kernel_optimize_test/drivers/md/bcache/journal.c

1008 lines
24 KiB
C
Raw Normal View History

License cleanup: add SPDX GPL-2.0 license identifier to files with no license Many source files in the tree are missing licensing information, which makes it harder for compliance tools to determine the correct license. By default all files without license information are under the default license of the kernel, which is GPL version 2. Update the files which contain no license information with the 'GPL-2.0' SPDX license identifier. The SPDX identifier is a legally binding shorthand, which can be used instead of the full boiler plate text. This patch is based on work done by Thomas Gleixner and Kate Stewart and Philippe Ombredanne. How this work was done: Patches were generated and checked against linux-4.14-rc6 for a subset of the use cases: - file had no licensing information it it. - file was a */uapi/* one with no licensing information in it, - file was a */uapi/* one with existing licensing information, Further patches will be generated in subsequent months to fix up cases where non-standard license headers were used, and references to license had to be inferred by heuristics based on keywords. The analysis to determine which SPDX License Identifier to be applied to a file was done in a spreadsheet of side by side results from of the output of two independent scanners (ScanCode & Windriver) producing SPDX tag:value files created by Philippe Ombredanne. Philippe prepared the base worksheet, and did an initial spot review of a few 1000 files. The 4.13 kernel was the starting point of the analysis with 60,537 files assessed. Kate Stewart did a file by file comparison of the scanner results in the spreadsheet to determine which SPDX license identifier(s) to be applied to the file. She confirmed any determination that was not immediately clear with lawyers working with the Linux Foundation. Criteria used to select files for SPDX license identifier tagging was: - Files considered eligible had to be source code files. - Make and config files were included as candidates if they contained >5 lines of source - File already had some variant of a license header in it (even if <5 lines). All documentation files were explicitly excluded. The following heuristics were used to determine which SPDX license identifiers to apply. - when both scanners couldn't find any license traces, file was considered to have no license information in it, and the top level COPYING file license applied. For non */uapi/* files that summary was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 11139 and resulted in the first patch in this series. If that file was a */uapi/* path one, it was "GPL-2.0 WITH Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was: SPDX license identifier # files ---------------------------------------------------|------- GPL-2.0 WITH Linux-syscall-note 930 and resulted in the second patch in this series. - if a file had some form of licensing information in it, and was one of the */uapi/* ones, it was denoted with the Linux-syscall-note if any GPL family license was found in the file or had no licensing in it (per prior point). Results summary: SPDX license identifier # files ---------------------------------------------------|------ GPL-2.0 WITH Linux-syscall-note 270 GPL-2.0+ WITH Linux-syscall-note 169 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21 ((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17 LGPL-2.1+ WITH Linux-syscall-note 15 GPL-1.0+ WITH Linux-syscall-note 14 ((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5 LGPL-2.0+ WITH Linux-syscall-note 4 LGPL-2.1 WITH Linux-syscall-note 3 ((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3 ((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1 and that resulted in the third patch in this series. - when the two scanners agreed on the detected license(s), that became the concluded license(s). - when there was disagreement between the two scanners (one detected a license but the other didn't, or they both detected different licenses) a manual inspection of the file occurred. - In most cases a manual inspection of the information in the file resulted in a clear resolution of the license that should apply (and which scanner probably needed to revisit its heuristics). - When it was not immediately clear, the license identifier was confirmed with lawyers working with the Linux Foundation. - If there was any question as to the appropriate license identifier, the file was flagged for further research and to be revisited later in time. In total, over 70 hours of logged manual review was done on the spreadsheet to determine the SPDX license identifiers to apply to the source files by Kate, Philippe, Thomas and, in some cases, confirmation by lawyers working with the Linux Foundation. Kate also obtained a third independent scan of the 4.13 code base from FOSSology, and compared selected files where the other two scanners disagreed against that SPDX file, to see if there was new insights. The Windriver scanner is based on an older version of FOSSology in part, so they are related. Thomas did random spot checks in about 500 files from the spreadsheets for the uapi headers and agreed with SPDX license identifier in the files he inspected. For the non-uapi files Thomas did random spot checks in about 15000 files. In initial set of patches against 4.14-rc6, 3 files were found to have copy/paste license identifier errors, and have been fixed to reflect the correct identifier. Additionally Philippe spent 10 hours this week doing a detailed manual inspection and review of the 12,461 patched files from the initial patch version early this week with: - a full scancode scan run, collecting the matched texts, detected license ids and scores - reviewing anything where there was a license detected (about 500+ files) to ensure that the applied SPDX license was correct - reviewing anything where there was no detection but the patch license was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied SPDX license was correct This produced a worksheet with 20 files needing minor correction. This worksheet was then exported into 3 different .csv files for the different types of files to be modified. These .csv files were then reviewed by Greg. Thomas wrote a script to parse the csv files and add the proper SPDX tag to the file, in the format that the file expected. This script was further refined by Greg based on the output to detect more types of files automatically and to distinguish between header and source .c files (which need different comment types.) Finally Greg ran the script using the .csv files to generate the patches. Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com> Reviewed-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 22:07:57 +08:00
// SPDX-License-Identifier: GPL-2.0
/*
* bcache journalling code, for btree insertions
*
* Copyright 2012 Google, Inc.
*/
#include "bcache.h"
#include "btree.h"
#include "debug.h"
#include "extents.h"
#include <trace/events/bcache.h>
/*
* Journal replay/recovery:
*
* This code is all driven from run_cache_set(); we first read the journal
* entries, do some other stuff, then we mark all the keys in the journal
* entries (same as garbage collection would), then we replay them - reinserting
* them into the cache in precisely the same order as they appear in the
* journal.
*
* We only journal keys that go in leaf nodes, which simplifies things quite a
* bit.
*/
static void journal_read_endio(struct bio *bio)
{
struct closure *cl = bio->bi_private;
closure_put(cl);
}
static int journal_read_bucket(struct cache *ca, struct list_head *list,
unsigned int bucket_index)
{
struct journal_device *ja = &ca->journal;
struct bio *bio = &ja->bio;
struct journal_replay *i;
struct jset *j, *data = ca->set->journal.w[0].data;
struct closure cl;
unsigned int len, left, offset = 0;
int ret = 0;
sector_t bucket = bucket_to_sector(ca->set, ca->sb.d[bucket_index]);
closure_init_stack(&cl);
pr_debug("reading %u", bucket_index);
while (offset < ca->sb.bucket_size) {
reread: left = ca->sb.bucket_size - offset;
len = min_t(unsigned int, left, PAGE_SECTORS << JSET_BITS);
bio_reset(bio);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_sector = bucket + offset;
bio_set_dev(bio, ca->bdev);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_size = len << 9;
bio->bi_end_io = journal_read_endio;
bio->bi_private = &cl;
bio_set_op_attrs(bio, REQ_OP_READ, 0);
bch_bio_map(bio, data);
bcache: add CACHE_SET_IO_DISABLE to struct cache_set flags When too many I/Os failed on cache device, bch_cache_set_error() is called in the error handling code path to retire whole problematic cache set. If new I/O requests continue to come and take refcount dc->count, the cache set won't be retired immediately, this is a problem. Further more, there are several kernel thread and self-armed kernel work may still running after bch_cache_set_error() is called. It needs to wait quite a while for them to stop, or they won't stop at all. They also prevent the cache set from being retired. The solution in this patch is, to add per cache set flag to disable I/O request on this cache and all attached backing devices. Then new coming I/O requests can be rejected in *_make_request() before taking refcount, kernel threads and self-armed kernel worker can stop very fast when flags bit CACHE_SET_IO_DISABLE is set. Because bcache also do internal I/Os for writeback, garbage collection, bucket allocation, journaling, this kind of I/O should be disabled after bch_cache_set_error() is called. So closure_bio_submit() is modified to check whether CACHE_SET_IO_DISABLE is set on cache_set->flags. If set, closure_bio_submit() will set bio->bi_status to BLK_STS_IOERR and return, generic_make_request() won't be called. A sysfs interface is also added to set or clear CACHE_SET_IO_DISABLE bit from cache_set->flags, to disable or enable cache set I/O for debugging. It is helpful to trigger more corner case issues for failed cache device. Changelog v4, add wait_for_kthread_stop(), and call it before exits writeback and gc kernel threads. v3, change CACHE_SET_IO_DISABLE from 4 to 3, since it is bit index. remove "bcache: " prefix when printing out kernel message. v2, more changes by previous review, - Use CACHE_SET_IO_DISABLE of cache_set->flags, suggested by Junhui. - Check CACHE_SET_IO_DISABLE in bch_btree_gc() to stop a while-loop, this is reported and inspired from origal patch of Pavel Vazharov. v1, initial version. Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Michael Lyle <mlyle@lyle.org> Cc: Junhui Tang <tang.junhui@zte.com.cn> Cc: Michael Lyle <mlyle@lyle.org> Cc: Pavel Vazharov <freakpv@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-03-19 08:36:17 +08:00
closure_bio_submit(ca->set, bio, &cl);
closure_sync(&cl);
/* This function could be simpler now since we no longer write
* journal entries that overlap bucket boundaries; this means
* the start of a bucket will always have a valid journal entry
* if it has any journal entries at all.
*/
j = data;
while (len) {
struct list_head *where;
size_t blocks, bytes = set_bytes(j);
if (j->magic != jset_magic(&ca->sb)) {
pr_debug("%u: bad magic", bucket_index);
return ret;
}
if (bytes > left << 9 ||
bytes > PAGE_SIZE << JSET_BITS) {
pr_info("%u: too big, %zu bytes, offset %u",
bucket_index, bytes, offset);
return ret;
}
if (bytes > len << 9)
goto reread;
if (j->csum != csum_set(j)) {
pr_info("%u: bad csum, %zu bytes, offset %u",
bucket_index, bytes, offset);
return ret;
}
blocks = set_blocks(j, block_bytes(ca->set));
/*
* Nodes in 'list' are in linear increasing order of
* i->j.seq, the node on head has the smallest (oldest)
* journal seq, the node on tail has the biggest
* (latest) journal seq.
*/
/*
* Check from the oldest jset for last_seq. If
* i->j.seq < j->last_seq, it means the oldest jset
* in list is expired and useless, remove it from
* this list. Otherwise, j is a condidate jset for
* further following checks.
*/
while (!list_empty(list)) {
i = list_first_entry(list,
struct journal_replay, list);
if (i->j.seq >= j->last_seq)
break;
list_del(&i->list);
kfree(i);
}
/* iterate list in reverse order (from latest jset) */
list_for_each_entry_reverse(i, list, list) {
if (j->seq == i->j.seq)
goto next_set;
/*
* if j->seq is less than any i->j.last_seq
* in list, j is an expired and useless jset.
*/
if (j->seq < i->j.last_seq)
goto next_set;
/*
* 'where' points to first jset in list which
* is elder then j.
*/
if (j->seq > i->j.seq) {
where = &i->list;
goto add;
}
}
where = list;
add:
i = kmalloc(offsetof(struct journal_replay, j) +
bytes, GFP_KERNEL);
if (!i)
return -ENOMEM;
memcpy(&i->j, j, bytes);
/* Add to the location after 'where' points to */
list_add(&i->list, where);
ret = 1;
if (j->seq > ja->seq[bucket_index])
ja->seq[bucket_index] = j->seq;
next_set:
offset += blocks * ca->sb.block_size;
len -= blocks * ca->sb.block_size;
j = ((void *) j) + blocks * block_bytes(ca);
}
}
return ret;
}
int bch_journal_read(struct cache_set *c, struct list_head *list)
{
#define read_bucket(b) \
({ \
ret = journal_read_bucket(ca, list, b); \
__set_bit(b, bitmap); \
if (ret < 0) \
return ret; \
ret; \
})
struct cache *ca;
unsigned int iter;
int ret = 0;
for_each_cache(ca, c, iter) {
struct journal_device *ja = &ca->journal;
DECLARE_BITMAP(bitmap, SB_JOURNAL_BUCKETS);
unsigned int i, l, r, m;
uint64_t seq;
bitmap_zero(bitmap, SB_JOURNAL_BUCKETS);
pr_debug("%u journal buckets", ca->sb.njournal_buckets);
/*
* Read journal buckets ordered by golden ratio hash to quickly
* find a sequence of buckets with valid journal entries
*/
for (i = 0; i < ca->sb.njournal_buckets; i++) {
/*
* We must try the index l with ZERO first for
* correctness due to the scenario that the journal
* bucket is circular buffer which might have wrapped
*/
l = (i * 2654435769U) % ca->sb.njournal_buckets;
if (test_bit(l, bitmap))
break;
if (read_bucket(l))
goto bsearch;
}
/*
* If that fails, check all the buckets we haven't checked
* already
*/
pr_debug("falling back to linear search");
for (l = find_first_zero_bit(bitmap, ca->sb.njournal_buckets);
l < ca->sb.njournal_buckets;
l = find_next_zero_bit(bitmap, ca->sb.njournal_buckets,
l + 1))
if (read_bucket(l))
goto bsearch;
/* no journal entries on this device? */
if (l == ca->sb.njournal_buckets)
continue;
bsearch:
BUG_ON(list_empty(list));
/* Binary search */
m = l;
r = find_next_bit(bitmap, ca->sb.njournal_buckets, l + 1);
pr_debug("starting binary search, l %u r %u", l, r);
while (l + 1 < r) {
seq = list_entry(list->prev, struct journal_replay,
list)->j.seq;
m = (l + r) >> 1;
read_bucket(m);
if (seq != list_entry(list->prev, struct journal_replay,
list)->j.seq)
l = m;
else
r = m;
}
/*
* Read buckets in reverse order until we stop finding more
* journal entries
*/
pr_debug("finishing up: m %u njournal_buckets %u",
m, ca->sb.njournal_buckets);
l = m;
while (1) {
if (!l--)
l = ca->sb.njournal_buckets - 1;
if (l == m)
break;
if (test_bit(l, bitmap))
continue;
if (!read_bucket(l))
break;
}
seq = 0;
for (i = 0; i < ca->sb.njournal_buckets; i++)
if (ja->seq[i] > seq) {
seq = ja->seq[i];
/*
* When journal_reclaim() goes to allocate for
* the first time, it'll use the bucket after
* ja->cur_idx
*/
ja->cur_idx = i;
ja->last_idx = ja->discard_idx = (i + 1) %
ca->sb.njournal_buckets;
}
}
if (!list_empty(list))
c->journal.seq = list_entry(list->prev,
struct journal_replay,
list)->j.seq;
return 0;
#undef read_bucket
}
void bch_journal_mark(struct cache_set *c, struct list_head *list)
{
atomic_t p = { 0 };
struct bkey *k;
struct journal_replay *i;
struct journal *j = &c->journal;
uint64_t last = j->seq;
/*
* journal.pin should never fill up - we never write a journal
* entry when it would fill up. But if for some reason it does, we
* iterate over the list in reverse order so that we can just skip that
* refcount instead of bugging.
*/
list_for_each_entry_reverse(i, list, list) {
BUG_ON(last < i->j.seq);
i->pin = NULL;
while (last-- != i->j.seq)
if (fifo_free(&j->pin) > 1) {
fifo_push_front(&j->pin, p);
atomic_set(&fifo_front(&j->pin), 0);
}
if (fifo_free(&j->pin) > 1) {
fifo_push_front(&j->pin, p);
i->pin = &fifo_front(&j->pin);
atomic_set(i->pin, 1);
}
for (k = i->j.start;
k < bset_bkey_last(&i->j);
k = bkey_next(k))
if (!__bch_extent_invalid(c, k)) {
unsigned int j;
for (j = 0; j < KEY_PTRS(k); j++)
if (ptr_available(c, k, j))
atomic_inc(&PTR_BUCKET(c, k, j)->pin);
bch_initial_mark_key(c, 0, k);
}
}
}
static bool is_discard_enabled(struct cache_set *s)
{
struct cache *ca;
unsigned int i;
for_each_cache(ca, s, i)
if (ca->discard)
return true;
return false;
}
int bch_journal_replay(struct cache_set *s, struct list_head *list)
{
int ret = 0, keys = 0, entries = 0;
struct bkey *k;
struct journal_replay *i =
list_entry(list->prev, struct journal_replay, list);
uint64_t start = i->j.last_seq, end = i->j.seq, n = start;
struct keylist keylist;
list_for_each_entry(i, list, list) {
BUG_ON(i->pin && atomic_read(i->pin) != 1);
if (n != i->j.seq) {
if (n == start && is_discard_enabled(s))
pr_info("bcache: journal entries %llu-%llu may be discarded! (replaying %llu-%llu)",
n, i->j.seq - 1, start, end);
else {
pr_err("bcache: journal entries %llu-%llu missing! (replaying %llu-%llu)",
n, i->j.seq - 1, start, end);
ret = -EIO;
goto err;
}
}
for (k = i->j.start;
k < bset_bkey_last(&i->j);
k = bkey_next(k)) {
trace_bcache_journal_replay_key(k);
bch_keylist_init_single(&keylist, k);
ret = bch_btree_insert(s, &keylist, i->pin, NULL);
if (ret)
goto err;
BUG_ON(!bch_keylist_empty(&keylist));
keys++;
cond_resched();
}
if (i->pin)
atomic_dec(i->pin);
n = i->j.seq + 1;
entries++;
}
pr_info("journal replay done, %i keys in %i entries, seq %llu",
keys, entries, end);
err:
while (!list_empty(list)) {
i = list_first_entry(list, struct journal_replay, list);
list_del(&i->list);
kfree(i);
}
return ret;
}
/* Journalling */
static void btree_flush_write(struct cache_set *c)
{
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
struct btree *b, *t, *btree_nodes[BTREE_FLUSH_NR];
bcache: fix incorrect data type usage in btree_flush_write() Dan Carpenter points out that from commit 2aa8c529387c ("bcache: avoid unnecessary btree nodes flushing in btree_flush_write()"), there is a incorrect data type usage which leads to the following static checker warning: drivers/md/bcache/journal.c:444 btree_flush_write() warn: 'ref_nr' unsigned <= 0 drivers/md/bcache/journal.c 422 static void btree_flush_write(struct cache_set *c) 423 { 424 struct btree *b, *t, *btree_nodes[BTREE_FLUSH_NR]; 425 unsigned int i, nr, ref_nr; ^^^^^^ 426 atomic_t *fifo_front_p, *now_fifo_front_p; 427 size_t mask; 428 429 if (c->journal.btree_flushing) 430 return; 431 432 spin_lock(&c->journal.flush_write_lock); 433 if (c->journal.btree_flushing) { 434 spin_unlock(&c->journal.flush_write_lock); 435 return; 436 } 437 c->journal.btree_flushing = true; 438 spin_unlock(&c->journal.flush_write_lock); 439 440 /* get the oldest journal entry and check its refcount */ 441 spin_lock(&c->journal.lock); 442 fifo_front_p = &fifo_front(&c->journal.pin); 443 ref_nr = atomic_read(fifo_front_p); 444 if (ref_nr <= 0) { ^^^^^^^^^^^ Unsigned can't be less than zero. 445 /* 446 * do nothing if no btree node references 447 * the oldest journal entry 448 */ 449 spin_unlock(&c->journal.lock); 450 goto out; 451 } 452 spin_unlock(&c->journal.lock); As the warning information indicates, local varaible ref_nr in unsigned int type is wrong, which does not matche atomic_read() and the "<= 0" checking. This patch fixes the above error by defining local variable ref_nr as int type. Fixes: 2aa8c529387c ("bcache: avoid unnecessary btree nodes flushing in btree_flush_write()") Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-02-01 22:42:34 +08:00
unsigned int i, nr;
int ref_nr;
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
atomic_t *fifo_front_p, *now_fifo_front_p;
size_t mask;
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
if (c->journal.btree_flushing)
return;
spin_lock(&c->journal.flush_write_lock);
if (c->journal.btree_flushing) {
spin_unlock(&c->journal.flush_write_lock);
return;
}
c->journal.btree_flushing = true;
spin_unlock(&c->journal.flush_write_lock);
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
/* get the oldest journal entry and check its refcount */
spin_lock(&c->journal.lock);
fifo_front_p = &fifo_front(&c->journal.pin);
ref_nr = atomic_read(fifo_front_p);
if (ref_nr <= 0) {
/*
* do nothing if no btree node references
* the oldest journal entry
*/
spin_unlock(&c->journal.lock);
goto out;
}
spin_unlock(&c->journal.lock);
mask = c->journal.pin.mask;
nr = 0;
atomic_long_inc(&c->flush_write);
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
memset(btree_nodes, 0, sizeof(btree_nodes));
bcache: fix race in btree_flush_write() There is a race between mca_reap(), btree_node_free() and journal code btree_flush_write(), which results very rare and strange deadlock or panic and are very hard to reproduce. Let me explain how the race happens. In btree_flush_write() one btree node with oldest journal pin is selected, then it is flushed to cache device, the select-and-flush is a two steps operation. Between these two steps, there are something may happen inside the race window, - The selected btree node was reaped by mca_reap() and allocated to other requesters for other btree node. - The slected btree node was selected, flushed and released by mca shrink callback bch_mca_scan(). When btree_flush_write() tries to flush the selected btree node, firstly b->write_lock is held by mutex_lock(). If the race happens and the memory of selected btree node is allocated to other btree node, if that btree node's write_lock is held already, a deadlock very probably happens here. A worse case is the memory of the selected btree node is released, then all references to this btree node (e.g. b->write_lock) will trigger NULL pointer deference panic. This race was introduced in commit cafe56359144 ("bcache: A block layer cache"), and enlarged by commit c4dc2497d50d ("bcache: fix high CPU occupancy during journal"), which selected 128 btree nodes and flushed them one-by-one in a quite long time period. Such race is not easy to reproduce before. On a Lenovo SR650 server with 48 Xeon cores, and configure 1 NVMe SSD as cache device, a MD raid0 device assembled by 3 NVMe SSDs as backing device, this race can be observed around every 10,000 times btree_flush_write() gets called. Both deadlock and kernel panic all happened as aftermath of the race. The idea of the fix is to add a btree flag BTREE_NODE_journal_flush. It is set when selecting btree nodes, and cleared after btree nodes flushed. Then when mca_reap() selects a btree node with this bit set, this btree node will be skipped. Since mca_reap() only reaps btree node without BTREE_NODE_journal_flush flag, such race is avoided. Once corner case should be noticed, that is btree_node_free(). It might be called in some error handling code path. For example the following code piece from btree_split(), 2149 err_free2: 2150 bkey_put(b->c, &n2->key); 2151 btree_node_free(n2); 2152 rw_unlock(true, n2); 2153 err_free1: 2154 bkey_put(b->c, &n1->key); 2155 btree_node_free(n1); 2156 rw_unlock(true, n1); At line 2151 and 2155, the btree node n2 and n1 are released without mac_reap(), so BTREE_NODE_journal_flush also needs to be checked here. If btree_node_free() is called directly in such error handling path, and the selected btree node has BTREE_NODE_journal_flush bit set, just delay for 1 us and retry again. In this case this btree node won't be skipped, just retry until the BTREE_NODE_journal_flush bit cleared, and free the btree node memory. Fixes: cafe56359144 ("bcache: A block layer cache") Signed-off-by: Coly Li <colyli@suse.de> Reported-and-tested-by: kbuild test robot <lkp@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:58 +08:00
mutex_lock(&c->bucket_lock);
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
list_for_each_entry_safe_reverse(b, t, &c->btree_cache, list) {
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
/*
* It is safe to get now_fifo_front_p without holding
* c->journal.lock here, because we don't need to know
* the exactly accurate value, just check whether the
* front pointer of c->journal.pin is changed.
*/
now_fifo_front_p = &fifo_front(&c->journal.pin);
/*
* If the oldest journal entry is reclaimed and front
* pointer of c->journal.pin changes, it is unnecessary
* to scan c->btree_cache anymore, just quit the loop and
* flush out what we have already.
*/
if (now_fifo_front_p != fifo_front_p)
break;
/*
* quit this loop if all matching btree nodes are
* scanned and record in btree_nodes[] already.
*/
ref_nr = atomic_read(fifo_front_p);
if (nr >= ref_nr)
break;
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
if (btree_node_journal_flush(b))
pr_err("BUG: flush_write bit should not be set here!");
mutex_lock(&b->write_lock);
if (!btree_node_dirty(b)) {
mutex_unlock(&b->write_lock);
continue;
}
if (!btree_current_write(b)->journal) {
mutex_unlock(&b->write_lock);
continue;
}
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
/*
* Only select the btree node which exactly references
* the oldest journal entry.
*
* If the journal entry pointed by fifo_front_p is
* reclaimed in parallel, don't worry:
* - the list_for_each_xxx loop will quit when checking
* next now_fifo_front_p.
* - If there are matched nodes recorded in btree_nodes[],
* they are clean now (this is why and how the oldest
* journal entry can be reclaimed). These selected nodes
* will be ignored and skipped in the folowing for-loop.
*/
if (((btree_current_write(b)->journal - fifo_front_p) &
mask) != 0) {
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
mutex_unlock(&b->write_lock);
continue;
}
bcache: fix race in btree_flush_write() There is a race between mca_reap(), btree_node_free() and journal code btree_flush_write(), which results very rare and strange deadlock or panic and are very hard to reproduce. Let me explain how the race happens. In btree_flush_write() one btree node with oldest journal pin is selected, then it is flushed to cache device, the select-and-flush is a two steps operation. Between these two steps, there are something may happen inside the race window, - The selected btree node was reaped by mca_reap() and allocated to other requesters for other btree node. - The slected btree node was selected, flushed and released by mca shrink callback bch_mca_scan(). When btree_flush_write() tries to flush the selected btree node, firstly b->write_lock is held by mutex_lock(). If the race happens and the memory of selected btree node is allocated to other btree node, if that btree node's write_lock is held already, a deadlock very probably happens here. A worse case is the memory of the selected btree node is released, then all references to this btree node (e.g. b->write_lock) will trigger NULL pointer deference panic. This race was introduced in commit cafe56359144 ("bcache: A block layer cache"), and enlarged by commit c4dc2497d50d ("bcache: fix high CPU occupancy during journal"), which selected 128 btree nodes and flushed them one-by-one in a quite long time period. Such race is not easy to reproduce before. On a Lenovo SR650 server with 48 Xeon cores, and configure 1 NVMe SSD as cache device, a MD raid0 device assembled by 3 NVMe SSDs as backing device, this race can be observed around every 10,000 times btree_flush_write() gets called. Both deadlock and kernel panic all happened as aftermath of the race. The idea of the fix is to add a btree flag BTREE_NODE_journal_flush. It is set when selecting btree nodes, and cleared after btree nodes flushed. Then when mca_reap() selects a btree node with this bit set, this btree node will be skipped. Since mca_reap() only reaps btree node without BTREE_NODE_journal_flush flag, such race is avoided. Once corner case should be noticed, that is btree_node_free(). It might be called in some error handling code path. For example the following code piece from btree_split(), 2149 err_free2: 2150 bkey_put(b->c, &n2->key); 2151 btree_node_free(n2); 2152 rw_unlock(true, n2); 2153 err_free1: 2154 bkey_put(b->c, &n1->key); 2155 btree_node_free(n1); 2156 rw_unlock(true, n1); At line 2151 and 2155, the btree node n2 and n1 are released without mac_reap(), so BTREE_NODE_journal_flush also needs to be checked here. If btree_node_free() is called directly in such error handling path, and the selected btree node has BTREE_NODE_journal_flush bit set, just delay for 1 us and retry again. In this case this btree node won't be skipped, just retry until the BTREE_NODE_journal_flush bit cleared, and free the btree node memory. Fixes: cafe56359144 ("bcache: A block layer cache") Signed-off-by: Coly Li <colyli@suse.de> Reported-and-tested-by: kbuild test robot <lkp@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:58 +08:00
set_btree_node_journal_flush(b);
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
mutex_unlock(&b->write_lock);
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
btree_nodes[nr++] = b;
/*
* To avoid holding c->bucket_lock too long time,
* only scan for BTREE_FLUSH_NR matched btree nodes
* at most. If there are more btree nodes reference
* the oldest journal entry, try to flush them next
* time when btree_flush_write() is called.
*/
if (nr == BTREE_FLUSH_NR)
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
break;
}
bcache: fix race in btree_flush_write() There is a race between mca_reap(), btree_node_free() and journal code btree_flush_write(), which results very rare and strange deadlock or panic and are very hard to reproduce. Let me explain how the race happens. In btree_flush_write() one btree node with oldest journal pin is selected, then it is flushed to cache device, the select-and-flush is a two steps operation. Between these two steps, there are something may happen inside the race window, - The selected btree node was reaped by mca_reap() and allocated to other requesters for other btree node. - The slected btree node was selected, flushed and released by mca shrink callback bch_mca_scan(). When btree_flush_write() tries to flush the selected btree node, firstly b->write_lock is held by mutex_lock(). If the race happens and the memory of selected btree node is allocated to other btree node, if that btree node's write_lock is held already, a deadlock very probably happens here. A worse case is the memory of the selected btree node is released, then all references to this btree node (e.g. b->write_lock) will trigger NULL pointer deference panic. This race was introduced in commit cafe56359144 ("bcache: A block layer cache"), and enlarged by commit c4dc2497d50d ("bcache: fix high CPU occupancy during journal"), which selected 128 btree nodes and flushed them one-by-one in a quite long time period. Such race is not easy to reproduce before. On a Lenovo SR650 server with 48 Xeon cores, and configure 1 NVMe SSD as cache device, a MD raid0 device assembled by 3 NVMe SSDs as backing device, this race can be observed around every 10,000 times btree_flush_write() gets called. Both deadlock and kernel panic all happened as aftermath of the race. The idea of the fix is to add a btree flag BTREE_NODE_journal_flush. It is set when selecting btree nodes, and cleared after btree nodes flushed. Then when mca_reap() selects a btree node with this bit set, this btree node will be skipped. Since mca_reap() only reaps btree node without BTREE_NODE_journal_flush flag, such race is avoided. Once corner case should be noticed, that is btree_node_free(). It might be called in some error handling code path. For example the following code piece from btree_split(), 2149 err_free2: 2150 bkey_put(b->c, &n2->key); 2151 btree_node_free(n2); 2152 rw_unlock(true, n2); 2153 err_free1: 2154 bkey_put(b->c, &n1->key); 2155 btree_node_free(n1); 2156 rw_unlock(true, n1); At line 2151 and 2155, the btree node n2 and n1 are released without mac_reap(), so BTREE_NODE_journal_flush also needs to be checked here. If btree_node_free() is called directly in such error handling path, and the selected btree node has BTREE_NODE_journal_flush bit set, just delay for 1 us and retry again. In this case this btree node won't be skipped, just retry until the BTREE_NODE_journal_flush bit cleared, and free the btree node memory. Fixes: cafe56359144 ("bcache: A block layer cache") Signed-off-by: Coly Li <colyli@suse.de> Reported-and-tested-by: kbuild test robot <lkp@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:58 +08:00
mutex_unlock(&c->bucket_lock);
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
for (i = 0; i < nr; i++) {
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
b = btree_nodes[i];
if (!b) {
pr_err("BUG: btree_nodes[%d] is NULL", i);
continue;
}
/* safe to check without holding b->write_lock */
if (!btree_node_journal_flush(b)) {
pr_err("BUG: bnode %p: journal_flush bit cleaned", b);
continue;
}
mutex_lock(&b->write_lock);
if (!btree_current_write(b)->journal) {
bcache: fix race in btree_flush_write() There is a race between mca_reap(), btree_node_free() and journal code btree_flush_write(), which results very rare and strange deadlock or panic and are very hard to reproduce. Let me explain how the race happens. In btree_flush_write() one btree node with oldest journal pin is selected, then it is flushed to cache device, the select-and-flush is a two steps operation. Between these two steps, there are something may happen inside the race window, - The selected btree node was reaped by mca_reap() and allocated to other requesters for other btree node. - The slected btree node was selected, flushed and released by mca shrink callback bch_mca_scan(). When btree_flush_write() tries to flush the selected btree node, firstly b->write_lock is held by mutex_lock(). If the race happens and the memory of selected btree node is allocated to other btree node, if that btree node's write_lock is held already, a deadlock very probably happens here. A worse case is the memory of the selected btree node is released, then all references to this btree node (e.g. b->write_lock) will trigger NULL pointer deference panic. This race was introduced in commit cafe56359144 ("bcache: A block layer cache"), and enlarged by commit c4dc2497d50d ("bcache: fix high CPU occupancy during journal"), which selected 128 btree nodes and flushed them one-by-one in a quite long time period. Such race is not easy to reproduce before. On a Lenovo SR650 server with 48 Xeon cores, and configure 1 NVMe SSD as cache device, a MD raid0 device assembled by 3 NVMe SSDs as backing device, this race can be observed around every 10,000 times btree_flush_write() gets called. Both deadlock and kernel panic all happened as aftermath of the race. The idea of the fix is to add a btree flag BTREE_NODE_journal_flush. It is set when selecting btree nodes, and cleared after btree nodes flushed. Then when mca_reap() selects a btree node with this bit set, this btree node will be skipped. Since mca_reap() only reaps btree node without BTREE_NODE_journal_flush flag, such race is avoided. Once corner case should be noticed, that is btree_node_free(). It might be called in some error handling code path. For example the following code piece from btree_split(), 2149 err_free2: 2150 bkey_put(b->c, &n2->key); 2151 btree_node_free(n2); 2152 rw_unlock(true, n2); 2153 err_free1: 2154 bkey_put(b->c, &n1->key); 2155 btree_node_free(n1); 2156 rw_unlock(true, n1); At line 2151 and 2155, the btree node n2 and n1 are released without mac_reap(), so BTREE_NODE_journal_flush also needs to be checked here. If btree_node_free() is called directly in such error handling path, and the selected btree node has BTREE_NODE_journal_flush bit set, just delay for 1 us and retry again. In this case this btree node won't be skipped, just retry until the BTREE_NODE_journal_flush bit cleared, and free the btree node memory. Fixes: cafe56359144 ("bcache: A block layer cache") Signed-off-by: Coly Li <colyli@suse.de> Reported-and-tested-by: kbuild test robot <lkp@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:58 +08:00
clear_bit(BTREE_NODE_journal_flush, &b->flags);
mutex_unlock(&b->write_lock);
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
pr_debug("bnode %p: written by others", b);
continue;
}
if (!btree_node_dirty(b)) {
clear_bit(BTREE_NODE_journal_flush, &b->flags);
mutex_unlock(&b->write_lock);
pr_debug("bnode %p: dirty bit cleaned by others", b);
continue;
}
__bch_btree_node_write(b, NULL);
bcache: fix race in btree_flush_write() There is a race between mca_reap(), btree_node_free() and journal code btree_flush_write(), which results very rare and strange deadlock or panic and are very hard to reproduce. Let me explain how the race happens. In btree_flush_write() one btree node with oldest journal pin is selected, then it is flushed to cache device, the select-and-flush is a two steps operation. Between these two steps, there are something may happen inside the race window, - The selected btree node was reaped by mca_reap() and allocated to other requesters for other btree node. - The slected btree node was selected, flushed and released by mca shrink callback bch_mca_scan(). When btree_flush_write() tries to flush the selected btree node, firstly b->write_lock is held by mutex_lock(). If the race happens and the memory of selected btree node is allocated to other btree node, if that btree node's write_lock is held already, a deadlock very probably happens here. A worse case is the memory of the selected btree node is released, then all references to this btree node (e.g. b->write_lock) will trigger NULL pointer deference panic. This race was introduced in commit cafe56359144 ("bcache: A block layer cache"), and enlarged by commit c4dc2497d50d ("bcache: fix high CPU occupancy during journal"), which selected 128 btree nodes and flushed them one-by-one in a quite long time period. Such race is not easy to reproduce before. On a Lenovo SR650 server with 48 Xeon cores, and configure 1 NVMe SSD as cache device, a MD raid0 device assembled by 3 NVMe SSDs as backing device, this race can be observed around every 10,000 times btree_flush_write() gets called. Both deadlock and kernel panic all happened as aftermath of the race. The idea of the fix is to add a btree flag BTREE_NODE_journal_flush. It is set when selecting btree nodes, and cleared after btree nodes flushed. Then when mca_reap() selects a btree node with this bit set, this btree node will be skipped. Since mca_reap() only reaps btree node without BTREE_NODE_journal_flush flag, such race is avoided. Once corner case should be noticed, that is btree_node_free(). It might be called in some error handling code path. For example the following code piece from btree_split(), 2149 err_free2: 2150 bkey_put(b->c, &n2->key); 2151 btree_node_free(n2); 2152 rw_unlock(true, n2); 2153 err_free1: 2154 bkey_put(b->c, &n1->key); 2155 btree_node_free(n1); 2156 rw_unlock(true, n1); At line 2151 and 2155, the btree node n2 and n1 are released without mac_reap(), so BTREE_NODE_journal_flush also needs to be checked here. If btree_node_free() is called directly in such error handling path, and the selected btree node has BTREE_NODE_journal_flush bit set, just delay for 1 us and retry again. In this case this btree node won't be skipped, just retry until the BTREE_NODE_journal_flush bit cleared, and free the btree node memory. Fixes: cafe56359144 ("bcache: A block layer cache") Signed-off-by: Coly Li <colyli@suse.de> Reported-and-tested-by: kbuild test robot <lkp@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:58 +08:00
clear_bit(BTREE_NODE_journal_flush, &b->flags);
mutex_unlock(&b->write_lock);
}
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
bcache: avoid unnecessary btree nodes flushing in btree_flush_write() the commit 91be66e1318f ("bcache: performance improvement for btree_flush_write()") was an effort to flushing btree node with oldest btree node faster in following methods, - Only iterate dirty btree nodes in c->btree_cache, avoid scanning a lot of clean btree nodes. - Take c->btree_cache as a LRU-like list, aggressively flushing all dirty nodes from tail of c->btree_cache util the btree node with oldest journal entry is flushed. This is to reduce the time of holding c->bucket_lock. Guoju Fang and Shuang Li reported that they observe unexptected extra write I/Os on cache device after applying the above patch. Guoju Fang provideed more detailed diagnose information that the aggressive btree nodes flushing may cause 10x more btree nodes to flush in his workload. He points out when system memory is large enough to hold all btree nodes in memory, c->btree_cache is not a LRU-like list any more. Then the btree node with oldest journal entry is very probably not- close to the tail of c->btree_cache list. In such situation much more dirty btree nodes will be aggressively flushed before the target node is flushed. When slow SATA SSD is used as cache device, such over- aggressive flushing behavior will cause performance regression. After spending a lot of time on debug and diagnose, I find the real condition is more complicated, aggressive flushing dirty btree nodes from tail of c->btree_cache list is not a good solution. - When all btree nodes are cached in memory, c->btree_cache is not a LRU-like list, the btree nodes with oldest journal entry won't be close to the tail of the list. - There can be hundreds dirty btree nodes reference the oldest journal entry, before flushing all the nodes the oldest journal entry cannot be reclaimed. When the above two conditions mixed together, a simply flushing from tail of c->btree_cache list is really NOT a good idea. Fortunately there is still chance to make btree_flush_write() work better. Here is how this patch avoids unnecessary btree nodes flushing, - Only acquire c->journal.lock when getting oldest journal entry of fifo c->journal.pin. In rested locations check the journal entries locklessly, so their values can be changed on other cores in parallel. - In loop list_for_each_entry_safe_reverse(), checking latest front point of fifo c->journal.pin. If it is different from the original point which we get with locking c->journal.lock, it means the oldest journal entry is reclaim on other cores. At this moment, all selected dirty nodes recorded in array btree_nodes[] are all flushed and clean on other CPU cores, it is unncessary to iterate c->btree_cache any longer. Just quit the list_for_each_entry_safe_reverse() loop and the following for-loop will skip all the selected clean nodes. - Find a proper time to quit the list_for_each_entry_safe_reverse() loop. Check the refcount value of orignial fifo front point, if the value is larger than selected node number of btree_nodes[], it means more matching btree nodes should be scanned. Otherwise it means no more matching btee nodes in rest of c->btree_cache list, the loop can be quit. If the original oldest journal entry is reclaimed and fifo front point is updated, the refcount of original fifo front point will be 0, then the loop will be quit too. - Not hold c->bucket_lock too long time. c->bucket_lock is also required for space allocation for cached data, hold it for too long time will block regular I/O requests. When iterating list c->btree_cache, even there are a lot of maching btree nodes, in order to not holding c->bucket_lock for too long time, only BTREE_FLUSH_NR nodes are selected and to flush in following for-loop. With this patch, only btree nodes referencing oldest journal entry are flushed to cache device, no aggressive flushing for unnecessary btree node any more. And in order to avoid blocking regluar I/O requests, each time when btree_flush_write() called, at most only BTREE_FLUSH_NR btree nodes are selected to flush, even there are more maching btree nodes in list c->btree_cache. At last, one more thing to explain: Why it is safe to read front point of c->journal.pin without holding c->journal.lock inside the list_for_each_entry_safe_reverse() loop ? Here is my answer: When reading the front point of fifo c->journal.pin, we don't need to know the exact value of front point, we just want to check whether the value is different from the original front point (which is accurate value because we get it while c->jouranl.lock is held). For such purpose, it works as expected without holding c->journal.lock. Even the front point is changed on other CPU core and not updated to local core, and current iterating btree node has identical journal entry local as original fetched fifo front point, it is still safe. Because after holding mutex b->write_lock (with memory barrier) this btree node can be found as clean and skipped, the loop will quite latter when iterate on next node of list c->btree_cache. Fixes: 91be66e1318f ("bcache: performance improvement for btree_flush_write()") Reported-by: Guoju Fang <fangguoju@gmail.com> Reported-by: Shuang Li <psymon@bonuscloud.io> Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2020-01-24 01:01:37 +08:00
out:
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
spin_lock(&c->journal.flush_write_lock);
c->journal.btree_flushing = false;
spin_unlock(&c->journal.flush_write_lock);
}
#define last_seq(j) ((j)->seq - fifo_used(&(j)->pin) + 1)
static void journal_discard_endio(struct bio *bio)
{
struct journal_device *ja =
container_of(bio, struct journal_device, discard_bio);
struct cache *ca = container_of(ja, struct cache, journal);
atomic_set(&ja->discard_in_flight, DISCARD_DONE);
closure_wake_up(&ca->set->journal.wait);
closure_put(&ca->set->cl);
}
static void journal_discard_work(struct work_struct *work)
{
struct journal_device *ja =
container_of(work, struct journal_device, discard_work);
submit_bio(&ja->discard_bio);
}
static void do_journal_discard(struct cache *ca)
{
struct journal_device *ja = &ca->journal;
struct bio *bio = &ja->discard_bio;
if (!ca->discard) {
ja->discard_idx = ja->last_idx;
return;
}
switch (atomic_read(&ja->discard_in_flight)) {
case DISCARD_IN_FLIGHT:
return;
case DISCARD_DONE:
ja->discard_idx = (ja->discard_idx + 1) %
ca->sb.njournal_buckets;
atomic_set(&ja->discard_in_flight, DISCARD_READY);
/* fallthrough */
case DISCARD_READY:
if (ja->discard_idx == ja->last_idx)
return;
atomic_set(&ja->discard_in_flight, DISCARD_IN_FLIGHT);
bio_init(bio, bio->bi_inline_vecs, 1);
bio_set_op_attrs(bio, REQ_OP_DISCARD, 0);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_sector = bucket_to_sector(ca->set,
ca->sb.d[ja->discard_idx]);
bio_set_dev(bio, ca->bdev);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_size = bucket_bytes(ca);
bio->bi_end_io = journal_discard_endio;
closure_get(&ca->set->cl);
INIT_WORK(&ja->discard_work, journal_discard_work);
queue_work(bch_journal_wq, &ja->discard_work);
}
}
static void journal_reclaim(struct cache_set *c)
{
struct bkey *k = &c->journal.key;
struct cache *ca;
uint64_t last_seq;
unsigned int iter, n = 0;
atomic_t p __maybe_unused;
atomic_long_inc(&c->reclaim);
while (!atomic_read(&fifo_front(&c->journal.pin)))
fifo_pop(&c->journal.pin, p);
last_seq = last_seq(&c->journal);
/* Update last_idx */
for_each_cache(ca, c, iter) {
struct journal_device *ja = &ca->journal;
while (ja->last_idx != ja->cur_idx &&
ja->seq[ja->last_idx] < last_seq)
ja->last_idx = (ja->last_idx + 1) %
ca->sb.njournal_buckets;
}
for_each_cache(ca, c, iter)
do_journal_discard(ca);
if (c->journal.blocks_free)
goto out;
/*
* Allocate:
* XXX: Sort by free journal space
*/
for_each_cache(ca, c, iter) {
struct journal_device *ja = &ca->journal;
unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets;
/* No space available on this device */
if (next == ja->discard_idx)
continue;
ja->cur_idx = next;
k->ptr[n++] = MAKE_PTR(0,
bucket_to_sector(c, ca->sb.d[ja->cur_idx]),
ca->sb.nr_this_dev);
atomic_long_inc(&c->reclaimed_journal_buckets);
}
bcache: never set KEY_PTRS of journal key to 0 in journal_reclaim() In journal_reclaim() ja->cur_idx of each cache will be update to reclaim available journal buckets. Variable 'int n' is used to count how many cache is successfully reclaimed, then n is set to c->journal.key by SET_KEY_PTRS(). Later in journal_write_unlocked(), a for_each_cache() loop will write the jset data onto each cache. The problem is, if all jouranl buckets on each cache is full, the following code in journal_reclaim(), 529 for_each_cache(ca, c, iter) { 530 struct journal_device *ja = &ca->journal; 531 unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets; 532 533 /* No space available on this device */ 534 if (next == ja->discard_idx) 535 continue; 536 537 ja->cur_idx = next; 538 k->ptr[n++] = MAKE_PTR(0, 539 bucket_to_sector(c, ca->sb.d[ja->cur_idx]), 540 ca->sb.nr_this_dev); 541 } 542 543 bkey_init(k); 544 SET_KEY_PTRS(k, n); If there is no available bucket to reclaim, the if() condition at line 534 will always true, and n remains 0. Then at line 544, SET_KEY_PTRS() will set KEY_PTRS field of c->journal.key to 0. Setting KEY_PTRS field of c->journal.key to 0 is wrong. Because in journal_write_unlocked() the journal data is written in following loop, 649 for (i = 0; i < KEY_PTRS(k); i++) { 650-671 submit journal data to cache device 672 } If KEY_PTRS field is set to 0 in jouranl_reclaim(), the journal data won't be written to cache device here. If system crahed or rebooted before bkeys of the lost journal entries written into btree nodes, data corruption will be reported during bcache reload after rebooting the system. Indeed there is only one cache in a cache set, there is no need to set KEY_PTRS field in journal_reclaim() at all. But in order to keep the for_each_cache() logic consistent for now, this patch fixes the above problem by not setting 0 KEY_PTRS of journal key, if there is no bucket available to reclaim. Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-04-25 00:48:33 +08:00
if (n) {
bkey_init(k);
SET_KEY_PTRS(k, n);
c->journal.blocks_free = c->sb.bucket_size >> c->block_bits;
bcache: never set KEY_PTRS of journal key to 0 in journal_reclaim() In journal_reclaim() ja->cur_idx of each cache will be update to reclaim available journal buckets. Variable 'int n' is used to count how many cache is successfully reclaimed, then n is set to c->journal.key by SET_KEY_PTRS(). Later in journal_write_unlocked(), a for_each_cache() loop will write the jset data onto each cache. The problem is, if all jouranl buckets on each cache is full, the following code in journal_reclaim(), 529 for_each_cache(ca, c, iter) { 530 struct journal_device *ja = &ca->journal; 531 unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets; 532 533 /* No space available on this device */ 534 if (next == ja->discard_idx) 535 continue; 536 537 ja->cur_idx = next; 538 k->ptr[n++] = MAKE_PTR(0, 539 bucket_to_sector(c, ca->sb.d[ja->cur_idx]), 540 ca->sb.nr_this_dev); 541 } 542 543 bkey_init(k); 544 SET_KEY_PTRS(k, n); If there is no available bucket to reclaim, the if() condition at line 534 will always true, and n remains 0. Then at line 544, SET_KEY_PTRS() will set KEY_PTRS field of c->journal.key to 0. Setting KEY_PTRS field of c->journal.key to 0 is wrong. Because in journal_write_unlocked() the journal data is written in following loop, 649 for (i = 0; i < KEY_PTRS(k); i++) { 650-671 submit journal data to cache device 672 } If KEY_PTRS field is set to 0 in jouranl_reclaim(), the journal data won't be written to cache device here. If system crahed or rebooted before bkeys of the lost journal entries written into btree nodes, data corruption will be reported during bcache reload after rebooting the system. Indeed there is only one cache in a cache set, there is no need to set KEY_PTRS field in journal_reclaim() at all. But in order to keep the for_each_cache() logic consistent for now, this patch fixes the above problem by not setting 0 KEY_PTRS of journal key, if there is no bucket available to reclaim. Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-04-25 00:48:33 +08:00
}
out:
if (!journal_full(&c->journal))
__closure_wake_up(&c->journal.wait);
}
void bch_journal_next(struct journal *j)
{
atomic_t p = { 1 };
j->cur = (j->cur == j->w)
? &j->w[1]
: &j->w[0];
/*
* The fifo_push() needs to happen at the same time as j->seq is
* incremented for last_seq() to be calculated correctly
*/
BUG_ON(!fifo_push(&j->pin, p));
atomic_set(&fifo_back(&j->pin), 1);
j->cur->data->seq = ++j->seq;
j->cur->dirty = false;
j->cur->need_write = false;
j->cur->data->keys = 0;
if (fifo_full(&j->pin))
pr_debug("journal_pin full (%zu)", fifo_used(&j->pin));
}
static void journal_write_endio(struct bio *bio)
{
struct journal_write *w = bio->bi_private;
cache_set_err_on(bio->bi_status, w->c, "journal io error");
closure_put(&w->c->journal.io);
}
static void journal_write(struct closure *cl);
static void journal_write_done(struct closure *cl)
{
struct journal *j = container_of(cl, struct journal, io);
struct journal_write *w = (j->cur == j->w)
? &j->w[1]
: &j->w[0];
__closure_wake_up(&w->wait);
continue_at_nobarrier(cl, journal_write, bch_journal_wq);
}
static void journal_write_unlock(struct closure *cl)
__releases(&c->journal.lock)
{
struct cache_set *c = container_of(cl, struct cache_set, journal.io);
c->journal.io_in_flight = 0;
spin_unlock(&c->journal.lock);
}
static void journal_write_unlocked(struct closure *cl)
__releases(c->journal.lock)
{
struct cache_set *c = container_of(cl, struct cache_set, journal.io);
struct cache *ca;
struct journal_write *w = c->journal.cur;
struct bkey *k = &c->journal.key;
unsigned int i, sectors = set_blocks(w->data, block_bytes(c)) *
c->sb.block_size;
struct bio *bio;
struct bio_list list;
bio_list_init(&list);
if (!w->need_write) {
closure_return_with_destructor(cl, journal_write_unlock);
return;
} else if (journal_full(&c->journal)) {
journal_reclaim(c);
spin_unlock(&c->journal.lock);
btree_flush_write(c);
continue_at(cl, journal_write, bch_journal_wq);
return;
}
c->journal.blocks_free -= set_blocks(w->data, block_bytes(c));
w->data->btree_level = c->root->level;
bkey_copy(&w->data->btree_root, &c->root->key);
bkey_copy(&w->data->uuid_bucket, &c->uuid_bucket);
for_each_cache(ca, c, i)
w->data->prio_bucket[ca->sb.nr_this_dev] = ca->prio_buckets[0];
w->data->magic = jset_magic(&c->sb);
w->data->version = BCACHE_JSET_VERSION;
w->data->last_seq = last_seq(&c->journal);
w->data->csum = csum_set(w->data);
for (i = 0; i < KEY_PTRS(k); i++) {
ca = PTR_CACHE(c, k, i);
bio = &ca->journal.bio;
atomic_long_add(sectors, &ca->meta_sectors_written);
bio_reset(bio);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_sector = PTR_OFFSET(k, i);
bio_set_dev(bio, ca->bdev);
block: Abstract out bvec iterator Immutable biovecs are going to require an explicit iterator. To implement immutable bvecs, a later patch is going to add a bi_bvec_done member to this struct; for now, this patch effectively just renames things. Signed-off-by: Kent Overstreet <kmo@daterainc.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Geert Uytterhoeven <geert@linux-m68k.org> Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org> Cc: Paul Mackerras <paulus@samba.org> Cc: "Ed L. Cashin" <ecashin@coraid.com> Cc: Nick Piggin <npiggin@kernel.dk> Cc: Lars Ellenberg <drbd-dev@lists.linbit.com> Cc: Jiri Kosina <jkosina@suse.cz> Cc: Matthew Wilcox <willy@linux.intel.com> Cc: Geoff Levand <geoff@infradead.org> Cc: Yehuda Sadeh <yehuda@inktank.com> Cc: Sage Weil <sage@inktank.com> Cc: Alex Elder <elder@inktank.com> Cc: ceph-devel@vger.kernel.org Cc: Joshua Morris <josh.h.morris@us.ibm.com> Cc: Philip Kelleher <pjk1939@linux.vnet.ibm.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: "Michael S. Tsirkin" <mst@redhat.com> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Cc: Jeremy Fitzhardinge <jeremy@goop.org> Cc: Neil Brown <neilb@suse.de> Cc: Alasdair Kergon <agk@redhat.com> Cc: Mike Snitzer <snitzer@redhat.com> Cc: dm-devel@redhat.com Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> Cc: Heiko Carstens <heiko.carstens@de.ibm.com> Cc: linux390@de.ibm.com Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: Benny Halevy <bhalevy@tonian.com> Cc: "James E.J. Bottomley" <JBottomley@parallels.com> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: "Nicholas A. Bellinger" <nab@linux-iscsi.org> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Chris Mason <chris.mason@fusionio.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Andreas Dilger <adilger.kernel@dilger.ca> Cc: Jaegeuk Kim <jaegeuk.kim@samsung.com> Cc: Steven Whitehouse <swhiteho@redhat.com> Cc: Dave Kleikamp <shaggy@kernel.org> Cc: Joern Engel <joern@logfs.org> Cc: Prasad Joshi <prasadjoshi.linux@gmail.com> Cc: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: KONISHI Ryusuke <konishi.ryusuke@lab.ntt.co.jp> Cc: Mark Fasheh <mfasheh@suse.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Ben Myers <bpm@sgi.com> Cc: xfs@oss.sgi.com Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Len Brown <len.brown@intel.com> Cc: Pavel Machek <pavel@ucw.cz> Cc: "Rafael J. Wysocki" <rjw@sisk.pl> Cc: Herton Ronaldo Krzesinski <herton.krzesinski@canonical.com> Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Guo Chao <yan@linux.vnet.ibm.com> Cc: Tejun Heo <tj@kernel.org> Cc: Asai Thambi S P <asamymuthupa@micron.com> Cc: Selvan Mani <smani@micron.com> Cc: Sam Bradshaw <sbradshaw@micron.com> Cc: Wei Yongjun <yongjun_wei@trendmicro.com.cn> Cc: "Roger Pau Monné" <roger.pau@citrix.com> Cc: Jan Beulich <jbeulich@suse.com> Cc: Stefano Stabellini <stefano.stabellini@eu.citrix.com> Cc: Ian Campbell <Ian.Campbell@citrix.com> Cc: Sebastian Ott <sebott@linux.vnet.ibm.com> Cc: Christian Borntraeger <borntraeger@de.ibm.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Jiang Liu <jiang.liu@huawei.com> Cc: Nitin Gupta <ngupta@vflare.org> Cc: Jerome Marchand <jmarchand@redhat.com> Cc: Joe Perches <joe@perches.com> Cc: Peng Tao <tao.peng@emc.com> Cc: Andy Adamson <andros@netapp.com> Cc: fanchaoting <fanchaoting@cn.fujitsu.com> Cc: Jie Liu <jeff.liu@oracle.com> Cc: Sunil Mushran <sunil.mushran@gmail.com> Cc: "Martin K. Petersen" <martin.petersen@oracle.com> Cc: Namjae Jeon <namjae.jeon@samsung.com> Cc: Pankaj Kumar <pankaj.km@samsung.com> Cc: Dan Magenheimer <dan.magenheimer@oracle.com> Cc: Mel Gorman <mgorman@suse.de>6
2013-10-12 06:44:27 +08:00
bio->bi_iter.bi_size = sectors << 9;
bio->bi_end_io = journal_write_endio;
bio->bi_private = w;
bio_set_op_attrs(bio, REQ_OP_WRITE,
REQ_SYNC|REQ_META|REQ_PREFLUSH|REQ_FUA);
bch_bio_map(bio, w->data);
trace_bcache_journal_write(bio, w->data->keys);
bio_list_add(&list, bio);
SET_PTR_OFFSET(k, i, PTR_OFFSET(k, i) + sectors);
ca->journal.seq[ca->journal.cur_idx] = w->data->seq;
}
bcache: never set KEY_PTRS of journal key to 0 in journal_reclaim() In journal_reclaim() ja->cur_idx of each cache will be update to reclaim available journal buckets. Variable 'int n' is used to count how many cache is successfully reclaimed, then n is set to c->journal.key by SET_KEY_PTRS(). Later in journal_write_unlocked(), a for_each_cache() loop will write the jset data onto each cache. The problem is, if all jouranl buckets on each cache is full, the following code in journal_reclaim(), 529 for_each_cache(ca, c, iter) { 530 struct journal_device *ja = &ca->journal; 531 unsigned int next = (ja->cur_idx + 1) % ca->sb.njournal_buckets; 532 533 /* No space available on this device */ 534 if (next == ja->discard_idx) 535 continue; 536 537 ja->cur_idx = next; 538 k->ptr[n++] = MAKE_PTR(0, 539 bucket_to_sector(c, ca->sb.d[ja->cur_idx]), 540 ca->sb.nr_this_dev); 541 } 542 543 bkey_init(k); 544 SET_KEY_PTRS(k, n); If there is no available bucket to reclaim, the if() condition at line 534 will always true, and n remains 0. Then at line 544, SET_KEY_PTRS() will set KEY_PTRS field of c->journal.key to 0. Setting KEY_PTRS field of c->journal.key to 0 is wrong. Because in journal_write_unlocked() the journal data is written in following loop, 649 for (i = 0; i < KEY_PTRS(k); i++) { 650-671 submit journal data to cache device 672 } If KEY_PTRS field is set to 0 in jouranl_reclaim(), the journal data won't be written to cache device here. If system crahed or rebooted before bkeys of the lost journal entries written into btree nodes, data corruption will be reported during bcache reload after rebooting the system. Indeed there is only one cache in a cache set, there is no need to set KEY_PTRS field in journal_reclaim() at all. But in order to keep the for_each_cache() logic consistent for now, this patch fixes the above problem by not setting 0 KEY_PTRS of journal key, if there is no bucket available to reclaim. Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Cc: stable@vger.kernel.org Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-04-25 00:48:33 +08:00
/* If KEY_PTRS(k) == 0, this jset gets lost in air */
BUG_ON(i == 0);
atomic_dec_bug(&fifo_back(&c->journal.pin));
bch_journal_next(&c->journal);
journal_reclaim(c);
spin_unlock(&c->journal.lock);
while ((bio = bio_list_pop(&list)))
bcache: add CACHE_SET_IO_DISABLE to struct cache_set flags When too many I/Os failed on cache device, bch_cache_set_error() is called in the error handling code path to retire whole problematic cache set. If new I/O requests continue to come and take refcount dc->count, the cache set won't be retired immediately, this is a problem. Further more, there are several kernel thread and self-armed kernel work may still running after bch_cache_set_error() is called. It needs to wait quite a while for them to stop, or they won't stop at all. They also prevent the cache set from being retired. The solution in this patch is, to add per cache set flag to disable I/O request on this cache and all attached backing devices. Then new coming I/O requests can be rejected in *_make_request() before taking refcount, kernel threads and self-armed kernel worker can stop very fast when flags bit CACHE_SET_IO_DISABLE is set. Because bcache also do internal I/Os for writeback, garbage collection, bucket allocation, journaling, this kind of I/O should be disabled after bch_cache_set_error() is called. So closure_bio_submit() is modified to check whether CACHE_SET_IO_DISABLE is set on cache_set->flags. If set, closure_bio_submit() will set bio->bi_status to BLK_STS_IOERR and return, generic_make_request() won't be called. A sysfs interface is also added to set or clear CACHE_SET_IO_DISABLE bit from cache_set->flags, to disable or enable cache set I/O for debugging. It is helpful to trigger more corner case issues for failed cache device. Changelog v4, add wait_for_kthread_stop(), and call it before exits writeback and gc kernel threads. v3, change CACHE_SET_IO_DISABLE from 4 to 3, since it is bit index. remove "bcache: " prefix when printing out kernel message. v2, more changes by previous review, - Use CACHE_SET_IO_DISABLE of cache_set->flags, suggested by Junhui. - Check CACHE_SET_IO_DISABLE in bch_btree_gc() to stop a while-loop, this is reported and inspired from origal patch of Pavel Vazharov. v1, initial version. Signed-off-by: Coly Li <colyli@suse.de> Reviewed-by: Hannes Reinecke <hare@suse.com> Reviewed-by: Michael Lyle <mlyle@lyle.org> Cc: Junhui Tang <tang.junhui@zte.com.cn> Cc: Michael Lyle <mlyle@lyle.org> Cc: Pavel Vazharov <freakpv@gmail.com> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2018-03-19 08:36:17 +08:00
closure_bio_submit(c, bio, cl);
continue_at(cl, journal_write_done, NULL);
}
static void journal_write(struct closure *cl)
{
struct cache_set *c = container_of(cl, struct cache_set, journal.io);
spin_lock(&c->journal.lock);
journal_write_unlocked(cl);
}
static void journal_try_write(struct cache_set *c)
__releases(c->journal.lock)
{
struct closure *cl = &c->journal.io;
struct journal_write *w = c->journal.cur;
w->need_write = true;
if (!c->journal.io_in_flight) {
c->journal.io_in_flight = 1;
closure_call(cl, journal_write_unlocked, NULL, &c->cl);
} else {
spin_unlock(&c->journal.lock);
}
}
static struct journal_write *journal_wait_for_write(struct cache_set *c,
unsigned int nkeys)
__acquires(&c->journal.lock)
{
size_t sectors;
struct closure cl;
bool wait = false;
closure_init_stack(&cl);
spin_lock(&c->journal.lock);
while (1) {
struct journal_write *w = c->journal.cur;
sectors = __set_blocks(w->data, w->data->keys + nkeys,
block_bytes(c)) * c->sb.block_size;
if (sectors <= min_t(size_t,
c->journal.blocks_free * c->sb.block_size,
PAGE_SECTORS << JSET_BITS))
return w;
if (wait)
closure_wait(&c->journal.wait, &cl);
if (!journal_full(&c->journal)) {
if (wait)
trace_bcache_journal_entry_full(c);
/*
* XXX: If we were inserting so many keys that they
* won't fit in an _empty_ journal write, we'll
* deadlock. For now, handle this in
* bch_keylist_realloc() - but something to think about.
*/
BUG_ON(!w->data->keys);
journal_try_write(c); /* unlocks */
} else {
if (wait)
trace_bcache_journal_full(c);
journal_reclaim(c);
spin_unlock(&c->journal.lock);
btree_flush_write(c);
}
closure_sync(&cl);
spin_lock(&c->journal.lock);
wait = true;
}
}
static void journal_write_work(struct work_struct *work)
{
struct cache_set *c = container_of(to_delayed_work(work),
struct cache_set,
journal.work);
spin_lock(&c->journal.lock);
if (c->journal.cur->dirty)
journal_try_write(c);
else
spin_unlock(&c->journal.lock);
}
/*
* Entry point to the journalling code - bio_insert() and btree_invalidate()
* pass bch_journal() a list of keys to be journalled, and then
* bch_journal() hands those same keys off to btree_insert_async()
*/
atomic_t *bch_journal(struct cache_set *c,
struct keylist *keys,
struct closure *parent)
{
struct journal_write *w;
atomic_t *ret;
/* No journaling if CACHE_SET_IO_DISABLE set already */
if (unlikely(test_bit(CACHE_SET_IO_DISABLE, &c->flags)))
return NULL;
if (!CACHE_SYNC(&c->sb))
return NULL;
w = journal_wait_for_write(c, bch_keylist_nkeys(keys));
memcpy(bset_bkey_last(w->data), keys->keys, bch_keylist_bytes(keys));
w->data->keys += bch_keylist_nkeys(keys);
ret = &fifo_back(&c->journal.pin);
atomic_inc(ret);
if (parent) {
closure_wait(&w->wait, parent);
journal_try_write(c);
} else if (!w->dirty) {
w->dirty = true;
schedule_delayed_work(&c->journal.work,
msecs_to_jiffies(c->journal_delay_ms));
spin_unlock(&c->journal.lock);
} else {
spin_unlock(&c->journal.lock);
}
return ret;
}
void bch_journal_meta(struct cache_set *c, struct closure *cl)
{
struct keylist keys;
atomic_t *ref;
bch_keylist_init(&keys);
ref = bch_journal(c, &keys, cl);
if (ref)
atomic_dec_bug(ref);
}
void bch_journal_free(struct cache_set *c)
{
free_pages((unsigned long) c->journal.w[1].data, JSET_BITS);
free_pages((unsigned long) c->journal.w[0].data, JSET_BITS);
free_fifo(&c->journal.pin);
}
int bch_journal_alloc(struct cache_set *c)
{
struct journal *j = &c->journal;
spin_lock_init(&j->lock);
bcache: performance improvement for btree_flush_write() This patch improves performance for btree_flush_write() in following ways, - Use another spinlock journal.flush_write_lock to replace the very hot journal.lock. We don't have to use journal.lock here, selecting candidate btree nodes takes a lot of time, hold journal.lock here will block other jouranling threads and drop the overall I/O performance. - Only select flushing btree node from c->btree_cache list. When the machine has a large system memory, mca cache may have a huge number of cached btree nodes. Iterating all the cached nodes will take a lot of CPU time, and most of the nodes on c->btree_cache_freeable and c->btree_cache_freed lists are cleared and have need to flush. So only travel mca list c->btree_cache to select flushing btree node should be enough for most of the cases. - Don't iterate whole c->btree_cache list, only reversely select first BTREE_FLUSH_NR btree nodes to flush. Iterate all btree nodes from c->btree_cache and select the oldest journal pin btree nodes consumes huge number of CPU cycles if the list is huge (push and pop a node into/out of a heap is expensive). The last several dirty btree nodes on the tail of c->btree_cache list are earlest allocated and cached btree nodes, they are relative to the oldest journal pin btree nodes. Therefore only flushing BTREE_FLUSH_NR btree nodes from tail of c->btree_cache probably includes the oldest journal pin btree nodes. In my testing, the above change decreases 50%+ CPU consumption when journal space is full. Some times IOPS drops to 0 for 5-8 seconds, comparing blocking I/O for 120+ seconds in previous code, this is much better. Maybe there is room to improve in future, but at this momment the fix looks fine and performs well in my testing. Signed-off-by: Coly Li <colyli@suse.de> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-06-28 19:59:59 +08:00
spin_lock_init(&j->flush_write_lock);
INIT_DELAYED_WORK(&j->work, journal_write_work);
c->journal_delay_ms = 100;
j->w[0].c = c;
j->w[1].c = c;
if (!(init_fifo(&j->pin, JOURNAL_PIN, GFP_KERNEL)) ||
!(j->w[0].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)) ||
!(j->w[1].data = (void *) __get_free_pages(GFP_KERNEL, JSET_BITS)))
return -ENOMEM;
return 0;
}