[PATCH] blk-settings: make sure that max_sectors is aligned on "logical_block_size" boundary. (fwd)

---------- Forwarded message ----------
Date: Thu, 19 Nov 2020 15:36:51 -0500 (EST)
From: Mikulas Patocka <mpatocka@redhat.com>
To: David Teigland <teigland@redhat.com>, Jens Axboe <axboe@kernel.dk>
Cc: heinzm@redhat.com, Zdenek Kabelac <redacted>,
    Marian Csontos [off-list ref], linux-block@vger.kernel.org,
    dm-devel@redhat.com
Subject: [PATCH] blk-settings: make sure that max_sectors is aligned on
    "logical_block_size" boundary.

We get these I/O errors when we run md-raid1 on the top of dm-integrity on 
the top of ramdisk:
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8048, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8147, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8246, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8345, 0xbb

The ramdisk device has logical_block_size 512 and max_sectors 255. The 
dm-integrity device uses logical_block_size 4096 and it doesn't affect the 
"max_sectors" value - thus, it inherits 255 from the ramdisk. So, we have 
a device with max_sectors not aligned on logical_block_size.

The md-raid device sees that the underlying leg has max_sectors 255 and it
will split the bios on 255-sector boundary, making the bios unaligned on
logical_block_size.

In order to fix the bug, we round down max_sectors to logical_block_size.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org

---
 block/blk-settings.c |   10 ++++++++++
 1 file changed, 10 insertions(+)

Index: linux-2.6/block/blk-settings.c
===================================================================

--- linux-2.6.orig/block/blk-settings.c	2020-10-29 12:20:46.000000000 +0100
+++ linux-2.6/block/blk-settings.c	2020-11-19 21:20:18.000000000 +0100

@@ -591,6 +591,16 @@ int blk_stack_limits(struct queue_limits
 		ret = -1;
 	}
 
+	t->max_sectors = round_down(t->max_sectors, t->logical_block_size / 512);
+	if (t->max_sectors < PAGE_SIZE / 512)
+		t->max_sectors = PAGE_SIZE / 512;
+	t->max_hw_sectors = round_down(t->max_hw_sectors, t->logical_block_size / 512);
+	if (t->max_sectors < PAGE_SIZE / 512)

+		t->max_hw_sectors = PAGE_SIZE / 512;
+	t->max_dev_sectors = round_down(t->max_dev_sectors, t->logical_block_size / 512);
+	if (t->max_sectors < PAGE_SIZE / 512)

On Tue, 23 Feb 2021, Ming Lei wrote:

quoted

I'd suggest to add a helper(such as, blk_round_down_sectors()) to round_down each
one.

Yes - Here I'm sending the updated patch.

quoted

-- 
Ming

From: Mikulas Patocka <mpatocka@redhat.com>

We get I/O errors when we run md-raid1 on the top of dm-integrity on the
top of ramdisk.
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xff00, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0xffff, 0x1
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8048, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8147, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8246, 0xff
device-mapper: integrity: Bio not aligned on 8 sectors: 0x8345, 0xbb

The ramdisk device has logical_block_size 512 and max_sectors 255. The
dm-integrity device uses logical_block_size 4096 and it doesn't affect the
"max_sectors" value - thus, it inherits 255 from the ramdisk. So, we have
a device with max_sectors not aligned on logical_block_size.

The md-raid device sees that the underlying leg has max_sectors 255 and it
will split the bios on 255-sector boundary, making the bios unaligned on
logical_block_size.

In order to fix the bug, we round down max_sectors to logical_block_size.

Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org

---
 block/blk-settings.c |   12 ++++++++++++
 1 file changed, 12 insertions(+)

Index: linux-2.6/block/blk-settings.c
===================================================================

--- linux-2.6.orig/block/blk-settings.c	2021-02-23 17:18:59.000000000 +0100
+++ linux-2.6/block/blk-settings.c	2021-02-23 17:23:58.000000000 +0100

@@ -481,6 +481,14 @@ void blk_queue_io_opt(struct request_que
 }
 EXPORT_SYMBOL(blk_queue_io_opt);
 
+static unsigned int blk_round_down_sectors(unsigned int sectors, unsigned int lbs)
+{
+	sectors = round_down(sectors, lbs >> SECTOR_SHIFT);
+	if (sectors < PAGE_SIZE >> SECTOR_SHIFT)
+		sectors = PAGE_SIZE >> SECTOR_SHIFT;
+	return sectors;
+}
+
 /**
  * blk_stack_limits - adjust queue_limits for stacked devices
  * @t:	the stacking driver limits (top device)

@@ -607,6 +615,10 @@ int blk_stack_limits(struct queue_limits
 		ret = -1;
 	}
 
+	t->max_sectors = blk_round_down_sectors(t->max_sectors, t->logical_block_size);
+	t->max_hw_sectors = blk_round_down_sectors(t->max_hw_sectors, t->logical_block_size);
+	t->max_dev_sectors = blk_round_down_sectors(t->max_dev_sectors, t->logical_block_size);
+
 	/* Discard alignment and granularity */
 	if (b->discard_granularity) {
 		alignment = queue_limit_discard_alignment(b, start);