linux/fs/fscache/volume.c

521 lines
15 KiB
C
Raw Normal View History

fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
// SPDX-License-Identifier: GPL-2.0-or-later
/* Volume-level cache cookie handling.
*
* Copyright (C) 2021 Red Hat, Inc. All Rights Reserved.
* Written by David Howells (dhowells@redhat.com)
*/
#define FSCACHE_DEBUG_LEVEL COOKIE
#include <linux/export.h>
#include <linux/slab.h>
#include "internal.h"
#define fscache_volume_hash_shift 10
static struct hlist_bl_head fscache_volume_hash[1 << fscache_volume_hash_shift];
static atomic_t fscache_volume_debug_id;
static LIST_HEAD(fscache_volumes);
static void fscache_create_volume_work(struct work_struct *work);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
struct fscache_volume *fscache_get_volume(struct fscache_volume *volume,
enum fscache_volume_trace where)
{
int ref;
__refcount_inc(&volume->ref, &ref);
trace_fscache_volume(volume->debug_id, ref + 1, where);
return volume;
}
static void fscache_see_volume(struct fscache_volume *volume,
enum fscache_volume_trace where)
{
int ref = refcount_read(&volume->ref);
trace_fscache_volume(volume->debug_id, ref, where);
}
fscache: Implement volume-level access helpers Add a pair of helper functions to manage access to a volume, pinning the volume in place for the duration to prevent cache withdrawal from removing it: bool fscache_begin_volume_access(struct fscache_volume *volume, enum fscache_access_trace why); void fscache_end_volume_access(struct fscache_volume *volume, enum fscache_access_trace why); The way the access gate on the volume works/will work is: (1) If the cache tests as not live (state is not FSCACHE_CACHE_IS_ACTIVE), then we return false to indicate access was not permitted. (2) If the cache tests as live, then we increment the volume's n_accesses count and then recheck the cache liveness, ending the access if it ceased to be live. (3) When we end the access, we decrement the volume's n_accesses and wake up the any waiters if it reaches 0. (4) Whilst the cache is caching, the volume's n_accesses is kept artificially incremented to prevent wakeups from happening. (5) When the cache is taken offline, the state is changed to prevent new accesses, the volume's n_accesses is decremented and we wait for it to become 0. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/163819594158.215744.8285859817391683254.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906894315.143852.5454793807544710479.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967095028.1823006.9173132503876627466.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021501546.640689.9631510472149608443.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
/*
* Pin the cache behind a volume so that we can access it.
*/
static void __fscache_begin_volume_access(struct fscache_volume *volume,
struct fscache_cookie *cookie,
enum fscache_access_trace why)
{
int n_accesses;
n_accesses = atomic_inc_return(&volume->n_accesses);
smp_mb__after_atomic();
trace_fscache_access_volume(volume->debug_id, cookie ? cookie->debug_id : 0,
refcount_read(&volume->ref),
n_accesses, why);
}
/**
* fscache_begin_volume_access - Pin a cache so a volume can be accessed
* @volume: The volume cookie
* @cookie: A datafile cookie for a tracing reference (or NULL)
* @why: An indication of the circumstances of the access for tracing
*
* Attempt to pin the cache to prevent it from going away whilst we're
* accessing a volume and returns true if successful. This works as follows:
*
* (1) If the cache tests as not live (state is not FSCACHE_CACHE_IS_ACTIVE),
* then we return false to indicate access was not permitted.
*
* (2) If the cache tests as live, then we increment the volume's n_accesses
* count and then recheck the cache liveness, ending the access if it
* ceased to be live.
*
* (3) When we end the access, we decrement the volume's n_accesses and wake
* up the any waiters if it reaches 0.
*
* (4) Whilst the cache is caching, the volume's n_accesses is kept
* artificially incremented to prevent wakeups from happening.
*
* (5) When the cache is taken offline, the state is changed to prevent new
* accesses, the volume's n_accesses is decremented and we wait for it to
* become 0.
*
* The datafile @cookie and the @why indicator are merely provided for tracing
* purposes.
*/
bool fscache_begin_volume_access(struct fscache_volume *volume,
struct fscache_cookie *cookie,
enum fscache_access_trace why)
{
if (!fscache_cache_is_live(volume->cache))
return false;
__fscache_begin_volume_access(volume, cookie, why);
if (!fscache_cache_is_live(volume->cache)) {
fscache_end_volume_access(volume, cookie, fscache_access_unlive);
return false;
}
return true;
}
/**
* fscache_end_volume_access - Unpin a cache at the end of an access.
* @volume: The volume cookie
* @cookie: A datafile cookie for a tracing reference (or NULL)
* @why: An indication of the circumstances of the access for tracing
*
* Unpin a cache volume after we've accessed it. The datafile @cookie and the
* @why indicator are merely provided for tracing purposes.
*/
void fscache_end_volume_access(struct fscache_volume *volume,
struct fscache_cookie *cookie,
enum fscache_access_trace why)
{
int n_accesses;
smp_mb__before_atomic();
n_accesses = atomic_dec_return(&volume->n_accesses);
trace_fscache_access_volume(volume->debug_id, cookie ? cookie->debug_id : 0,
refcount_read(&volume->ref),
n_accesses, why);
if (n_accesses == 0)
wake_up_var(&volume->n_accesses);
}
EXPORT_SYMBOL(fscache_end_volume_access);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
static bool fscache_volume_same(const struct fscache_volume *a,
const struct fscache_volume *b)
{
size_t klen;
if (a->key_hash != b->key_hash ||
a->cache != b->cache ||
a->key[0] != b->key[0])
return false;
klen = round_up(a->key[0] + 1, sizeof(__le32));
return memcmp(a->key, b->key, klen) == 0;
}
static bool fscache_is_acquire_pending(struct fscache_volume *volume)
{
return test_bit(FSCACHE_VOLUME_ACQUIRE_PENDING, &volume->flags);
}
static void fscache_wait_on_volume_collision(struct fscache_volume *candidate,
unsigned int collidee_debug_id)
{
fscache: Use wait_on_bit() to wait for the freeing of relinquished volume The freeing of relinquished volume will wake up the pending volume acquisition by using wake_up_bit(), however it is mismatched with wait_var_event() used in fscache_wait_on_volume_collision() and it will never wake up the waiter in the wait-queue because these two functions operate on different wait-queues. According to the implementation in fscache_wait_on_volume_collision(), if the wake-up of pending acquisition is delayed longer than 20 seconds (e.g., due to the delay of on-demand fd closing), the first wait_var_event_timeout() will timeout and the following wait_var_event() will hang forever as shown below: FS-Cache: Potential volume collision new=00000024 old=00000022 ...... INFO: task mount:1148 blocked for more than 122 seconds. Not tainted 6.1.0-rc6+ #1 task:mount state:D stack:0 pid:1148 ppid:1 Call Trace: <TASK> __schedule+0x2f6/0xb80 schedule+0x67/0xe0 fscache_wait_on_volume_collision.cold+0x80/0x82 __fscache_acquire_volume+0x40d/0x4e0 erofs_fscache_register_volume+0x51/0xe0 [erofs] erofs_fscache_register_fs+0x19c/0x240 [erofs] erofs_fc_fill_super+0x746/0xaf0 [erofs] vfs_get_super+0x7d/0x100 get_tree_nodev+0x16/0x20 erofs_fc_get_tree+0x20/0x30 [erofs] vfs_get_tree+0x24/0xb0 path_mount+0x2fa/0xa90 do_mount+0x7c/0xa0 __x64_sys_mount+0x8b/0xe0 do_syscall_64+0x30/0x60 entry_SYSCALL_64_after_hwframe+0x46/0xb0 Considering that wake_up_bit() is more selective, so fix it by using wait_on_bit() instead of wait_var_event() to wait for the freeing of relinquished volume. In addition because waitqueue_active() is used in wake_up_bit() and clear_bit() doesn't imply any memory barrier, use clear_and_wake_up_bit() to add the missing memory barrier between cursor->flags and waitqueue_active(). Fixes: 62ab63352350 ("fscache: Implement volume registration") Reviewed-by: Jingbo Xu <jefflexu@linux.alibaba.com> Signed-off-by: Hou Tao <houtao1@huawei.com> Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Link: https://lore.kernel.org/r/20230113115211.2895845-2-houtao@huaweicloud.com/ # v3
2023-01-13 11:52:10 +00:00
wait_on_bit_timeout(&candidate->flags, FSCACHE_VOLUME_ACQUIRE_PENDING,
TASK_UNINTERRUPTIBLE, 20 * HZ);
if (fscache_is_acquire_pending(candidate)) {
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
pr_notice("Potential volume collision new=%08x old=%08x",
candidate->debug_id, collidee_debug_id);
fscache_stat(&fscache_n_volumes_collision);
fscache: Use wait_on_bit() to wait for the freeing of relinquished volume The freeing of relinquished volume will wake up the pending volume acquisition by using wake_up_bit(), however it is mismatched with wait_var_event() used in fscache_wait_on_volume_collision() and it will never wake up the waiter in the wait-queue because these two functions operate on different wait-queues. According to the implementation in fscache_wait_on_volume_collision(), if the wake-up of pending acquisition is delayed longer than 20 seconds (e.g., due to the delay of on-demand fd closing), the first wait_var_event_timeout() will timeout and the following wait_var_event() will hang forever as shown below: FS-Cache: Potential volume collision new=00000024 old=00000022 ...... INFO: task mount:1148 blocked for more than 122 seconds. Not tainted 6.1.0-rc6+ #1 task:mount state:D stack:0 pid:1148 ppid:1 Call Trace: <TASK> __schedule+0x2f6/0xb80 schedule+0x67/0xe0 fscache_wait_on_volume_collision.cold+0x80/0x82 __fscache_acquire_volume+0x40d/0x4e0 erofs_fscache_register_volume+0x51/0xe0 [erofs] erofs_fscache_register_fs+0x19c/0x240 [erofs] erofs_fc_fill_super+0x746/0xaf0 [erofs] vfs_get_super+0x7d/0x100 get_tree_nodev+0x16/0x20 erofs_fc_get_tree+0x20/0x30 [erofs] vfs_get_tree+0x24/0xb0 path_mount+0x2fa/0xa90 do_mount+0x7c/0xa0 __x64_sys_mount+0x8b/0xe0 do_syscall_64+0x30/0x60 entry_SYSCALL_64_after_hwframe+0x46/0xb0 Considering that wake_up_bit() is more selective, so fix it by using wait_on_bit() instead of wait_var_event() to wait for the freeing of relinquished volume. In addition because waitqueue_active() is used in wake_up_bit() and clear_bit() doesn't imply any memory barrier, use clear_and_wake_up_bit() to add the missing memory barrier between cursor->flags and waitqueue_active(). Fixes: 62ab63352350 ("fscache: Implement volume registration") Reviewed-by: Jingbo Xu <jefflexu@linux.alibaba.com> Signed-off-by: Hou Tao <houtao1@huawei.com> Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Link: https://lore.kernel.org/r/20230113115211.2895845-2-houtao@huaweicloud.com/ # v3
2023-01-13 11:52:10 +00:00
wait_on_bit(&candidate->flags, FSCACHE_VOLUME_ACQUIRE_PENDING,
TASK_UNINTERRUPTIBLE);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
}
}
/*
* Attempt to insert the new volume into the hash. If there's a collision, we
* wait for the old volume to complete if it's being relinquished and an error
* otherwise.
*/
static bool fscache_hash_volume(struct fscache_volume *candidate)
{
struct fscache_volume *cursor;
struct hlist_bl_head *h;
struct hlist_bl_node *p;
unsigned int bucket, collidee_debug_id = 0;
bucket = candidate->key_hash & (ARRAY_SIZE(fscache_volume_hash) - 1);
h = &fscache_volume_hash[bucket];
hlist_bl_lock(h);
hlist_bl_for_each_entry(cursor, p, h, hash_link) {
if (fscache_volume_same(candidate, cursor)) {
if (!test_bit(FSCACHE_VOLUME_RELINQUISHED, &cursor->flags))
goto collision;
fscache_see_volume(cursor, fscache_volume_get_hash_collision);
set_bit(FSCACHE_VOLUME_COLLIDED_WITH, &cursor->flags);
set_bit(FSCACHE_VOLUME_ACQUIRE_PENDING, &candidate->flags);
collidee_debug_id = cursor->debug_id;
break;
}
}
hlist_bl_add_head(&candidate->hash_link, h);
hlist_bl_unlock(h);
if (fscache_is_acquire_pending(candidate))
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
fscache_wait_on_volume_collision(candidate, collidee_debug_id);
return true;
collision:
fscache_see_volume(cursor, fscache_volume_collision);
hlist_bl_unlock(h);
return false;
}
/*
* Allocate and initialise a volume representation cookie.
*/
static struct fscache_volume *fscache_alloc_volume(const char *volume_key,
const char *cache_name,
const void *coherency_data,
size_t coherency_len)
{
struct fscache_volume *volume;
struct fscache_cache *cache;
size_t klen, hlen;
fscache: fix OOB Read in __fscache_acquire_volume The type of a->key[0] is char in fscache_volume_same(). If the length of cache volume key is greater than 127, the value of a->key[0] is less than 0. In this case, klen becomes much larger than 255 after type conversion, because the type of klen is size_t. As a result, memcmp() is read out of bounds. This causes a slab-out-of-bounds Read in __fscache_acquire_volume(), as reported by Syzbot. Fix this by changing the type of the stored key to "u8 *" rather than "char *" (it isn't a simple string anyway). Also put in a check that the volume name doesn't exceed NAME_MAX. BUG: KASAN: slab-out-of-bounds in memcmp+0x16f/0x1c0 lib/string.c:757 Read of size 8 at addr ffff888016f3aa90 by task syz-executor344/3613 Call Trace: memcmp+0x16f/0x1c0 lib/string.c:757 memcmp include/linux/fortify-string.h:420 [inline] fscache_volume_same fs/fscache/volume.c:133 [inline] fscache_hash_volume fs/fscache/volume.c:171 [inline] __fscache_acquire_volume+0x76c/0x1080 fs/fscache/volume.c:328 fscache_acquire_volume include/linux/fscache.h:204 [inline] v9fs_cache_session_get_cookie+0x143/0x240 fs/9p/cache.c:34 v9fs_session_init+0x1166/0x1810 fs/9p/v9fs.c:473 v9fs_mount+0xba/0xc90 fs/9p/vfs_super.c:126 legacy_get_tree+0x105/0x220 fs/fs_context.c:610 vfs_get_tree+0x89/0x2f0 fs/super.c:1530 do_new_mount fs/namespace.c:3040 [inline] path_mount+0x1326/0x1e20 fs/namespace.c:3370 do_mount fs/namespace.c:3383 [inline] __do_sys_mount fs/namespace.c:3591 [inline] __se_sys_mount fs/namespace.c:3568 [inline] __x64_sys_mount+0x27f/0x300 fs/namespace.c:3568 Fixes: 62ab63352350 ("fscache: Implement volume registration") Reported-by: syzbot+a76f6a6e524cf2080aa3@syzkaller.appspotmail.com Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Zhang Peng <zhangpeng362@huawei.com> Reviewed-by: Jingbo Xu <jefflexu@linux.alibaba.com> cc: Dominique Martinet <asmadeus@codewreck.org> cc: Jeff Layton <jlayton@kernel.org> cc: v9fs-developer@lists.sourceforge.net cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/Y3OH+Dmi0QIOK18n@codewreck.org/ # Zhang Peng's v1 fix Link: https://lore.kernel.org/r/20221115140447.2971680-1-zhangpeng362@huawei.com/ # Zhang Peng's v2 fix Link: https://lore.kernel.org/r/166869954095.3793579.8500020902371015443.stgit@warthog.procyon.org.uk/ # v1 Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-11-21 16:31:34 +00:00
u8 *key;
klen = strlen(volume_key);
if (klen > NAME_MAX)
return NULL;
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
if (!coherency_data)
coherency_len = 0;
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
cache = fscache_lookup_cache(cache_name, false);
if (IS_ERR(cache))
return NULL;
volume = kzalloc(struct_size(volume, coherency, coherency_len),
GFP_KERNEL);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
if (!volume)
goto err_cache;
volume->cache = cache;
volume->coherency_len = coherency_len;
if (coherency_data)
memcpy(volume->coherency, coherency_data, coherency_len);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
INIT_LIST_HEAD(&volume->proc_link);
INIT_WORK(&volume->work, fscache_create_volume_work);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
refcount_set(&volume->ref, 1);
spin_lock_init(&volume->lock);
/* Stick the length on the front of the key and pad it out to make
* hashing easier.
*/
hlen = round_up(1 + klen + 1, sizeof(__le32));
key = kzalloc(hlen, GFP_KERNEL);
if (!key)
goto err_vol;
key[0] = klen;
memcpy(key + 1, volume_key, klen);
volume->key = key;
volume->key_hash = fscache_hash(0, key, hlen);
volume->debug_id = atomic_inc_return(&fscache_volume_debug_id);
down_write(&fscache_addremove_sem);
atomic_inc(&cache->n_volumes);
list_add_tail(&volume->proc_link, &fscache_volumes);
fscache_see_volume(volume, fscache_volume_new_acquire);
fscache_stat(&fscache_n_volumes);
up_write(&fscache_addremove_sem);
_leave(" = v=%x", volume->debug_id);
return volume;
err_vol:
kfree(volume);
err_cache:
fscache_put_cache(cache, fscache_cache_put_alloc_volume);
fscache_stat(&fscache_n_volumes_nomem);
return NULL;
}
/*
* Create a volume's representation on disk. Have a volume ref and a cache
* access we have to release.
*/
static void fscache_create_volume_work(struct work_struct *work)
{
const struct fscache_cache_ops *ops;
struct fscache_volume *volume =
container_of(work, struct fscache_volume, work);
fscache_see_volume(volume, fscache_volume_see_create_work);
ops = volume->cache->ops;
if (ops->acquire_volume)
ops->acquire_volume(volume);
fscache_end_cache_access(volume->cache,
fscache_access_acquire_volume_end);
clear_and_wake_up_bit(FSCACHE_VOLUME_CREATING, &volume->flags);
fscache_put_volume(volume, fscache_volume_put_create_work);
}
/*
* Dispatch a worker thread to create a volume's representation on disk.
*/
void fscache_create_volume(struct fscache_volume *volume, bool wait)
{
if (test_and_set_bit(FSCACHE_VOLUME_CREATING, &volume->flags))
goto maybe_wait;
if (volume->cache_priv)
goto no_wait; /* We raced */
if (!fscache_begin_cache_access(volume->cache,
fscache_access_acquire_volume))
goto no_wait;
fscache_get_volume(volume, fscache_volume_get_create_work);
if (!schedule_work(&volume->work))
fscache_put_volume(volume, fscache_volume_put_create_work);
maybe_wait:
if (wait) {
fscache_see_volume(volume, fscache_volume_wait_create_work);
wait_on_bit(&volume->flags, FSCACHE_VOLUME_CREATING,
TASK_UNINTERRUPTIBLE);
}
return;
no_wait:
clear_bit_unlock(FSCACHE_VOLUME_CREATING, &volume->flags);
wake_up_bit(&volume->flags, FSCACHE_VOLUME_CREATING);
}
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
/*
* Acquire a volume representation cookie and link it to a (proposed) cache.
*/
struct fscache_volume *__fscache_acquire_volume(const char *volume_key,
const char *cache_name,
const void *coherency_data,
size_t coherency_len)
{
struct fscache_volume *volume;
volume = fscache_alloc_volume(volume_key, cache_name,
coherency_data, coherency_len);
if (!volume)
return ERR_PTR(-ENOMEM);
if (!fscache_hash_volume(volume)) {
fscache_put_volume(volume, fscache_volume_put_hash_collision);
return ERR_PTR(-EBUSY);
}
fscache_create_volume(volume, false);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
return volume;
}
EXPORT_SYMBOL(__fscache_acquire_volume);
static void fscache_wake_pending_volume(struct fscache_volume *volume,
struct hlist_bl_head *h)
{
struct fscache_volume *cursor;
struct hlist_bl_node *p;
hlist_bl_for_each_entry(cursor, p, h, hash_link) {
if (fscache_volume_same(cursor, volume)) {
fscache_see_volume(cursor, fscache_volume_see_hash_wake);
fscache: Use wait_on_bit() to wait for the freeing of relinquished volume The freeing of relinquished volume will wake up the pending volume acquisition by using wake_up_bit(), however it is mismatched with wait_var_event() used in fscache_wait_on_volume_collision() and it will never wake up the waiter in the wait-queue because these two functions operate on different wait-queues. According to the implementation in fscache_wait_on_volume_collision(), if the wake-up of pending acquisition is delayed longer than 20 seconds (e.g., due to the delay of on-demand fd closing), the first wait_var_event_timeout() will timeout and the following wait_var_event() will hang forever as shown below: FS-Cache: Potential volume collision new=00000024 old=00000022 ...... INFO: task mount:1148 blocked for more than 122 seconds. Not tainted 6.1.0-rc6+ #1 task:mount state:D stack:0 pid:1148 ppid:1 Call Trace: <TASK> __schedule+0x2f6/0xb80 schedule+0x67/0xe0 fscache_wait_on_volume_collision.cold+0x80/0x82 __fscache_acquire_volume+0x40d/0x4e0 erofs_fscache_register_volume+0x51/0xe0 [erofs] erofs_fscache_register_fs+0x19c/0x240 [erofs] erofs_fc_fill_super+0x746/0xaf0 [erofs] vfs_get_super+0x7d/0x100 get_tree_nodev+0x16/0x20 erofs_fc_get_tree+0x20/0x30 [erofs] vfs_get_tree+0x24/0xb0 path_mount+0x2fa/0xa90 do_mount+0x7c/0xa0 __x64_sys_mount+0x8b/0xe0 do_syscall_64+0x30/0x60 entry_SYSCALL_64_after_hwframe+0x46/0xb0 Considering that wake_up_bit() is more selective, so fix it by using wait_on_bit() instead of wait_var_event() to wait for the freeing of relinquished volume. In addition because waitqueue_active() is used in wake_up_bit() and clear_bit() doesn't imply any memory barrier, use clear_and_wake_up_bit() to add the missing memory barrier between cursor->flags and waitqueue_active(). Fixes: 62ab63352350 ("fscache: Implement volume registration") Reviewed-by: Jingbo Xu <jefflexu@linux.alibaba.com> Signed-off-by: Hou Tao <houtao1@huawei.com> Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Link: https://lore.kernel.org/r/20230113115211.2895845-2-houtao@huaweicloud.com/ # v3
2023-01-13 11:52:10 +00:00
clear_and_wake_up_bit(FSCACHE_VOLUME_ACQUIRE_PENDING,
&cursor->flags);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
return;
}
}
}
/*
* Remove a volume cookie from the hash table.
*/
static void fscache_unhash_volume(struct fscache_volume *volume)
{
struct hlist_bl_head *h;
unsigned int bucket;
bucket = volume->key_hash & (ARRAY_SIZE(fscache_volume_hash) - 1);
h = &fscache_volume_hash[bucket];
hlist_bl_lock(h);
hlist_bl_del(&volume->hash_link);
if (test_bit(FSCACHE_VOLUME_COLLIDED_WITH, &volume->flags))
fscache_wake_pending_volume(volume, h);
hlist_bl_unlock(h);
}
/*
* Drop a cache's volume attachments.
*/
static void fscache_free_volume(struct fscache_volume *volume)
{
struct fscache_cache *cache = volume->cache;
if (volume->cache_priv) {
__fscache_begin_volume_access(volume, NULL,
fscache_access_relinquish_volume);
if (volume->cache_priv)
cache->ops->free_volume(volume);
fscache_end_volume_access(volume, NULL,
fscache_access_relinquish_volume_end);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
}
down_write(&fscache_addremove_sem);
list_del_init(&volume->proc_link);
atomic_dec(&volume->cache->n_volumes);
up_write(&fscache_addremove_sem);
if (!hlist_bl_unhashed(&volume->hash_link))
fscache_unhash_volume(volume);
trace_fscache_volume(volume->debug_id, 0, fscache_volume_free);
kfree(volume->key);
kfree(volume);
fscache_stat_d(&fscache_n_volumes);
fscache_put_cache(cache, fscache_cache_put_volume);
}
/*
* Drop a reference to a volume cookie.
*/
void fscache_put_volume(struct fscache_volume *volume,
enum fscache_volume_trace where)
{
if (volume) {
unsigned int debug_id = volume->debug_id;
bool zero;
int ref;
zero = __refcount_dec_and_test(&volume->ref, &ref);
trace_fscache_volume(debug_id, ref - 1, where);
if (zero)
fscache_free_volume(volume);
}
}
/*
* Relinquish a volume representation cookie.
*/
void __fscache_relinquish_volume(struct fscache_volume *volume,
const void *coherency_data,
bool invalidate)
{
if (WARN_ON(test_and_set_bit(FSCACHE_VOLUME_RELINQUISHED, &volume->flags)))
return;
if (invalidate) {
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
set_bit(FSCACHE_VOLUME_INVALIDATE, &volume->flags);
} else if (coherency_data) {
memcpy(volume->coherency, coherency_data, volume->coherency_len);
}
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
fscache_put_volume(volume, fscache_volume_put_relinquish);
}
EXPORT_SYMBOL(__fscache_relinquish_volume);
/**
* fscache_withdraw_volume - Withdraw a volume from being cached
* @volume: Volume cookie
*
* Withdraw a cache volume from service, waiting for all accesses to complete
* before returning.
*/
void fscache_withdraw_volume(struct fscache_volume *volume)
{
int n_accesses;
_debug("withdraw V=%x", volume->debug_id);
/* Allow wakeups on dec-to-0 */
n_accesses = atomic_dec_return(&volume->n_accesses);
trace_fscache_access_volume(volume->debug_id, 0,
refcount_read(&volume->ref),
n_accesses, fscache_access_cache_unpin);
wait_var_event(&volume->n_accesses,
atomic_read(&volume->n_accesses) == 0);
}
EXPORT_SYMBOL(fscache_withdraw_volume);
fscache: Implement volume registration Add functions to the fscache API to allow volumes to be acquired and relinquished by the network filesystem. A volume is an index of data storage cache objects. A volume is represented by a volume cookie in the API. A filesystem would typically create a volume for a superblock and then create per-inode cookies within it. To request a volume, the filesystem calls: struct fscache_volume * fscache_acquire_volume(const char *volume_key, const char *cache_name, const void *coherency_data, size_t coherency_len) The volume_key is a printable string used to match the volume in the cache. It should not contain any '/' characters. For AFS, for example, this would be "afs,<cellname>,<volume_id>", e.g. "afs,example.com,523001". The cache_name can be NULL, but if not it should be a string indicating the name of the cache to use if there's more than one available. The coherency data, if given, is an arbitrarily-sized blob that's attached to the volume and is compared when the volume is looked up. If it doesn't match, the old volume is judged to be out of date and it and everything within it is discarded. Acquiring a volume twice concurrently is disallowed, though the function will wait if an old volume cookie is being relinquishing. When a network filesystem has finished with a volume, it should return the volume cookie by calling: void fscache_relinquish_volume(struct fscache_volume *volume, const void *coherency_data, bool invalidate) If invalidate is true, the entire volume will be discarded; if false, the volume will be synced and the coherency data will be updated. Changes ======= ver #4: - Removed an extraneous param from kdoc on fscache_relinquish_volume()[3]. ver #3: - fscache_hash()'s size parameter is now in bytes. Use __le32 as the unit to round up to. - When comparing cookies, simply see if the attributes are the same rather than subtracting them to produce a strcmp-style return[2]. - Make the coherency data an arbitrary blob rather than a u64, but don't store it for the moment. ver #2: - Fix error check[1]. - Make a fscache_acquire_volume() return errors, including EBUSY if a conflicting volume cookie already exists. No error is printed now - that's left to the netfs. Signed-off-by: David Howells <dhowells@redhat.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/20211203095608.GC2480@kili/ [1] Link: https://lore.kernel.org/r/CAHk-=whtkzB446+hX0zdLsdcUJsJ=8_-0S1mE_R+YurThfUbLA@mail.gmail.com/ [2] Link: https://lore.kernel.org/r/20211220224646.30e8205c@canb.auug.org.au/ [3] Link: https://lore.kernel.org/r/163819588944.215744.1629085755564865996.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906890630.143852.13972180614535611154.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967086836.1823006.8191672796841981763.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021495816.640689.4403156093668590217.stgit@warthog.procyon.org.uk/ # v4
2021-10-20 14:26:17 +00:00
#ifdef CONFIG_PROC_FS
/*
* Generate a list of volumes in /proc/fs/fscache/volumes
*/
static int fscache_volumes_seq_show(struct seq_file *m, void *v)
{
struct fscache_volume *volume;
if (v == &fscache_volumes) {
seq_puts(m,
"VOLUME REF nCOOK ACC FL CACHE KEY\n"
"======== ===== ===== === == =============== ================\n");
return 0;
}
volume = list_entry(v, struct fscache_volume, proc_link);
seq_printf(m,
"%08x %5d %5d %3d %02lx %-15.15s %s\n",
volume->debug_id,
refcount_read(&volume->ref),
atomic_read(&volume->n_cookies),
atomic_read(&volume->n_accesses),
volume->flags,
volume->cache->name ?: "-",
volume->key + 1);
return 0;
}
static void *fscache_volumes_seq_start(struct seq_file *m, loff_t *_pos)
__acquires(&fscache_addremove_sem)
{
down_read(&fscache_addremove_sem);
return seq_list_start_head(&fscache_volumes, *_pos);
}
static void *fscache_volumes_seq_next(struct seq_file *m, void *v, loff_t *_pos)
{
return seq_list_next(v, &fscache_volumes, _pos);
}
static void fscache_volumes_seq_stop(struct seq_file *m, void *v)
__releases(&fscache_addremove_sem)
{
up_read(&fscache_addremove_sem);
}
const struct seq_operations fscache_volumes_seq_ops = {
.start = fscache_volumes_seq_start,
.next = fscache_volumes_seq_next,
.stop = fscache_volumes_seq_stop,
.show = fscache_volumes_seq_show,
};
#endif /* CONFIG_PROC_FS */