[PATCH v5 0/3] fix runtime PM for SD card readers

[PATCH v5 0/3] fix runtime PM for SD card readers

[Martin Kepplinger]

hi,

(According to Alan Stern, "as far as I know, all") SD card readers send
MEDIA_CHANGED unit attention notification on (runtime) resume. We cannot
use runtime PM with these devices as I/O always fails in that case.

This fixes runtime PM for SD cardreaders. I'd appreciate any feedback.

To enable runtime PM for an SD cardreader number 0:0:0:0, do:
    
echo 0 > /sys/module/block/parameters/events_dfl_poll_msecs
echo 1000 > /sys/bus/scsi/devices/0:0:0:0/power/autosuspend_delay_ms
echo auto > /sys/bus/scsi/devices/0:0:0:0/power/control

thank you,
                           martin

revision history
----------------
v5: (thank you Bart)
* simplify the sense request itself and remove unnecessary code

v4: (thank you Bart and Alan)
* send SENSE REQUEST in sd instead of adding a global scsi error flag.
https://lore.kernel.org/linux-scsi/20210628133412.1172068-1-martin.kepplinger@puri.sm/T/#t

v3: (thank you Bart)
 * create a new BLIST entry to mark affected devices instead of the
   sysfs module parameter for sd only. still, only sd implements handling
   the flag for now.
 * cc linux-pm list
https://lore.kernel.org/linux-scsi/20210328102531.1114535-1-martin.kepplinger@puri.sm/

v2:
https://lore.kernel.org/linux-scsi/20210112093329.3639-1-martin.kepplinger@puri.sm/
 * move module parameter to sd
 * add Documentation
v1:
https://lore.kernel.org/linux-scsi/20210111152029.28426-1-martin.kepplinger@puri.sm/T/

For the full background, the discussion started in June 2020 here:
https://lore.kernel.org/linux-scsi/20200623111018.31954-1-martin.kepplinger@puri.sm/


Martin Kepplinger (3):
  scsi: devinfo: add new flag BLIST_MEDIA_CHANGE
  scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in
    runtime_resume()
  scsi: devinfo: add BLIST_MEDIA_CHANGE for Ultra HS-SD/MMC usb
    cardreaders

 drivers/scsi/scsi_devinfo.c |  1 +
 drivers/scsi/sd.c           | 26 +++++++++++++++++++++++++-
 include/scsi/scsi_devinfo.h |  6 +++---
 3 files changed, 29 insertions(+), 4 deletions(-)

-- 
2.30.2

[PATCH v5 1/3] scsi: devinfo: add new flag BLIST_MEDIA_CHANGE

[Martin Kepplinger]

add a new flag for devices that issue MEDIA CHANGE unit attentions
when actually no medium changed. Drivers can then act accordingly in
case they need to work around it, i.e. in resume().

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
---
 include/scsi/scsi_devinfo.h | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/include/scsi/scsi_devinfo.h b/include/scsi/scsi_devinfo.h
index 3fdb322d4c4b..766fa876598e 100644
--- a/include/scsi/scsi_devinfo.h
+++ b/include/scsi/scsi_devinfo.h
@@ -28,7 +28,8 @@
 #define BLIST_LARGELUN		((__force blist_flags_t)(1ULL << 9))
 /* override additional length field */
 #define BLIST_INQUIRY_36	((__force blist_flags_t)(1ULL << 10))
-#define __BLIST_UNUSED_11	((__force blist_flags_t)(1ULL << 11))
+/* ignore one MEDIA CHANGE unit attention after resuming from runtime suspend */
+#define BLIST_MEDIA_CHANGE	((__force blist_flags_t)(1ULL << 11))
 /* do not do automatic start on add */
 #define BLIST_NOSTARTONADD	((__force blist_flags_t)(1ULL << 12))
 #define __BLIST_UNUSED_13	((__force blist_flags_t)(1ULL << 13))
@@ -73,8 +74,7 @@
 #define __BLIST_HIGH_UNUSED (~(__BLIST_LAST_USED | \
 			       (__force blist_flags_t) \
 			       ((__force __u64)__BLIST_LAST_USED - 1ULL)))
-#define __BLIST_UNUSED_MASK (__BLIST_UNUSED_11 | \
-			     __BLIST_UNUSED_13 | \
+#define __BLIST_UNUSED_MASK (__BLIST_UNUSED_13 | \
 			     __BLIST_UNUSED_14 | \
 			     __BLIST_UNUSED_15 | \
 			     __BLIST_UNUSED_16 | \
-- 
2.30.2

[PATCH v5 2/3] scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in runtime_resume()

[Martin Kepplinger]

For SD cardreader devices that have the BLIST_MEDIA_CHANGE flag set,
a MEDIA CHANGE unit attention is received after resuming from runtime
suspend. Send a REQUEST SENSE to avoid that.

The "downside" is that for these devices we now rely on users not to
really change the medium (SD card) *during* a runtime suspend/resume
cycle, i.e. when not unmounting.

To enable runtime PM for an SD cardreader (device number 0:0:0:0), do:

echo 0 > /sys/module/block/parameters/events_dfl_poll_msecs
echo 1000 > /sys/bus/scsi/devices/0:0:0:0/power/autosuspend_delay_ms
echo auto > /sys/bus/scsi/devices/0:0:0:0/power/control

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
---
 drivers/scsi/sd.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index 6d2d63629a90..b02378f40620 100644
--- a/drivers/scsi/sd.c
+++ b/drivers/scsi/sd.c
@@ -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>
@@ -114,6 +115,7 @@ static void sd_shutdown(struct device *);
 static int sd_suspend_system(struct device *);
 static int sd_suspend_runtime(struct device *);
 static int sd_resume(struct device *);
+static int sd_resume_runtime(struct device *);
 static void sd_rescan(struct device *);
 static blk_status_t sd_init_command(struct scsi_cmnd *SCpnt);
 static void sd_uninit_command(struct scsi_cmnd *SCpnt);
@@ -608,7 +610,7 @@ static const struct dev_pm_ops sd_pm_ops = {
 	.poweroff		= sd_suspend_system,
 	.restore		= sd_resume,
 	.runtime_suspend	= sd_suspend_runtime,
-	.runtime_resume		= sd_resume,
+	.runtime_resume		= sd_resume_runtime,
 };
 
 static struct scsi_driver sd_template = {
@@ -3720,6 +3722,28 @@ static int sd_resume(struct device *dev)
 	return ret;
 }
 
+static int sd_resume_runtime(struct device *dev)
+{
+	struct scsi_disk *sdkp = dev_get_drvdata(dev);
+	struct scsi_device *sdp = sdkp->device;
+	int timeout, res;
+
+	timeout = sdp->request_queue->rq_timeout * SD_FLUSH_TIMEOUT_MULTIPLIER;
+
+	if (sdp->sdev_bflags & BLIST_MEDIA_CHANGE) {
+		/* clear the devices' sense data */
+		static const u8 cmd[10] = { REQUEST_SENSE };
+
+		res = scsi_execute(sdp, cmd, DMA_NONE, NULL, 0, NULL,
+				   NULL, timeout, 1, 0, RQF_PM, NULL);
+		if (res)
+			sd_printk(KERN_NOTICE, sdkp,
+				  "Failed to clear sense data\n");
+	}
+
+	return sd_resume(dev);
+}
+
 /**
  *	init_sd - entry point for this driver (both when built in or when
  *	a module).
-- 
2.30.2

[PATCH v5 3/3] scsi: devinfo: add BLIST_MEDIA_CHANGE for Ultra HS-SD/MMC usb cardreaders

[Martin Kepplinger]

These cardreader devices issue a MEDIA CHANGE unit attention not only
when actually a medium changed but also simply when resuming from suspend.

Setting the BLIST_MEDIA_CHANGE flag enables using runtime pm for SD cardreaders.

Signed-off-by: Martin Kepplinger <martin.kepplinger@puri.sm>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
---
 drivers/scsi/scsi_devinfo.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/scsi/scsi_devinfo.c b/drivers/scsi/scsi_devinfo.c
index d33355ab6e14..8ac0a11ac541 100644
--- a/drivers/scsi/scsi_devinfo.c
+++ b/drivers/scsi/scsi_devinfo.c
@@ -171,6 +171,7 @@ static struct {
 	{"FUJITSU", "ETERNUS_DXM", "*", BLIST_RETRY_ASC_C1},
 	{"Generic", "USB SD Reader", "1.00", BLIST_FORCELUN | BLIST_INQUIRY_36},
 	{"Generic", "USB Storage-SMC", NULL, BLIST_FORCELUN | BLIST_INQUIRY_36}, /* FW: 0180 and 0207 */
+	{"Generic", "Ultra HS-SD/MMC", "2.09", BLIST_MEDIA_CHANGE | BLIST_INQUIRY_36},
 	{"HITACHI", "DF400", "*", BLIST_REPORTLUN2},
 	{"HITACHI", "DF500", "*", BLIST_REPORTLUN2},
 	{"HITACHI", "DISK-SUBSYSTEM", "*", BLIST_REPORTLUN2},
-- 
2.30.2

Re: [PATCH v5 2/3] scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in runtime_resume()

[Bart Van Assche]

On 6/30/21 1:44 AM, Martin Kepplinger wrote:
> For SD cardreader devices that have the BLIST_MEDIA_CHANGE flag set,
> a MEDIA CHANGE unit attention is received after resuming from runtime
> suspend. Send a REQUEST SENSE to avoid that.

Reviewed-by: Bart Van Assche <bvanassche@acm.org>

Re: [PATCH v5 2/3] scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in runtime_resume()

[Christoph Hellwig]

On Wed, Jun 30, 2021 at 10:44:52AM +0200, Martin Kepplinger wrote:
> +	struct scsi_disk *sdkp = dev_get_drvdata(dev);
> +	struct scsi_device *sdp = sdkp->device;
> +	int timeout, res;
> +
> +	timeout = sdp->request_queue->rq_timeout * SD_FLUSH_TIMEOUT_MULTIPLIER;

Is REQUEST SENSE reqlly a so slow operation on these devices that
we need to override the timeout?

Re: [PATCH v5 2/3] scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in runtime_resume()

[Martin Kepplinger]

Am Donnerstag, dem 01.07.2021 um 15:49 +0100 schrieb Christoph Hellwig:
> On Wed, Jun 30, 2021 at 10:44:52AM +0200, Martin Kepplinger wrote:
> > +       struct scsi_disk *sdkp = dev_get_drvdata(dev);
> > +       struct scsi_device *sdp = sdkp->device;
> > +       int timeout, res;
> > +
> > +       timeout = sdp->request_queue->rq_timeout *
> > SD_FLUSH_TIMEOUT_MULTIPLIER;
> 
> Is REQUEST SENSE reqlly a so slow operation on these devices that
> we need to override the timeout?

using SD_TIMEOUT works equally fine for me. Is that what you'd rather
like to see?

Bart, is SD_TIMEOUT equally ok for you? If so, I'll resend with your
reviewed-by.

thank you for reviewing!

                            martin

Re: [PATCH v5 2/3] scsi: sd: send REQUEST SENSE for BLIST_MEDIA_CHANGE devices in runtime_resume()

[Bart Van Assche]

On 7/2/21 1:04 AM, Martin Kepplinger wrote:
> Am Donnerstag, dem 01.07.2021 um 15:49 +0100 schrieb Christoph Hellwig:
>> On Wed, Jun 30, 2021 at 10:44:52AM +0200, Martin Kepplinger wrote:
>>> +       struct scsi_disk *sdkp = dev_get_drvdata(dev);
>>> +       struct scsi_device *sdp = sdkp->device;
>>> +       int timeout, res;
>>> +
>>> +       timeout = sdp->request_queue->rq_timeout *
>>> SD_FLUSH_TIMEOUT_MULTIPLIER;
>>
>> Is REQUEST SENSE reqlly a so slow operation on these devices that
>> we need to override the timeout?
> 
> using SD_TIMEOUT works equally fine for me. Is that what you'd rather
> like to see?
> 
> Bart, is SD_TIMEOUT equally ok for you? If so, I'll resend with your
> reviewed-by.

Hi Martin,

I prefer sdp->request_queue->rq_timeout instead of SD_TIMEOUT since the 
former is configurable via sysfs.

Thanks,

Bart.