cdb897e327
* a large series of fixes and improvements to the snapshot-handling code (Zheng Yan) * individual read/write OSD requests passed down to libceph are now limited to 16M in size to avoid hitting OSD-side limits (Zheng Yan) * encode MStatfs v2 message to allow for more accurate space usage reporting (Douglas Fuller) * switch to the new writeback error tracking infrastructure (Jeff Layton) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAABCAAGBQJZuAC0AAoJEEp/3jgCEfOLb14H/REYq4fDDkUa70L4leKWWdCa n71ipkKeoorfivts71iOtGMJfK+Z6ax+dq1PvBWMy6PtzXS/+2B+t2XwILvLiwWH h87i44bY68aLWRTSusgTfB+I7gyVrWN0WMLznZ5rfM9XuyPv+RPyJYh3EhxWI5+U 2kOHFEc+cPL6mAshGmB8lIzKOWTfmBiw28ulICwlcazm79hh39aNBQE546lS8gA3 kXuJ55odojPgXOYh+vs60raIBnm6flek1jLxBGYG3MU4gv0VVWOyW0eWeuqW+EcR 6dVYlzg1xGlPp+vRmDZQuv/E2MafBxdcil/RrdLeqcx/Hf1KJBzcLgUzIMbnOAI= =YDZP -----END PGP SIGNATURE----- Merge tag 'ceph-for-4.14-rc1' of git://github.com/ceph/ceph-client Pull ceph updates from Ilya Dryomov: "The highlights include: - a large series of fixes and improvements to the snapshot-handling code (Zheng Yan) - individual read/write OSD requests passed down to libceph are now limited to 16M in size to avoid hitting OSD-side limits (Zheng Yan) - encode MStatfs v2 message to allow for more accurate space usage reporting (Douglas Fuller) - switch to the new writeback error tracking infrastructure (Jeff Layton)" * tag 'ceph-for-4.14-rc1' of git://github.com/ceph/ceph-client: (35 commits) ceph: stop on-going cached readdir if mds revokes FILE_SHARED cap ceph: wait on writeback after writing snapshot data ceph: fix capsnap dirty pages accounting ceph: ignore wbc->range_{start,end} when write back snapshot data ceph: fix "range cyclic" mode writepages ceph: cleanup local variables in ceph_writepages_start() ceph: optimize pagevec iterating in ceph_writepages_start() ceph: make writepage_nounlock() invalidate page that beyonds EOF ceph: properly get capsnap's size in get_oldest_context() ceph: remove stale check in ceph_invalidatepage() ceph: queue cap snap only when snap realm's context changes ceph: handle race between vmtruncate and queuing cap snap ceph: fix message order check in handle_cap_export() ceph: fix NULL pointer dereference in ceph_flush_snaps() ceph: adjust 36 checks for NULL pointers ceph: delete an unnecessary return statement in update_dentry_lease() ceph: ENOMEM pr_err in __get_or_create_frag() is redundant ceph: check negative offsets in ceph_llseek() ceph: more accurate statfs ceph: properly set snap follows for cap reconnect ...
412 lines
10 KiB
C
412 lines
10 KiB
C
/*
|
|
* Ceph cache definitions.
|
|
*
|
|
* Copyright (C) 2013 by Adfin Solutions, Inc. All Rights Reserved.
|
|
* Written by Milosz Tanski (milosz@adfin.com)
|
|
*
|
|
* This program is free software; you can redistribute it and/or modify
|
|
* it under the terms of the GNU General Public License version 2
|
|
* as published by the Free Software Foundation.
|
|
*
|
|
* This program is distributed in the hope that it will be useful,
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
* GNU General Public License for more details.
|
|
*
|
|
* You should have received a copy of the GNU General Public License
|
|
* along with this program; if not, write to:
|
|
* Free Software Foundation
|
|
* 51 Franklin Street, Fifth Floor
|
|
* Boston, MA 02111-1301 USA
|
|
*
|
|
*/
|
|
|
|
#include "super.h"
|
|
#include "cache.h"
|
|
|
|
struct ceph_aux_inode {
|
|
u64 version;
|
|
struct timespec mtime;
|
|
loff_t size;
|
|
};
|
|
|
|
struct fscache_netfs ceph_cache_netfs = {
|
|
.name = "ceph",
|
|
.version = 0,
|
|
};
|
|
|
|
static DEFINE_MUTEX(ceph_fscache_lock);
|
|
static LIST_HEAD(ceph_fscache_list);
|
|
|
|
struct ceph_fscache_entry {
|
|
struct list_head list;
|
|
struct fscache_cookie *fscache;
|
|
struct ceph_fsid fsid;
|
|
size_t uniq_len;
|
|
char uniquifier[0];
|
|
};
|
|
|
|
static uint16_t ceph_fscache_session_get_key(const void *cookie_netfs_data,
|
|
void *buffer, uint16_t maxbuf)
|
|
{
|
|
const struct ceph_fs_client* fsc = cookie_netfs_data;
|
|
const char *fscache_uniq = fsc->mount_options->fscache_uniq;
|
|
uint16_t fsid_len, uniq_len;
|
|
|
|
fsid_len = sizeof(fsc->client->fsid);
|
|
uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0;
|
|
if (fsid_len + uniq_len > maxbuf)
|
|
return 0;
|
|
|
|
memcpy(buffer, &fsc->client->fsid, fsid_len);
|
|
if (uniq_len)
|
|
memcpy(buffer + fsid_len, fscache_uniq, uniq_len);
|
|
|
|
return fsid_len + uniq_len;
|
|
}
|
|
|
|
static const struct fscache_cookie_def ceph_fscache_fsid_object_def = {
|
|
.name = "CEPH.fsid",
|
|
.type = FSCACHE_COOKIE_TYPE_INDEX,
|
|
.get_key = ceph_fscache_session_get_key,
|
|
};
|
|
|
|
int ceph_fscache_register(void)
|
|
{
|
|
return fscache_register_netfs(&ceph_cache_netfs);
|
|
}
|
|
|
|
void ceph_fscache_unregister(void)
|
|
{
|
|
fscache_unregister_netfs(&ceph_cache_netfs);
|
|
}
|
|
|
|
int ceph_fscache_register_fs(struct ceph_fs_client* fsc)
|
|
{
|
|
const struct ceph_fsid *fsid = &fsc->client->fsid;
|
|
const char *fscache_uniq = fsc->mount_options->fscache_uniq;
|
|
size_t uniq_len = fscache_uniq ? strlen(fscache_uniq) : 0;
|
|
struct ceph_fscache_entry *ent;
|
|
int err = 0;
|
|
|
|
mutex_lock(&ceph_fscache_lock);
|
|
list_for_each_entry(ent, &ceph_fscache_list, list) {
|
|
if (memcmp(&ent->fsid, fsid, sizeof(*fsid)))
|
|
continue;
|
|
if (ent->uniq_len != uniq_len)
|
|
continue;
|
|
if (uniq_len && memcmp(ent->uniquifier, fscache_uniq, uniq_len))
|
|
continue;
|
|
|
|
pr_err("fscache cookie already registered for fsid %pU\n", fsid);
|
|
pr_err(" use fsc=%%s mount option to specify a uniquifier\n");
|
|
err = -EBUSY;
|
|
goto out_unlock;
|
|
}
|
|
|
|
ent = kzalloc(sizeof(*ent) + uniq_len, GFP_KERNEL);
|
|
if (!ent) {
|
|
err = -ENOMEM;
|
|
goto out_unlock;
|
|
}
|
|
|
|
fsc->fscache = fscache_acquire_cookie(ceph_cache_netfs.primary_index,
|
|
&ceph_fscache_fsid_object_def,
|
|
fsc, true);
|
|
|
|
if (fsc->fscache) {
|
|
memcpy(&ent->fsid, fsid, sizeof(*fsid));
|
|
if (uniq_len > 0) {
|
|
memcpy(&ent->uniquifier, fscache_uniq, uniq_len);
|
|
ent->uniq_len = uniq_len;
|
|
}
|
|
ent->fscache = fsc->fscache;
|
|
list_add_tail(&ent->list, &ceph_fscache_list);
|
|
} else {
|
|
kfree(ent);
|
|
pr_err("unable to register fscache cookie for fsid %pU\n",
|
|
fsid);
|
|
/* all other fs ignore this error */
|
|
}
|
|
out_unlock:
|
|
mutex_unlock(&ceph_fscache_lock);
|
|
return err;
|
|
}
|
|
|
|
static uint16_t ceph_fscache_inode_get_key(const void *cookie_netfs_data,
|
|
void *buffer, uint16_t maxbuf)
|
|
{
|
|
const struct ceph_inode_info* ci = cookie_netfs_data;
|
|
uint16_t klen;
|
|
|
|
/* use ceph virtual inode (id + snapshot) */
|
|
klen = sizeof(ci->i_vino);
|
|
if (klen > maxbuf)
|
|
return 0;
|
|
|
|
memcpy(buffer, &ci->i_vino, klen);
|
|
return klen;
|
|
}
|
|
|
|
static uint16_t ceph_fscache_inode_get_aux(const void *cookie_netfs_data,
|
|
void *buffer, uint16_t bufmax)
|
|
{
|
|
struct ceph_aux_inode aux;
|
|
const struct ceph_inode_info* ci = cookie_netfs_data;
|
|
const struct inode* inode = &ci->vfs_inode;
|
|
|
|
memset(&aux, 0, sizeof(aux));
|
|
aux.version = ci->i_version;
|
|
aux.mtime = inode->i_mtime;
|
|
aux.size = i_size_read(inode);
|
|
|
|
memcpy(buffer, &aux, sizeof(aux));
|
|
|
|
return sizeof(aux);
|
|
}
|
|
|
|
static void ceph_fscache_inode_get_attr(const void *cookie_netfs_data,
|
|
uint64_t *size)
|
|
{
|
|
const struct ceph_inode_info* ci = cookie_netfs_data;
|
|
*size = i_size_read(&ci->vfs_inode);
|
|
}
|
|
|
|
static enum fscache_checkaux ceph_fscache_inode_check_aux(
|
|
void *cookie_netfs_data, const void *data, uint16_t dlen)
|
|
{
|
|
struct ceph_aux_inode aux;
|
|
struct ceph_inode_info* ci = cookie_netfs_data;
|
|
struct inode* inode = &ci->vfs_inode;
|
|
|
|
if (dlen != sizeof(aux))
|
|
return FSCACHE_CHECKAUX_OBSOLETE;
|
|
|
|
memset(&aux, 0, sizeof(aux));
|
|
aux.version = ci->i_version;
|
|
aux.mtime = inode->i_mtime;
|
|
aux.size = i_size_read(inode);
|
|
|
|
if (memcmp(data, &aux, sizeof(aux)) != 0)
|
|
return FSCACHE_CHECKAUX_OBSOLETE;
|
|
|
|
dout("ceph inode 0x%p cached okay", ci);
|
|
return FSCACHE_CHECKAUX_OKAY;
|
|
}
|
|
|
|
static const struct fscache_cookie_def ceph_fscache_inode_object_def = {
|
|
.name = "CEPH.inode",
|
|
.type = FSCACHE_COOKIE_TYPE_DATAFILE,
|
|
.get_key = ceph_fscache_inode_get_key,
|
|
.get_attr = ceph_fscache_inode_get_attr,
|
|
.get_aux = ceph_fscache_inode_get_aux,
|
|
.check_aux = ceph_fscache_inode_check_aux,
|
|
};
|
|
|
|
void ceph_fscache_register_inode_cookie(struct inode *inode)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
|
|
|
|
/* No caching for filesystem */
|
|
if (!fsc->fscache)
|
|
return;
|
|
|
|
/* Only cache for regular files that are read only */
|
|
if (!S_ISREG(inode->i_mode))
|
|
return;
|
|
|
|
inode_lock_nested(inode, I_MUTEX_CHILD);
|
|
if (!ci->fscache) {
|
|
ci->fscache = fscache_acquire_cookie(fsc->fscache,
|
|
&ceph_fscache_inode_object_def,
|
|
ci, false);
|
|
}
|
|
inode_unlock(inode);
|
|
}
|
|
|
|
void ceph_fscache_unregister_inode_cookie(struct ceph_inode_info* ci)
|
|
{
|
|
struct fscache_cookie* cookie;
|
|
|
|
if ((cookie = ci->fscache) == NULL)
|
|
return;
|
|
|
|
ci->fscache = NULL;
|
|
|
|
fscache_uncache_all_inode_pages(cookie, &ci->vfs_inode);
|
|
fscache_relinquish_cookie(cookie, 0);
|
|
}
|
|
|
|
static bool ceph_fscache_can_enable(void *data)
|
|
{
|
|
struct inode *inode = data;
|
|
return !inode_is_open_for_write(inode);
|
|
}
|
|
|
|
void ceph_fscache_file_set_cookie(struct inode *inode, struct file *filp)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
|
|
if (!fscache_cookie_valid(ci->fscache))
|
|
return;
|
|
|
|
if (inode_is_open_for_write(inode)) {
|
|
dout("fscache_file_set_cookie %p %p disabling cache\n",
|
|
inode, filp);
|
|
fscache_disable_cookie(ci->fscache, false);
|
|
fscache_uncache_all_inode_pages(ci->fscache, inode);
|
|
} else {
|
|
fscache_enable_cookie(ci->fscache, ceph_fscache_can_enable,
|
|
inode);
|
|
if (fscache_cookie_enabled(ci->fscache)) {
|
|
dout("fscache_file_set_cookie %p %p enabling cache\n",
|
|
inode, filp);
|
|
}
|
|
}
|
|
}
|
|
|
|
static void ceph_readpage_from_fscache_complete(struct page *page, void *data, int error)
|
|
{
|
|
if (!error)
|
|
SetPageUptodate(page);
|
|
|
|
unlock_page(page);
|
|
}
|
|
|
|
static inline bool cache_valid(struct ceph_inode_info *ci)
|
|
{
|
|
return ci->i_fscache_gen == ci->i_rdcache_gen;
|
|
}
|
|
|
|
|
|
/* Atempt to read from the fscache,
|
|
*
|
|
* This function is called from the readpage_nounlock context. DO NOT attempt to
|
|
* unlock the page here (or in the callback).
|
|
*/
|
|
int ceph_readpage_from_fscache(struct inode *inode, struct page *page)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
int ret;
|
|
|
|
if (!cache_valid(ci))
|
|
return -ENOBUFS;
|
|
|
|
ret = fscache_read_or_alloc_page(ci->fscache, page,
|
|
ceph_readpage_from_fscache_complete, NULL,
|
|
GFP_KERNEL);
|
|
|
|
switch (ret) {
|
|
case 0: /* Page found */
|
|
dout("page read submitted\n");
|
|
return 0;
|
|
case -ENOBUFS: /* Pages were not found, and can't be */
|
|
case -ENODATA: /* Pages were not found */
|
|
dout("page/inode not in cache\n");
|
|
return ret;
|
|
default:
|
|
dout("%s: unknown error ret = %i\n", __func__, ret);
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
int ceph_readpages_from_fscache(struct inode *inode,
|
|
struct address_space *mapping,
|
|
struct list_head *pages,
|
|
unsigned *nr_pages)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
int ret;
|
|
|
|
if (!cache_valid(ci))
|
|
return -ENOBUFS;
|
|
|
|
ret = fscache_read_or_alloc_pages(ci->fscache, mapping, pages, nr_pages,
|
|
ceph_readpage_from_fscache_complete,
|
|
NULL, mapping_gfp_mask(mapping));
|
|
|
|
switch (ret) {
|
|
case 0: /* All pages found */
|
|
dout("all-page read submitted\n");
|
|
return 0;
|
|
case -ENOBUFS: /* Some pages were not found, and can't be */
|
|
case -ENODATA: /* some pages were not found */
|
|
dout("page/inode not in cache\n");
|
|
return ret;
|
|
default:
|
|
dout("%s: unknown error ret = %i\n", __func__, ret);
|
|
return ret;
|
|
}
|
|
}
|
|
|
|
void ceph_readpage_to_fscache(struct inode *inode, struct page *page)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
int ret;
|
|
|
|
if (!PageFsCache(page))
|
|
return;
|
|
|
|
if (!cache_valid(ci))
|
|
return;
|
|
|
|
ret = fscache_write_page(ci->fscache, page, GFP_KERNEL);
|
|
if (ret)
|
|
fscache_uncache_page(ci->fscache, page);
|
|
}
|
|
|
|
void ceph_invalidate_fscache_page(struct inode* inode, struct page *page)
|
|
{
|
|
struct ceph_inode_info *ci = ceph_inode(inode);
|
|
|
|
if (!PageFsCache(page))
|
|
return;
|
|
|
|
fscache_wait_on_page_write(ci->fscache, page);
|
|
fscache_uncache_page(ci->fscache, page);
|
|
}
|
|
|
|
void ceph_fscache_unregister_fs(struct ceph_fs_client* fsc)
|
|
{
|
|
if (fscache_cookie_valid(fsc->fscache)) {
|
|
struct ceph_fscache_entry *ent;
|
|
bool found = false;
|
|
|
|
mutex_lock(&ceph_fscache_lock);
|
|
list_for_each_entry(ent, &ceph_fscache_list, list) {
|
|
if (ent->fscache == fsc->fscache) {
|
|
list_del(&ent->list);
|
|
kfree(ent);
|
|
found = true;
|
|
break;
|
|
}
|
|
}
|
|
WARN_ON_ONCE(!found);
|
|
mutex_unlock(&ceph_fscache_lock);
|
|
|
|
__fscache_relinquish_cookie(fsc->fscache, 0);
|
|
}
|
|
fsc->fscache = NULL;
|
|
}
|
|
|
|
/*
|
|
* caller should hold CEPH_CAP_FILE_{RD,CACHE}
|
|
*/
|
|
void ceph_fscache_revalidate_cookie(struct ceph_inode_info *ci)
|
|
{
|
|
if (cache_valid(ci))
|
|
return;
|
|
|
|
/* resue i_truncate_mutex. There should be no pending
|
|
* truncate while the caller holds CEPH_CAP_FILE_RD */
|
|
mutex_lock(&ci->i_truncate_mutex);
|
|
if (!cache_valid(ci)) {
|
|
if (fscache_check_consistency(ci->fscache))
|
|
fscache_invalidate(ci->fscache);
|
|
spin_lock(&ci->i_ceph_lock);
|
|
ci->i_fscache_gen = ci->i_rdcache_gen;
|
|
spin_unlock(&ci->i_ceph_lock);
|
|
}
|
|
mutex_unlock(&ci->i_truncate_mutex);
|
|
}
|