From e636c8d336489bf3eed5878299e6cc045bbad077 Mon Sep 17 00:00:00 2001
From: hc <hc@nodka.com>
Date: Tue, 20 Feb 2024 01:17:29 +0000
Subject: [PATCH] debug lk
---
kernel/fs/ceph/quota.c | 286 +++++++++++++++++++++++++++++++++++++++++++++++++-------
1 files changed, 249 insertions(+), 37 deletions(-)
diff --git a/kernel/fs/ceph/quota.c b/kernel/fs/ceph/quota.c
index 03f4d24..9b785f1 100644
--- a/kernel/fs/ceph/quota.c
+++ b/kernel/fs/ceph/quota.c
@@ -3,19 +3,6 @@
* quota.c - CephFS quota
*
* Copyright (C) 2017-2018 SUSE
- *
- * This program is free software; you can redistribute it and/or
- * modify it under the terms of the GNU General Public License
- * as published by the Free Software Foundation; either version 2
- * of the License, or (at your option) any later version.
- *
- * 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, see <http://www.gnu.org/licenses/>.
*/
#include <linux/statfs.h>
@@ -25,7 +12,7 @@
void ceph_adjust_quota_realms_count(struct inode *inode, bool inc)
{
- struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
+ struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb);
if (inc)
atomic64_inc(&mdsc->quotarealms_count);
else
@@ -34,8 +21,17 @@
static inline bool ceph_has_realms_with_quotas(struct inode *inode)
{
- struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
- return atomic64_read(&mdsc->quotarealms_count) > 0;
+ struct super_block *sb = inode->i_sb;
+ struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(sb);
+ struct inode *root = d_inode(sb->s_root);
+
+ if (atomic64_read(&mdsc->quotarealms_count) > 0)
+ return true;
+ /* if root is the real CephFS root, we don't have quota realms */
+ if (root && ceph_ino(root) == CEPH_INO_ROOT)
+ return false;
+ /* otherwise, we can't know for sure */
+ return true;
}
void ceph_handle_quota(struct ceph_mds_client *mdsc,
@@ -57,7 +53,7 @@
/* increment msg sequence number */
mutex_lock(&session->s_mutex);
- session->s_seq++;
+ inc_session_sequence(session);
mutex_unlock(&session->s_mutex);
/* lookup inode */
@@ -78,7 +74,121 @@
le64_to_cpu(h->max_files));
spin_unlock(&ci->i_ceph_lock);
- iput(inode);
+ /* avoid calling iput_final() in dispatch thread */
+ ceph_async_iput(inode);
+}
+
+static struct ceph_quotarealm_inode *
+find_quotarealm_inode(struct ceph_mds_client *mdsc, u64 ino)
+{
+ struct ceph_quotarealm_inode *qri = NULL;
+ struct rb_node **node, *parent = NULL;
+
+ mutex_lock(&mdsc->quotarealms_inodes_mutex);
+ node = &(mdsc->quotarealms_inodes.rb_node);
+ while (*node) {
+ parent = *node;
+ qri = container_of(*node, struct ceph_quotarealm_inode, node);
+
+ if (ino < qri->ino)
+ node = &((*node)->rb_left);
+ else if (ino > qri->ino)
+ node = &((*node)->rb_right);
+ else
+ break;
+ }
+ if (!qri || (qri->ino != ino)) {
+ /* Not found, create a new one and insert it */
+ qri = kmalloc(sizeof(*qri), GFP_KERNEL);
+ if (qri) {
+ qri->ino = ino;
+ qri->inode = NULL;
+ qri->timeout = 0;
+ mutex_init(&qri->mutex);
+ rb_link_node(&qri->node, parent, node);
+ rb_insert_color(&qri->node, &mdsc->quotarealms_inodes);
+ } else
+ pr_warn("Failed to alloc quotarealms_inode\n");
+ }
+ mutex_unlock(&mdsc->quotarealms_inodes_mutex);
+
+ return qri;
+}
+
+/*
+ * This function will try to lookup a realm inode which isn't visible in the
+ * filesystem mountpoint. A list of these kind of inodes (not visible) is
+ * maintained in the mdsc and freed only when the filesystem is umounted.
+ *
+ * Note that these inodes are kept in this list even if the lookup fails, which
+ * allows to prevent useless lookup requests.
+ */
+static struct inode *lookup_quotarealm_inode(struct ceph_mds_client *mdsc,
+ struct super_block *sb,
+ struct ceph_snap_realm *realm)
+{
+ struct ceph_quotarealm_inode *qri;
+ struct inode *in;
+
+ qri = find_quotarealm_inode(mdsc, realm->ino);
+ if (!qri)
+ return NULL;
+
+ mutex_lock(&qri->mutex);
+ if (qri->inode && ceph_is_any_caps(qri->inode)) {
+ /* A request has already returned the inode */
+ mutex_unlock(&qri->mutex);
+ return qri->inode;
+ }
+ /* Check if this inode lookup has failed recently */
+ if (qri->timeout &&
+ time_before_eq(jiffies, qri->timeout)) {
+ mutex_unlock(&qri->mutex);
+ return NULL;
+ }
+ if (qri->inode) {
+ /* get caps */
+ int ret = __ceph_do_getattr(qri->inode, NULL,
+ CEPH_STAT_CAP_INODE, true);
+ if (ret >= 0)
+ in = qri->inode;
+ else
+ in = ERR_PTR(ret);
+ } else {
+ in = ceph_lookup_inode(sb, realm->ino);
+ }
+
+ if (IS_ERR(in)) {
+ dout("Can't lookup inode %llx (err: %ld)\n",
+ realm->ino, PTR_ERR(in));
+ qri->timeout = jiffies + msecs_to_jiffies(60 * 1000); /* XXX */
+ } else {
+ qri->timeout = 0;
+ qri->inode = in;
+ }
+ mutex_unlock(&qri->mutex);
+
+ return in;
+}
+
+void ceph_cleanup_quotarealms_inodes(struct ceph_mds_client *mdsc)
+{
+ struct ceph_quotarealm_inode *qri;
+ struct rb_node *node;
+
+ /*
+ * It should now be safe to clean quotarealms_inode tree without holding
+ * mdsc->quotarealms_inodes_mutex...
+ */
+ mutex_lock(&mdsc->quotarealms_inodes_mutex);
+ while (!RB_EMPTY_ROOT(&mdsc->quotarealms_inodes)) {
+ node = rb_first(&mdsc->quotarealms_inodes);
+ qri = rb_entry(node, struct ceph_quotarealm_inode, node);
+ rb_erase(node, &mdsc->quotarealms_inodes);
+ iput(qri->inode);
+ kfree(qri);
+ }
+ mutex_unlock(&mdsc->quotarealms_inodes_mutex);
}
/*
@@ -89,9 +199,15 @@
*
* Note that the caller is responsible for calling ceph_put_snap_realm() on the
* returned realm.
+ *
+ * Callers of this function need to hold mdsc->snap_rwsem. However, if there's
+ * a need to do an inode lookup, this rwsem will be temporarily dropped. Hence
+ * the 'retry' argument: if rwsem needs to be dropped and 'retry' is 'false'
+ * this function will return -EAGAIN; otherwise, the snaprealms walk-through
+ * will be restarted.
*/
static struct ceph_snap_realm *get_quota_realm(struct ceph_mds_client *mdsc,
- struct inode *inode)
+ struct inode *inode, bool retry)
{
struct ceph_inode_info *ci = NULL;
struct ceph_snap_realm *realm, *next;
@@ -101,6 +217,7 @@
if (ceph_snap(inode) != CEPH_NOSNAP)
return NULL;
+restart:
realm = ceph_inode(inode)->i_snap_realm;
if (realm)
ceph_get_snap_realm(mdsc, realm);
@@ -108,15 +225,30 @@
pr_err_ratelimited("get_quota_realm: ino (%llx.%llx) "
"null i_snap_realm\n", ceph_vinop(inode));
while (realm) {
+ bool has_inode;
+
spin_lock(&realm->inodes_with_caps_lock);
- in = realm->inode ? igrab(realm->inode) : NULL;
+ has_inode = realm->inode;
+ in = has_inode ? igrab(realm->inode) : NULL;
spin_unlock(&realm->inodes_with_caps_lock);
- if (!in)
+ if (has_inode && !in)
break;
+ if (!in) {
+ up_read(&mdsc->snap_rwsem);
+ in = lookup_quotarealm_inode(mdsc, inode->i_sb, realm);
+ down_read(&mdsc->snap_rwsem);
+ if (IS_ERR_OR_NULL(in))
+ break;
+ ceph_put_snap_realm(mdsc, realm);
+ if (!retry)
+ return ERR_PTR(-EAGAIN);
+ goto restart;
+ }
ci = ceph_inode(in);
has_quota = __ceph_has_any_quota(ci);
- iput(in);
+ /* avoid calling iput_final() while holding mdsc->snap_rwsem */
+ ceph_async_iput(in);
next = realm->parent;
if (has_quota || !next)
@@ -132,15 +264,28 @@
return NULL;
}
-bool ceph_quota_is_same_realm(struct inode *old, struct inode *new)
+static bool ceph_quota_is_same_realm(struct inode *old, struct inode *new)
{
- struct ceph_mds_client *mdsc = ceph_inode_to_client(old)->mdsc;
+ struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(old->i_sb);
struct ceph_snap_realm *old_realm, *new_realm;
bool is_same;
+restart:
+ /*
+ * We need to lookup 2 quota realms atomically, i.e. with snap_rwsem.
+ * However, get_quota_realm may drop it temporarily. By setting the
+ * 'retry' parameter to 'false', we'll get -EAGAIN if the rwsem was
+ * dropped and we can then restart the whole operation.
+ */
down_read(&mdsc->snap_rwsem);
- old_realm = get_quota_realm(mdsc, old);
- new_realm = get_quota_realm(mdsc, new);
+ old_realm = get_quota_realm(mdsc, old, true);
+ new_realm = get_quota_realm(mdsc, new, false);
+ if (PTR_ERR(new_realm) == -EAGAIN) {
+ up_read(&mdsc->snap_rwsem);
+ if (old_realm)
+ ceph_put_snap_realm(mdsc, old_realm);
+ goto restart;
+ }
is_same = (old_realm == new_realm);
up_read(&mdsc->snap_rwsem);
@@ -168,7 +313,7 @@
static bool check_quota_exceeded(struct inode *inode, enum quota_check_op op,
loff_t delta)
{
- struct ceph_mds_client *mdsc = ceph_inode_to_client(inode)->mdsc;
+ struct ceph_mds_client *mdsc = ceph_sb_to_mdsc(inode->i_sb);
struct ceph_inode_info *ci;
struct ceph_snap_realm *realm, *next;
struct inode *in;
@@ -179,6 +324,7 @@
return false;
down_read(&mdsc->snap_rwsem);
+restart:
realm = ceph_inode(inode)->i_snap_realm;
if (realm)
ceph_get_snap_realm(mdsc, realm);
@@ -186,12 +332,23 @@
pr_err_ratelimited("check_quota_exceeded: ino (%llx.%llx) "
"null i_snap_realm\n", ceph_vinop(inode));
while (realm) {
- spin_lock(&realm->inodes_with_caps_lock);
- in = realm->inode ? igrab(realm->inode) : NULL;
- spin_unlock(&realm->inodes_with_caps_lock);
- if (!in)
- break;
+ bool has_inode;
+ spin_lock(&realm->inodes_with_caps_lock);
+ has_inode = realm->inode;
+ in = has_inode ? igrab(realm->inode) : NULL;
+ spin_unlock(&realm->inodes_with_caps_lock);
+ if (has_inode && !in)
+ break;
+ if (!in) {
+ up_read(&mdsc->snap_rwsem);
+ in = lookup_quotarealm_inode(mdsc, inode->i_sb, realm);
+ down_read(&mdsc->snap_rwsem);
+ if (IS_ERR_OR_NULL(in))
+ break;
+ ceph_put_snap_realm(mdsc, realm);
+ goto restart;
+ }
ci = ceph_inode(in);
spin_lock(&ci->i_ceph_lock);
if (op == QUOTA_CHECK_MAX_FILES_OP) {
@@ -204,8 +361,6 @@
spin_unlock(&ci->i_ceph_lock);
switch (op) {
case QUOTA_CHECK_MAX_FILES_OP:
- exceeded = (max && (rvalue >= max));
- break;
case QUOTA_CHECK_MAX_BYTES_OP:
exceeded = (max && (rvalue + delta > max));
break;
@@ -228,7 +383,8 @@
pr_warn("Invalid quota check op (%d)\n", op);
exceeded = true; /* Just break the loop */
}
- iput(in);
+ /* avoid calling iput_final() while holding mdsc->snap_rwsem */
+ ceph_async_iput(in);
next = realm->parent;
if (exceeded || !next)
@@ -259,7 +415,7 @@
WARN_ON(!S_ISDIR(inode->i_mode));
- return check_quota_exceeded(inode, QUOTA_CHECK_MAX_FILES_OP, 0);
+ return check_quota_exceeded(inode, QUOTA_CHECK_MAX_FILES_OP, 1);
}
/*
@@ -327,7 +483,7 @@
bool is_updated = false;
down_read(&mdsc->snap_rwsem);
- realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root));
+ realm = get_quota_realm(mdsc, d_inode(fsc->sb->s_root), true);
up_read(&mdsc->snap_rwsem);
if (!realm)
return false;
@@ -360,3 +516,59 @@
return is_updated;
}
+/*
+ * ceph_quota_check_rename - check if a rename can be executed
+ * @mdsc: MDS client instance
+ * @old: inode to be copied
+ * @new: destination inode (directory)
+ *
+ * This function verifies if a rename (e.g. moving a file or directory) can be
+ * executed. It forces an rstat update in the @new target directory (and in the
+ * source @old as well, if it's a directory). The actual check is done both for
+ * max_files and max_bytes.
+ *
+ * This function returns 0 if it's OK to do the rename, or, if quotas are
+ * exceeded, -EXDEV (if @old is a directory) or -EDQUOT.
+ */
+int ceph_quota_check_rename(struct ceph_mds_client *mdsc,
+ struct inode *old, struct inode *new)
+{
+ struct ceph_inode_info *ci_old = ceph_inode(old);
+ int ret = 0;
+
+ if (ceph_quota_is_same_realm(old, new))
+ return 0;
+
+ /*
+ * Get the latest rstat for target directory (and for source, if a
+ * directory)
+ */
+ ret = ceph_do_getattr(new, CEPH_STAT_RSTAT, false);
+ if (ret)
+ return ret;
+
+ if (S_ISDIR(old->i_mode)) {
+ ret = ceph_do_getattr(old, CEPH_STAT_RSTAT, false);
+ if (ret)
+ return ret;
+ ret = check_quota_exceeded(new, QUOTA_CHECK_MAX_BYTES_OP,
+ ci_old->i_rbytes);
+ if (!ret)
+ ret = check_quota_exceeded(new,
+ QUOTA_CHECK_MAX_FILES_OP,
+ ci_old->i_rfiles +
+ ci_old->i_rsubdirs);
+ if (ret)
+ ret = -EXDEV;
+ } else {
+ ret = check_quota_exceeded(new, QUOTA_CHECK_MAX_BYTES_OP,
+ i_size_read(old));
+ if (!ret)
+ ret = check_quota_exceeded(new,
+ QUOTA_CHECK_MAX_FILES_OP, 1);
+ if (ret)
+ ret = -EDQUOT;
+ }
+
+ return ret;
+}
--
Gitblit v1.6.2