linux-toradex.git/drivers/scsi/ufs, branch master

scsi: ufs: Split the drivers/scsi/ufs directory

2022-05-20T00:27:37+00:00

Split the drivers/scsi/ufs directory into 'core' and 'host' directories
under the drivers/ufs/ directory. Move shared header files into the
include/ufs/ directory. This separation makes it clear which header files
UFS drivers are allowed to include (include/ufs/*.h) and which header files
UFS drivers are not allowed to include (drivers/ufs/core/*.h).

Update the MAINTAINERS file. Add myself as a UFS reviewer.

Link: https://lore.kernel.org/r/20220511212552.655341-1-bvanassche@acm.org
Cc: Adrian Hunter 
Cc: Avri Altman 
Cc: Bean Huo 
Cc: Bjorn Andersson 
Cc: Keoseong Park 
Tested-by: Bean Huo 
Tested-by: Adrian Hunter 
Reviewed-by: Bean Huo 
Acked-by: Avri Altman 
Acked-by: Adrian Hunter 
Signed-off-by: Bart Van Assche 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Clean up ufshpb_suspend()/resume()

2022-05-11T02:01:52+00:00

ufshpb_resume() is only called when the HPB state is HPB_SUSPEND, so the
check statement for "ufshpb_get_state(hpb) != HPB_PRESENT" is useless.

Link: https://lore.kernel.org/r/20220505134707.35929-7-huobean@gmail.com
Reviewed-by: Keoseong Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Add handing of device reset regions in HPB device mode

2022-05-11T02:01:52+00:00

In UFS HPB Spec JESD220-3A,

"5.8. Active and inactive information upon power cycle
...
When the device is powered off by the host, the device may restore L2P map
data upon power up or build from the host's HPB READ command. In case
device powered up and lost HPB information, device can signal to the host
through HPB Sense data, by setting HPB Operation as '2' which will inform
the host that device reset HPB information."

Therefore, for HPB device control mode, if the UFS device is reset via the
RST_N pin, the active region information in the device will be reset. If
the host side receives this notification from the device side, it is
recommended to inactivate all active regions in the host's HPB cache.

Link: https://lore.kernel.org/r/20220505134707.35929-6-huobean@gmail.com
Reviewed-by: Keoseong Park 
Reviewed-by: Daejun Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Change sysfs node hpb_stats/rb_* prefix to start with rcmd_*

2022-05-11T02:01:52+00:00

According to the documentation of the sysfs nodes rb_noti_cnt,
rb_active_cnt and rb_inactive_cnt, these are all related to HPB
recommendation in UPIU response packet. 'rcmd' (recommendation) should be
the correct abbreviation.

Change the sysfs documentation about these sysfs nodes to highlight what
they mean under different HPB control modes.

Link: https://lore.kernel.org/r/20220505134707.35929-5-huobean@gmail.com
Reviewed-by: Keoseong Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Clean up the handler when device resets HPB information

2022-05-11T02:01:51+00:00

"When the device is powered off by the host, the device may restore L2P map
data upon power up or build from the host's HPB READ command. In case
device powered up and lost HPB information, device can signal to the host
through HPB Sense data, by setting HPB Operation as '2' which will inform
the host that device reset HPB information."

Clean up the handler and make the intent of this handler more readable, no
functional change.

Link: https://lore.kernel.org/r/20220505134707.35929-4-huobean@gmail.com
Reviewed-by: Keoseong Park 
Reviewed-by: Daejun Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Remove enum initialization value

2022-05-11T02:01:51+00:00

If the first enumerator has no initializer, the value of the corresponding
constant is zero.

Link: https://lore.kernel.org/r/20220505134707.35929-3-huobean@gmail.com
Reviewed-by: Keoseong Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: ufshpb: Merge ufshpb_reset() and ufshpb_reset_host()

2022-05-11T02:01:51+00:00

There is no functional change in this patch, just merge ufshpb_reset() and
ufshpb_reset_host() into one function ufshpb_toggle_state().

Link: https://lore.kernel.org/r/20220505134707.35929-2-huobean@gmail.com
Reviewed-by: Keoseong Park 
Signed-off-by: Bean Huo 
Signed-off-by: Martin K. Petersen

scsi: ufs: qcom: Enable RPM_AUTOSUSPEND for runtime PM

2022-05-11T01:52:53+00:00

In order to allow the block devices to enter autosuspend mode during
runtime, thereby allowing the ufshcd host driver to also runtime suspend,
let's make use of the RPM_AUTOSUSPEND flag.

Without this flag, userspace needs to enable the autosuspend feature of
the block devices through sysfs.

Link: https://lore.kernel.org/r/20220504084212.11605-6-manivannan.sadhasivam@linaro.org
Reviewed-by: Bjorn Andersson 
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Martin K. Petersen

scsi: ufs: core: Remove redundant wmb() in ufshcd_send_command()

2022-05-11T01:52:53+00:00

The wmb() inside ufshcd_send_command() is added to make sure that the
doorbell is committed immediately. This leads to couple of expectations:

 1. The doorbell write should complete before the function return.

 2. The doorbell write should not cross the function boundary.

2nd expectation is fullfilled by the Linux memory model as there is a
guarantee that the critical section won't cross the unlock (release)
operation.

1st expectation is not really needed here as there is no following read/
write that depends on the doorbell to be complete implicitly. Even if the
doorbell write is in a CPUs Write Buffer (WB), wmb() won't flush it. And
there is no real need of a WB flush here as well.

So let's get rid of the wmb() that seems redundant.

Link: https://lore.kernel.org/r/20220504084212.11605-5-manivannan.sadhasivam@linaro.org
Reviewed-by: Bart Van Assche 
Reviewed-by: Bjorn Andersson 
Acked-by: Bean Huo 
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Martin K. Petersen

scsi: ufs: qcom: Add a readl() to make sure ref_clk gets enabled

2022-05-11T01:52:53+00:00

In ufs_qcom_dev_ref_clk_ctrl(), it was noted that the ref_clk needs to be
stable for at least 1us. Even though there is wmb() to make sure the write
gets "completed", there is no guarantee that the write actually reached the
UFS device. There is a good chance that the write could be stored in a
Write Buffer (WB). In that case, even though the CPU waits for 1us, the
ref_clk might not be stable for that period.

So lets do a readl() to make sure that the previous write has reached the
UFS device before udelay().

Also, the wmb() after writel_relaxed() is not really needed. Both writel()
and readl() are ordered on all architectures and the CPU won't speculate
instructions after readl() due to the in-built control dependency with read
value on weakly ordered architectures. So it can be safely removed.

Link: https://lore.kernel.org/r/20220504084212.11605-4-manivannan.sadhasivam@linaro.org
Fixes: f06fcc7155dc ("scsi: ufs-qcom: add QUniPro hardware support and power optimizations")
Cc: stable@vger.kernel.org
Reviewed-by: Bjorn Andersson 
Signed-off-by: Manivannan Sadhasivam 
Signed-off-by: Martin K. Petersen