Do not read the IO hints VPD page from USB storage devices
Subsystems (5)
v2 cover Do not read the IO hints VPD page from USB storage devices
Hi Martin,
Recently it was reported that reading the IO hints VPD page makes at least two USB storage devices crash. Hence this patch series that disables reading the IO hints VPD page from USB storage devices. Please consider this patch series for your scsi-fixes branch.
Thanks,
Bart.
Changes compared to v1:
- Also set the BLIST_SKIP_IO_HINTS for USB Attached SCSI devices.
Bart Van Assche (2):
scsi: core: Introduce the BLIST_SKIP_IO_HINTS flag
scsi: core: Do not query IO hints for USB devices
drivers/scsi/sd.c | 4 ++++ drivers/usb/storage/scsiglue.c | 6 ++++++ drivers/usb/storage/uas.c | 7 +++++++ include/scsi/scsi_devinfo.h | 4 +++- 4 files changed, 20 insertions(+), 1 deletion(-)
PATCH v2 1/1 scsi: core: Fix an incorrect comment
The comment that scsi_static_device_list would go away was added more than 18 years ago. Today, that list is still there and a large number of additional entries have been added. This shows that this comment is incorrect. Hence fix that comment.
---
Changes compared to v1: addressed Christoph's review feedback.
drivers/scsi/scsi_devinfo.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-)
drivers/scsi/scsi_devinfo.c +5 -6
@@ -39,13 +39,12 @@ static LIST_HEAD(scsi_dev_info_list);static char scsi_dev_flags[256 ];/* -* scsi_static_device_list: deprecated list of devices that require -* settings that differ from the default, includes black-listed (broken) -* devices. The entries here are added to the tail of scsi_dev_info_list -* via scsi_dev_info_list_init. +* scsi_static_device_list: list of devices that require settings that differ +* from the default, includes black-listed (broken) devices. The entries here +* are added to the tail of scsi_dev_info_list via scsi_dev_info_list_init. * -* Do not add to this list, use the command line or proc interface to add -* to the scsi_dev_info_list. This table will eventually go away. +* If possible, set the BLIST_* flags from inside a SCSI LLD rather than +* adding an entry to this list. */ static struct { char *vendor;
Re: [PATCH v2] scsi: core: Fix an incorrect comment
Quoted 4 lines · Bart Van Assche
The comment that
scsi_static_device_listwould go away was added more than 18 years ago. Today, that list is still there and a large number of additional entries have been added. This shows that this comment is incorrect. Hence fix that comment.
Please ignore this patch - this patch was already posted earlier today.
Thanks,
Bart.
PATCH v2 1/2 scsi: core: Introduce the BLIST_SKIP_IO_HINTS flag
In reply to Bart Van AsschePrepare for skipping reading the IO hints VPD page for USB storage devices.
--- drivers/scsi/sd.c | 4 ++++ include/scsi/scsi_devinfo.h | 4 +++- 2 files changed, 7 insertions(+), 1 deletion(-)
drivers/scsi/sd.c +4 -0
@@ -63,6 +63,7 @@#include <scsi/scsi_cmnd.h> #include <scsi/scsi_dbg.h> #include <scsi/scsi_device.h> +#include <scsi/scsi_devinfo.h> #include <scsi/scsi_driver.h> #include <scsi/scsi_eh.h> #include <scsi/scsi_host.h>
@@ -3117,6 +3118,9 @@ static void sd_read_io_hints(struct scsi_disk *sdkp, unsigned char *buffer)struct scsi_mode_data data;int res; +if (sdp->sdev_bflags &BLIST_SKIP_IO_HINTS ) +return ; + res =scsi_mode_sense (sdp,/*dbd=*/ 0x8 ,/*modepage=*/ 0x0a ,/*subpage=*/ 0x05 , buffer,SD_BUF_SIZE ,SD_TIMEOUT , sdkp->max_retries, &data, &sshdr);
include/scsi/scsi_devinfo.h +3 -1
@@ -69,8 +69,10 @@#define BLIST_RETRY_ITF ((__force blist_flags_t)(1ULL << 32))/* Always retry ABORTED_COMMAND with ASC 0xc1 */ #define BLIST_RETRY_ASC_C1 ((__force blist_flags_t)(1ULL << 33)) +/* Do not read the I/O hints mode page */ +#define BLIST_SKIP_IO_HINTS ((__force blist_flags_t)(1ULL << 34)) -#define __BLIST_LAST_USED BLIST_RETRY_ASC_C1 +#define __BLIST_LAST_USED BLIST_SKIP_IO_HINTS#define __BLIST_HIGH_UNUSED (~(__BLIST_LAST_USED | \ (__force blist_flags_t) \
Re: [PATCH v2 1/2] scsi: core: Introduce the BLIST_SKIP_IO_HINTS flag
Hi Bart!
Quoted 1 line
Prepare for skipping reading the IO hints VPD page for USB storage
^^^^^^^^^^^^^^^^^
[...]
drivers/scsi/sd.c static void sd_read_io_hints(struct scsi_disk *sdkp, unsigned char *buffer) · line 3126
Quoted 10 lines
drivers/scsi/sd.c +3
-0@@ -3117,6 +3118,9 @@ static void sd_read_io_hints(struct scsi_disk *sdkp, unsigned char *buffer)struct scsi_mode_data data;int res; +if (sdp->sdev_bflags &BLIST_SKIP_IO_HINTS ) +return ; + res =scsi_mode_sense (sdp,/*dbd=*/ 0x8 ,/*modepage=*/ 0x0a ,/*subpage=*/ 0x05 , buffer,SD_BUF_SIZE ,SD_TIMEOUT , sdkp->max_retries, &data, &sshdr);
^^^^^^^^^^^^^^^
s/IO hints VPD page/IO Advice Hints Grouping mode page/g
?
PS. I'm not fussy about Cc. But I generally avoid listing anybody who will be automatically copied by virtue of any *-by: tags.
I tend to use Cc as an indicator that this entity needs to act upon the patch in question. Ack, review, test, respond to comments, or merge in case of stable@.
If a person listed in Cc subsequently responds with a tag, their name may be listed more than once in the commit description. But I view that as documentation that the person whose feedback was requested actually responded. That's useful information as far as I'm concerned...
Signature · 1 line
Martin K. Petersen Oracle Linux Engineering
Re: [PATCH v2 1/2] scsi: core: Introduce the BLIST_SKIP_IO_HINTS flag
On Thu, Jun 13, 2024 at 9:40 PM Martin K. Petersen <[email protected]> wrote:
...
Quoted 6 lines
PS. I'm not fussy about Cc. But I generally avoid listing anybody who will be automatically copied by virtue of any *-by: tags.
I tend to use Cc as an indicator that this entity needs to act upon the patch in question. Ack, review, test, respond to comments, or merge in case of stable@.
And nobody asks to remove them! The only outcome of what I see is beneficial, i.e. cleaning up the commit message. Everything else will work as expected, e.g., all listed people will be informed as you want to.
Quoted 4 lines
If a person listed in Cc subsequently responds with a tag, their name may be listed more than once in the commit description. But I view that as documentation that the person whose feedback was requested actually responded. That's useful information as far as I'm concerned...
Again, feel free to have a duplicate, but please do it after --- line.
Signature · 2 lines
With Best Regards, Andy Shevchenko
PATCH v2 2/2 scsi: core: Do not query IO hints for USB devices
In reply to Bart Van AsscheRecently it was reported that the following USB storage devices are unusable with Linux kernel 6.9:
* Kingston DataTraveler G2
* Garmin FR35
This is because attempting to read the IO hint VPD page causes these devices to reset. Hence do not read the IO hint VPD page from USB storage devices.
--- drivers/usb/storage/scsiglue.c | 6 ++++++ drivers/usb/storage/uas.c | 7 +++++++ 2 files changed, 13 insertions(+)
drivers/usb/storage/scsiglue.c +6 -0
@@ -79,6 +79,12 @@ static int slave_alloc (struct scsi_device *sdev)if (us->protocol ==USB_PR_BULK && us->max_lun >0 ) sdev->sdev_bflags |=BLIST_FORCELUN ; +/* +* Some USB storage devices reset if the IO hints VPD page is queried. +* Hence skip that VPD page. +*/ + sdev->sdev_bflags |=BLIST_SKIP_IO_HINTS ; +return 0 ; }
drivers/usb/storage/uas.c +7 -0
@@ -21,6 +21,7 @@#include <scsi/scsi.h> #include <scsi/scsi_eh.h> #include <scsi/scsi_dbg.h> +#include <scsi/scsi_devinfo.h> #include <scsi/scsi_cmnd.h> #include <scsi/scsi_device.h> #include <scsi/scsi_host.h>
@@ -820,6 +821,12 @@ static int uas_slave_alloc(struct scsi_device *sdev)struct uas_dev_info *devinfo = (struct uas_dev_info *)sdev->host->hostdata; +/* +* Some USB storage devices reset if the IO hints VPD page is queried. +* Hence skip that VPD page. +*/ + sdev->sdev_bflags |=BLIST_SKIP_IO_HINTS ; + sdev->hostdata = devinfo;return 0 ; }
Re: [PATCH v2 2/2] scsi: core: Do not query IO hints for USB devices
Quoted 22 lines · Bart Van Assche
Recently it was reported that the following USB storage devices are unusable with Linux kernel 6.9:
* Kingston DataTraveler G2
* Garmin FR35This is because attempting to read the IO hint VPD page causes these devices to reset. Hence do not read the IO hint VPD page from USB storage devices.
Cc Alan Stern <[email protected]>Cc Joao Machado <[email protected]>Cc Andy Shevchenko <[email protected]>Cc Christian Heusel <[email protected]>Fixes 4f53138fffc2 ("scsi: sd: Translate data lifetime information")Reported-by Joao Machado <[email protected]>Closes https://lore.kernel.org/linux-scsi/[email protected]/T/#mf4e3410d8f210454d7e4c3d1fb5c0f41e651b85f open hereTested-by Andy Shevchenko <[email protected]>Bisected-by Christian Heusel <[email protected]>Reported-by Andy Shevchenko <[email protected]>Closes https://lore.kernel.org/linux-scsi/CACLx9VdpUanftfPo2jVAqXdcWe8Y43MsDeZmMPooTzVaVJAh2w@mail.gmail.com/ open hereSigned-off-by Bart Van Assche <[email protected]>
Quoted 1 line
---
Quoted 46 lines
drivers/usb/storage/scsiglue.c | 6 ++++++ drivers/usb/storage/uas.c | 7 +++++++ 2 files changed, 13 insertions(+)drivers/usb/storage/scsiglue.c +6
-0@@ -79,6 +79,12 @@ static int slave_alloc (struct scsi_device *sdev)if (us->protocol ==USB_PR_BULK && us->max_lun >0 ) sdev->sdev_bflags |=BLIST_FORCELUN ; +/* +* Some USB storage devices reset if the IO hints VPD page is queried. +* Hence skip that VPD page. +*/ + sdev->sdev_bflags |=BLIST_SKIP_IO_HINTS ; +return 0 ; }drivers/usb/storage/uas.c +7
-0@@ -21,6 +21,7 @@#include <scsi/scsi.h> #include <scsi/scsi_eh.h> #include <scsi/scsi_dbg.h> +#include <scsi/scsi_devinfo.h> #include <scsi/scsi_cmnd.h> #include <scsi/scsi_device.h> #include <scsi/scsi_host.h> @@ -820,6 +821,12 @@ static int uas_slave_alloc(struct scsi_device *sdev)struct uas_dev_info *devinfo = (struct uas_dev_info *)sdev->host->hostdata; +/* +* Some USB storage devices reset if the IO hints VPD page is queried. +* Hence skip that VPD page. +*/ + sdev->sdev_bflags |=BLIST_SKIP_IO_HINTS ; + sdev->hostdata = devinfo;return 0 ; }
Re: [PATCH v2 2/2] scsi: core: Do not query IO hints for USB devices
In reply to Bart Van AsscheQuoted 8 lines · Bart Van Assche
Recently it was reported that the following USB storage devices are unusable with Linux kernel 6.9:
* Kingston DataTraveler G2
* Garmin FR35This is because attempting to read the IO hint VPD page causes these devices to reset. Hence do not read the IO hint VPD page from USB storage devices.
I have commented on v1, same applicable here. Not that it's a big deal for this change, but in general can you follow the advice given there?
Signature · 2 lines
With Best Regards, Andy Shevchenko