[PATCH] fs: guard_bio_eod() needs to consider partitions

[PATCH] fs: guard_bio_eod() needs to consider partitions

[Greg Edwards]

guard_bio_eod() needs to look at the partition capacity, not just the
capacity of the whole device, when determining if truncation is
necessary.

[   60.268688] attempt to access beyond end of device
[   60.268690] unknown-block(9,1): rw=0, want=67103509, limit=67103506
[   60.268693] buffer_io_error: 2 callbacks suppressed
[   60.268696] Buffer I/O error on dev md1p7, logical block 4524305, async page read

Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
Signed-off-by: Greg Edwards <gedwards@ddn.com>
---
 fs/buffer.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index 170df856bdb9..989d3b085efe 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -3055,8 +3055,15 @@ void guard_bio_eod(int op, struct bio *bio)
 	sector_t maxsector;
 	struct bio_vec *bvec = &bio->bi_io_vec[bio->bi_vcnt - 1];
 	unsigned truncated_bytes;
+	struct hd_struct *part;
+
+	part = disk_get_part(bio->bi_disk, bio->bi_partno);
+	if (part)
+		maxsector = part_nr_sects_read(part);
+	else
+		maxsector = get_capacity(bio->bi_disk);
+	disk_put_part(part);
 
-	maxsector = get_capacity(bio->bi_disk);
 	if (!maxsector)
 		return;
 
-- 
2.13.6

Re: [PATCH] fs: guard_bio_eod() needs to consider partitions

[Al Viro]

On Mon, Oct 23, 2017 at 07:10:07PM -0600, Greg Edwards wrote:
> guard_bio_eod() needs to look at the partition capacity, not just the
> capacity of the whole device, when determining if truncation is
> necessary.
> 
> [   60.268688] attempt to access beyond end of device
> [   60.268690] unknown-block(9,1): rw=0, want=67103509, limit=67103506
> [   60.268693] buffer_io_error: 2 callbacks suppressed
> [   60.268696] Buffer I/O error on dev md1p7, logical block 4524305, async page read
> 
> Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
> Signed-off-by: Greg Edwards <gedwards@ddn.com>

That smells like a nasty source of overhead, especially on SMP boxen...

Re: [PATCH] fs: guard_bio_eod() needs to consider partitions

[Christoph Hellwig]

On Tue, Oct 24, 2017 at 03:16:36AM +0100, Al Viro wrote:
> > Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
> > Signed-off-by: Greg Edwards <gedwards@ddn.com>
> 
> That smells like a nasty source of overhead, especially on SMP boxen...

We could just use __disk_get_part similar to the partition remapping,
or just pass the block_device to guard_bio_eod.

Re: [PATCH] fs: guard_bio_eod() needs to consider partitions

[Greg Edwards]

On Tue, Oct 24, 2017 at 09:46:15AM +0200, Christoph Hellwig wrote:
> On Tue, Oct 24, 2017 at 03:16:36AM +0100, Al Viro wrote:
>>> Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
>>> Signed-off-by: Greg Edwards <gedwards@ddn.com>
>>
>> That smells like a nasty source of overhead, especially on SMP boxen...
>
> We could just use __disk_get_part similar to the partition remapping,
> or just pass the block_device to guard_bio_eod.

Do you have a preference?  __disk_get_part would be isolated to
guard_bio_eod.  If we pass in the block_device, it would also touch all
the mpage_bio_submit callers.

[PATCH v2] fs: guard_bio_eod() needs to consider partitions

[Greg Edwards]

guard_bio_eod() needs to look at the partition capacity, not just the
capacity of the whole device, when determining if truncation is
necessary.

[   60.268688] attempt to access beyond end of device
[   60.268690] unknown-block(9,1): rw=0, want=67103509, limit=67103506
[   60.268693] buffer_io_error: 2 callbacks suppressed
[   60.268696] Buffer I/O error on dev md1p7, logical block 4524305, async page read

Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
Signed-off-by: Greg Edwards <gedwards@ddn.com>
---
Changes from v1:
  * use __disk_get_part instead of disk_get_part, similar to what
    blk_partition_remap does

 fs/buffer.c           | 10 +++++++++-
 include/linux/genhd.h |  1 +
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/fs/buffer.c b/fs/buffer.c
index 170df856bdb9..b96f3b98a6ef 100644
--- a/fs/buffer.c
+++ b/fs/buffer.c
@@ -3055,8 +3055,16 @@ void guard_bio_eod(int op, struct bio *bio)
 	sector_t maxsector;
 	struct bio_vec *bvec = &bio->bi_io_vec[bio->bi_vcnt - 1];
 	unsigned truncated_bytes;
+	struct hd_struct *part;
+
+	rcu_read_lock();
+	part = __disk_get_part(bio->bi_disk, bio->bi_partno);
+	if (part)
+		maxsector = part_nr_sects_read(part);
+	else
+		maxsector = get_capacity(bio->bi_disk);
+	rcu_read_unlock();
 
-	maxsector = get_capacity(bio->bi_disk);
 	if (!maxsector)
 		return;
 
diff --git a/include/linux/genhd.h b/include/linux/genhd.h
index ea652bfcd675..cb8c5fd74b71 100644
--- a/include/linux/genhd.h
+++ b/include/linux/genhd.h
@@ -242,6 +242,7 @@ static inline dev_t part_devt(struct hd_struct *part)
 	return part_to_dev(part)->devt;
 }
 
+extern struct hd_struct *__disk_get_part(struct gendisk *disk, int partno);
 extern struct hd_struct *disk_get_part(struct gendisk *disk, int partno);
 
 static inline void disk_put_part(struct hd_struct *part)
-- 
2.13.6

Re: [PATCH v2] fs: guard_bio_eod() needs to consider partitions

[Greg Edwards]

On Tue, Oct 24, 2017 at 11:21:48AM -0600, Greg Edwards wrote:
> guard_bio_eod() needs to look at the partition capacity, not just the
> capacity of the whole device, when determining if truncation is
> necessary.
>
> [   60.268688] attempt to access beyond end of device
> [   60.268690] unknown-block(9,1): rw=0, want=67103509, limit=67103506
> [   60.268693] buffer_io_error: 2 callbacks suppressed
> [   60.268696] Buffer I/O error on dev md1p7, logical block 4524305, async page read
>
> Fixes: 74d46992e0d9 ("block: replace bi_bdev with a gendisk pointer and partitions index")
> Signed-off-by: Greg Edwards <gedwards@ddn.com>
> ---
> Changes from v1:
>   * use __disk_get_part instead of disk_get_part, similar to what
>     blk_partition_remap does

Al, Christoph,

Any thoughts on this version?

It would be nice to get this fixed before 4.14 goes out, as this is a
regression from previous releases.

Greg

Re: [PATCH v2] fs: guard_bio_eod() needs to consider partitions

[Christoph Hellwig]

This looks good to me:

Reviewed-by: Christoph Hellwig <hch@lst.de>

Jens, can you pick it up through the block tree given that is where
the issue was introduced?

Re: [PATCH v2] fs: guard_bio_eod() needs to consider partitions

[Jens Axboe]

On 11/10/2017 01:46 AM, Christoph Hellwig wrote:
> This looks good to me:
> 
> Reviewed-by: Christoph Hellwig <hch@lst.de>
> 
> Jens, can you pick it up through the block tree given that is where
> the issue was introduced?

I picked it up, and marked it for 4.13 stable as well.

-- 
Jens Axboe