[PATCH v1] ufs: core: complete wl runtime resume after SCSI EH

[Hongjie Fang]

A PM START STOP sent from the UFS well-known LU resume path can race with
SCSI EH.

The "wl resume" task flow is:
  __ufshcd_wl_resume()
    ufshcd_set_dev_pwr_mode(UFS_ACTIVE_PWR_MODE)
      ufshcd_execute_start_stop()
        scsi_execute_cmd()
          blk_execute_rq           <-- wait
          scsi_check_passthrough() <-- may retry START STOP

If the first START STOP time out, SCSI EH may already recover the link and
reset the device before scsi_execute_cmd() returns:
  scsi_timeout()
    scsi_eh_scmd_add()
      scsi_error_handler()
        scsi_unjam_host()
          scsi_eh_ready_devs()
            scsi_eh_host_reset()
              ufshcd_eh_host_reset_handler()
                if (hba->pm_op_in_progress)
                  ufshcd_link_recovery()
                    ufshcd_device_reset()
                    ufshcd_host_reset_and_restore()
          ...
          scsi_eh_flush_done_q()   <-- wakeup "wl resume" task
        ...                        <-- host still in SHOST_RECOVERY
        scsi_restart_operations()

A later passthrough retry can then run while the host is still in
SHOST_RECOVERY and hit the SCMD_FAIL_IF_RECOVERING path:
  scsi_queue_rq()
    if (scsi_host_in_recovery(shost) &&
        cmd->flags & SCMD_FAIL_IF_RECOVERING)
      return BLK_STS_OFFLINE

That retry completes with DID_ERROR or DID_NO_CONNECT even though EH may
already have restored the device to an operational ACTIVE state.

In this case __ufshcd_wl_resume() can return -EIO even though recovery
has already restored the device to an operational ACTIVE state. This is
especially harmful for runtime PM because callback errors other than
-EAGAIN and -EBUSY are latched by the PM core as runtime_error.

After a runtime-PM START STOP returns -EIO, wait for any ongoing SCSI EH
round to finish and treat the resume as successful if recovery has
already restored:

  - an online device
  - UFSHCD_STATE_OPERATIONAL
  - an active link
  - cached ACTIVE device power mode

Fixes: b294ff3e3449 ("scsi: ufs: core: Enable power management for wlun")
Signed-off-by: Hongjie Fang <hongjiefang@xxxxxxxxxxxx>
---
 drivers/ufs/core/ufshcd.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
index c3f08957d179..e7517cf23f06 100644
--- a/drivers/ufs/core/ufshcd.c
+++ b/drivers/ufs/core/ufshcd.c
@@ -10368,6 +10368,22 @@ static int __ufshcd_wl_suspend(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 }
 
 #ifdef CONFIG_PM
+static int ufshcd_wl_resume_pm_recovered(struct ufs_hba *hba)
+{
+	int ret = 0;
+	struct scsi_device *sdp = hba->ufs_device_wlun;
+
+	if (!sdp || !scsi_block_when_processing_errors(sdp))
+		return 0;
+
+	if (hba->ufshcd_state == UFSHCD_STATE_OPERATIONAL &&
+	    ufshcd_is_link_active(hba) &&
+	    ufshcd_is_ufs_dev_active(hba))
+		ret = 1;
+
+	return ret;
+}
+
 static int __ufshcd_wl_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 {
 	int ret;
@@ -10422,6 +10438,9 @@ static int __ufshcd_wl_resume(struct ufs_hba *hba, enum ufs_pm_op pm_op)
 
 	if (!ufshcd_is_ufs_dev_active(hba)) {
 		ret = ufshcd_set_dev_pwr_mode(hba, UFS_ACTIVE_PWR_MODE);
+		if (pm_op == UFS_RUNTIME_PM && ret == -EIO &&
+		    ufshcd_wl_resume_pm_recovered(hba))
+			ret = 0;
 		if (ret)
 			goto set_old_link_state;
 		ufshcd_set_timestamp_attr(hba);
-- 
2.25.1