[PATCH 0/2] scsi: vpd sanity checks

[PATCH 0/2] scsi: vpd sanity checks

[Hannes Reinecke]

Hi all,

here's a combined patchset to clear up the sanity checks for VPD page
support. The first patch streamlines the function to check for VPD page
support, and the second one lowers the minimal SCSI version for VPD
page support to SPC-2, to avoid having to whitelist a lot of devices.

As usual, comments and reviews are welcome.

Hannes Reinecke (2):
  scsi: Do not attach VPD to devices that don't support it
  scsi: vpd pages are mandatory for SPC-2

 drivers/scsi/scsi.c        |  3 ++-
 drivers/scsi/sd.c          | 19 +------------------
 include/scsi/scsi_device.h | 25 +++++++++++++++++++++++++
 3 files changed, 28 insertions(+), 19 deletions(-)

-- 
2.6.2

[PATCH 1/2] scsi: Do not attach VPD to devices that don't support it

[Hannes Reinecke]

The patch "scsi: rescan VPD attributes" introduced a regression in which
devices that don't support VPD were being scanned for VPD attributes
anyway.  This could cause issues for this parts and should be avoided so
the check for scsi_level has been moved out of scsi_add_lun and into
scsi_attach_vpd so that all callers will not scan VPD for devices that
don't support it.

Fixes: 09e2b0b14690 ("scsi: rescan VPD attributes")

Suggested-by: Alexander Duyck <aduyck@mirantis.com>
Signed-off-by: Hannes Reinecke <hare@suse.com>
---
 drivers/scsi/scsi.c        |  3 ++-
 drivers/scsi/sd.c          | 19 +------------------
 include/scsi/scsi_device.h | 25 +++++++++++++++++++++++++
 3 files changed, 28 insertions(+), 19 deletions(-)

diff --git a/drivers/scsi/scsi.c b/drivers/scsi/scsi.c
index b1bf42b..1deb6ad 100644
--- a/drivers/scsi/scsi.c
+++ b/drivers/scsi/scsi.c
@@ -784,8 +784,9 @@ void scsi_attach_vpd(struct scsi_device *sdev)
 	int pg83_supported = 0;
 	unsigned char __rcu *vpd_buf, *orig_vpd_buf = NULL;
 
-	if (sdev->skip_vpd_pages)
+	if (!scsi_device_supports_vpd(sdev))
 		return;
+
 retry_pg0:
 	vpd_buf = kmalloc(vpd_len, GFP_KERNEL);
 	if (!vpd_buf)
diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index 5a5457a..70042f2 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -2795,23 +2795,6 @@ static void sd_read_write_same(struct scsi_disk *sdkp, unsigned char *buffer)
 		sdkp->ws10 = 1;
 }
 
-static int sd_try_extended_inquiry(struct scsi_device *sdp)
-{
-	/* Attempt VPD inquiry if the device blacklist explicitly calls
-	 * for it.
-	 */
-	if (sdp->try_vpd_pages)
-		return 1;
-	/*
-	 * Although VPD inquiries can go to SCSI-2 type devices,
-	 * some USB ones crash on receiving them, and the pages
-	 * we currently ask for are for SPC-3 and beyond
-	 */
-	if (sdp->scsi_level > SCSI_SPC_2 && !sdp->skip_vpd_pages)
-		return 1;
-	return 0;
-}
-
 static inline u32 logical_to_sectors(struct scsi_device *sdev, u32 blocks)
 {
 	return blocks << (ilog2(sdev->sector_size) - 9);
@@ -2856,7 +2839,7 @@ static int sd_revalidate_disk(struct gendisk *disk)
 	if (sdkp->media_present) {
 		sd_read_capacity(sdkp, buffer);
 
-		if (sd_try_extended_inquiry(sdp)) {
+		if (scsi_device_supports_vpd(sdp)) {
 			sd_read_block_provisioning(sdkp);
 			sd_read_block_limits(sdkp);
 			sd_read_block_characteristics(sdkp);
diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index 61341d3..1d4a329 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -523,6 +523,31 @@ static inline int scsi_device_tpgs(struct scsi_device *sdev)
 	return sdev->inquiry ? (sdev->inquiry[5] >> 4) & 0x3 : 0;
 }
 
+/**
+ * scsi_device_supports_vpd - test if a device supports VPD pages
+ * @sdev: the &struct scsi_device to test
+ *
+ * If the 'try_vpd_pages' flag is set it takes precedence.
+ * Otherwise we will assume VPD pages are supported if the
+ * SCSI level is at least SPC-3 and 'skip_vpd_pages' is not set.
+ */
+static inline int scsi_device_supports_vpd(struct scsi_device *sdev)
+{
+	/* Attempt VPD inquiry if the device blacklist explicitly calls
+	 * for it.
+	 */
+	if (sdev->try_vpd_pages)
+		return 1;
+	/*
+	 * Although VPD inquiries can go to SCSI-2 type devices,
+	 * some USB ones crash on receiving them, and the pages
+	 * we currently ask for are for SPC-3 and beyond
+	 */
+	if (sdev->scsi_level > SCSI_SPC_2 && !sdev->skip_vpd_pages)
+		return 1;
+	return 0;
+}
+
 #define MODULE_ALIAS_SCSI_DEVICE(type) \
 	MODULE_ALIAS("scsi:t-" __stringify(type) "*")
 #define SCSI_DEVICE_MODALIAS_FMT "scsi:t-0x%02x"
-- 
2.6.2

Re: [PATCH 1/2] scsi: Do not attach VPD to devices that don't support it

[Johannes Thumshirn]

On 2016-04-01 08:57, Hannes Reinecke wrote:
> The patch "scsi: rescan VPD attributes" introduced a regression in 
> which
> devices that don't support VPD were being scanned for VPD attributes
> anyway.  This could cause issues for this parts and should be avoided 
> so
> the check for scsi_level has been moved out of scsi_add_lun and into
> scsi_attach_vpd so that all callers will not scan VPD for devices that
> don't support it.
> 
> Fixes: 09e2b0b14690 ("scsi: rescan VPD attributes")
> 
> Suggested-by: Alexander Duyck <aduyck@mirantis.com>
> Signed-off-by: Hannes Reinecke <hare@suse.com>

Looks good,
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>

Re: [PATCH 1/2] scsi: Do not attach VPD to devices that don't support it

[Martin K. Petersen]

>>>>> "Hannes" == Hannes Reinecke <hare@suse.de> writes:

Hannes> The patch "scsi: rescan VPD attributes" introduced a regression
Hannes> in which devices that don't support VPD were being scanned for
Hannes> VPD attributes anyway.  This could cause issues for this parts
Hannes> and should be avoided so the check for scsi_level has been moved
Hannes> out of scsi_add_lun and into scsi_attach_vpd so that all callers
Hannes> will not scan VPD for devices that don't support it.

Applied to 4.6/scsi-fixes, CC:stable 4.5+.

Thanks!

-- 
Martin K. Petersen	Oracle Linux Engineering

[PATCH 2/2] scsi: vpd pages are mandatory for SPC-2

[Hannes Reinecke]

VPD pages 0x0 and 0x83 are mandatory even for SPC-2, so we should
be lowering the restriction to avoid having to whitelist every
SPC-2 compliant device.

Signed-off-by: Hannes Reinecke <hare@suse.com>
---
 include/scsi/scsi_device.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/include/scsi/scsi_device.h b/include/scsi/scsi_device.h
index 1d4a329..40a789f 100644
--- a/include/scsi/scsi_device.h
+++ b/include/scsi/scsi_device.h
@@ -541,9 +541,9 @@ static inline int scsi_device_supports_vpd(struct scsi_device *sdev)
 	/*
 	 * Although VPD inquiries can go to SCSI-2 type devices,
 	 * some USB ones crash on receiving them, and the pages
-	 * we currently ask for are for SPC-3 and beyond
+	 * we currently ask for are mandatory for SPC-2 and beyond
 	 */
-	if (sdev->scsi_level > SCSI_SPC_2 && !sdev->skip_vpd_pages)
+	if (sdev->scsi_level >= SCSI_SPC_2 && !sdev->skip_vpd_pages)
 		return 1;
 	return 0;
 }
-- 
2.6.2

Re: [PATCH 2/2] scsi: vpd pages are mandatory for SPC-2

[Johannes Thumshirn]

On 2016-04-01 08:57, Hannes Reinecke wrote:
> VPD pages 0x0 and 0x83 are mandatory even for SPC-2, so we should
> be lowering the restriction to avoid having to whitelist every
> SPC-2 compliant device.
> 
> Signed-off-by: Hannes Reinecke <hare@suse.com>

Looks good,
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>

Re: [PATCH 2/2] scsi: vpd pages are mandatory for SPC-2

[Martin K. Petersen]

>>>>> "Hannes" == Hannes Reinecke <hare@suse.de> writes:

Hannes> VPD pages 0x0 and 0x83 are mandatory even for SPC-2, so we
Hannes> should be lowering the restriction to avoid having to whitelist
Hannes> every SPC-2 compliant device.

Linus took patch 1/2 for 4.6rc3 so I have rebased 4.7/scsi-queue and
applied patch 2/2.

-- 
Martin K. Petersen	Oracle Linux Engineering