linux-stable/fs/gfs2/super.c

1560 lines
38 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved.
* Copyright (C) 2004-2007 Red Hat, Inc. All rights reserved.
*/
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
#include <linux/bio.h>
#include <linux/sched/signal.h>
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/completion.h>
#include <linux/buffer_head.h>
#include <linux/statfs.h>
#include <linux/seq_file.h>
#include <linux/mount.h>
#include <linux/kthread.h>
#include <linux/delay.h>
#include <linux/gfs2_ondisk.h>
#include <linux/crc32.h>
#include <linux/time.h>
#include <linux/wait.h>
#include <linux/writeback.h>
#include <linux/backing-dev.h>
#include <linux/kernel.h>
#include "gfs2.h"
#include "incore.h"
#include "bmap.h"
#include "dir.h"
#include "glock.h"
#include "glops.h"
#include "inode.h"
#include "log.h"
#include "meta_io.h"
#include "quota.h"
#include "recovery.h"
#include "rgrp.h"
#include "super.h"
#include "trans.h"
#include "util.h"
#include "sys.h"
#include "xattr.h"
#include "lops.h"
enum dinode_demise {
SHOULD_DELETE_DINODE,
SHOULD_NOT_DELETE_DINODE,
SHOULD_DEFER_EVICTION,
};
/**
* gfs2_jindex_free - Clear all the journal index information
* @sdp: The GFS2 superblock
*
*/
void gfs2_jindex_free(struct gfs2_sbd *sdp)
{
struct list_head list;
struct gfs2_jdesc *jd;
spin_lock(&sdp->sd_jindex_spin);
list_add(&list, &sdp->sd_jindex_list);
list_del_init(&sdp->sd_jindex_list);
sdp->sd_journals = 0;
spin_unlock(&sdp->sd_jindex_spin);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
sdp->sd_jdesc = NULL;
while (!list_empty(&list)) {
jd = list_first_entry(&list, struct gfs2_jdesc, jd_list);
gfs2_free_journal_extents(jd);
list_del(&jd->jd_list);
iput(jd->jd_inode);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
jd->jd_inode = NULL;
kfree(jd);
}
}
static struct gfs2_jdesc *jdesc_find_i(struct list_head *head, unsigned int jid)
{
struct gfs2_jdesc *jd;
list_for_each_entry(jd, head, jd_list) {
if (jd->jd_jid == jid)
return jd;
}
return NULL;
}
struct gfs2_jdesc *gfs2_jdesc_find(struct gfs2_sbd *sdp, unsigned int jid)
{
struct gfs2_jdesc *jd;
spin_lock(&sdp->sd_jindex_spin);
jd = jdesc_find_i(&sdp->sd_jindex_list, jid);
spin_unlock(&sdp->sd_jindex_spin);
return jd;
}
int gfs2_jdesc_check(struct gfs2_jdesc *jd)
{
struct gfs2_inode *ip = GFS2_I(jd->jd_inode);
struct gfs2_sbd *sdp = GFS2_SB(jd->jd_inode);
u64 size = i_size_read(jd->jd_inode);
if (gfs2_check_internal_file_size(jd->jd_inode, 8 << 20, BIT(30)))
return -EIO;
jd->jd_blocks = size >> sdp->sd_sb.sb_bsize_shift;
if (gfs2_write_alloc_required(ip, 0, size)) {
gfs2_consist_inode(ip);
return -EIO;
}
return 0;
}
static int init_threads(struct gfs2_sbd *sdp)
{
struct task_struct *p;
int error = 0;
p = kthread_run(gfs2_logd, sdp, "gfs2_logd");
if (IS_ERR(p)) {
error = PTR_ERR(p);
fs_err(sdp, "can't start logd thread: %d\n", error);
return error;
}
sdp->sd_logd_process = p;
p = kthread_run(gfs2_quotad, sdp, "gfs2_quotad");
if (IS_ERR(p)) {
error = PTR_ERR(p);
fs_err(sdp, "can't start quotad thread: %d\n", error);
goto fail;
}
sdp->sd_quotad_process = p;
return 0;
fail:
kthread_stop(sdp->sd_logd_process);
sdp->sd_logd_process = NULL;
return error;
}
/**
* gfs2_make_fs_rw - Turn a Read-Only FS into a Read-Write one
* @sdp: the filesystem
*
* Returns: errno
*/
int gfs2_make_fs_rw(struct gfs2_sbd *sdp)
{
struct gfs2_inode *ip = GFS2_I(sdp->sd_jdesc->jd_inode);
struct gfs2_glock *j_gl = ip->i_gl;
struct gfs2_log_header_host head;
int error;
error = init_threads(sdp);
if (error) {
gfs2_withdraw_delayed(sdp);
return error;
}
j_gl->gl_ops->go_inval(j_gl, DIO_METADATA);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (gfs2_withdrawn(sdp)) {
error = -EIO;
goto fail;
}
error = gfs2_find_jhead(sdp->sd_jdesc, &head, false);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (error || gfs2_withdrawn(sdp))
goto fail;
if (!(head.lh_flags & GFS2_LOG_HEAD_UNMOUNT)) {
gfs2_consist(sdp);
error = -EIO;
goto fail;
}
/* Initialize some head of the log stuff */
sdp->sd_log_sequence = head.lh_sequence + 1;
gfs2_log_pointers_init(sdp, head.lh_blkno);
error = gfs2_quota_init(sdp);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (error || gfs2_withdrawn(sdp))
goto fail;
set_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);
return 0;
fail:
if (sdp->sd_quotad_process)
kthread_stop(sdp->sd_quotad_process);
sdp->sd_quotad_process = NULL;
if (sdp->sd_logd_process)
kthread_stop(sdp->sd_logd_process);
sdp->sd_logd_process = NULL;
return error;
}
void gfs2_statfs_change_in(struct gfs2_statfs_change_host *sc, const void *buf)
{
const struct gfs2_statfs_change *str = buf;
sc->sc_total = be64_to_cpu(str->sc_total);
sc->sc_free = be64_to_cpu(str->sc_free);
sc->sc_dinodes = be64_to_cpu(str->sc_dinodes);
}
void gfs2_statfs_change_out(const struct gfs2_statfs_change_host *sc, void *buf)
{
struct gfs2_statfs_change *str = buf;
str->sc_total = cpu_to_be64(sc->sc_total);
str->sc_free = cpu_to_be64(sc->sc_free);
str->sc_dinodes = cpu_to_be64(sc->sc_dinodes);
}
int gfs2_statfs_init(struct gfs2_sbd *sdp)
{
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
struct buffer_head *m_bh, *l_bh;
struct gfs2_holder gh;
int error;
error = gfs2_glock_nq_init(m_ip->i_gl, LM_ST_EXCLUSIVE, GL_NOCACHE,
&gh);
if (error)
return error;
error = gfs2_meta_inode_buffer(m_ip, &m_bh);
if (error)
goto out;
if (sdp->sd_args.ar_spectator) {
spin_lock(&sdp->sd_statfs_spin);
gfs2_statfs_change_in(m_sc, m_bh->b_data +
sizeof(struct gfs2_dinode));
spin_unlock(&sdp->sd_statfs_spin);
} else {
error = gfs2_meta_inode_buffer(l_ip, &l_bh);
if (error)
goto out_m_bh;
spin_lock(&sdp->sd_statfs_spin);
gfs2_statfs_change_in(m_sc, m_bh->b_data +
sizeof(struct gfs2_dinode));
gfs2_statfs_change_in(l_sc, l_bh->b_data +
sizeof(struct gfs2_dinode));
spin_unlock(&sdp->sd_statfs_spin);
brelse(l_bh);
}
out_m_bh:
brelse(m_bh);
out:
gfs2_glock_dq_uninit(&gh);
return 0;
}
void gfs2_statfs_change(struct gfs2_sbd *sdp, s64 total, s64 free,
s64 dinodes)
{
struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
struct buffer_head *l_bh;
s64 x, y;
int need_sync = 0;
int error;
error = gfs2_meta_inode_buffer(l_ip, &l_bh);
if (error)
return;
gfs2_trans_add_meta(l_ip->i_gl, l_bh);
spin_lock(&sdp->sd_statfs_spin);
l_sc->sc_total += total;
l_sc->sc_free += free;
l_sc->sc_dinodes += dinodes;
gfs2_statfs_change_out(l_sc, l_bh->b_data + sizeof(struct gfs2_dinode));
if (sdp->sd_args.ar_statfs_percent) {
x = 100 * l_sc->sc_free;
y = m_sc->sc_free * sdp->sd_args.ar_statfs_percent;
if (x >= y || x <= -y)
need_sync = 1;
}
spin_unlock(&sdp->sd_statfs_spin);
brelse(l_bh);
if (need_sync)
gfs2_wake_up_statfs(sdp);
}
void update_statfs(struct gfs2_sbd *sdp, struct buffer_head *m_bh,
struct buffer_head *l_bh)
{
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
gfs2_trans_add_meta(l_ip->i_gl, l_bh);
gfs2_trans_add_meta(m_ip->i_gl, m_bh);
spin_lock(&sdp->sd_statfs_spin);
m_sc->sc_total += l_sc->sc_total;
m_sc->sc_free += l_sc->sc_free;
m_sc->sc_dinodes += l_sc->sc_dinodes;
memset(l_sc, 0, sizeof(struct gfs2_statfs_change));
memset(l_bh->b_data + sizeof(struct gfs2_dinode),
0, sizeof(struct gfs2_statfs_change));
gfs2_statfs_change_out(m_sc, m_bh->b_data + sizeof(struct gfs2_dinode));
spin_unlock(&sdp->sd_statfs_spin);
}
int gfs2_statfs_sync(struct super_block *sb, int type)
{
struct gfs2_sbd *sdp = sb->s_fs_info;
struct gfs2_inode *m_ip = GFS2_I(sdp->sd_statfs_inode);
struct gfs2_inode *l_ip = GFS2_I(sdp->sd_sc_inode);
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
struct gfs2_holder gh;
struct buffer_head *m_bh, *l_bh;
int error;
error = gfs2_glock_nq_init(m_ip->i_gl, LM_ST_EXCLUSIVE, GL_NOCACHE,
&gh);
if (error)
goto out;
error = gfs2_meta_inode_buffer(m_ip, &m_bh);
if (error)
goto out_unlock;
spin_lock(&sdp->sd_statfs_spin);
gfs2_statfs_change_in(m_sc, m_bh->b_data +
sizeof(struct gfs2_dinode));
if (!l_sc->sc_total && !l_sc->sc_free && !l_sc->sc_dinodes) {
spin_unlock(&sdp->sd_statfs_spin);
goto out_bh;
}
spin_unlock(&sdp->sd_statfs_spin);
error = gfs2_meta_inode_buffer(l_ip, &l_bh);
if (error)
goto out_bh;
error = gfs2_trans_begin(sdp, 2 * RES_DINODE, 0);
if (error)
goto out_bh2;
update_statfs(sdp, m_bh, l_bh);
sdp->sd_statfs_force_sync = 0;
gfs2_trans_end(sdp);
out_bh2:
brelse(l_bh);
out_bh:
brelse(m_bh);
out_unlock:
gfs2_glock_dq_uninit(&gh);
out:
return error;
}
struct lfcc {
struct list_head list;
struct gfs2_holder gh;
};
/**
* gfs2_lock_fs_check_clean - Stop all writes to the FS and check that all
* journals are clean
* @sdp: the file system
*
* Returns: errno
*/
static int gfs2_lock_fs_check_clean(struct gfs2_sbd *sdp)
{
struct gfs2_inode *ip;
struct gfs2_jdesc *jd;
struct lfcc *lfcc;
LIST_HEAD(list);
struct gfs2_log_header_host lh;
int error;
list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
lfcc = kmalloc(sizeof(struct lfcc), GFP_KERNEL);
if (!lfcc) {
error = -ENOMEM;
goto out;
}
ip = GFS2_I(jd->jd_inode);
error = gfs2_glock_nq_init(ip->i_gl, LM_ST_SHARED, 0, &lfcc->gh);
if (error) {
kfree(lfcc);
goto out;
}
list_add(&lfcc->list, &list);
}
GFS2: remove transaction glock GFS2 has a transaction glock, which must be grabbed for every transaction, whose purpose is to deal with freezing the filesystem. Aside from this involving a large amount of locking, it is very easy to make the current fsfreeze code hang on unfreezing. This patch rewrites how gfs2 handles freezing the filesystem. The transaction glock is removed. In it's place is a freeze glock, which is cached (but not held) in a shared state by every node in the cluster when the filesystem is mounted. This lock only needs to be grabbed on freezing, and actions which need to be safe from freezing, like recovery. When a node wants to freeze the filesystem, it grabs this glock exclusively. When the freeze glock state changes on the nodes (either from shared to unlocked, or shared to exclusive), the filesystem does a special log flush. gfs2_log_flush() does all the work for flushing out the and shutting down the incore log, and then it tries to grab the freeze glock in a shared state again. Since the filesystem is stuck in gfs2_log_flush, no new transaction can start, and nothing can be written to disk. Unfreezing the filesytem simply involes dropping the freeze glock, allowing gfs2_log_flush() to grab and then release the shared lock, so it is cached for next time. However, in order for the unfreezing ioctl to occur, gfs2 needs to get a shared lock on the filesystem root directory inode to check permissions. If that glock has already been grabbed exclusively, fsfreeze will be unable to get the shared lock and unfreeze the filesystem. In order to allow the unfreeze, this patch makes gfs2 grab a shared lock on the filesystem root directory during the freeze, and hold it until it unfreezes the filesystem. The functions which need to grab a shared lock in order to allow the unfreeze ioctl to be issued now use the lock grabbed by the freeze code instead. The freeze and unfreeze code take care to make sure that this shared lock will not be dropped while another process is using it. Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2014-05-02 03:26:55 +00:00
error = gfs2_glock_nq_init(sdp->sd_freeze_gl, LM_ST_EXCLUSIVE,
LM_FLAG_NOEXP, &sdp->sd_freeze_gh);
if (error)
goto out;
list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
error = gfs2_jdesc_check(jd);
if (error)
break;
error = gfs2_find_jhead(jd, &lh, false);
if (error)
break;
if (!(lh.lh_flags & GFS2_LOG_HEAD_UNMOUNT)) {
error = -EBUSY;
break;
}
}
if (error)
gfs2_freeze_unlock(&sdp->sd_freeze_gh);
out:
while (!list_empty(&list)) {
lfcc = list_first_entry(&list, struct lfcc, list);
list_del(&lfcc->list);
gfs2_glock_dq_uninit(&lfcc->gh);
kfree(lfcc);
}
return error;
}
void gfs2_dinode_out(const struct gfs2_inode *ip, void *buf)
{
struct gfs2_dinode *str = buf;
str->di_header.mh_magic = cpu_to_be32(GFS2_MAGIC);
str->di_header.mh_type = cpu_to_be32(GFS2_METATYPE_DI);
str->di_header.mh_format = cpu_to_be32(GFS2_FORMAT_DI);
str->di_num.no_addr = cpu_to_be64(ip->i_no_addr);
str->di_num.no_formal_ino = cpu_to_be64(ip->i_no_formal_ino);
str->di_mode = cpu_to_be32(ip->i_inode.i_mode);
str->di_uid = cpu_to_be32(i_uid_read(&ip->i_inode));
str->di_gid = cpu_to_be32(i_gid_read(&ip->i_inode));
str->di_nlink = cpu_to_be32(ip->i_inode.i_nlink);
str->di_size = cpu_to_be64(i_size_read(&ip->i_inode));
str->di_blocks = cpu_to_be64(gfs2_get_inode_blocks(&ip->i_inode));
str->di_atime = cpu_to_be64(ip->i_inode.i_atime.tv_sec);
str->di_mtime = cpu_to_be64(ip->i_inode.i_mtime.tv_sec);
str->di_ctime = cpu_to_be64(ip->i_inode.i_ctime.tv_sec);
str->di_goal_meta = cpu_to_be64(ip->i_goal);
str->di_goal_data = cpu_to_be64(ip->i_goal);
str->di_generation = cpu_to_be64(ip->i_generation);
str->di_flags = cpu_to_be32(ip->i_diskflags);
str->di_height = cpu_to_be16(ip->i_height);
str->di_payload_format = cpu_to_be32(S_ISDIR(ip->i_inode.i_mode) &&
!(ip->i_diskflags & GFS2_DIF_EXHASH) ?
GFS2_FORMAT_DE : 0);
str->di_depth = cpu_to_be16(ip->i_depth);
str->di_entries = cpu_to_be32(ip->i_entries);
str->di_eattr = cpu_to_be64(ip->i_eattr);
str->di_atime_nsec = cpu_to_be32(ip->i_inode.i_atime.tv_nsec);
str->di_mtime_nsec = cpu_to_be32(ip->i_inode.i_mtime.tv_nsec);
str->di_ctime_nsec = cpu_to_be32(ip->i_inode.i_ctime.tv_nsec);
}
/**
* gfs2_write_inode - Make sure the inode is stable on the disk
* @inode: The inode
* @wbc: The writeback control structure
*
* Returns: errno
*/
static int gfs2_write_inode(struct inode *inode, struct writeback_control *wbc)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct address_space *metamapping = gfs2_glock2aspace(ip->i_gl);
struct backing_dev_info *bdi = inode_to_bdi(metamapping->host);
int ret = 0;
bool flush_all = (wbc->sync_mode == WB_SYNC_ALL || gfs2_is_jdata(ip));
if (flush_all)
gfs2_log_flush(GFS2_SB(inode), ip->i_gl,
GFS2_LOG_HEAD_FLUSH_NORMAL |
GFS2_LFC_WRITE_INODE);
writeback: move bandwidth related fields from backing_dev_info into bdi_writeback Currently, a bdi (backing_dev_info) embeds single wb (bdi_writeback) and the role of the separation is unclear. For cgroup support for writeback IOs, a bdi will be updated to host multiple wb's where each wb serves writeback IOs of a different cgroup on the bdi. To achieve that, a wb should carry all states necessary for servicing writeback IOs for a cgroup independently. This patch moves bandwidth related fields from backing_dev_info into bdi_writeback. * The moved fields are: bw_time_stamp, dirtied_stamp, written_stamp, write_bandwidth, avg_write_bandwidth, dirty_ratelimit, balanced_dirty_ratelimit, completions and dirty_exceeded. * writeback_chunk_size() and over_bground_thresh() now take @wb instead of @bdi. * bdi_writeout_fraction(bdi, ...) -> wb_writeout_fraction(wb, ...) bdi_dirty_limit(bdi, ...) -> wb_dirty_limit(wb, ...) bdi_position_ration(bdi, ...) -> wb_position_ratio(wb, ...) bdi_update_writebandwidth(bdi, ...) -> wb_update_write_bandwidth(wb, ...) [__]bdi_update_bandwidth(bdi, ...) -> [__]wb_update_bandwidth(wb, ...) bdi_{max|min}_pause(bdi, ...) -> wb_{max|min}_pause(wb, ...) bdi_dirty_limits(bdi, ...) -> wb_dirty_limits(wb, ...) * Init/exits of the relocated fields are moved to bdi_wb_init/exit() respectively. Note that explicit zeroing is dropped in the process as wb's are cleared in entirety anyway. * As there's still only one bdi_writeback per backing_dev_info, all uses of bdi->stat[] are mechanically replaced with bdi->wb.stat[] introducing no behavior changes. v2: Typo in description fixed as suggested by Jan. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Jan Kara <jack@suse.cz> Cc: Jens Axboe <axboe@kernel.dk> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Steven Whitehouse <swhiteho@redhat.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:28 +00:00
if (bdi->wb.dirty_exceeded)
gfs2_ail1_flush(sdp, wbc);
else
filemap_fdatawrite(metamapping);
if (flush_all)
ret = filemap_fdatawait(metamapping);
if (ret)
mark_inode_dirty_sync(inode);
else {
spin_lock(&inode->i_lock);
if (!(inode->i_flags & I_DIRTY))
gfs2_ordered_del_inode(ip);
spin_unlock(&inode->i_lock);
}
return ret;
}
/**
* gfs2_dirty_inode - check for atime updates
* @inode: The inode in question
* @flags: The type of dirty
*
* Unfortunately it can be called under any combination of inode
* glock and transaction lock, so we have to check carefully.
*
* At the moment this deals only with atime - it should be possible
* to expand that role in future, once a review of the locking has
* been carried out.
*/
static void gfs2_dirty_inode(struct inode *inode, int flags)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct buffer_head *bh;
struct gfs2_holder gh;
int need_unlock = 0;
int need_endtrans = 0;
int ret;
if (unlikely(gfs2_withdrawn(sdp)))
return;
if (!gfs2_glock_is_locked_by_me(ip->i_gl)) {
ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, 0, &gh);
if (ret) {
fs_err(sdp, "dirty_inode: glock %d\n", ret);
gfs2_dump_glock(NULL, ip->i_gl, true);
return;
}
need_unlock = 1;
} else if (WARN_ON_ONCE(ip->i_gl->gl_state != LM_ST_EXCLUSIVE))
return;
if (current->journal_info == NULL) {
ret = gfs2_trans_begin(sdp, RES_DINODE, 0);
if (ret) {
fs_err(sdp, "dirty_inode: gfs2_trans_begin %d\n", ret);
goto out;
}
need_endtrans = 1;
}
ret = gfs2_meta_inode_buffer(ip, &bh);
if (ret == 0) {
gfs2_trans_add_meta(ip->i_gl, bh);
gfs2_dinode_out(ip, bh->b_data);
brelse(bh);
}
if (need_endtrans)
gfs2_trans_end(sdp);
out:
if (need_unlock)
gfs2_glock_dq_uninit(&gh);
}
/**
* gfs2_make_fs_ro - Turn a Read-Write FS into a Read-Only one
* @sdp: the filesystem
*
* Returns: errno
*/
void gfs2_make_fs_ro(struct gfs2_sbd *sdp)
{
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
int log_write_allowed = test_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);
gfs2_flush_delete_work(sdp);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (!log_write_allowed && current == sdp->sd_quotad_process)
fs_warn(sdp, "The quotad daemon is withdrawing.\n");
else if (sdp->sd_quotad_process)
kthread_stop(sdp->sd_quotad_process);
sdp->sd_quotad_process = NULL;
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (!log_write_allowed && current == sdp->sd_logd_process)
fs_warn(sdp, "The logd daemon is withdrawing.\n");
else if (sdp->sd_logd_process)
kthread_stop(sdp->sd_logd_process);
sdp->sd_logd_process = NULL;
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (log_write_allowed) {
gfs2_quota_sync(sdp->sd_vfs, 0);
gfs2_statfs_sync(sdp->sd_vfs, 0);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
gfs2_log_flush(sdp, NULL, GFS2_LOG_HEAD_FLUSH_SHUTDOWN |
GFS2_LFC_MAKE_FS_RO);
wait_event_timeout(sdp->sd_log_waitq,
gfs2_log_is_empty(sdp),
HZ * 5);
gfs2_assert_warn(sdp, gfs2_log_is_empty(sdp));
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
} else {
wait_event_timeout(sdp->sd_log_waitq,
gfs2_log_is_empty(sdp),
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
HZ * 5);
}
gfs2_quota_cleanup(sdp);
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (!log_write_allowed)
sdp->sd_vfs->s_flags |= SB_RDONLY;
}
/**
* gfs2_put_super - Unmount the filesystem
* @sb: The VFS superblock
*
*/
static void gfs2_put_super(struct super_block *sb)
{
struct gfs2_sbd *sdp = sb->s_fs_info;
struct gfs2_jdesc *jd;
/* No more recovery requests */
set_bit(SDF_NORECOVERY, &sdp->sd_flags);
smp_mb();
/* Wait on outstanding recovery */
restart:
spin_lock(&sdp->sd_jindex_spin);
list_for_each_entry(jd, &sdp->sd_jindex_list, jd_list) {
if (!test_bit(JDF_RECOVERY, &jd->jd_flags))
continue;
spin_unlock(&sdp->sd_jindex_spin);
wait_on_bit(&jd->jd_flags, JDF_RECOVERY,
sched: Remove proliferation of wait_on_bit() action functions The current "wait_on_bit" interface requires an 'action' function to be provided which does the actual waiting. There are over 20 such functions, many of them identical. Most cases can be satisfied by one of just two functions, one which uses io_schedule() and one which just uses schedule(). So: Rename wait_on_bit and wait_on_bit_lock to wait_on_bit_action and wait_on_bit_lock_action to make it explicit that they need an action function. Introduce new wait_on_bit{,_lock} and wait_on_bit{,_lock}_io which are *not* given an action function but implicitly use a standard one. The decision to error-out if a signal is pending is now made based on the 'mode' argument rather than being encoded in the action function. All instances of the old wait_on_bit and wait_on_bit_lock which can use the new version have been changed accordingly and their action functions have been discarded. wait_on_bit{_lock} does not return any specific error code in the event of a signal so the caller must check for non-zero and interpolate their own error code as appropriate. The wait_on_bit() call in __fscache_wait_on_invalidate() was ambiguous as it specified TASK_UNINTERRUPTIBLE but used fscache_wait_bit_interruptible as an action function. David Howells confirms this should be uniformly "uninterruptible" The main remaining user of wait_on_bit{,_lock}_action is NFS which needs to use a freezer-aware schedule() call. A comment in fs/gfs2/glock.c notes that having multiple 'action' functions is useful as they display differently in the 'wchan' field of 'ps'. (and /proc/$PID/wchan). As the new bit_wait{,_io} functions are tagged "__sched", they will not show up at all, but something higher in the stack. So the distinction will still be visible, only with different function names (gds2_glock_wait versus gfs2_glock_dq_wait in the gfs2/glock.c case). Since first version of this patch (against 3.15) two new action functions appeared, on in NFS and one in CIFS. CIFS also now uses an action function that makes the same freezer aware schedule call as NFS. Signed-off-by: NeilBrown <neilb@suse.de> Acked-by: David Howells <dhowells@redhat.com> (fscache, keys) Acked-by: Steven Whitehouse <swhiteho@redhat.com> (gfs2) Acked-by: Peter Zijlstra <peterz@infradead.org> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Steve French <sfrench@samba.org> Cc: Linus Torvalds <torvalds@linux-foundation.org> Link: http://lkml.kernel.org/r/20140707051603.28027.72349.stgit@notabene.brown Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-07-07 05:16:04 +00:00
TASK_UNINTERRUPTIBLE);
goto restart;
}
spin_unlock(&sdp->sd_jindex_spin);
if (!sb_rdonly(sb)) {
gfs2_make_fs_ro(sdp);
}
WARN_ON(gfs2_withdrawing(sdp));
/* At this point, we're through modifying the disk */
/* Release stuff */
iput(sdp->sd_jindex);
iput(sdp->sd_statfs_inode);
iput(sdp->sd_rindex);
iput(sdp->sd_quota_inode);
gfs2_glock_put(sdp->sd_rename_gl);
GFS2: remove transaction glock GFS2 has a transaction glock, which must be grabbed for every transaction, whose purpose is to deal with freezing the filesystem. Aside from this involving a large amount of locking, it is very easy to make the current fsfreeze code hang on unfreezing. This patch rewrites how gfs2 handles freezing the filesystem. The transaction glock is removed. In it's place is a freeze glock, which is cached (but not held) in a shared state by every node in the cluster when the filesystem is mounted. This lock only needs to be grabbed on freezing, and actions which need to be safe from freezing, like recovery. When a node wants to freeze the filesystem, it grabs this glock exclusively. When the freeze glock state changes on the nodes (either from shared to unlocked, or shared to exclusive), the filesystem does a special log flush. gfs2_log_flush() does all the work for flushing out the and shutting down the incore log, and then it tries to grab the freeze glock in a shared state again. Since the filesystem is stuck in gfs2_log_flush, no new transaction can start, and nothing can be written to disk. Unfreezing the filesytem simply involes dropping the freeze glock, allowing gfs2_log_flush() to grab and then release the shared lock, so it is cached for next time. However, in order for the unfreezing ioctl to occur, gfs2 needs to get a shared lock on the filesystem root directory inode to check permissions. If that glock has already been grabbed exclusively, fsfreeze will be unable to get the shared lock and unfreeze the filesystem. In order to allow the unfreeze, this patch makes gfs2 grab a shared lock on the filesystem root directory during the freeze, and hold it until it unfreezes the filesystem. The functions which need to grab a shared lock in order to allow the unfreeze ioctl to be issued now use the lock grabbed by the freeze code instead. The freeze and unfreeze code take care to make sure that this shared lock will not be dropped while another process is using it. Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>
2014-05-02 03:26:55 +00:00
gfs2_glock_put(sdp->sd_freeze_gl);
if (!sdp->sd_args.ar_spectator) {
gfs2: Force withdraw to replay journals and wait for it to finish When a node withdraws from a file system, it often leaves its journal in an incomplete state. This is especially true when the withdraw is caused by io errors writing to the journal. Before this patch, a withdraw would try to write a "shutdown" record to the journal, tell dlm it's done with the file system, and none of the other nodes know about the problem. Later, when the problem is fixed and the withdrawn node is rebooted, it would then discover that its own journal was incomplete, and replay it. However, replaying it at this point is almost guaranteed to introduce corruption because the other nodes are likely to have used affected resource groups that appeared in the journal since the time of the withdraw. Replaying the journal later will overwrite any changes made, and not through any fault of dlm, which was instructed during the withdraw to release those resources. This patch makes file system withdraws seen by the entire cluster. Withdrawing nodes dequeue their journal glock to allow recovery. The remaining nodes check all the journals to see if they are clean or in need of replay. They try to replay dirty journals, but only the journals of withdrawn nodes will be "not busy" and therefore available for replay. Until the journal replay is complete, no i/o related glocks may be given out, to ensure that the replay does not cause the aforementioned corruption: We cannot allow any journal replay to overwrite blocks associated with a glock once it is held. The "live" glock which is now used to signal when a withdraw occurs. When a withdraw occurs, the node signals its withdraw by dequeueing the "live" glock and trying to enqueue it in EX mode, thus forcing the other nodes to all see a demote request, by way of a "1CB" (one callback) try lock. The "live" glock is not granted in EX; the callback is only just used to indicate a withdraw has occurred. Note that all nodes in the cluster must wait for the recovering node to finish replaying the withdrawing node's journal before continuing. To this end, it checks that the journals are clean multiple times in a retry loop. Also note that the withdraw function may be called from a wide variety of situations, and therefore, we need to take extra precautions to make sure pointers are valid before using them in many circumstances. We also need to take care when glocks decide to withdraw, since the withdraw code now uses glocks. Also, before this patch, if a process encountered an error and decided to withdraw, if another process was already withdrawing, the second withdraw would be silently ignored, which set it free to unlock its glocks. That's correct behavior if the original withdrawer encounters further errors down the road. But if secondary waiters don't wait for the journal replay, unlocking glocks will allow other nodes to use them, despite the fact that the journal containing those blocks is being replayed. The replay needs to finish before our glocks are released to other nodes. IOW, secondary withdraws need to wait for the first withdraw to finish. For example, if an rgrp glock is unlocked by a process that didn't wait for the first withdraw, a journal replay could introduce file system corruption by replaying a rgrp block that has already been granted to a different cluster node. Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2020-01-28 19:23:45 +00:00
if (gfs2_holder_initialized(&sdp->sd_journal_gh))
gfs2_glock_dq_uninit(&sdp->sd_journal_gh);
if (gfs2_holder_initialized(&sdp->sd_jinode_gh))
gfs2_glock_dq_uninit(&sdp->sd_jinode_gh);
gfs2_glock_dq_uninit(&sdp->sd_sc_gh);
gfs2_glock_dq_uninit(&sdp->sd_qc_gh);
free_local_statfs_inodes(sdp);
iput(sdp->sd_qc_inode);
}
gfs2_glock_dq_uninit(&sdp->sd_live_gh);
gfs2_clear_rgrpd(sdp);
gfs2_jindex_free(sdp);
/* Take apart glock structures and buffer lists */
gfs2_gl_hash_clear(sdp);
truncate_inode_pages_final(&sdp->sd_aspace);
gfs2_delete_debugfs_file(sdp);
/* Unmount the locking protocol */
gfs2_lm_unmount(sdp);
/* At this point, we're through participating in the lockspace */
gfs2_sys_fs_del(sdp);
gfs2: use-after-free in sysfs deregistration syzkaller found the following splat with CONFIG_DEBUG_KOBJECT_RELEASE=y: Read of size 1 at addr ffff000028e896b8 by task kworker/1:2/228 CPU: 1 PID: 228 Comm: kworker/1:2 Tainted: G S 5.9.0-rc8+ #101 Hardware name: linux,dummy-virt (DT) Workqueue: events kobject_delayed_cleanup Call trace: dump_backtrace+0x0/0x4d8 show_stack+0x34/0x48 dump_stack+0x174/0x1f8 print_address_description.constprop.0+0x5c/0x550 kasan_report+0x13c/0x1c0 __asan_report_load1_noabort+0x34/0x60 memcmp+0xd0/0xd8 gfs2_uevent+0xc4/0x188 kobject_uevent_env+0x54c/0x1240 kobject_uevent+0x2c/0x40 __kobject_del+0x190/0x1d8 kobject_delayed_cleanup+0x2bc/0x3b8 process_one_work+0x96c/0x18c0 worker_thread+0x3f0/0xc30 kthread+0x390/0x498 ret_from_fork+0x10/0x18 Allocated by task 1110: kasan_save_stack+0x28/0x58 __kasan_kmalloc.isra.0+0xc8/0xe8 kasan_kmalloc+0x10/0x20 kmem_cache_alloc_trace+0x1d8/0x2f0 alloc_super+0x64/0x8c0 sget_fc+0x110/0x620 get_tree_bdev+0x190/0x648 gfs2_get_tree+0x50/0x228 vfs_get_tree+0x84/0x2e8 path_mount+0x1134/0x1da8 do_mount+0x124/0x138 __arm64_sys_mount+0x164/0x238 el0_svc_common.constprop.0+0x15c/0x598 do_el0_svc+0x60/0x150 el0_svc+0x34/0xb0 el0_sync_handler+0xc8/0x5b4 el0_sync+0x15c/0x180 Freed by task 228: kasan_save_stack+0x28/0x58 kasan_set_track+0x28/0x40 kasan_set_free_info+0x24/0x48 __kasan_slab_free+0x118/0x190 kasan_slab_free+0x14/0x20 slab_free_freelist_hook+0x6c/0x210 kfree+0x13c/0x460 Use the same pattern as f2fs + ext4 where the kobject destruction must complete before allowing the FS itself to be freed. This means that we need an explicit free_sbd in the callers. Cc: Bob Peterson <rpeterso@redhat.com> Cc: Andreas Gruenbacher <agruenba@redhat.com> Signed-off-by: Jamie Iles <jamie@nuviainc.com> [Also go to fail_free when init_names fails.] Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
2020-10-12 13:13:09 +00:00
free_sbd(sdp);
}
/**
* gfs2_sync_fs - sync the filesystem
* @sb: the superblock
* @wait: true to wait for completion
*
* Flushes the log to disk.
*/
static int gfs2_sync_fs(struct super_block *sb, int wait)
{
struct gfs2_sbd *sdp = sb->s_fs_info;
gfs2_quota_sync(sb, -1);
if (wait)
gfs2_log_flush(sdp, NULL, GFS2_LOG_HEAD_FLUSH_NORMAL |
GFS2_LFC_SYNC_FS);
return sdp->sd_log_error;
}
void gfs2_freeze_func(struct work_struct *work)
{
int error;
struct gfs2_holder freeze_gh;
struct gfs2_sbd *sdp = container_of(work, struct gfs2_sbd, sd_freeze_work);
struct super_block *sb = sdp->sd_vfs;
atomic_inc(&sb->s_active);
error = gfs2_freeze_lock(sdp, &freeze_gh, 0);
if (error) {
gfs2_assert_withdraw(sdp, 0);
} else {
atomic_set(&sdp->sd_freeze_state, SFS_UNFROZEN);
error = thaw_super(sb);
if (error) {
fs_info(sdp, "GFS2: couldn't thaw filesystem: %d\n",
error);
gfs2_assert_withdraw(sdp, 0);
}
gfs2_freeze_unlock(&freeze_gh);
}
deactivate_super(sb);
clear_bit_unlock(SDF_FS_FROZEN, &sdp->sd_flags);
wake_up_bit(&sdp->sd_flags, SDF_FS_FROZEN);
return;
}
/**
* gfs2_freeze - prevent further writes to the filesystem
* @sb: the VFS structure for the filesystem
*
*/
static int gfs2_freeze(struct super_block *sb)
{
struct gfs2_sbd *sdp = sb->s_fs_info;
int error;
mutex_lock(&sdp->sd_freeze_mutex);
if (atomic_read(&sdp->sd_freeze_state) != SFS_UNFROZEN) {
error = -EBUSY;
goto out;
}
for (;;) {
if (gfs2_withdrawn(sdp)) {
error = -EINVAL;
goto out;
}
error = gfs2_lock_fs_check_clean(sdp);
if (!error)
break;
if (error == -EBUSY)
fs_err(sdp, "waiting for recovery before freeze\n");
else if (error == -EIO) {
fs_err(sdp, "Fatal IO error: cannot freeze gfs2 due "
"to recovery error.\n");
goto out;
} else {
fs_err(sdp, "error freezing FS: %d\n", error);
}
fs_err(sdp, "retrying...\n");
msleep(1000);
}
set_bit(SDF_FS_FROZEN, &sdp->sd_flags);
out:
mutex_unlock(&sdp->sd_freeze_mutex);
return error;
}
/**
* gfs2_unfreeze - reallow writes to the filesystem
* @sb: the VFS structure for the filesystem
*
*/
static int gfs2_unfreeze(struct super_block *sb)
{
struct gfs2_sbd *sdp = sb->s_fs_info;
mutex_lock(&sdp->sd_freeze_mutex);
if (atomic_read(&sdp->sd_freeze_state) != SFS_FROZEN ||
!gfs2_holder_initialized(&sdp->sd_freeze_gh)) {
mutex_unlock(&sdp->sd_freeze_mutex);
return -EINVAL;
}
gfs2_freeze_unlock(&sdp->sd_freeze_gh);
mutex_unlock(&sdp->sd_freeze_mutex);
return wait_on_bit(&sdp->sd_flags, SDF_FS_FROZEN, TASK_INTERRUPTIBLE);
}
/**
* statfs_slow_fill - fill in the sg for a given RG
* @rgd: the RG
* @sc: the sc structure
*
* Returns: 0 on success, -ESTALE if the LVB is invalid
*/
static int statfs_slow_fill(struct gfs2_rgrpd *rgd,
struct gfs2_statfs_change_host *sc)
{
gfs2_rgrp_verify(rgd);
sc->sc_total += rgd->rd_data;
sc->sc_free += rgd->rd_free;
sc->sc_dinodes += rgd->rd_dinodes;
return 0;
}
/**
* gfs2_statfs_slow - Stat a filesystem using asynchronous locking
* @sdp: the filesystem
* @sc: the sc info that will be returned
*
* Any error (other than a signal) will cause this routine to fall back
* to the synchronous version.
*
* FIXME: This really shouldn't busy wait like this.
*
* Returns: errno
*/
static int gfs2_statfs_slow(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
{
struct gfs2_rgrpd *rgd_next;
struct gfs2_holder *gha, *gh;
unsigned int slots = 64;
unsigned int x;
int done;
int error = 0, err;
memset(sc, 0, sizeof(struct gfs2_statfs_change_host));
treewide: kmalloc() -> kmalloc_array() The kmalloc() function has a 2-factor argument form, kmalloc_array(). This patch replaces cases of: kmalloc(a * b, gfp) with: kmalloc_array(a * b, gfp) as well as handling cases of: kmalloc(a * b * c, gfp) with: kmalloc(array3_size(a, b, c), gfp) as it's slightly less ugly than: kmalloc_array(array_size(a, b), c, gfp) This does, however, attempt to ignore constant size factors like: kmalloc(4 * 1024, gfp) though any constants defined via macros get caught up in the conversion. Any factors with a sizeof() of "unsigned char", "char", and "u8" were dropped, since they're redundant. The tools/ directory was manually excluded, since it has its own implementation of kmalloc(). The Coccinelle script used for this was: // Fix redundant parens around sizeof(). @@ type TYPE; expression THING, E; @@ ( kmalloc( - (sizeof(TYPE)) * E + sizeof(TYPE) * E , ...) | kmalloc( - (sizeof(THING)) * E + sizeof(THING) * E , ...) ) // Drop single-byte sizes and redundant parens. @@ expression COUNT; typedef u8; typedef __u8; @@ ( kmalloc( - sizeof(u8) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(__u8) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(char) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(unsigned char) * (COUNT) + COUNT , ...) | kmalloc( - sizeof(u8) * COUNT + COUNT , ...) | kmalloc( - sizeof(__u8) * COUNT + COUNT , ...) | kmalloc( - sizeof(char) * COUNT + COUNT , ...) | kmalloc( - sizeof(unsigned char) * COUNT + COUNT , ...) ) // 2-factor product with sizeof(type/expression) and identifier or constant. @@ type TYPE; expression THING; identifier COUNT_ID; constant COUNT_CONST; @@ ( - kmalloc + kmalloc_array ( - sizeof(TYPE) * (COUNT_ID) + COUNT_ID, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * COUNT_ID + COUNT_ID, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * (COUNT_CONST) + COUNT_CONST, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * COUNT_CONST + COUNT_CONST, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (COUNT_ID) + COUNT_ID, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * COUNT_ID + COUNT_ID, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (COUNT_CONST) + COUNT_CONST, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * COUNT_CONST + COUNT_CONST, sizeof(THING) , ...) ) // 2-factor product, only identifiers. @@ identifier SIZE, COUNT; @@ - kmalloc + kmalloc_array ( - SIZE * COUNT + COUNT, SIZE , ...) // 3-factor product with 1 sizeof(type) or sizeof(expression), with // redundant parens removed. @@ expression THING; identifier STRIDE, COUNT; type TYPE; @@ ( kmalloc( - sizeof(TYPE) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(TYPE) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(TYPE)) , ...) | kmalloc( - sizeof(THING) * (COUNT) * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * (COUNT) * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * COUNT * (STRIDE) + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) | kmalloc( - sizeof(THING) * COUNT * STRIDE + array3_size(COUNT, STRIDE, sizeof(THING)) , ...) ) // 3-factor product with 2 sizeof(variable), with redundant parens removed. @@ expression THING1, THING2; identifier COUNT; type TYPE1, TYPE2; @@ ( kmalloc( - sizeof(TYPE1) * sizeof(TYPE2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(TYPE2)) , ...) | kmalloc( - sizeof(THING1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | kmalloc( - sizeof(THING1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(THING1), sizeof(THING2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * COUNT + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) | kmalloc( - sizeof(TYPE1) * sizeof(THING2) * (COUNT) + array3_size(COUNT, sizeof(TYPE1), sizeof(THING2)) , ...) ) // 3-factor product, only identifiers, with redundant parens removed. @@ identifier STRIDE, SIZE, COUNT; @@ ( kmalloc( - (COUNT) * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * (STRIDE) * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * STRIDE * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - (COUNT) * (STRIDE) * (SIZE) + array3_size(COUNT, STRIDE, SIZE) , ...) | kmalloc( - COUNT * STRIDE * SIZE + array3_size(COUNT, STRIDE, SIZE) , ...) ) // Any remaining multi-factor products, first at least 3-factor products, // when they're not all constants... @@ expression E1, E2, E3; constant C1, C2, C3; @@ ( kmalloc(C1 * C2 * C3, ...) | kmalloc( - (E1) * E2 * E3 + array3_size(E1, E2, E3) , ...) | kmalloc( - (E1) * (E2) * E3 + array3_size(E1, E2, E3) , ...) | kmalloc( - (E1) * (E2) * (E3) + array3_size(E1, E2, E3) , ...) | kmalloc( - E1 * E2 * E3 + array3_size(E1, E2, E3) , ...) ) // And then all remaining 2 factors products when they're not all constants, // keeping sizeof() as the second factor argument. @@ expression THING, E1, E2; type TYPE; constant C1, C2, C3; @@ ( kmalloc(sizeof(THING) * C2, ...) | kmalloc(sizeof(TYPE) * C2, ...) | kmalloc(C1 * C2 * C3, ...) | kmalloc(C1 * C2, ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * (E2) + E2, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(TYPE) * E2 + E2, sizeof(TYPE) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * (E2) + E2, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - sizeof(THING) * E2 + E2, sizeof(THING) , ...) | - kmalloc + kmalloc_array ( - (E1) * E2 + E1, E2 , ...) | - kmalloc + kmalloc_array ( - (E1) * (E2) + E1, E2 , ...) | - kmalloc + kmalloc_array ( - E1 * E2 + E1, E2 , ...) ) Signed-off-by: Kees Cook <keescook@chromium.org>
2018-06-12 20:55:00 +00:00
gha = kmalloc_array(slots, sizeof(struct gfs2_holder), GFP_KERNEL);
if (!gha)
return -ENOMEM;
for (x = 0; x < slots; x++)
gfs2_holder_mark_uninitialized(gha + x);
rgd_next = gfs2_rgrpd_get_first(sdp);
for (;;) {
done = 1;
for (x = 0; x < slots; x++) {
gh = gha + x;
if (gfs2_holder_initialized(gh) && gfs2_glock_poll(gh)) {
err = gfs2_glock_wait(gh);
if (err) {
gfs2_holder_uninit(gh);
error = err;
} else {
if (!error) {
struct gfs2_rgrpd *rgd =
gfs2_glock2rgrp(gh->gh_gl);
error = statfs_slow_fill(rgd, sc);
}
gfs2_glock_dq_uninit(gh);
}
}
if (gfs2_holder_initialized(gh))
done = 0;
else if (rgd_next && !error) {
error = gfs2_glock_nq_init(rgd_next->rd_gl,
LM_ST_SHARED,
GL_ASYNC,
gh);
rgd_next = gfs2_rgrpd_get_next(rgd_next);
done = 0;
}
if (signal_pending(current))
error = -ERESTARTSYS;
}
if (done)
break;
yield();
}
kfree(gha);
return error;
}
/**
* gfs2_statfs_i - Do a statfs
* @sdp: the filesystem
* @sc: the sc structure
*
* Returns: errno
*/
static int gfs2_statfs_i(struct gfs2_sbd *sdp, struct gfs2_statfs_change_host *sc)
{
struct gfs2_statfs_change_host *m_sc = &sdp->sd_statfs_master;
struct gfs2_statfs_change_host *l_sc = &sdp->sd_statfs_local;
spin_lock(&sdp->sd_statfs_spin);
*sc = *m_sc;
sc->sc_total += l_sc->sc_total;
sc->sc_free += l_sc->sc_free;
sc->sc_dinodes += l_sc->sc_dinodes;
spin_unlock(&sdp->sd_statfs_spin);
if (sc->sc_free < 0)
sc->sc_free = 0;
if (sc->sc_free > sc->sc_total)
sc->sc_free = sc->sc_total;
if (sc->sc_dinodes < 0)
sc->sc_dinodes = 0;
return 0;
}
/**
* gfs2_statfs - Gather and return stats about the filesystem
* @dentry: The name of the link
* @buf: The buffer
*
* Returns: 0 on success or error code
*/
static int gfs2_statfs(struct dentry *dentry, struct kstatfs *buf)
{
struct super_block *sb = dentry->d_sb;
struct gfs2_sbd *sdp = sb->s_fs_info;
struct gfs2_statfs_change_host sc;
int error;
error = gfs2_rindex_update(sdp);
if (error)
return error;
if (gfs2_tune_get(sdp, gt_statfs_slow))
error = gfs2_statfs_slow(sdp, &sc);
else
error = gfs2_statfs_i(sdp, &sc);
if (error)
return error;
buf->f_type = GFS2_MAGIC;
buf->f_bsize = sdp->sd_sb.sb_bsize;
buf->f_blocks = sc.sc_total;
buf->f_bfree = sc.sc_free;
buf->f_bavail = sc.sc_free;
buf->f_files = sc.sc_dinodes + sc.sc_free;
buf->f_ffree = sc.sc_free;
buf->f_namelen = GFS2_FNAMESIZE;
return 0;
}
/**
* gfs2_drop_inode - Drop an inode (test for remote unlink)
* @inode: The inode to drop
*
* If we've received a callback on an iopen lock then it's because a
* remote node tried to deallocate the inode but failed due to this node
* still having the inode open. Here we mark the link count zero
* since we know that it must have reached zero if the GLF_DEMOTE flag
* is set on the iopen glock. If we didn't do a disk read since the
* remote node removed the final link then we might otherwise miss
* this event. This check ensures that this node will deallocate the
* inode's blocks, or alternatively pass the baton on to another
* node for later deallocation.
*/
static int gfs2_drop_inode(struct inode *inode)
{
struct gfs2_inode *ip = GFS2_I(inode);
if (!test_bit(GIF_FREE_VFS_INODE, &ip->i_flags) &&
inode->i_nlink &&
gfs2_holder_initialized(&ip->i_iopen_gh)) {
struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
if (test_bit(GLF_DEMOTE, &gl->gl_flags))
clear_nlink(inode);
}
/*
* When under memory pressure when an inode's link count has dropped to
* zero, defer deleting the inode to the delete workqueue. This avoids
* calling into DLM under memory pressure, which can deadlock.
*/
if (!inode->i_nlink &&
unlikely(current->flags & PF_MEMALLOC) &&
gfs2_holder_initialized(&ip->i_iopen_gh)) {
struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
gfs2_glock_hold(gl);
if (!gfs2_queue_delete_work(gl, 0))
gfs2_glock_queue_put(gl);
return false;
}
return generic_drop_inode(inode);
}
static int is_ancestor(const struct dentry *d1, const struct dentry *d2)
{
do {
if (d1 == d2)
return 1;
d1 = d1->d_parent;
} while (!IS_ROOT(d1));
return 0;
}
/**
* gfs2_show_options - Show mount options for /proc/mounts
* @s: seq_file structure
* @root: root of this (sub)tree
*
* Returns: 0 on success or error code
*/
static int gfs2_show_options(struct seq_file *s, struct dentry *root)
{
struct gfs2_sbd *sdp = root->d_sb->s_fs_info;
struct gfs2_args *args = &sdp->sd_args;
int val;
if (is_ancestor(root, sdp->sd_master_dir))
seq_puts(s, ",meta");
if (args->ar_lockproto[0])
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:44:57 +00:00
seq_show_option(s, "lockproto", args->ar_lockproto);
if (args->ar_locktable[0])
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:44:57 +00:00
seq_show_option(s, "locktable", args->ar_locktable);
if (args->ar_hostdata[0])
fs: create and use seq_show_option for escaping Many file systems that implement the show_options hook fail to correctly escape their output which could lead to unescaped characters (e.g. new lines) leaking into /proc/mounts and /proc/[pid]/mountinfo files. This could lead to confusion, spoofed entries (resulting in things like systemd issuing false d-bus "mount" notifications), and who knows what else. This looks like it would only be the root user stepping on themselves, but it's possible weird things could happen in containers or in other situations with delegated mount privileges. Here's an example using overlay with setuid fusermount trusting the contents of /proc/mounts (via the /etc/mtab symlink). Imagine the use of "sudo" is something more sneaky: $ BASE="ovl" $ MNT="$BASE/mnt" $ LOW="$BASE/lower" $ UP="$BASE/upper" $ WORK="$BASE/work/ 0 0 none /proc fuse.pwn user_id=1000" $ mkdir -p "$LOW" "$UP" "$WORK" $ sudo mount -t overlay -o "lowerdir=$LOW,upperdir=$UP,workdir=$WORK" none /mnt $ cat /proc/mounts none /root/ovl/mnt overlay rw,relatime,lowerdir=ovl/lower,upperdir=ovl/upper,workdir=ovl/work/ 0 0 none /proc fuse.pwn user_id=1000 0 0 $ fusermount -u /proc $ cat /proc/mounts cat: /proc/mounts: No such file or directory This fixes the problem by adding new seq_show_option and seq_show_option_n helpers, and updating the vulnerable show_option handlers to use them as needed. Some, like SELinux, need to be open coded due to unusual existing escape mechanisms. [akpm@linux-foundation.org: add lost chunk, per Kees] [keescook@chromium.org: seq_show_option should be using const parameters] Signed-off-by: Kees Cook <keescook@chromium.org> Acked-by: Serge Hallyn <serge.hallyn@canonical.com> Acked-by: Jan Kara <jack@suse.com> Acked-by: Paul Moore <paul@paul-moore.com> Cc: J. R. Okajima <hooanon05g@gmail.com> Signed-off-by: Kees Cook <keescook@chromium.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:44:57 +00:00
seq_show_option(s, "hostdata", args->ar_hostdata);
if (args->ar_spectator)
seq_puts(s, ",spectator");
if (args->ar_localflocks)
seq_puts(s, ",localflocks");
if (args->ar_debug)
seq_puts(s, ",debug");
if (args->ar_posix_acl)
seq_puts(s, ",acl");
if (args->ar_quota != GFS2_QUOTA_DEFAULT) {
char *state;
switch (args->ar_quota) {
case GFS2_QUOTA_OFF:
state = "off";
break;
case GFS2_QUOTA_ACCOUNT:
state = "account";
break;
case GFS2_QUOTA_ON:
state = "on";
break;
default:
state = "unknown";
break;
}
seq_printf(s, ",quota=%s", state);
}
if (args->ar_suiddir)
seq_puts(s, ",suiddir");
if (args->ar_data != GFS2_DATA_DEFAULT) {
char *state;
switch (args->ar_data) {
case GFS2_DATA_WRITEBACK:
state = "writeback";
break;
case GFS2_DATA_ORDERED:
state = "ordered";
break;
default:
state = "unknown";
break;
}
seq_printf(s, ",data=%s", state);
}
if (args->ar_discard)
seq_puts(s, ",discard");
val = sdp->sd_tune.gt_logd_secs;
if (val != 30)
seq_printf(s, ",commit=%d", val);
val = sdp->sd_tune.gt_statfs_quantum;
if (val != 30)
seq_printf(s, ",statfs_quantum=%d", val);
else if (sdp->sd_tune.gt_statfs_slow)
seq_puts(s, ",statfs_quantum=0");
val = sdp->sd_tune.gt_quota_quantum;
if (val != 60)
seq_printf(s, ",quota_quantum=%d", val);
if (args->ar_statfs_percent)
seq_printf(s, ",statfs_percent=%d", args->ar_statfs_percent);
if (args->ar_errors != GFS2_ERRORS_DEFAULT) {
const char *state;
switch (args->ar_errors) {
case GFS2_ERRORS_WITHDRAW:
state = "withdraw";
break;
case GFS2_ERRORS_PANIC:
state = "panic";
break;
default:
state = "unknown";
break;
}
seq_printf(s, ",errors=%s", state);
}
if (test_bit(SDF_NOBARRIERS, &sdp->sd_flags))
seq_puts(s, ",nobarrier");
if (test_bit(SDF_DEMOTE, &sdp->sd_flags))
seq_puts(s, ",demote_interface_used");
if (args->ar_rgrplvb)
seq_puts(s, ",rgrplvb");
gfs2: change gfs2 readdir cookie gfs2 currently returns 31 bits of filename hash as a cookie that readdir uses for an offset into the directory. When there are a large number of directory entries, the likelihood of a collision goes up way too quickly. GFS2 will now return cookies that are guaranteed unique for a while, and then fail back to using 30 bits of filename hash. Specifically, the directory leaf blocks are divided up into chunks based on the minimum size of a gfs2 directory entry (48 bytes). Each entry's cookie is based off the chunk where it starts, in the linked list of leaf blocks that it hashes to (there are 131072 hash buckets). Directory entries will have unique names until they take reach chunk 8192. Assuming the largest filenames possible, and the least efficient spacing possible, this new method will still be able to return unique names when the previous method has statistically more than a 99% chance of a collision. The non-unique names it fails back to are guaranteed to not collide with the unique names. unique cookies will be in this format: - 1 bit "0" to make sure the the returned cookie is positive - 17 bits for the hash table index - 1 bit for the mode "0" - 13 bits for the offset non-unique cookies will be in this format: - 1 bit "0" to make sure the the returned cookie is positive - 17 bits for the hash table index - 1 bit for the mode "1" - 13 more bits of the name hash Another benefit of location based cookies, is that once a directory's exhash table is fully extended (so that multiple hash table indexs do not use the same leaf blocks), gfs2 can skip sorting the directory entries until it reaches the non-unique ones, and then it only needs to sort these. This provides a significant speed up for directory reads of very large directories. The only issue is that for these cookies to continue to point to the correct entry as files are added and removed from the directory, gfs2 must keep the entries at the same offset in the leaf block when they are split (see my previous patch). This means that until all the nodes in a cluster are running with code that will split the directory leaf blocks this way, none of the nodes can use the new cookie code. To deal with this, gfs2 now has the mount option loccookie, which, if set, will make it return these new location based cookies. This option must not be set until all nodes in the cluster are at least running this version of the kernel code, and you have guaranteed that there are no outstanding cookies required by other software, such as NFS. gfs2 uses some of the extra space at the end of the gfs2_dirent structure to store the calculated readdir cookies. This keeps us from needing to allocate a seperate array to hold these values. gfs2 recomputes the cookie stored in de_cookie for every readdir call. The time it takes to do so is small, and if gfs2 expected this value to be saved on disk, the new code wouldn't work correctly on filesystems created with an earlier version of gfs2. One issue with adding de_cookie to the union in the gfs2_dirent structure is that it caused the union to align itself to a 4 byte boundary, instead of its previous 2 byte boundary. This changed the offset of de_rahead. To solve that, I pulled de_rahead out of the union, since it does not need to be there. Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Bob Peterson <rpeterso@redhat.com>
2015-12-01 14:46:55 +00:00
if (args->ar_loccookie)
seq_puts(s, ",loccookie");
return 0;
}
static void gfs2_final_release_pages(struct gfs2_inode *ip)
{
struct inode *inode = &ip->i_inode;
struct gfs2_glock *gl = ip->i_gl;
truncate_inode_pages(gfs2_glock2aspace(ip->i_gl), 0);
truncate_inode_pages(&inode->i_data, 0);
if (atomic_read(&gl->gl_revokes) == 0) {
clear_bit(GLF_LFLUSH, &gl->gl_flags);
clear_bit(GLF_DIRTY, &gl->gl_flags);
}
}
static int gfs2_dinode_dealloc(struct gfs2_inode *ip)
{
struct gfs2_sbd *sdp = GFS2_SB(&ip->i_inode);
struct gfs2_rgrpd *rgd;
struct gfs2_holder gh;
int error;
if (gfs2_get_inode_blocks(&ip->i_inode) != 1) {
gfs2_consist_inode(ip);
return -EIO;
}
error = gfs2_rindex_update(sdp);
if (error)
return error;
error = gfs2_quota_hold(ip, NO_UID_QUOTA_CHANGE, NO_GID_QUOTA_CHANGE);
if (error)
return error;
rgd = gfs2_blk2rgrpd(sdp, ip->i_no_addr, 1);
if (!rgd) {
gfs2_consist_inode(ip);
error = -EIO;
goto out_qs;
}
error = gfs2_glock_nq_init(rgd->rd_gl, LM_ST_EXCLUSIVE,
LM_FLAG_NODE_SCOPE, &gh);
if (error)
goto out_qs;
error = gfs2_trans_begin(sdp, RES_RG_BIT + RES_STATFS + RES_QUOTA,
sdp->sd_jdesc->jd_blocks);
if (error)
goto out_rg_gunlock;
gfs2_free_di(rgd, ip);
gfs2_final_release_pages(ip);
gfs2_trans_end(sdp);
out_rg_gunlock:
gfs2_glock_dq_uninit(&gh);
out_qs:
gfs2_quota_unhold(ip);
return error;
}
/**
* gfs2_glock_put_eventually
* @gl: The glock to put
*
* When under memory pressure, trigger a deferred glock put to make sure we
* won't call into DLM and deadlock. Otherwise, put the glock directly.
*/
static void gfs2_glock_put_eventually(struct gfs2_glock *gl)
{
if (current->flags & PF_MEMALLOC)
gfs2_glock_queue_put(gl);
else
gfs2_glock_put(gl);
}
static bool gfs2_upgrade_iopen_glock(struct inode *inode)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_sbd *sdp = GFS2_SB(inode);
struct gfs2_holder *gh = &ip->i_iopen_gh;
long timeout = 5 * HZ;
int error;
gh->gh_flags |= GL_NOCACHE;
gfs2_glock_dq_wait(gh);
/*
* If there are no other lock holders, we'll get the lock immediately.
* Otherwise, the other nodes holding the lock will be notified about
* our locking request. If they don't have the inode open, they'll
* evict the cached inode and release the lock. Otherwise, if they
* poke the inode glock, we'll take this as an indication that they
* still need the iopen glock and that they'll take care of deleting
* the inode when they're done. As a last resort, if another node
* keeps holding the iopen glock without showing any activity on the
* inode glock, we'll eventually time out.
*
* Note that we're passing the LM_FLAG_TRY_1CB flag to the first
* locking request as an optimization to notify lock holders as soon as
* possible. Without that flag, they'd be notified implicitly by the
* second locking request.
*/
gfs2_holder_reinit(LM_ST_EXCLUSIVE, LM_FLAG_TRY_1CB | GL_NOCACHE, gh);
error = gfs2_glock_nq(gh);
if (error != GLR_TRYFAILED)
return !error;
gfs2_holder_reinit(LM_ST_EXCLUSIVE, GL_ASYNC | GL_NOCACHE, gh);
error = gfs2_glock_nq(gh);
if (error)
return false;
timeout = wait_event_interruptible_timeout(sdp->sd_async_glock_wait,
!test_bit(HIF_WAIT, &gh->gh_iflags) ||
test_bit(GLF_DEMOTE, &ip->i_gl->gl_flags),
timeout);
if (!test_bit(HIF_HOLDER, &gh->gh_iflags)) {
gfs2_glock_dq(gh);
return false;
}
return true;
}
/**
* evict_should_delete - determine whether the inode is eligible for deletion
* @inode: The inode to evict
* @gh: The glock holder structure
*
* This function determines whether the evicted inode is eligible to be deleted
* and locks the inode glock.
*
* Returns: the fate of the dinode
*/
static enum dinode_demise evict_should_delete(struct inode *inode,
struct gfs2_holder *gh)
{
struct gfs2_inode *ip = GFS2_I(inode);
struct super_block *sb = inode->i_sb;
struct gfs2_sbd *sdp = sb->s_fs_info;
int ret;
if (test_bit(GIF_ALLOC_FAILED, &ip->i_flags)) {
BUG_ON(!gfs2_glock_is_locked_by_me(ip->i_gl));
goto should_delete;
}
if (test_bit(GIF_DEFERRED_DELETE, &ip->i_flags))
return SHOULD_DEFER_EVICTION;
/* Deletes should never happen under memory pressure anymore. */
if (WARN_ON_ONCE(current->flags & PF_MEMALLOC))
return SHOULD_DEFER_EVICTION;
/* Must not read inode block until block type has been verified */
ret = gfs2_glock_nq_init(ip->i_gl, LM_ST_EXCLUSIVE, GL_SKIP, gh);
if (unlikely(ret)) {
glock_clear_object(ip->i_iopen_gh.gh_gl, ip);
ip->i_iopen_gh.gh_flags |= GL_NOCACHE;
gfs2_glock_dq_uninit(&ip->i_iopen_gh);
return SHOULD_DEFER_EVICTION;
}
if (gfs2_inode_already_deleted(ip->i_gl, ip->i_no_formal_ino))
return SHOULD_NOT_DELETE_DINODE;
ret = gfs2_check_blk_type(sdp, ip->i_no_addr, GFS2_BLKST_UNLINKED);
if (ret)
return SHOULD_NOT_DELETE_DINODE;
if (test_bit(GIF_INVALID, &ip->i_flags)) {
ret = gfs2_inode_refresh(ip);
if (ret)
return SHOULD_NOT_DELETE_DINODE;
}
/*
* The inode may have been recreated in the meantime.
*/
if (inode->i_nlink)
return SHOULD_NOT_DELETE_DINODE;
should_delete:
if (gfs2_holder_initialized(&ip->i_iopen_gh) &&
test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags)) {
if (!gfs2_upgrade_iopen_glock(inode)) {
gfs2_holder_uninit(&ip->i_iopen_gh);
return SHOULD_NOT_DELETE_DINODE;
}
}
return SHOULD_DELETE_DINODE;
}
/**
* evict_unlinked_inode - delete the pieces of an unlinked evicted inode
* @inode: The inode to evict
*/
static int evict_unlinked_inode(struct inode *inode)
{
struct gfs2_inode *ip = GFS2_I(inode);
int ret;
if (S_ISDIR(inode->i_mode) &&
(ip->i_diskflags & GFS2_DIF_EXHASH)) {
ret = gfs2_dir_exhash_dealloc(ip);
if (ret)
goto out;
}
if (ip->i_eattr) {
ret = gfs2_ea_dealloc(ip);
if (ret)
goto out;
}
if (!gfs2_is_stuffed(ip)) {
ret = gfs2_file_dealloc(ip);
if (ret)
goto out;
}
/* We're about to clear the bitmap for the dinode, but as soon as we
do, gfs2_create_inode can create another inode at the same block
location and try to set gl_object again. We clear gl_object here so
that subsequent inode creates don't see an old gl_object. */
glock_clear_object(ip->i_gl, ip);
ret = gfs2_dinode_dealloc(ip);
gfs2_inode_remember_delete(ip->i_gl, ip->i_no_formal_ino);
out:
return ret;
}
/*
* evict_linked_inode - evict an inode whose dinode has not been unlinked
* @inode: The inode to evict
*/
static int evict_linked_inode(struct inode *inode)
{
struct super_block *sb = inode->i_sb;
struct gfs2_sbd *sdp = sb->s_fs_info;
struct gfs2_inode *ip = GFS2_I(inode);
struct address_space *metamapping;
int ret;
gfs2_log_flush(sdp, ip->i_gl, GFS2_LOG_HEAD_FLUSH_NORMAL |
GFS2_LFC_EVICT_INODE);
metamapping = gfs2_glock2aspace(ip->i_gl);
if (test_bit(GLF_DIRTY, &ip->i_gl->gl_flags)) {
filemap_fdatawrite(metamapping);
filemap_fdatawait(metamapping);
}
write_inode_now(inode, 1);
gfs2_ail_flush(ip->i_gl, 0);
ret = gfs2_trans_begin(sdp, 0, sdp->sd_jdesc->jd_blocks);
if (ret)
return ret;
/* Needs to be done before glock release & also in a transaction */
truncate_inode_pages(&inode->i_data, 0);
truncate_inode_pages(metamapping, 0);
gfs2_trans_end(sdp);
return 0;
}
/**
* gfs2_evict_inode - Remove an inode from cache
* @inode: The inode to evict
*
* There are three cases to consider:
* 1. i_nlink == 0, we are final opener (and must deallocate)
* 2. i_nlink == 0, we are not the final opener (and cannot deallocate)
* 3. i_nlink > 0
*
* If the fs is read only, then we have to treat all cases as per #3
* since we are unable to do any deallocation. The inode will be
* deallocated by the next read/write node to attempt an allocation
* in the same resource group
*
* We have to (at the moment) hold the inodes main lock to cover
* the gap between unlocking the shared lock on the iopen lock and
* taking the exclusive lock. I'd rather do a shared -> exclusive
* conversion on the iopen lock, but we can change that later. This
* is safe, just less efficient.
*/
static void gfs2_evict_inode(struct inode *inode)
{
struct super_block *sb = inode->i_sb;
struct gfs2_sbd *sdp = sb->s_fs_info;
struct gfs2_inode *ip = GFS2_I(inode);
struct gfs2_holder gh;
int ret;
if (test_bit(GIF_FREE_VFS_INODE, &ip->i_flags)) {
clear_inode(inode);
return;
}
if (inode->i_nlink || sb_rdonly(sb))
goto out;
gfs2_holder_mark_uninitialized(&gh);
ret = evict_should_delete(inode, &gh);
if (ret == SHOULD_DEFER_EVICTION)
goto out;
if (ret == SHOULD_DELETE_DINODE)
ret = evict_unlinked_inode(inode);
else
ret = evict_linked_inode(inode);
if (gfs2_rs_active(&ip->i_res))
gfs2_rs_deltree(&ip->i_res);
if (gfs2_holder_initialized(&gh)) {
glock_clear_object(ip->i_gl, ip);
gfs2_glock_dq_uninit(&gh);
}
if (ret && ret != GLR_TRYFAILED && ret != -EROFS)
fs_warn(sdp, "gfs2_evict_inode: %d\n", ret);
out:
mm + fs: store shadow entries in page cache Reclaim will be leaving shadow entries in the page cache radix tree upon evicting the real page. As those pages are found from the LRU, an iput() can lead to the inode being freed concurrently. At this point, reclaim must no longer install shadow pages because the inode freeing code needs to ensure the page tree is really empty. Add an address_space flag, AS_EXITING, that the inode freeing code sets under the tree lock before doing the final truncate. Reclaim will check for this flag before installing shadow pages. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Rik van Riel <riel@redhat.com> Reviewed-by: Minchan Kim <minchan@kernel.org> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Bob Liu <bob.liu@oracle.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Dave Chinner <david@fromorbit.com> Cc: Greg Thelen <gthelen@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: Jan Kara <jack@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Luigi Semenzato <semenzato@google.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Metin Doslu <metin@citusdata.com> Cc: Michel Lespinasse <walken@google.com> Cc: Ozgun Erdogan <ozgun@citusdata.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Roman Gushchin <klamm@yandex-team.ru> Cc: Ryan Mallon <rmallon@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vlastimil Babka <vbabka@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-04-03 21:47:49 +00:00
truncate_inode_pages_final(&inode->i_data);
if (ip->i_qadata)
gfs2_assert_warn(sdp, ip->i_qadata->qa_ref == 0);
gfs2_rs_delete(ip, NULL);
gfs2_ordered_del_inode(ip);
clear_inode(inode);
gfs2_dir_hash_inval(ip);
if (ip->i_gl) {
glock_clear_object(ip->i_gl, ip);
wait_on_bit_io(&ip->i_flags, GIF_GLOP_PENDING, TASK_UNINTERRUPTIBLE);
gfs2_glock_add_to_lru(ip->i_gl);
gfs2_glock_put_eventually(ip->i_gl);
ip->i_gl = NULL;
}
if (gfs2_holder_initialized(&ip->i_iopen_gh)) {
struct gfs2_glock *gl = ip->i_iopen_gh.gh_gl;
glock_clear_object(gl, ip);
if (test_bit(HIF_HOLDER, &ip->i_iopen_gh.gh_iflags)) {
ip->i_iopen_gh.gh_flags |= GL_NOCACHE;
gfs2_glock_dq(&ip->i_iopen_gh);
}
gfs2_glock_hold(gl);
gfs2_holder_uninit(&ip->i_iopen_gh);
gfs2_glock_put_eventually(gl);
}
}
static struct inode *gfs2_alloc_inode(struct super_block *sb)
{
struct gfs2_inode *ip;
ip = kmem_cache_alloc(gfs2_inode_cachep, GFP_KERNEL);
if (!ip)
return NULL;
ip->i_flags = 0;
ip->i_gl = NULL;
gfs2_holder_mark_uninitialized(&ip->i_iopen_gh);
memset(&ip->i_res, 0, sizeof(ip->i_res));
RB_CLEAR_NODE(&ip->i_res.rs_node);
ip->i_rahead = 0;
return &ip->i_inode;
}
static void gfs2_free_inode(struct inode *inode)
{
kmem_cache_free(gfs2_inode_cachep, GFS2_I(inode));
2011-01-07 06:49:49 +00:00
}
extern void free_local_statfs_inodes(struct gfs2_sbd *sdp)
{
struct local_statfs_inode *lsi, *safe;
/* Run through the statfs inodes list to iput and free memory */
list_for_each_entry_safe(lsi, safe, &sdp->sd_sc_inodes_list, si_list) {
if (lsi->si_jid == sdp->sd_jdesc->jd_jid)
sdp->sd_sc_inode = NULL; /* belongs to this node */
if (lsi->si_sc_inode)
iput(lsi->si_sc_inode);
list_del(&lsi->si_list);
kfree(lsi);
}
}
extern struct inode *find_local_statfs_inode(struct gfs2_sbd *sdp,
unsigned int index)
{
struct local_statfs_inode *lsi;
/* Return the local (per node) statfs inode in the
* sdp->sd_sc_inodes_list corresponding to the 'index'. */
list_for_each_entry(lsi, &sdp->sd_sc_inodes_list, si_list) {
if (lsi->si_jid == index)
return lsi->si_sc_inode;
}
return NULL;
}
const struct super_operations gfs2_super_ops = {
.alloc_inode = gfs2_alloc_inode,
.free_inode = gfs2_free_inode,
.write_inode = gfs2_write_inode,
.dirty_inode = gfs2_dirty_inode,
.evict_inode = gfs2_evict_inode,
.put_super = gfs2_put_super,
.sync_fs = gfs2_sync_fs,
.freeze_super = gfs2_freeze,
.thaw_super = gfs2_unfreeze,
.statfs = gfs2_statfs,
.drop_inode = gfs2_drop_inode,
.show_options = gfs2_show_options,
};