(FSCACHE_OBJECT_PENDING_WRITE, &obj->flags),
PENDWR, NOPENDWR);
FILTER(atomic_read(&obj->n_reads),
READS, NOREADS);
FILTER(obj->events & obj->event_mask,
EVENTS, NOEVENTS);
FILTER(work_busy(&obj->work), WORK, NOWORK);
}
seq_printf(m,
"%8x %8x %s %5u %3u %3u %3u %2u %5u %2lx %2lx %2lx %1x | ",
obj->debug_id,
obj->parent ? obj->parent->debug_id : -1,
obj->state->short_name,
obj->n_children,
obj->n_ops,
obj->n_obj_ops,
obj->n_in_progress,
obj->n_exclusive,
atomic_read(&obj->n_reads),
obj->event_mask,
obj->events,
obj->flags,
work_busy(&obj->work));
if (fscache_use_cookie(obj)) {
uint16_t keylen = 0, auxlen = 0;
switch (cookie->def->type) {
case 0:
type = "IX";
break;
case 1:
type = "DT";
break;
default:
sprintf(_type, "%02u", cookie->def->type);
type = _type;
break;
}
seq_printf(m, "%-16s %s %2lx %16p",
cookie->def->name,
type,
cookie->flags,
cookie->netfs_data);
if (cookie->def->get_key &&
config & FSCACHE_OBJLIST_CONFIG_KEY)
keylen = cookie->def->get_key(cookie->netfs_data,
buf, 400);
if (cookie->def->get_aux &&
config & FSCACHE_OBJLIST_CONFIG_AUX)
auxlen = cookie->def->get_aux(cookie->netfs_data,
buf + keylen, 512 - keylen);
fscache_unuse_cookie(obj);
if (keylen > 0 || auxlen > 0) {
seq_printf(m, " ");
for (p = buf; keylen > 0; keylen--)
seq_printf(m, "%02x", *p++);
if (auxlen > 0) {
if (config & FSCACHE_OBJLIST_CONFIG_KEY)
seq_printf(m, ", ");
for (; auxlen > 0; auxlen--)
seq_printf(m, "%02x", *p++);
}
}
seq_printf(m, "\n");
} else {
seq_printf(m, "<no_netfs>\n");
}
return 0;
}
static const struct seq_operations fscache_objlist_ops = {
.start = fscache_objlist_start,
.stop = fscache_objlist_stop,
.next = fscache_objlist_next,
.show = fscache_objlist_show,
};
/*
* get the configuration for filtering the list
*/
static void fscache_objlist_config(struct fscache_objlist_data *data)
{
#ifdef CONFIG_KEYS
struct user_key_payload *confkey;
unsigned long config;
struct key *key;
const char *buf;
int len;
key = request_key(&key_type_user, "fscache:objlist", NULL);
if (IS_ERR(key))
goto no_config;
config = 0;
rcu_read_lock();
confkey = key->payload.data;
buf = confkey->data;
for (len = confkey->datalen - 1; len >= 0; len--) {
switch (buf[len]) {
case 'K': config |= FSCACHE_OBJLIST_CONFIG_KEY; break;
case 'A': config |= FSCACHE_OBJLIST_CONFIG_AUX; break;
case 'C': config |= FSCACHE_OBJLIST_CONFIG_COOKIE; break;
case 'c': config |= FSCACHE_OBJLIST_CONFIG_NOCOOKIE; break;
case 'B': config |= FSCACHE_OBJLIST_CONFIG_BUSY; break;
case 'b': config |= FSCACHE_OBJLIST_CONFIG_IDLE; break;
case 'W': config |= FSCACHE_OBJLIST_CONFIG_PENDWR; break;
case 'w': config |= FSCACHE_OBJLIST_CONFIG_NOPENDWR; break;
case 'R': config |= FSCACHE_OBJLIST_CONFIG_READS; break;
case 'r': config |= FSCACHE_OBJLIST_CONFIG_NOREADS; break;
case 'S': config |= FSCACHE_OBJLIST_CONFIG_WORK; break;
case 's': config |= FSCACHE_OBJLIST_CONFIG_NOWORK; break;
}
}
rcu_read_unlock();
key_put(key);
if (!(config & (FSCACHE_OBJLIST_CONFIG_COOKIE | FSCACHE_OBJLIST_CONFIG_NOCOOKIE)))
config |= FSCACHE_OBJLIST_CONFIG_COOKIE | FSCACHE_OBJLIST_CONFIG_NOCOOKIE;
if (!(config & (FSCACHE_OBJLIST_CONFIG_BUSY | FSCACHE_OBJLIST_CONFIG_IDLE)))
config |= FSCACHE_OBJLIST_CONFIG_BUSY | FSCACHE_OBJLIST_CONFIG_IDLE;
if (!(config & (FSCACHE_OBJLIST_CONFIG_PENDWR | FSCACHE_OBJLIST_CONFIG_NOPENDWR)))
config |= FSCACHE_OBJLIST_CONFIG_PENDWR | FSCACHE_OBJLIST_CONFIG_NOPENDWR;
if (!(config & (FSCACHE_OBJLIST_CONFIG_READS | FSCACHE_OBJLIST_CONFIG_NOREADS)))
config |= FSCACHE_OBJLIST_CONFIG_READS | FSCACHE_OBJLIST_CONFIG_NOREADS;
if (!(config & (FSCACHE_OBJLIST_CONFIG_EVENTS | FSCACHE_OBJLIST_CONFIG_NOEVENTS)))
config |= FSCACHE_OBJLIST_CONFIG_EVENTS | FSCACHE_OBJLIST_CONFIG_NOEVENTS;
if (!(config & (FSCACHE_OBJLIST_CONFIG_WORK | FSCACHE_OBJLIST_CONFIG_NOWORK)))
config |= FSCACHE_OBJLIST_CONFIG_WORK | FSCACHE_OBJLIST_CONFIG_NOWORK;
data->config = config;
return;
no_config:
#endif
data->config = ULONG_MAX;
}
/*
* open "/proc/fs/fscache/objects" to provide a list of active objects
* - can be configured by a user-defined key added to the caller's keyrings
*/
static int fscache_objlist_open(struct inode *inode, struct file *file)
{
struct fscache_objlist_data *data;
struct seq_file *m;
int ret;
ret = seq_open(file, &fscache_objlist_ops);
if (ret < 0)
return ret;
m = file->private_data;
/* buffer for key extraction */
data = kmalloc(sizeof(struct fscache_objlist_data), GFP_KERNEL);
if (!data) {
seq_release(inode, file);
return -ENOMEM;
}
/* get the configuration key */
fscache_objlist_config(data);
m->private = data;
return 0;
}
/*
* clean up on close
*/
static int fscache_objlist_release(struct inode *inode, struct file *file)
{
struct seq_file *m = file->private_data;
kfree(m->private);
m->private = NULL;
return seq_release(inode, file);
}
const struct file_operations fscache_objlist_fops = {
.owner = THIS_MODULE,
.open = fscache_objlist_open,
.read = seq_read,
.llseek = seq_lseek,
.release = fscache_objlist_release,
};