linux/fs/ubifs/commit.c

723 lines
20 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* This file is part of UBIFS.
*
* Copyright (C) 2006-2008 Nokia Corporation.
*
* Authors: Adrian Hunter
* Artem Bityutskiy (Битюцкий Артём)
*/
/*
* This file implements functions that manage the running of the commit process.
* Each affected module has its own functions to accomplish their part in the
* commit and those functions are called here.
*
* The commit is the process whereby all updates to the index and LEB properties
* are written out together and the journal becomes empty. This keeps the
* file system consistent - at all times the state can be recreated by reading
* the index and LEB properties and then replaying the journal.
*
* The commit is split into two parts named "commit start" and "commit end".
* During commit start, the commit process has exclusive access to the journal
* by holding the commit semaphore down for writing. As few I/O operations as
* possible are performed during commit start, instead the nodes that are to be
* written are merely identified. During commit end, the commit semaphore is no
* longer held and the journal is again in operation, allowing users to continue
* to use the file system while the bulk of the commit I/O is performed. The
* purpose of this two-step approach is to prevent the commit from causing any
* latency blips. Note that in any case, the commit does not prevent lookups
* (as permitted by the TNC mutex), or access to VFS data structures e.g. page
* cache.
*/
#include <linux/freezer.h>
#include <linux/kthread.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h percpu.h is included by sched.h and module.h and thus ends up being included when building most .c files. percpu.h includes slab.h which in turn includes gfp.h making everything defined by the two files universally available and complicating inclusion dependencies. percpu.h -> slab.h dependency is about to be removed. Prepare for this change by updating users of gfp and slab facilities include those headers directly instead of assuming availability. As this conversion needs to touch large number of source files, the following script is used as the basis of conversion. http://userweb.kernel.org/~tj/misc/slabh-sweep.py The script does the followings. * Scan files for gfp and slab usages and update includes such that only the necessary includes are there. ie. if only gfp is used, gfp.h, if slab is used, slab.h. * When the script inserts a new include, it looks at the include blocks and try to put the new include such that its order conforms to its surrounding. It's put in the include block which contains core kernel includes, in the same order that the rest are ordered - alphabetical, Christmas tree, rev-Xmas-tree or at the end if there doesn't seem to be any matching order. * If the script can't find a place to put a new include (mostly because the file doesn't have fitting include block), it prints out an error message indicating which .h file needs to be added to the file. The conversion was done in the following steps. 1. The initial automatic conversion of all .c files updated slightly over 4000 files, deleting around 700 includes and adding ~480 gfp.h and ~3000 slab.h inclusions. The script emitted errors for ~400 files. 2. Each error was manually checked. Some didn't need the inclusion, some needed manual addition while adding it to implementation .h or embedding .c file was more appropriate for others. This step added inclusions to around 150 files. 3. The script was run again and the output was compared to the edits from #2 to make sure no file was left behind. 4. Several build tests were done and a couple of problems were fixed. e.g. lib/decompress_*.c used malloc/free() wrappers around slab APIs requiring slab.h to be added manually. 5. The script was run on all .h files but without automatically editing them as sprinkling gfp.h and slab.h inclusions around .h files could easily lead to inclusion dependency hell. Most gfp.h inclusion directives were ignored as stuff from gfp.h was usually wildly available and often used in preprocessor macros. Each slab.h inclusion directive was examined and added manually as necessary. 6. percpu.h was updated not to include slab.h. 7. Build test were done on the following configurations and failures were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my distributed build env didn't work with gcov compiles) and a few more options had to be turned off depending on archs to make things build (like ipr on powerpc/64 which failed due to missing writeq). * x86 and x86_64 UP and SMP allmodconfig and a custom test config. * powerpc and powerpc64 SMP allmodconfig * sparc and sparc64 SMP allmodconfig * ia64 SMP allmodconfig * s390 SMP allmodconfig * alpha SMP allmodconfig * um on x86_64 SMP allmodconfig 8. percpu.h modifications were reverted so that it could be applied as a separate patch and serve as bisection point. Given the fact that I had only a couple of failures from tests on step 6, I'm fairly confident about the coverage of this conversion patch. If there is a breakage, it's likely to be something in one of the arch headers which should be easily discoverable easily on most builds of the specific arch. Signed-off-by: Tejun Heo <tj@kernel.org> Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 08:04:11 +00:00
#include <linux/slab.h>
#include "ubifs.h"
/*
* nothing_to_commit - check if there is nothing to commit.
* @c: UBIFS file-system description object
*
* This is a helper function which checks if there is anything to commit. It is
* used as an optimization to avoid starting the commit if it is not really
* necessary. Indeed, the commit operation always assumes flash I/O (e.g.,
* writing the commit start node to the log), and it is better to avoid doing
* this unnecessarily. E.g., 'ubifs_sync_fs()' runs the commit, but if there is
* nothing to commit, it is more optimal to avoid any flash I/O.
*
* This function has to be called with @c->commit_sem locked for writing -
* this function does not take LPT/TNC locks because the @c->commit_sem
* guarantees that we have exclusive access to the TNC and LPT data structures.
*
* This function returns %1 if there is nothing to commit and %0 otherwise.
*/
static int nothing_to_commit(struct ubifs_info *c)
{
/*
* During mounting or remounting from R/O mode to R/W mode we may
* commit for various recovery-related reasons.
*/
if (c->mounting || c->remounting_rw)
return 0;
/*
* If the root TNC node is dirty, we definitely have something to
* commit.
*/
if (c->zroot.znode && ubifs_zn_dirty(c->zroot.znode))
return 0;
/*
* Even though the TNC is clean, the LPT tree may have dirty nodes. For
* example, this may happen if the budgeting subsystem invoked GC to
* make some free space, and the GC found an LEB with only dirty and
* free space. In this case GC would just change the lprops of this
* LEB (by turning all space into free space) and unmap it.
*/
if (c->nroot && test_bit(DIRTY_CNODE, &c->nroot->flags))
return 0;
ubifs_assert(c, atomic_long_read(&c->dirty_zn_cnt) == 0);
ubifs_assert(c, c->dirty_pn_cnt == 0);
ubifs_assert(c, c->dirty_nn_cnt == 0);
return 1;
}
/**
* do_commit - commit the journal.
* @c: UBIFS file-system description object
*
* This function implements UBIFS commit. It has to be called with commit lock
* locked. Returns zero in case of success and a negative error code in case of
* failure.
*/
static int do_commit(struct ubifs_info *c)
{
int err, new_ltail_lnum, old_ltail_lnum, i;
struct ubifs_zbranch zroot;
struct ubifs_lp_stats lst;
dbg_cmt("start");
ubifs_assert(c, !c->ro_media && !c->ro_mount);
if (c->ro_error) {
err = -EROFS;
goto out_up;
}
if (nothing_to_commit(c)) {
up_write(&c->commit_sem);
err = 0;
goto out_cancel;
}
/* Sync all write buffers (necessary for recovery) */
for (i = 0; i < c->jhead_cnt; i++) {
err = ubifs_wbuf_sync(&c->jheads[i].wbuf);
if (err)
goto out_up;
}
c->cmt_no += 1;
err = ubifs_gc_start_commit(c);
if (err)
goto out_up;
err = dbg_check_lprops(c);
if (err)
goto out_up;
err = ubifs_log_start_commit(c, &new_ltail_lnum);
if (err)
goto out_up;
err = ubifs_tnc_start_commit(c, &zroot);
if (err)
goto out_up;
err = ubifs_lpt_start_commit(c);
if (err)
goto out_up;
err = ubifs_orphan_start_commit(c);
if (err)
goto out_up;
ubifs_get_lp_stats(c, &lst);
up_write(&c->commit_sem);
err = ubifs_tnc_end_commit(c);
if (err)
goto out;
err = ubifs_lpt_end_commit(c);
if (err)
goto out;
err = ubifs_orphan_end_commit(c);
if (err)
goto out;
err = dbg_check_old_index(c, &zroot);
if (err)
goto out;
c->mst_node->cmt_no = cpu_to_le64(c->cmt_no);
c->mst_node->log_lnum = cpu_to_le32(new_ltail_lnum);
c->mst_node->root_lnum = cpu_to_le32(zroot.lnum);
c->mst_node->root_offs = cpu_to_le32(zroot.offs);
c->mst_node->root_len = cpu_to_le32(zroot.len);
c->mst_node->ihead_lnum = cpu_to_le32(c->ihead_lnum);
c->mst_node->ihead_offs = cpu_to_le32(c->ihead_offs);
c->mst_node->index_size = cpu_to_le64(c->bi.old_idx_sz);
c->mst_node->lpt_lnum = cpu_to_le32(c->lpt_lnum);
c->mst_node->lpt_offs = cpu_to_le32(c->lpt_offs);
c->mst_node->nhead_lnum = cpu_to_le32(c->nhead_lnum);
c->mst_node->nhead_offs = cpu_to_le32(c->nhead_offs);
c->mst_node->ltab_lnum = cpu_to_le32(c->ltab_lnum);
c->mst_node->ltab_offs = cpu_to_le32(c->ltab_offs);
c->mst_node->lsave_lnum = cpu_to_le32(c->lsave_lnum);
c->mst_node->lsave_offs = cpu_to_le32(c->lsave_offs);
c->mst_node->lscan_lnum = cpu_to_le32(c->lscan_lnum);
c->mst_node->empty_lebs = cpu_to_le32(lst.empty_lebs);
c->mst_node->idx_lebs = cpu_to_le32(lst.idx_lebs);
c->mst_node->total_free = cpu_to_le64(lst.total_free);
c->mst_node->total_dirty = cpu_to_le64(lst.total_dirty);
c->mst_node->total_used = cpu_to_le64(lst.total_used);
c->mst_node->total_dead = cpu_to_le64(lst.total_dead);
c->mst_node->total_dark = cpu_to_le64(lst.total_dark);
if (c->no_orphs)
c->mst_node->flags |= cpu_to_le32(UBIFS_MST_NO_ORPHS);
else
c->mst_node->flags &= ~cpu_to_le32(UBIFS_MST_NO_ORPHS);
UBIFS: fix a race condition Hu (hujianyang@huawei.com) discovered a race condition which may lead to a situation when UBIFS is unable to mount the file-system after an unclean reboot. The problem is theoretical, though. In UBIFS, we have the log, which basically a set of LEBs in a certain area. The log has the tail and the head. Every time user writes data to the file-system, the UBIFS journal grows, and the log grows as well, because we append new reference nodes to the head of the log. So the head moves forward all the time, while the log tail stays at the same position. At any time, the UBIFS master node points to the tail of the log. When we mount the file-system, we scan the log, and we always start from its tail, because this is where the master node points to. The only occasion when the tail of the log changes is the commit operation. The commit operation has 2 phases - "commit start" and "commit end". The former is relatively short, and does not involve much I/O. During this phase we mostly just build various in-memory lists of the things which have to be written to the flash media during "commit end" phase. During the commit start phase, what we do is we "clean" the log. Indeed, the commit operation will index all the data in the journal, so the entire journal "disappears", and therefore the data in the log become unneeded. So we just move the head of the log to the next LEB, and write the CS node there. This LEB will be the tail of the new log when the commit operation finishes. When the "commit start" phase finishes, users may write more data to the file-system, in parallel with the ongoing "commit end" operation. At this point the log tail was not changed yet, it is the same as it had been before we started the commit. The log head keeps moving forward, though. The commit operation now needs to write the new master node, and the new master node should point to the new log tail. After this the LEBs between the old log tail and the new log tail can be unmapped and re-used again. And here is the possible problem. We do 2 operations: (a) We first update the log tail position in memory (see 'ubifs_log_end_commit()'). (b) And then we write the master node (see the big lock of code in 'do_commit()'). But nothing prevents the log head from moving forward between (a) and (b), and the log head may "wrap" now to the old log tail. And when the "wrap" happens, the contends of the log tail gets erased. Now a power cut happens and we are in trouble. We end up with the old master node pointing to the old tail, which was erased. And replay fails because it expects the master node to point to the correct log tail at all times. This patch merges the abovementioned (a) and (b) operations by moving the master node change code to the 'ubifs_log_end_commit()' function, so that it runs with the log mutex locked, which will prevent the log from being changed benween operations (a) and (b). Cc: stable@vger.kernel.org # 07e19df UBIFS: remove mst_mutex Cc: stable@vger.kernel.org Reported-by: hujianyang <hujianyang@huawei.com> Tested-by: hujianyang <hujianyang@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2014-06-29 14:00:45 +00:00
old_ltail_lnum = c->ltail_lnum;
err = ubifs_log_end_commit(c, new_ltail_lnum);
if (err)
goto out;
err = ubifs_log_post_commit(c, old_ltail_lnum);
if (err)
goto out;
err = ubifs_gc_end_commit(c);
if (err)
goto out;
err = ubifs_lpt_post_commit(c);
if (err)
goto out;
out_cancel:
spin_lock(&c->cs_lock);
c->cmt_state = COMMIT_RESTING;
wake_up(&c->cmt_wq);
dbg_cmt("commit end");
spin_unlock(&c->cs_lock);
return 0;
out_up:
up_write(&c->commit_sem);
out:
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_err(c, "commit failed, error %d", err);
spin_lock(&c->cs_lock);
c->cmt_state = COMMIT_BROKEN;
wake_up(&c->cmt_wq);
spin_unlock(&c->cs_lock);
ubifs_ro_mode(c, err);
return err;
}
/**
* run_bg_commit - run background commit if it is needed.
* @c: UBIFS file-system description object
*
* This function runs background commit if it is needed. Returns zero in case
* of success and a negative error code in case of failure.
*/
static int run_bg_commit(struct ubifs_info *c)
{
spin_lock(&c->cs_lock);
/*
* Run background commit only if background commit was requested or if
* commit is required.
*/
if (c->cmt_state != COMMIT_BACKGROUND &&
c->cmt_state != COMMIT_REQUIRED)
goto out;
spin_unlock(&c->cs_lock);
down_write(&c->commit_sem);
spin_lock(&c->cs_lock);
if (c->cmt_state == COMMIT_REQUIRED)
c->cmt_state = COMMIT_RUNNING_REQUIRED;
else if (c->cmt_state == COMMIT_BACKGROUND)
c->cmt_state = COMMIT_RUNNING_BACKGROUND;
else
goto out_cmt_unlock;
spin_unlock(&c->cs_lock);
return do_commit(c);
out_cmt_unlock:
up_write(&c->commit_sem);
out:
spin_unlock(&c->cs_lock);
return 0;
}
/**
* ubifs_bg_thread - UBIFS background thread function.
* @info: points to the file-system description object
*
* This function implements various file-system background activities:
* o when a write-buffer timer expires it synchronizes the appropriate
* write-buffer;
* o when the journal is about to be full, it starts in-advance commit.
*
* Note, other stuff like background garbage collection may be added here in
* future.
*/
int ubifs_bg_thread(void *info)
{
int err;
struct ubifs_info *c = info;
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_msg(c, "background thread \"%s\" started, PID %d",
c->bgt_name, current->pid);
set_freezable();
while (1) {
if (kthread_should_stop())
break;
if (try_to_freeze())
continue;
set_current_state(TASK_INTERRUPTIBLE);
/* Check if there is something to do */
if (!c->need_bgt) {
/*
* Nothing prevents us from going sleep now and
* be never woken up and block the task which
* could wait in 'kthread_stop()' forever.
*/
if (kthread_should_stop())
break;
schedule();
continue;
} else
__set_current_state(TASK_RUNNING);
c->need_bgt = 0;
err = ubifs_bg_wbufs_sync(c);
if (err)
ubifs_ro_mode(c, err);
run_bg_commit(c);
cond_resched();
}
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_msg(c, "background thread \"%s\" stops", c->bgt_name);
return 0;
}
/**
* ubifs_commit_required - set commit state to "required".
* @c: UBIFS file-system description object
*
* This function is called if a commit is required but cannot be done from the
* calling function, so it is just flagged instead.
*/
void ubifs_commit_required(struct ubifs_info *c)
{
spin_lock(&c->cs_lock);
switch (c->cmt_state) {
case COMMIT_RESTING:
case COMMIT_BACKGROUND:
dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
dbg_cstate(COMMIT_REQUIRED));
c->cmt_state = COMMIT_REQUIRED;
break;
case COMMIT_RUNNING_BACKGROUND:
dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
dbg_cstate(COMMIT_RUNNING_REQUIRED));
c->cmt_state = COMMIT_RUNNING_REQUIRED;
break;
case COMMIT_REQUIRED:
case COMMIT_RUNNING_REQUIRED:
case COMMIT_BROKEN:
break;
}
spin_unlock(&c->cs_lock);
}
/**
* ubifs_request_bg_commit - notify the background thread to do a commit.
* @c: UBIFS file-system description object
*
* This function is called if the journal is full enough to make a commit
* worthwhile, so background thread is kicked to start it.
*/
void ubifs_request_bg_commit(struct ubifs_info *c)
{
spin_lock(&c->cs_lock);
if (c->cmt_state == COMMIT_RESTING) {
dbg_cmt("old: %s, new: %s", dbg_cstate(c->cmt_state),
dbg_cstate(COMMIT_BACKGROUND));
c->cmt_state = COMMIT_BACKGROUND;
spin_unlock(&c->cs_lock);
ubifs_wake_up_bgt(c);
} else
spin_unlock(&c->cs_lock);
}
/**
* wait_for_commit - wait for commit.
* @c: UBIFS file-system description object
*
* This function sleeps until the commit operation is no longer running.
*/
static int wait_for_commit(struct ubifs_info *c)
{
dbg_cmt("pid %d goes sleep", current->pid);
/*
* The following sleeps if the condition is false, and will be woken
* when the commit ends. It is possible, although very unlikely, that we
* will wake up and see the subsequent commit running, rather than the
* one we were waiting for, and go back to sleep. However, we will be
* woken again, so there is no danger of sleeping forever.
*/
wait_event(c->cmt_wq, c->cmt_state != COMMIT_RUNNING_BACKGROUND &&
c->cmt_state != COMMIT_RUNNING_REQUIRED);
dbg_cmt("commit finished, pid %d woke up", current->pid);
return 0;
}
/**
* ubifs_run_commit - run or wait for commit.
* @c: UBIFS file-system description object
*
* This function runs commit and returns zero in case of success and a negative
* error code in case of failure.
*/
int ubifs_run_commit(struct ubifs_info *c)
{
int err = 0;
spin_lock(&c->cs_lock);
if (c->cmt_state == COMMIT_BROKEN) {
err = -EROFS;
goto out;
}
if (c->cmt_state == COMMIT_RUNNING_BACKGROUND)
/*
* We set the commit state to 'running required' to indicate
* that we want it to complete as quickly as possible.
*/
c->cmt_state = COMMIT_RUNNING_REQUIRED;
if (c->cmt_state == COMMIT_RUNNING_REQUIRED) {
spin_unlock(&c->cs_lock);
return wait_for_commit(c);
}
spin_unlock(&c->cs_lock);
/* Ok, the commit is indeed needed */
down_write(&c->commit_sem);
spin_lock(&c->cs_lock);
/*
* Since we unlocked 'c->cs_lock', the state may have changed, so
* re-check it.
*/
if (c->cmt_state == COMMIT_BROKEN) {
err = -EROFS;
goto out_cmt_unlock;
}
if (c->cmt_state == COMMIT_RUNNING_BACKGROUND)
c->cmt_state = COMMIT_RUNNING_REQUIRED;
if (c->cmt_state == COMMIT_RUNNING_REQUIRED) {
up_write(&c->commit_sem);
spin_unlock(&c->cs_lock);
return wait_for_commit(c);
}
c->cmt_state = COMMIT_RUNNING_REQUIRED;
spin_unlock(&c->cs_lock);
err = do_commit(c);
return err;
out_cmt_unlock:
up_write(&c->commit_sem);
out:
spin_unlock(&c->cs_lock);
return err;
}
/**
* ubifs_gc_should_commit - determine if it is time for GC to run commit.
* @c: UBIFS file-system description object
*
* This function is called by garbage collection to determine if commit should
* be run. If commit state is @COMMIT_BACKGROUND, which means that the journal
* is full enough to start commit, this function returns true. It is not
* absolutely necessary to commit yet, but it feels like this should be better
* then to keep doing GC. This function returns %1 if GC has to initiate commit
* and %0 if not.
*/
int ubifs_gc_should_commit(struct ubifs_info *c)
{
int ret = 0;
spin_lock(&c->cs_lock);
if (c->cmt_state == COMMIT_BACKGROUND) {
dbg_cmt("commit required now");
c->cmt_state = COMMIT_REQUIRED;
} else
dbg_cmt("commit not requested");
if (c->cmt_state == COMMIT_REQUIRED)
ret = 1;
spin_unlock(&c->cs_lock);
return ret;
}
/*
* Everything below is related to debugging.
*/
/**
* struct idx_node - hold index nodes during index tree traversal.
* @list: list
* @iip: index in parent (slot number of this indexing node in the parent
* indexing node)
* @upper_key: all keys in this indexing node have to be less or equivalent to
* this key
* @idx: index node (8-byte aligned because all node structures must be 8-byte
* aligned)
*/
struct idx_node {
struct list_head list;
int iip;
union ubifs_key upper_key;
struct ubifs_idx_node idx __aligned(8);
};
/**
* dbg_old_index_check_init - get information for the next old index check.
* @c: UBIFS file-system description object
* @zroot: root of the index
*
* This function records information about the index that will be needed for the
* next old index check i.e. 'dbg_check_old_index()'.
*
* This function returns %0 on success and a negative error code on failure.
*/
int dbg_old_index_check_init(struct ubifs_info *c, struct ubifs_zbranch *zroot)
{
struct ubifs_idx_node *idx;
int lnum, offs, len, err = 0;
struct ubifs_debug_info *d = c->dbg;
d->old_zroot = *zroot;
lnum = d->old_zroot.lnum;
offs = d->old_zroot.offs;
len = d->old_zroot.len;
idx = kmalloc(c->max_idx_node_sz, GFP_NOFS);
if (!idx)
return -ENOMEM;
err = ubifs_read_node(c, idx, UBIFS_IDX_NODE, len, lnum, offs);
if (err)
goto out;
d->old_zroot_level = le16_to_cpu(idx->level);
d->old_zroot_sqnum = le64_to_cpu(idx->ch.sqnum);
out:
kfree(idx);
return err;
}
/**
* dbg_check_old_index - check the old copy of the index.
* @c: UBIFS file-system description object
* @zroot: root of the new index
*
* In order to be able to recover from an unclean unmount, a complete copy of
* the index must exist on flash. This is the "old" index. The commit process
* must write the "new" index to flash without overwriting or destroying any
* part of the old index. This function is run at commit end in order to check
* that the old index does indeed exist completely intact.
*
* This function returns %0 on success and a negative error code on failure.
*/
int dbg_check_old_index(struct ubifs_info *c, struct ubifs_zbranch *zroot)
{
treewide: Remove uninitialized_var() usage Using uninitialized_var() is dangerous as it papers over real bugs[1] (or can in the future), and suppresses unrelated compiler warnings (e.g. "unused variable"). If the compiler thinks it is uninitialized, either simply initialize the variable or make compiler changes. In preparation for removing[2] the[3] macro[4], remove all remaining needless uses with the following script: git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \ xargs perl -pi -e \ 's/\buninitialized_var\(([^\)]+)\)/\1/g; s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;' drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid pathological white-space. No outstanding warnings were found building allmodconfig with GCC 9.3.0 for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64, alpha, and m68k. [1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/ [2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/ [3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/ [4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/ Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5 Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-03 20:09:38 +00:00
int lnum, offs, len, err = 0, last_level, child_cnt;
int first = 1, iip;
struct ubifs_debug_info *d = c->dbg;
treewide: Remove uninitialized_var() usage Using uninitialized_var() is dangerous as it papers over real bugs[1] (or can in the future), and suppresses unrelated compiler warnings (e.g. "unused variable"). If the compiler thinks it is uninitialized, either simply initialize the variable or make compiler changes. In preparation for removing[2] the[3] macro[4], remove all remaining needless uses with the following script: git grep '\buninitialized_var\b' | cut -d: -f1 | sort -u | \ xargs perl -pi -e \ 's/\buninitialized_var\(([^\)]+)\)/\1/g; s:\s*/\* (GCC be quiet|to make compiler happy) \*/$::g;' drivers/video/fbdev/riva/riva_hw.c was manually tweaked to avoid pathological white-space. No outstanding warnings were found building allmodconfig with GCC 9.3.0 for x86_64, i386, arm64, arm, powerpc, powerpc64le, s390x, mips, sparc64, alpha, and m68k. [1] https://lore.kernel.org/lkml/20200603174714.192027-1-glider@google.com/ [2] https://lore.kernel.org/lkml/CA+55aFw+Vbj0i=1TGqCR5vQkCzWJ0QxK6CernOU6eedsudAixw@mail.gmail.com/ [3] https://lore.kernel.org/lkml/CA+55aFwgbgqhbp1fkxvRKEpzyR5J8n1vKT1VZdz9knmPuXhOeg@mail.gmail.com/ [4] https://lore.kernel.org/lkml/CA+55aFz2500WfbKXAx8s67wrm9=yVJu65TpLgN_ybYNv0VEOKA@mail.gmail.com/ Reviewed-by: Leon Romanovsky <leonro@mellanox.com> # drivers/infiniband and mlx4/mlx5 Acked-by: Jason Gunthorpe <jgg@mellanox.com> # IB Acked-by: Kalle Valo <kvalo@codeaurora.org> # wireless drivers Reviewed-by: Chao Yu <yuchao0@huawei.com> # erofs Signed-off-by: Kees Cook <keescook@chromium.org>
2020-06-03 20:09:38 +00:00
union ubifs_key lower_key, upper_key, l_key, u_key;
unsigned long long last_sqnum;
struct ubifs_idx_node *idx;
struct list_head list;
struct idx_node *i;
size_t sz;
if (!dbg_is_chk_index(c))
return 0;
INIT_LIST_HEAD(&list);
sz = sizeof(struct idx_node) + ubifs_idx_node_sz(c, c->fanout) -
UBIFS_IDX_NODE_SZ;
/* Start at the old zroot */
lnum = d->old_zroot.lnum;
offs = d->old_zroot.offs;
len = d->old_zroot.len;
iip = 0;
/*
* Traverse the index tree preorder depth-first i.e. do a node and then
* its subtrees from left to right.
*/
while (1) {
struct ubifs_branch *br;
/* Get the next index node */
i = kmalloc(sz, GFP_NOFS);
if (!i) {
err = -ENOMEM;
goto out_free;
}
i->iip = iip;
/* Keep the index nodes on our path in a linked list */
list_add_tail(&i->list, &list);
/* Read the index node */
idx = &i->idx;
err = ubifs_read_node(c, idx, UBIFS_IDX_NODE, len, lnum, offs);
if (err)
goto out_free;
/* Validate index node */
child_cnt = le16_to_cpu(idx->child_cnt);
if (child_cnt < 1 || child_cnt > c->fanout) {
err = 1;
goto out_dump;
}
if (first) {
first = 0;
/* Check root level and sqnum */
if (le16_to_cpu(idx->level) != d->old_zroot_level) {
err = 2;
goto out_dump;
}
if (le64_to_cpu(idx->ch.sqnum) != d->old_zroot_sqnum) {
err = 3;
goto out_dump;
}
/* Set last values as though root had a parent */
last_level = le16_to_cpu(idx->level) + 1;
last_sqnum = le64_to_cpu(idx->ch.sqnum) + 1;
key_read(c, ubifs_idx_key(c, idx), &lower_key);
highest_ino_key(c, &upper_key, INUM_WATERMARK);
}
key_copy(c, &upper_key, &i->upper_key);
if (le16_to_cpu(idx->level) != last_level - 1) {
err = 3;
goto out_dump;
}
/*
* The index is always written bottom up hence a child's sqnum
* is always less than the parents.
*/
if (le64_to_cpu(idx->ch.sqnum) >= last_sqnum) {
err = 4;
goto out_dump;
}
/* Check key range */
key_read(c, ubifs_idx_key(c, idx), &l_key);
br = ubifs_idx_branch(c, idx, child_cnt - 1);
key_read(c, &br->key, &u_key);
if (keys_cmp(c, &lower_key, &l_key) > 0) {
err = 5;
goto out_dump;
}
if (keys_cmp(c, &upper_key, &u_key) < 0) {
err = 6;
goto out_dump;
}
if (keys_cmp(c, &upper_key, &u_key) == 0)
if (!is_hash_key(c, &u_key)) {
err = 7;
goto out_dump;
}
/* Go to next index node */
if (le16_to_cpu(idx->level) == 0) {
/* At the bottom, so go up until can go right */
while (1) {
/* Drop the bottom of the list */
list_del(&i->list);
kfree(i);
/* No more list means we are done */
if (list_empty(&list))
goto out;
/* Look at the new bottom */
i = list_entry(list.prev, struct idx_node,
list);
idx = &i->idx;
/* Can we go right */
if (iip + 1 < le16_to_cpu(idx->child_cnt)) {
iip = iip + 1;
break;
} else
/* Nope, so go up again */
iip = i->iip;
}
} else
/* Go down left */
iip = 0;
/*
* We have the parent in 'idx' and now we set up for reading the
* child pointed to by slot 'iip'.
*/
last_level = le16_to_cpu(idx->level);
last_sqnum = le64_to_cpu(idx->ch.sqnum);
br = ubifs_idx_branch(c, idx, iip);
lnum = le32_to_cpu(br->lnum);
offs = le32_to_cpu(br->offs);
len = le32_to_cpu(br->len);
key_read(c, &br->key, &lower_key);
if (iip + 1 < le16_to_cpu(idx->child_cnt)) {
br = ubifs_idx_branch(c, idx, iip + 1);
key_read(c, &br->key, &upper_key);
} else
key_copy(c, &i->upper_key, &upper_key);
}
out:
err = dbg_old_index_check_init(c, zroot);
if (err)
goto out_free;
return 0;
out_dump:
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_err(c, "dumping index node (iip=%d)", i->iip);
ubifs_dump_node(c, idx, ubifs_idx_node_sz(c, c->fanout));
list_del(&i->list);
kfree(i);
if (!list_empty(&list)) {
i = list_entry(list.prev, struct idx_node, list);
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_err(c, "dumping parent index node");
ubifs_dump_node(c, &i->idx, ubifs_idx_node_sz(c, c->fanout));
}
out_free:
while (!list_empty(&list)) {
i = list_entry(list.next, struct idx_node, list);
list_del(&i->list);
kfree(i);
}
UBIFS: extend debug/message capabilities In the case where we have more than one volumes on different UBI devices, it may be not that easy to tell which volume prints the messages. Add ubi number and volume id in ubifs_msg/warn/error to help debug. These two values are passed by struct ubifs_info. For those where ubifs_info is not initialized yet, ubifs_* is replaced by pr_*. For those where ubifs_info is not avaliable, ubifs_info is passed to the calling function as a const parameter. The output looks like, [ 95.444879] UBIFS (ubi0:1): background thread "ubifs_bgt0_1" started, PID 696 [ 95.484688] UBIFS (ubi0:1): UBIFS: mounted UBI device 0, volume 1, name "test1" [ 95.484694] UBIFS (ubi0:1): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.484699] UBIFS (ubi0:1): FS size: 30220288 bytes (28 MiB, 238 LEBs), journal size 1523712 bytes (1 MiB, 12 LEBs) [ 95.484703] UBIFS (ubi0:1): reserved for root: 1427378 bytes (1393 KiB) [ 95.484709] UBIFS (ubi0:1): media format: w4/r0 (latest is w4/r0), UUID 40DFFC0E-70BE-4193-8905-F7D6DFE60B17, small LPT model [ 95.489875] UBIFS (ubi1:0): background thread "ubifs_bgt1_0" started, PID 699 [ 95.529713] UBIFS (ubi1:0): UBIFS: mounted UBI device 1, volume 0, name "test2" [ 95.529718] UBIFS (ubi1:0): LEB size: 126976 bytes (124 KiB), min./max. I/O unit sizes: 2048 bytes/2048 bytes [ 95.529724] UBIFS (ubi1:0): FS size: 19808256 bytes (18 MiB, 156 LEBs), journal size 1015809 bytes (0 MiB, 8 LEBs) [ 95.529727] UBIFS (ubi1:0): reserved for root: 935592 bytes (913 KiB) [ 95.529733] UBIFS (ubi1:0): media format: w4/r0 (latest is w4/r0), UUID EEB7779D-F419-4CA9-811B-831CAC7233D4, small LPT model [ 954.264767] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node type (255 but expected 6) [ 954.367030] UBIFS error (ubi1:0 pid 756): ubifs_read_node: bad node at LEB 0:0, LEB mapping status 1 Signed-off-by: Sheng Yong <shengyong1@huawei.com> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@linux.intel.com>
2015-03-20 10:39:42 +00:00
ubifs_err(c, "failed, error %d", err);
if (err > 0)
err = -EINVAL;
return err;
}