[PATCH 00/15] bcache patches for 4.20

[PATCH 00/15] bcache patches for 4.20

[Coly Li]

Hi Jens,

Here are patches for Linux 4.20. In this run we don't have big changes,
most of the patches for fixes or code cleanup.

Ben Peddell resends a patch which had no response for a while, I pick
it for this version. Our new developer Dongbo Cao contributes a bug fix
to avoid kernel panic for too small cache set size, with other code
cleanup changes. Shenghui and me have several code cleanup and minor
fixes, and Junhui again contributes helpful fixes for several real bugs.

Please pull them for 4.20 merge window. Thanks in advance.

Coly Li
---

Ben Peddell (1):
  bcache: Populate writeback_rate_minimum attribute

Coly Li (3):
  bcache: use REQ_PRIO to indicate bio for metadata
  bcache: fix typo in code comments of closure_return_with_destructor()
  bcache: replace hard coded number with BUCKET_GC_GEN_MAX

Dongbo Cao (3):
  bcache: remove useless parameter of bch_debug_init()
  bcache: split combined if-condition code into separate ones
  bcache: panic fix for making cache device

Shenghui Wang (4):
  bcache: account size of buckets used in uuid write to
    ca->meta_sectors_written
  bcache: recal cached_dev_sectors on detach
  bcache: remove unused bch_passthrough_cache
  bcache: use MAX_CACHES_PER_SET instead of magic number 8 in
    __bch_bucket_alloc_set

Tang Junhui (4):
  bcache: trace missed reading by cache_missed
  bcache: fix ioctl in flash device
  bcache: correct dirty data statistics
  bcache: fix miss key refill->end in writeback

 drivers/md/bcache/alloc.c   |   2 +-
 drivers/md/bcache/bcache.h  |   2 +-
 drivers/md/bcache/btree.c   |   2 +-
 drivers/md/bcache/closure.h |   3 +-
 drivers/md/bcache/debug.c   |   2 +-
 drivers/md/bcache/extents.c |   2 +-
 drivers/md/bcache/request.c |   9 ++-
 drivers/md/bcache/request.h |   2 +-
 drivers/md/bcache/super.c   | 113 +++++++++++++++++++++++++++++-------
 drivers/md/bcache/sysfs.c   |   2 +
 10 files changed, 109 insertions(+), 30 deletions(-)

-- 
2.19.0

[PATCH 01/15] bcache: account size of buckets used in uuid write to ca->meta_sectors_written

[Coly Li]

From: Shenghui Wang <shhuiw@foxmail.com>

UUIDs are considered as metadata. __uuid_write should add the number
of buckets (in sectors) written to disk to ca->meta_sectors_written.
Currently only 1 bucket is used in uuid write.

Steps to test:
1) create a fresh backing device and a fresh cache device separately.
   The backing device didn't attach to any cache set.
2) cd /sys/block/<cache device>/bcache
   cat metadata_written      // record the output value
   cat bucket_size
3) attach the backing device to cache set
4) cat metadata_written
   The output value is almost the same as the value in step 2
   before the change.
   After the change, the value is bigger about 1 bucket size.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Reviewed-by: Tang Junhui <tang.junhui.linux@gmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/super.c | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 94c756c66bd7..448e531e8c2d 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -417,6 +417,7 @@ static int __uuid_write(struct cache_set *c)
 {
 	BKEY_PADDED(key) k;
 	struct closure cl;
+	struct cache *ca;
 
 	closure_init_stack(&cl);
 	lockdep_assert_held(&bch_register_lock);
@@ -428,6 +429,10 @@ static int __uuid_write(struct cache_set *c)
 	uuid_io(c, REQ_OP_WRITE, 0, &k.key, &cl);
 	closure_sync(&cl);
 
+	/* Only one bucket used for uuid write */
+	ca = PTR_CACHE(c, &k.key, 0);
+	atomic_long_add(ca->sb.bucket_size, &ca->meta_sectors_written);
+
 	bkey_copy(&c->uuid_bucket, &k.key);
 	bkey_put(c, &k.key);
 	return 0;
-- 
2.19.0

[PATCH 02/15] bcache: trace missed reading by cache_missed

[Coly Li]

From: Tang Junhui <tang.junhui.linux@gmail.com>

Missed reading IOs are identified by s->cache_missed, not the
s->cache_miss, so in trace_bcache_read() using trace_bcache_read
to identify whether the IO is missed or not.

Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/request.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 51be355a3309..4946d486f734 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -850,7 +850,7 @@ static void cached_dev_read_done_bh(struct closure *cl)
 
 	bch_mark_cache_accounting(s->iop.c, s->d,
 				  !s->cache_missed, s->iop.bypass);
-	trace_bcache_read(s->orig_bio, !s->cache_miss, s->iop.bypass);
+	trace_bcache_read(s->orig_bio, !s->cache_missed, s->iop.bypass);
 
 	if (s->iop.status)
 		continue_at_nobarrier(cl, cached_dev_read_error, bcache_wq);
-- 
2.19.0

[PATCH 03/15] bcache: use REQ_PRIO to indicate bio for metadata

[Coly Li]

In cached_dev_cache_miss() and check_should_bypass(), REQ_META is used
to check whether a bio is for metadata request. REQ_META is used for
blktrace, the correct REQ_ flag should be REQ_PRIO. This flag means the
bio should be prior to other bio, and frequently be used to indicate
metadata io in file system code.

This patch replaces REQ_META with correct flag REQ_PRIO.

CC Adam Manzanares because he explains to me what REQ_PRIO is for.

Signed-off-by: Coly Li <colyli@suse.de>
Cc: Adam Manzanares <adam.manzanares@wdc.com>
---
 drivers/md/bcache/request.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index 4946d486f734..ee15fb039fd0 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -395,7 +395,7 @@ static bool check_should_bypass(struct cached_dev *dc, struct bio *bio)
 	 * unless the read-ahead request is for metadata (eg, for gfs2).
 	 */
 	if (bio->bi_opf & (REQ_RAHEAD|REQ_BACKGROUND) &&
-	    !(bio->bi_opf & REQ_META))
+	    !(bio->bi_opf & REQ_PRIO))
 		goto skip;
 
 	if (bio->bi_iter.bi_sector & (c->sb.block_size - 1) ||
@@ -877,7 +877,7 @@ static int cached_dev_cache_miss(struct btree *b, struct search *s,
 	}
 
 	if (!(bio->bi_opf & REQ_RAHEAD) &&
-	    !(bio->bi_opf & REQ_META) &&
+	    !(bio->bi_opf & REQ_PRIO) &&
 	    s->iop.c->gc_stats.in_use < CUTOFF_CACHE_READA)
 		reada = min_t(sector_t, dc->readahead >> 9,
 			      get_capacity(bio->bi_disk) - bio_end_sector(bio));
-- 
2.19.0

[PATCH 04/15] bcache: fix ioctl in flash device

[Coly Li]

From: Tang Junhui <tang.junhui.linux@gmail.com>

When doing ioctl in flash device, it will call ioctl_dev() in super.c,
then we should not to get cached device since flash only device has
no backend device. This patch just move the jugement dc->io_disable
to cached_dev_ioctl() to make ioctl in flash device correctly.

Fixes: 0f0709e6bfc3c ("bcache: stop bcache device when backing device is offline")
Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/request.c | 3 +++
 drivers/md/bcache/super.c   | 4 ----
 2 files changed, 3 insertions(+), 4 deletions(-)

diff --git a/drivers/md/bcache/request.c b/drivers/md/bcache/request.c
index ee15fb039fd0..3bf35914bb57 100644
--- a/drivers/md/bcache/request.c
+++ b/drivers/md/bcache/request.c
@@ -1218,6 +1218,9 @@ static int cached_dev_ioctl(struct bcache_device *d, fmode_t mode,
 {
 	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
 
+	if (dc->io_disable)
+		return -EIO;
+
 	return __blkdev_driver_ioctl(dc->bdev, mode, cmd, arg);
 }
 
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 448e531e8c2d..a99af19d2f91 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -647,10 +647,6 @@ static int ioctl_dev(struct block_device *b, fmode_t mode,
 		     unsigned int cmd, unsigned long arg)
 {
 	struct bcache_device *d = b->bd_disk->private_data;
-	struct cached_dev *dc = container_of(d, struct cached_dev, disk);
-
-	if (dc->io_disable)
-		return -EIO;
 
 	return d->ioctl(d, mode, cmd, arg);
 }
-- 
2.19.0

[PATCH 05/15] bcache: fix typo in code comments of closure_return_with_destructor()

[Coly Li]

The code comments of closure_return_with_destructor() in closure.h makrs
function name as closure_return(). This patch fixes this type with the
correct name - closure_return_with_destructor.

Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/closure.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/md/bcache/closure.h b/drivers/md/bcache/closure.h
index eca0d496b686..c88cdc4ae4ec 100644
--- a/drivers/md/bcache/closure.h
+++ b/drivers/md/bcache/closure.h
@@ -345,7 +345,8 @@ do {									\
 } while (0)
 
 /**
- * closure_return - finish execution of a closure, with destructor
+ * closure_return_with_destructor - finish execution of a closure,
+ *				    with destructor
  *
  * Works like closure_return(), except @destructor will be called when all
  * outstanding refs on @cl have been dropped; @destructor may be used to safely
-- 
2.19.0

[PATCH 06/15] bcache: correct dirty data statistics

[Coly Li]

From: Tang Junhui <tang.junhui.linux@gmail.com>

When bcache device is clean, dirty keys may still exist after
journal replay, so we need to count these dirty keys even
device in clean status, otherwise after writeback, the amount
of dirty data would be incorrect.

Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/super.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index a99af19d2f91..4989c7d4d4d0 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1152,11 +1152,12 @@ int bch_cached_dev_attach(struct cached_dev *dc, struct cache_set *c,
 	}
 
 	if (BDEV_STATE(&dc->sb) == BDEV_STATE_DIRTY) {
-		bch_sectors_dirty_init(&dc->disk);
 		atomic_set(&dc->has_dirty, 1);
 		bch_writeback_queue(dc);
 	}
 
+	bch_sectors_dirty_init(&dc->disk);
+
 	bch_cached_dev_run(dc);
 	bcache_device_link(&dc->disk, c, "bdev");
 	atomic_inc(&c->attached_dev_nr);
-- 
2.19.0

[PATCH 07/15] bcache: Populate writeback_rate_minimum attribute

[Coly Li]

From: Ben Peddell <klightspeed@killerwolves.net>

Forgot to include the maintainers with my first email.

Somewhere between Michael Lyle's original
"bcache: PI controller for writeback rate V2" patch dated 07 Sep 2017
and 1d316e6 bcache: implement PI controller for writeback rate,
the mapping of the writeback_rate_minimum attribute was dropped.

Re-add the missing sysfs writeback_rate_minimum attribute mapping to
"allow the user to specify a minimum rate at which dirty blocks are
retired."

Fixes: 1d316e6 bcache: implement PI controller for writeback rate
Signed-off-by: Ben Peddell <klightspeed@killerwolves.net>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/sysfs.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/md/bcache/sysfs.c b/drivers/md/bcache/sysfs.c
index 150cf4f4cf74..26f035a0c5b9 100644
--- a/drivers/md/bcache/sysfs.c
+++ b/drivers/md/bcache/sysfs.c
@@ -285,6 +285,7 @@ STORE(__cached_dev)
 			    1, WRITEBACK_RATE_UPDATE_SECS_MAX);
 	d_strtoul(writeback_rate_i_term_inverse);
 	d_strtoul_nonzero(writeback_rate_p_term_inverse);
+	d_strtoul_nonzero(writeback_rate_minimum);
 
 	sysfs_strtoul_clamp(io_error_limit, dc->error_limit, 0, INT_MAX);
 
@@ -412,6 +413,7 @@ static struct attribute *bch_cached_dev_files[] = {
 	&sysfs_writeback_rate_update_seconds,
 	&sysfs_writeback_rate_i_term_inverse,
 	&sysfs_writeback_rate_p_term_inverse,
+	&sysfs_writeback_rate_minimum,
 	&sysfs_writeback_rate_debug,
 	&sysfs_errors,
 	&sysfs_io_error_limit,
-- 
2.19.0

Re: [PATCH 07/15] bcache: Populate writeback_rate_minimum attribute

[Jens Axboe]

On 10/8/18 6:41 AM, Coly Li wrote:
> From: Ben Peddell <klightspeed@killerwolves.net>
> 
> Forgot to include the maintainers with my first email.
> 
> Somewhere between Michael Lyle's original
> "bcache: PI controller for writeback rate V2" patch dated 07 Sep 2017
> and 1d316e6 bcache: implement PI controller for writeback rate,
> the mapping of the writeback_rate_minimum attribute was dropped.
> 
> Re-add the missing sysfs writeback_rate_minimum attribute mapping to
> "allow the user to specify a minimum rate at which dirty blocks are
> retired."
> 
> Fixes: 1d316e6 bcache: implement PI controller for writeback rate

Remember to use the correct format for the Fixes line.

-- 
Jens Axboe

Re: [PATCH 07/15] bcache: Populate writeback_rate_minimum attribute

[Coly Li]

On 2018/10/8 下午10:19, Jens Axboe wrote:
> On 10/8/18 6:41 AM, Coly Li wrote:
>> From: Ben Peddell <klightspeed@killerwolves.net>
>>
>> Forgot to include the maintainers with my first email.
>>
>> Somewhere between Michael Lyle's original
>> "bcache: PI controller for writeback rate V2" patch dated 07 Sep 2017
>> and 1d316e6 bcache: implement PI controller for writeback rate,
>> the mapping of the writeback_rate_minimum attribute was dropped.
>>
>> Re-add the missing sysfs writeback_rate_minimum attribute mapping to
>> "allow the user to specify a minimum rate at which dirty blocks are
>> retired."
>>
>> Fixes: 1d316e6 bcache: implement PI controller for writeback rate
> 
> Remember to use the correct format for the Fixes line.
> 

Hi Jens,

This is my fault, it slips from my eyes. I'd like to resend a fixed up
version to you.

BTW, now I only use checkpatches.pl to check patch format, do you use
extra scripts as format checking hooks ?

Thanks.

Coly Li

Re: [PATCH 07/15] bcache: Populate writeback_rate_minimum attribute

[Jens Axboe]

On 10/8/18 9:02 AM, Coly Li wrote:
> On 2018/10/8 下午10:19, Jens Axboe wrote:
>> On 10/8/18 6:41 AM, Coly Li wrote:
>>> From: Ben Peddell <klightspeed@killerwolves.net>
>>>
>>> Forgot to include the maintainers with my first email.
>>>
>>> Somewhere between Michael Lyle's original
>>> "bcache: PI controller for writeback rate V2" patch dated 07 Sep 2017
>>> and 1d316e6 bcache: implement PI controller for writeback rate,
>>> the mapping of the writeback_rate_minimum attribute was dropped.
>>>
>>> Re-add the missing sysfs writeback_rate_minimum attribute mapping to
>>> "allow the user to specify a minimum rate at which dirty blocks are
>>> retired."
>>>
>>> Fixes: 1d316e6 bcache: implement PI controller for writeback rate
>>
>> Remember to use the correct format for the Fixes line.
>>
> 
> Hi Jens,
> 
> This is my fault, it slips from my eyes. I'd like to resend a fixed up
> version to you.

I fixed it up when I applied it.

> BTW, now I only use checkpatches.pl to check patch format, do you use
> extra scripts as format checking hooks ?

I don't use checkpatch, I just use my eye balls :-)

-- 
Jens Axboe

[PATCH 08/15] bcache: fix miss key refill->end in writeback

[Coly Li]

From: Tang Junhui <tang.junhui.linux@gmail.com>

refill->end record the last key of writeback, for example, at the first
time, keys (1,128K) to (1,1024K) are flush to the backend device, but
the end key (1,1024K) is not included, since the bellow code:
	if (bkey_cmp(k, refill->end) >= 0) {
		ret = MAP_DONE;
		goto out;
	}
And in the next time when we refill writeback keybuf again, we searched
key start from (1,1024K), and got a key bigger than it, so the key
(1,1024K) missed.
This patch modify the above code, and let the end key to be included to
the writeback key buffer.

Signed-off-by: Tang Junhui <tang.junhui.linux@gmail.com>
Cc: stable@vger.kernel.org
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/btree.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/bcache/btree.c b/drivers/md/bcache/btree.c
index e7d4817681f2..3f4211b5cd33 100644
--- a/drivers/md/bcache/btree.c
+++ b/drivers/md/bcache/btree.c
@@ -2434,7 +2434,7 @@ static int refill_keybuf_fn(struct btree_op *op, struct btree *b,
 	struct keybuf *buf = refill->buf;
 	int ret = MAP_CONTINUE;
 
-	if (bkey_cmp(k, refill->end) >= 0) {
+	if (bkey_cmp(k, refill->end) > 0) {
 		ret = MAP_DONE;
 		goto out;
 	}
-- 
2.19.0

[PATCH 09/15] bcache: recal cached_dev_sectors on detach

[Coly Li]

From: Shenghui Wang <shhuiw@foxmail.com>

Recal cached_dev_sectors on cached_dev detached, as recal done on
cached_dev attached.

Update the cached_dev_sectors before bcache_device_detach called
as bcache_device_detach will set bcache_device->c to NULL.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/super.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 4989c7d4d4d0..dc392a485269 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -1008,6 +1008,7 @@ static void cached_dev_detach_finish(struct work_struct *w)
 	bch_write_bdev_super(dc, &cl);
 	closure_sync(&cl);
 
+	calc_cached_dev_sectors(dc->disk.c);
 	bcache_device_detach(&dc->disk);
 	list_move(&dc->list, &uncached_devices);
 
-- 
2.19.0

[PATCH 10/15] bcache: remove unused bch_passthrough_cache

[Coly Li]

From: Shenghui Wang <shhuiw@foxmail.com>

struct kmem_cache *bch_passthrough_cache is not used in
bcache code. Remove it.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/request.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/bcache/request.h b/drivers/md/bcache/request.h
index aa055cfeb099..721bf336ed1a 100644
--- a/drivers/md/bcache/request.h
+++ b/drivers/md/bcache/request.h
@@ -39,6 +39,6 @@ void bch_data_insert(struct closure *cl);
 void bch_cached_dev_request_init(struct cached_dev *dc);
 void bch_flash_dev_request_init(struct bcache_device *d);
 
-extern struct kmem_cache *bch_search_cache, *bch_passthrough_cache;
+extern struct kmem_cache *bch_search_cache;
 
 #endif /* _BCACHE_REQUEST_H_ */
-- 
2.19.0

[PATCH 11/15] bcache: remove useless parameter of bch_debug_init()

[Coly Li]

From: Dongbo Cao <cdbdyx@163.com>

Parameter "struct kobject *kobj" in bch_debug_init() is useless,
remove it in this patch.

Signed-off-by: Dongbo Cao <cdbdyx@163.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/bcache.h | 2 +-
 drivers/md/bcache/debug.c  | 2 +-
 drivers/md/bcache/super.c  | 2 +-
 3 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/md/bcache/bcache.h b/drivers/md/bcache/bcache.h
index 83504dd8100a..5cfa1ec1b7a1 100644
--- a/drivers/md/bcache/bcache.h
+++ b/drivers/md/bcache/bcache.h
@@ -1003,7 +1003,7 @@ void bch_open_buckets_free(struct cache_set *c);
 int bch_cache_allocator_start(struct cache *ca);
 
 void bch_debug_exit(void);
-void bch_debug_init(struct kobject *kobj);
+void bch_debug_init(void);
 void bch_request_exit(void);
 int bch_request_init(void);
 
diff --git a/drivers/md/bcache/debug.c b/drivers/md/bcache/debug.c
index 06da66b2488a..8f448b9c96a1 100644
--- a/drivers/md/bcache/debug.c
+++ b/drivers/md/bcache/debug.c
@@ -253,7 +253,7 @@ void bch_debug_exit(void)
 		debugfs_remove_recursive(bcache_debug);
 }
 
-void __init bch_debug_init(struct kobject *kobj)
+void __init bch_debug_init(void)
 {
 	/*
 	 * it is unnecessary to check return value of
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index dc392a485269..719686ea5e27 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -2381,7 +2381,7 @@ static int __init bcache_init(void)
 	    sysfs_create_files(bcache_kobj, files))
 		goto err;
 
-	bch_debug_init(bcache_kobj);
+	bch_debug_init();
 	closure_debug_init();
 
 	return 0;
-- 
2.19.0

[PATCH 12/15] bcache: replace hard coded number with BUCKET_GC_GEN_MAX

[Coly Li]

In extents.c:bch_extent_bad(), number 96 is used as parameter to call
btree_bug_on(). The purpose is to check whether stale gen value exceeds
BUCKET_GC_GEN_MAX, so it is better to use macro BUCKET_GC_GEN_MAX to
make the code more understandable.

Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/extents.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/bcache/extents.c b/drivers/md/bcache/extents.c
index c809724e6571..956004366699 100644
--- a/drivers/md/bcache/extents.c
+++ b/drivers/md/bcache/extents.c
@@ -553,7 +553,7 @@ static bool bch_extent_bad(struct btree_keys *bk, const struct bkey *k)
 	for (i = 0; i < KEY_PTRS(k); i++) {
 		stale = ptr_stale(b->c, k, i);
 
-		btree_bug_on(stale > 96, b,
+		btree_bug_on(stale > BUCKET_GC_GEN_MAX, b,
 			     "key too stale: %i, need_gc %u",
 			     stale, b->c->need_gc);
 
-- 
2.19.0

[PATCH 13/15] bcache: use MAX_CACHES_PER_SET instead of magic number 8 in __bch_bucket_alloc_set

[Coly Li]

From: Shenghui Wang <shhuiw@foxmail.com>

Current cache_set has MAX_CACHES_PER_SET caches most, and the macro
is used for
"
	struct cache *cache_by_alloc[MAX_CACHES_PER_SET];
"
in the define of struct cache_set.

Use MAX_CACHES_PER_SET instead of magic number 8 in
__bch_bucket_alloc_set.

Signed-off-by: Shenghui Wang <shhuiw@foxmail.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/alloc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/md/bcache/alloc.c b/drivers/md/bcache/alloc.c
index 7a28232d868b..5002838ea476 100644
--- a/drivers/md/bcache/alloc.c
+++ b/drivers/md/bcache/alloc.c
@@ -484,7 +484,7 @@ int __bch_bucket_alloc_set(struct cache_set *c, unsigned int reserve,
 	int i;
 
 	lockdep_assert_held(&c->bucket_lock);
-	BUG_ON(!n || n > c->caches_loaded || n > 8);
+	BUG_ON(!n || n > c->caches_loaded || n > MAX_CACHES_PER_SET);
 
 	bkey_init(k);
 
-- 
2.19.0

[PATCH 14/15] bcache: split combined if-condition code into separate ones

[Coly Li]

From: Dongbo Cao <cdbdyx@163.com>

Split the combined '||' statements in if() check, to make the code easier
for debug.

Signed-off-by: Dongbo Cao <cdbdyx@163.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/super.c | 90 +++++++++++++++++++++++++++++++++------
 1 file changed, 76 insertions(+), 14 deletions(-)

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 719686ea5e27..2a362a7ba995 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -2051,6 +2051,8 @@ static int cache_alloc(struct cache *ca)
 	size_t free;
 	size_t btree_buckets;
 	struct bucket *b;
+	int ret = -ENOMEM;
+	const char *err = NULL;
 
 	__module_get(THIS_MODULE);
 	kobject_init(&ca->kobj, &bch_cache_ktype);
@@ -2069,26 +2071,86 @@ static int cache_alloc(struct cache *ca)
 	btree_buckets = ca->sb.njournal_buckets ?: 8;
 	free = roundup_pow_of_two(ca->sb.nbuckets) >> 10;
 
-	if (!init_fifo(&ca->free[RESERVE_BTREE], btree_buckets, GFP_KERNEL) ||
-	    !init_fifo_exact(&ca->free[RESERVE_PRIO], prio_buckets(ca), GFP_KERNEL) ||
-	    !init_fifo(&ca->free[RESERVE_MOVINGGC], free, GFP_KERNEL) ||
-	    !init_fifo(&ca->free[RESERVE_NONE], free, GFP_KERNEL) ||
-	    !init_fifo(&ca->free_inc,	free << 2, GFP_KERNEL) ||
-	    !init_heap(&ca->heap,	free << 3, GFP_KERNEL) ||
-	    !(ca->buckets	= vzalloc(array_size(sizeof(struct bucket),
-						     ca->sb.nbuckets))) ||
-	    !(ca->prio_buckets	= kzalloc(array3_size(sizeof(uint64_t),
-						      prio_buckets(ca), 2),
-					  GFP_KERNEL)) ||
-	    !(ca->disk_buckets	= alloc_bucket_pages(GFP_KERNEL, ca)))
-		return -ENOMEM;
+	if (!init_fifo(&ca->free[RESERVE_BTREE], btree_buckets,
+						GFP_KERNEL)) {
+		err = "ca->free[RESERVE_BTREE] alloc failed";
+		goto err_btree_alloc;
+	}
+
+	if (!init_fifo_exact(&ca->free[RESERVE_PRIO], prio_buckets(ca),
+							GFP_KERNEL)) {
+		err = "ca->free[RESERVE_PRIO] alloc failed";
+		goto err_prio_alloc;
+	}
+
+	if (!init_fifo(&ca->free[RESERVE_MOVINGGC], free, GFP_KERNEL)) {
+		err = "ca->free[RESERVE_MOVINGGC] alloc failed";
+		goto err_movinggc_alloc;
+	}
+
+	if (!init_fifo(&ca->free[RESERVE_NONE], free, GFP_KERNEL)) {
+		err = "ca->free[RESERVE_NONE] alloc failed";
+		goto err_none_alloc;
+	}
+
+	if (!init_fifo(&ca->free_inc, free << 2, GFP_KERNEL)) {
+		err = "ca->free_inc alloc failed";
+		goto err_free_inc_alloc;
+	}
+
+	if (!init_heap(&ca->heap, free << 3, GFP_KERNEL)) {
+		err = "ca->heap alloc failed";
+		goto err_heap_alloc;
+	}
+
+	ca->buckets = vzalloc(array_size(sizeof(struct bucket),
+			      ca->sb.nbuckets));
+	if (!ca->buckets) {
+		err = "ca->buckets alloc failed";
+		goto err_buckets_alloc;
+	}
+
+	ca->prio_buckets = kzalloc(array3_size(sizeof(uint64_t),
+				   prio_buckets(ca), 2),
+				   GFP_KERNEL);
+	if (!ca->prio_buckets) {
+		err = "ca->prio_buckets alloc failed";
+		goto err_prio_buckets_alloc;
+	}
+
+	ca->disk_buckets = alloc_bucket_pages(GFP_KERNEL, ca);
+	if (!ca->disk_buckets) {
+		err = "ca->disk_buckets alloc failed";
+		goto err_disk_buckets_alloc;
+	}
 
 	ca->prio_last_buckets = ca->prio_buckets + prio_buckets(ca);
 
 	for_each_bucket(b, ca)
 		atomic_set(&b->pin, 0);
-
 	return 0;
+
+err_disk_buckets_alloc:
+	kfree(ca->prio_buckets);
+err_prio_buckets_alloc:
+	vfree(ca->buckets);
+err_buckets_alloc:
+	free_heap(&ca->heap);
+err_heap_alloc:
+	free_fifo(&ca->free_inc);
+err_free_inc_alloc:
+	free_fifo(&ca->free[RESERVE_NONE]);
+err_none_alloc:
+	free_fifo(&ca->free[RESERVE_MOVINGGC]);
+err_movinggc_alloc:
+	free_fifo(&ca->free[RESERVE_PRIO]);
+err_prio_alloc:
+	free_fifo(&ca->free[RESERVE_BTREE]);
+err_btree_alloc:
+	module_put(THIS_MODULE);
+	if (err)
+		pr_notice("error %s: %s", ca->cache_dev_name, err);
+	return ret;
 }
 
 static int register_cache(struct cache_sb *sb, struct page *sb_page,
-- 
2.19.0

[PATCH 15/15] bcache: panic fix for making cache device

[Coly Li]

From: Dongbo Cao <cdbdyx@163.com>

when the nbuckets of cache device is smaller than 1024, making cache
device will trigger BUG_ON in kernel, add a condition to avoid this.

Reported-by: nitroxis <n@nxs.re>
Signed-off-by: Dongbo Cao <cdbdyx@163.com>
Signed-off-by: Coly Li <colyli@suse.de>
---
 drivers/md/bcache/super.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index 2a362a7ba995..563f75686100 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -2070,6 +2070,11 @@ static int cache_alloc(struct cache *ca)
 	 */
 	btree_buckets = ca->sb.njournal_buckets ?: 8;
 	free = roundup_pow_of_two(ca->sb.nbuckets) >> 10;
+	if (!free) {
+		ret = -EPERM;
+		err = "ca->sb.nbuckets is too small";
+		goto err_free;
+	}
 
 	if (!init_fifo(&ca->free[RESERVE_BTREE], btree_buckets,
 						GFP_KERNEL)) {
@@ -2147,6 +2152,7 @@ static int cache_alloc(struct cache *ca)
 err_prio_alloc:
 	free_fifo(&ca->free[RESERVE_BTREE]);
 err_btree_alloc:
+err_free:
 	module_put(THIS_MODULE);
 	if (err)
 		pr_notice("error %s: %s", ca->cache_dev_name, err);
@@ -2176,6 +2182,8 @@ static int register_cache(struct cache_sb *sb, struct page *sb_page,
 		blkdev_put(bdev, FMODE_READ|FMODE_WRITE|FMODE_EXCL);
 		if (ret == -ENOMEM)
 			err = "cache_alloc(): -ENOMEM";
+		else if (ret == -EPERM)
+			err = "cache_alloc(): cache device is too small";
 		else
 			err = "cache_alloc(): unknown error";
 		goto err;
-- 
2.19.0

Re: [PATCH 00/15] bcache patches for 4.20

[Jens Axboe]

On 10/8/18 6:41 AM, Coly Li wrote:
> Hi Jens,
> 
> Here are patches for Linux 4.20. In this run we don't have big changes,
> most of the patches for fixes or code cleanup.
> 
> Ben Peddell resends a patch which had no response for a while, I pick
> it for this version. Our new developer Dongbo Cao contributes a bug fix
> to avoid kernel panic for too small cache set size, with other code
> cleanup changes. Shenghui and me have several code cleanup and minor
> fixes, and Junhui again contributes helpful fixes for several real bugs.
> 
> Please pull them for 4.20 merge window. Thanks in advance.

Applied, thanks.

-- 
Jens Axboe