mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
* [PATCH] usb: uas: expose US_FL_NO_ATA_1X in sysfs as deny_ata_passthrough
@ 2026-11-15 12:40 Mateusz Jończyk
  2026-11-15 15:52 ` Alan Stern
  0 siblings, 1 reply; 3+ messages in thread
From: Mateusz Jończyk @ 2026-11-15 12:40 UTC (permalink / raw)
  To: Oliver Neukum
  Cc: Alan Stern, Greg Kroah-Hartman, linux-usb, linux-scsi,
	usb-storage, linux-kernel, smartmontools-support,
	Christian Franke, Mateusz Jończyk

The uas driver blocks access to ATA passthrough commands for some
devices it considers buggy (via the US_FL_NO_ATA_1X flag in
unusual_uas.h and uas_detect.h). However, it does not expose any
meaningful information whether the command was blocked by the driver or
is genuinly not supported by the hardware. This creates a problem for
the smartmontools program and makes it difficult for it to provide a
useful error message (with instructions to disable UAS on the drive to
get SMART readings).

So, expose the information in sysfs as deny_ata_passthrough. Other
quirks may be added in the future if there is any demand.

Link: https://www.smartmontools.org/wiki/SAT-with-UAS-Linux
Signed-off-by: Mateusz Jończyk <mat.jonczyk@o2.pl>
---
 drivers/usb/storage/uas.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/usb/storage/uas.c b/drivers/usb/storage/uas.c
index 8655edbd66b1..aaea74297026 100644
--- a/drivers/usb/storage/uas.c
+++ b/drivers/usb/storage/uas.c
@@ -910,6 +910,25 @@ static int uas_sdev_configure(struct scsi_device *sdev,
 	return 0;
 }
 
+/* Sysfs interface */
+static ssize_t deny_ata_passthrough_show(struct device *dev, struct device_attribute *attr,
+					 char *buf)
+{
+	struct scsi_device *sdev = to_scsi_device(dev);
+	struct uas_dev_info *devinfo = sdev->hostdata;
+
+	return sprintf(buf, "%d\n", !!(devinfo->flags & US_FL_NO_ATA_1X));
+}
+
+static DEVICE_ATTR_RO(deny_ata_passthrough);
+
+static struct attribute *uas_sdev_attrs[] = {
+	&dev_attr_deny_ata_passthrough.attr,
+	NULL,
+};
+
+ATTRIBUTE_GROUPS(uas_sdev);
+
 static const struct scsi_host_template uas_host_template = {
 	.module = THIS_MODULE,
 	.name = "uas",
@@ -929,6 +948,7 @@ static const struct scsi_host_template uas_host_template = {
 	.dma_alignment = 511,
 	.dma_boundary = PAGE_SIZE - 1,
 	.cmd_size = sizeof(struct uas_cmd_info),
+	.sdev_groups = uas_sdev_groups,
 };
 
 #define UNUSUAL_DEV(id_vendor, id_product, bcdDeviceMin, bcdDeviceMax, \
-- 
2.47.3


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] usb: uas: expose US_FL_NO_ATA_1X in sysfs as deny_ata_passthrough
  2026-11-15 12:40 [PATCH] usb: uas: expose US_FL_NO_ATA_1X in sysfs as deny_ata_passthrough Mateusz Jończyk
@ 2026-11-15 15:52 ` Alan Stern
  2026-11-15 18:26   ` Mateusz Jończyk
  0 siblings, 1 reply; 3+ messages in thread
From: Alan Stern @ 2026-11-15 15:52 UTC (permalink / raw)
  To: Mateusz Jończyk
  Cc: Oliver Neukum, Greg Kroah-Hartman, linux-usb, linux-scsi,
	usb-storage, linux-kernel, smartmontools-support,
	Christian Franke

On Sun, Nov 15, 2026 at 01:40:46PM +0100, Mateusz Jończyk wrote:
> The uas driver blocks access to ATA passthrough commands for some
> devices it considers buggy (via the US_FL_NO_ATA_1X flag in
> unusual_uas.h and uas_detect.h). However, it does not expose any
> meaningful information whether the command was blocked by the driver or
> is genuinly not supported by the hardware. This creates a problem for
> the smartmontools program and makes it difficult for it to provide a
> useful error message (with instructions to disable UAS on the drive to
> get SMART readings).
> 
> So, expose the information in sysfs as deny_ata_passthrough. Other
> quirks may be added in the future if there is any demand.
> 
> Link: https://www.smartmontools.org/wiki/SAT-with-UAS-Linux
> Signed-off-by: Mateusz Jończyk <mat.jonczyk@o2.pl>

Why was this change made to uas but not to usb-storage?  Was it because 
some devices can handle ATA passthrough commands over the Bulk-Only 
transport but not over the UAS transport?  Note that the 
US_FL_NO_ATA_1X flag applies to both transports.

Alan Stern

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [PATCH] usb: uas: expose US_FL_NO_ATA_1X in sysfs as deny_ata_passthrough
  2026-11-15 15:52 ` Alan Stern
@ 2026-11-15 18:26   ` Mateusz Jończyk
  0 siblings, 0 replies; 3+ messages in thread
From: Mateusz Jończyk @ 2026-11-15 18:26 UTC (permalink / raw)
  To: Alan Stern
  Cc: Oliver Neukum, Greg Kroah-Hartman, linux-usb, linux-scsi,
	usb-storage, linux-kernel, smartmontools-support,
	Christian Franke

W dniu 15.11.2026 o 16:52, Alan Stern pisze:
> On Sun, Nov 15, 2026 at 01:40:46PM +0100, Mateusz Jończyk wrote:
>> The uas driver blocks access to ATA passthrough commands for some
>> devices it considers buggy (via the US_FL_NO_ATA_1X flag in
>> unusual_uas.h and uas_detect.h). However, it does not expose any
>> meaningful information whether the command was blocked by the driver or
>> is genuinly not supported by the hardware. This creates a problem for
>> the smartmontools program and makes it difficult for it to provide a
>> useful error message (with instructions to disable UAS on the drive to
>> get SMART readings).
>>
>> So, expose the information in sysfs as deny_ata_passthrough. Other
>> quirks may be added in the future if there is any demand.
>>
>> Link: https://www.smartmontools.org/wiki/SAT-with-UAS-Linux
>> Signed-off-by: Mateusz Jończyk <mat.jonczyk@o2.pl>
> Why was this change made to uas but not to usb-storage?  Was it because
> some devices can handle ATA passthrough commands over the Bulk-Only
> transport but not over the UAS transport?  Note that the
> US_FL_NO_ATA_1X flag applies to both transports.
>
> Alan Stern

There are only very few devices that cannot handle ATA passthrough commands
over the Bulk-Only transfer (and with them it's impossible to get ATA
passthrough by disabling UAS). But that's a fair point, will fix it in v2.

Considering the US_FL_NO_ATA_1X flag:

- I suspect that many more devices would accept the ATA passthrough 
commands in
   UAS mode if these would be issued in isolation (if uas_queuecommand_lck()
   would stop all other transfers while doing them). After all, Windows
   apparently does not carry such quirks for modern devices. I do not 
have the
   hardware to test, though.

- Since 2017, it is set unequivocally for all Seagate drives, even 
though most
   likely modern drives / firmware versions do not need it. I do not see 
a way
   out, though.

Greetings,

Mateusz


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2026-11-15 18:26 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2026-11-15 12:40 [PATCH] usb: uas: expose US_FL_NO_ATA_1X in sysfs as deny_ata_passthrough Mateusz Jończyk
2026-11-15 15:52 ` Alan Stern
2026-11-15 18:26   ` Mateusz Jończyk

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

all inboxes | Powered by JetHome®