Linux SCSI subsystem development
 help / color / mirror / Atom feed
* [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX
@ 2026-11-16 21:06 Tal Zussman
  2026-11-16 21:06 ` [PATCH v2 1/2] scsi: scsi_debug: wrap around the store when zeroing a reset zone Tal Zussman
                   ` (2 more replies)
  0 siblings, 3 replies; 4+ messages in thread
From: Tal Zussman @ 2026-11-16 21:06 UTC (permalink / raw)
  To: James E.J. Bottomley, Martin K. Petersen, Damien Le Moal,
	Douglas Gilbert, Shin'ichiro Kawasaki
  Cc: linux-scsi, linux-kernel, Tal Zussman

Resetting a written zone on scsi_debug zeroes the zone's blocks in the
store, but does so with a raw offset that doesn't wrap around the store
like writes do, and without resetting the blocks' protection
information. With virtual_gb set, resetting a zone that starts past the
end of the store oopses. With DIX enabled, a reset zone can't be read
back. The first patch makes the reset wrap around the end of the store
like writes do. The second resets the protection information of the
zeroed blocks.

The fixes were proposed by an LLM.

---
Changes in v2:
- Zero the at most two ranges with two memset() calls rather than a
  loop, per Damien
- Point the Fixes tags to 2d62253eb1b6, per Sashiko
- Link to v1: https://patch.msgid.link/20261110-scsi-debug-zone-reset-v1-0-28cb9184b5ea@columbia.edu

---
Tal Zussman (2):
      scsi: scsi_debug: wrap around the store when zeroing a reset zone
      scsi: scsi_debug: reset protection information on zone reset

 drivers/scsi/scsi_debug.c | 34 ++++++++++++++++++++++++++++++++--
 1 file changed, 32 insertions(+), 2 deletions(-)
---
base-commit: 6628e7d4c6b25c8d18f1345e3a20d21829aae7e4
change-id: 20261109-scsi-debug-zone-reset-14fc587c9d45

Best regards,
--  
Tal Zussman <tz2294@columbia.edu>


^ permalink raw reply	[flat|nested] 4+ messages in thread

* [PATCH v2 1/2] scsi: scsi_debug: wrap around the store when zeroing a reset zone
  2026-11-16 21:06 [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Tal Zussman
@ 2026-11-16 21:06 ` Tal Zussman
  2026-11-16 21:06 ` [PATCH v2 2/2] scsi: scsi_debug: reset protection information on zone reset Tal Zussman
  2026-11-16 21:42 ` [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Damien Le Moal
  2 siblings, 0 replies; 4+ messages in thread
From: Tal Zussman @ 2026-11-16 21:06 UTC (permalink / raw)
  To: James E.J. Bottomley, Martin K. Petersen, Damien Le Moal,
	Douglas Gilbert, Shin'ichiro Kawasaki
  Cc: linux-scsi, linux-kernel, Tal Zussman

zbc_rwp_zone() zeroes the blocks of a zone that is reset, using the
start of the zone as an offset into the store. With virtual_gb set, the
device and its zones can extend past the end of the store, which is
only dev_size_mb in size. Writes wrap around in do_device_access(), but
the memset() doesn't, so resetting a zone that has been written to can
write past the end of the store.

On a device whose fifth zone starts exactly at the end of the store,
write to that zone and reset it:

  # modprobe scsi_debug zbc=managed sector_size=4096 dev_size_mb=16 \
        virtual_gb=1 zone_size_mb=4
  # dd if=/dev/urandom of=/dev/sda bs=1M count=1 seek=16 oflag=direct
  # blkzone reset -o 32768 -c 1 /dev/sda

The memset() starts at the first byte past the store, in the guard page
after the vmalloc'ed area, for the 1 MiB that was written:

BUG: unable to handle page fault for address: ffffc90002c00000
Oops: Oops: 0002 [#1] SMP KASAN NOPTI
RIP: 0010:memset_orig+0x33/0xb0
RDX: 0000000000100000 RSI: 0000000000000000 RDI: ffffc90002c00000
Call Trace:
 zbc_rwp_zone+0x1a7/0x2a0 [scsi_debug]
 resp_rwp_zone+0x19c/0x450 [scsi_debug]
 schedule_resp+0x2dd/0xbc0 [scsi_debug]
 scsi_debug_queuecommand+0x1286/0x2950 [scsi_debug]
 scsi_queue_rq+0x1d74/0x3200
 blk_mq_dispatch_rq_list+0x3ad/0x23c0

The command never completes, so the reset returns EIO to userspace and
the device is unusable afterwards.

Zero the blocks in a helper that wraps around at the end of the store,
and that stops after one pass over it for a zone that is larger than
the store.

Fixes: 2d62253eb1b6 ("scsi: scsi_debug: Zero clear zones at reset write pointer")
Signed-off-by: Tal Zussman <tz2294@columbia.edu>
---
 drivers/scsi/scsi_debug.c | 25 +++++++++++++++++++++++--
 1 file changed, 23 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index a922c8318a2f..7e78c158f736 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -6612,6 +6612,28 @@ static int resp_finish_zone(struct scsi_cmnd *scp,
 	return res;
 }
 
+/*
+ * Zero the blocks of a zone that is being reset. The store can be smaller
+ * than the device, so wrap around at its end like do_device_access() does.
+ */
+static void zbc_zero_blocks(struct sdeb_store_info *sip, sector_t lba,
+			    sector_t num)
+{
+	u64 block, rest = 0;
+
+	/* One pass over the store zeroes all of it */
+	num = min_t(sector_t, num, sdebug_store_sectors);
+
+	block = sector_div(lba, sdebug_store_sectors);
+	if (block + num > sdebug_store_sectors)
+		rest = block + num - sdebug_store_sectors;
+
+	memset(sip->storep + block * sdebug_sector_size, 0,
+	       (num - rest) * sdebug_sector_size);
+	if (rest)
+		memset(sip->storep, 0, rest * sdebug_sector_size);
+}
+
 static void zbc_rwp_zone(struct sdebug_dev_info *devip,
 			 struct sdeb_zone_state *zsp)
 {
@@ -6629,8 +6651,7 @@ static void zbc_rwp_zone(struct sdebug_dev_info *devip,
 		devip->nr_closed--;
 
 	if (zsp->z_wp > zsp->z_start)
-		memset(sip->storep + zsp->z_start * sdebug_sector_size, 0,
-		       (zsp->z_wp - zsp->z_start) * sdebug_sector_size);
+		zbc_zero_blocks(sip, zsp->z_start, zsp->z_wp - zsp->z_start);
 
 	zsp->z_non_seq_resource = false;
 	zsp->z_wp = zsp->z_start;

-- 
2.39.5


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* [PATCH v2 2/2] scsi: scsi_debug: reset protection information on zone reset
  2026-11-16 21:06 [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Tal Zussman
  2026-11-16 21:06 ` [PATCH v2 1/2] scsi: scsi_debug: wrap around the store when zeroing a reset zone Tal Zussman
@ 2026-11-16 21:06 ` Tal Zussman
  2026-11-16 21:42 ` [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Damien Le Moal
  2 siblings, 0 replies; 4+ messages in thread
From: Tal Zussman @ 2026-11-16 21:06 UTC (permalink / raw)
  To: James E.J. Bottomley, Martin K. Petersen, Damien Le Moal,
	Douglas Gilbert, Shin'ichiro Kawasaki
  Cc: linux-scsi, linux-kernel, Tal Zussman

zbc_rwp_zone() zeroes the data of a zone that is reset, but leaves the
protection information of its previous contents in place. Reads of a
zoned device are not limited to the blocks below the write pointer, so
with DIX enabled reading those blocks back fails the guard check.

Write to a zone, reset it and read it back:

  # modprobe scsi_debug zbc=managed sector_size=4096 dev_size_mb=64 \
        zone_size_mb=4 dif=1 dix=1
  # dd if=/dev/urandom of=/dev/sda bs=1M count=1 seek=8 oflag=direct
  # blkzone reset -o 16384 -c 1 /dev/sda
  # dd if=/dev/sda of=/dev/null bs=1M count=1 skip=8 iflag=direct
  dd: error reading '/dev/sda': Invalid or incomplete multibyte or wide character

The read fails with EILSEQ, and the guard tag that the device reports
is still that of the data that was written, while the block now holds
zeroes:

scsi_debug:dif_verify: GUARD check failed on sector 2048 rcvd 0x288f, data 0x0000
protection error, dev sda, sector 16384 op 0x0:(READ) flags 0x10800 phys_seg 256 prio class 2

Reset the protection information of the zeroed blocks to 0xff, as
unmap_region() does, so that they are not checked until they are
written again.

Fixes: 2d62253eb1b6 ("scsi: scsi_debug: Zero clear zones at reset write pointer")
Signed-off-by: Tal Zussman <tz2294@columbia.edu>
---
 drivers/scsi/scsi_debug.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/drivers/scsi/scsi_debug.c b/drivers/scsi/scsi_debug.c
index 7e78c158f736..f64cd701d64b 100644
--- a/drivers/scsi/scsi_debug.c
+++ b/drivers/scsi/scsi_debug.c
@@ -6613,8 +6613,9 @@ static int resp_finish_zone(struct scsi_cmnd *scp,
 }
 
 /*
- * Zero the blocks of a zone that is being reset. The store can be smaller
- * than the device, so wrap around at its end like do_device_access() does.
+ * Zero the blocks of a zone that is being reset, and reset their protection
+ * information. The store can be smaller than the device, so wrap around at
+ * its end like do_device_access() does.
  */
 static void zbc_zero_blocks(struct sdeb_store_info *sip, sector_t lba,
 			    sector_t num)
@@ -6632,6 +6633,14 @@ static void zbc_zero_blocks(struct sdeb_store_info *sip, sector_t lba,
 	       (num - rest) * sdebug_sector_size);
 	if (rest)
 		memset(sip->storep, 0, rest * sdebug_sector_size);
+
+	if (!sip->dif_storep)
+		return;
+
+	memset(sip->dif_storep + block, 0xff,
+	       (num - rest) * sizeof(*sip->dif_storep));
+	if (rest)
+		memset(sip->dif_storep, 0xff, rest * sizeof(*sip->dif_storep));
 }
 
 static void zbc_rwp_zone(struct sdebug_dev_info *devip,

-- 
2.39.5


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX
  2026-11-16 21:06 [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Tal Zussman
  2026-11-16 21:06 ` [PATCH v2 1/2] scsi: scsi_debug: wrap around the store when zeroing a reset zone Tal Zussman
  2026-11-16 21:06 ` [PATCH v2 2/2] scsi: scsi_debug: reset protection information on zone reset Tal Zussman
@ 2026-11-16 21:42 ` Damien Le Moal
  2 siblings, 0 replies; 4+ messages in thread
From: Damien Le Moal @ 2026-11-16 21:42 UTC (permalink / raw)
  To: Tal Zussman, James E.J. Bottomley, Martin K. Petersen,
	Douglas Gilbert, Shin'ichiro Kawasaki
  Cc: linux-scsi, linux-kernel

On 11/16/26 13:06, Tal Zussman wrote:
> Resetting a written zone on scsi_debug zeroes the zone's blocks in the
> store, but does so with a raw offset that doesn't wrap around the store
> like writes do, and without resetting the blocks' protection
> information. With virtual_gb set, resetting a zone that starts past the
> end of the store oopses. With DIX enabled, a reset zone can't be read
> back. The first patch makes the reset wrap around the end of the store
> like writes do. The second resets the protection information of the
> zeroed blocks.

Looks OK.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>


-- 
Damien Le Moal
Western Digital Research

^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2026-11-16 21:42 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-11-16 21:06 [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Tal Zussman
2026-11-16 21:06 ` [PATCH v2 1/2] scsi: scsi_debug: wrap around the store when zeroing a reset zone Tal Zussman
2026-11-16 21:06 ` [PATCH v2 2/2] scsi: scsi_debug: reset protection information on zone reset Tal Zussman
2026-11-16 21:42 ` [PATCH v2 0/2] scsi: scsi_debug: fix zone reset with virtual_gb and DIX Damien Le Moal

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox