platform/kernel/linux-rpi.git
2 years agoscsi: smartpqi: Stop logging spurious PQI reset failures
Kevin Barnett [Fri, 8 Jul 2022 18:47:31 +0000 (13:47 -0500)]
scsi: smartpqi: Stop logging spurious PQI reset failures

Change method used to detect controller firmware crash during PQI reset.

PQI reset can fail with error -6 if firmware takes > 100ms to complete
reset.

Method used by driver to detect controller firmware crash during PQI was
incorrect in some cases.

Link: https://lore.kernel.org/r/165730605108.177165.1132931838384767071.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Signed-off-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Add PCI IDs for Lenovo controllers
Mike McGowen [Fri, 8 Jul 2022 18:47:26 +0000 (13:47 -0500)]
scsi: smartpqi: Add PCI IDs for Lenovo controllers

Add PCI IDs for Lenovo controllers (values in hex):

                                        VID  / DID  / SVID / SDID
                                        ----   ----   ----   ----
Lenovo 4350-8i HBA                      9005 / 028f / 1d49 / 0220
Lenovo 4350-16i HBA                     9005 / 028f / 1d49 / 0221
Lenovo 5350-8i RAID                     9005 / 028f / 1d49 / 0520
Lenovo 5350-8i Internal RAID            9005 / 028f / 1d49 / 0522
Lenovo 9350-8i RAID                     9005 / 028f / 1d49 / 0620
Lenovo 9350-8i Internal RAID            9005 / 028f / 1d49 / 0621
Lenovo 9350-16i RAID                    9005 / 028f / 1d49 / 0622
Lenovo 9350-16i Internal RAID           9005 / 028f / 1d49 / 0623

Link: https://lore.kernel.org/r/165730604598.177165.9910276232981721083.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Mike McGowen <Mike.McGowen@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Add PCI ID for Adaptec SmartHBA 2100-8i
Mike McGowen [Fri, 8 Jul 2022 18:47:20 +0000 (13:47 -0500)]
scsi: smartpqi: Add PCI ID for Adaptec SmartHBA 2100-8i

Add the PCI ID for (values in hex):
                                        VID  / DID  / SVID / SDID
                                        ----   ----   ----   ----
Adaptec SmartHBA 2100-8i-o              9005 / 0285 / 9005 / 0659

Link: https://lore.kernel.org/r/165730604089.177165.17257514581321583667.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Mike McGowen <Mike.McGowen@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Fix PCI control linkdown system hang
Sagar Biradar [Fri, 8 Jul 2022 18:47:15 +0000 (13:47 -0500)]
scsi: smartpqi: Fix PCI control linkdown system hang

Fail all outstanding requests after a PCI linkdown.

Block access to device SCSI attributes during the following conditions:

  "Cable pull" is called PQI_CTRL_SURPRISE_REMOVAL.

  "PCIe Link Down" is called PQI_CTRL_GRACEFUL_REMOVAL.

Block access to device SCSI attributes during and in rare instances when
the controller goes offline.

Either outstanding requests or the access of SCSI attributes post linkdown
can lead to a hang.

Post linkdown, driver does not fail the outstanding requests leading to
long wait time before all the IOs eventually fail.

Also access of the SCSI attributes by host applications can lead to a
system hang.

Link: https://lore.kernel.org/r/165730603578.177165.4699352086827187263.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Sagar Biradar <sagar.biradar@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Add driver support for multi-LUN devices
Kumar Meiyappan [Fri, 8 Jul 2022 18:47:10 +0000 (13:47 -0500)]
scsi: smartpqi: Add driver support for multi-LUN devices

Add driver support for up to 256 LUNs per device.

Update AIO path to pass the appropriate LUN number for base-code to target
the correct LUN.

Update RAID IO path to pass the appropriate LUN number for FW to target the
correct LUN.

Pass the correct LUN number while doing a LUN reset.

Count the outstanding commands based on LUN number.  While removing a
Multi-LUN device, wait for all outstanding commands to complete for all
LUNs.

Add Feature bit support.

Link: https://lore.kernel.org/r/165730603067.177165.14016422176841798336.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Kumar Meiyappan <Kumar.Meiyappan@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Close write read holes
Mike McGowen [Fri, 8 Jul 2022 18:47:05 +0000 (13:47 -0500)]
scsi: smartpqi: Close write read holes

Insert a minimum 1 millisecond delay after writing to a register before
reading from it.

SIS and PQI registers that can be both written to and read from can return
stale data if read from too soon after having been written to.

There is no read/write ordering or hazard detection on the inbound path to
the MSGU from the PCIe bus, therefore reads could pass writes.

Link: https://lore.kernel.org/r/165730602555.177165.11181012469428348394.stgit@brunhilda
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Signed-off-by: Mike McGowen <mike.mcgowen@microchip.com>
Co-developed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Add PCI IDs for ramaxel controllers
Murthy Bhat [Fri, 8 Jul 2022 18:47:00 +0000 (13:47 -0500)]
scsi: smartpqi: Add PCI IDs for ramaxel controllers

Add the following controllers (values in hex):

                               VID  / DID  / SVID / SDID
                               ---- / ---- / ---- / ----
Ramaxel FBGF-RAD PM8204        9005 / 028F / 1CC4 / 0101
Ramaxel FBGF-RAD PM8222        9005 / 028F / 1CC4 / 0201

Link: https://lore.kernel.org/r/165730602045.177165.3720208650043407285.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Murthy Bhat <Murthy.Bhat@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Add controller fw version to console log
Gilbert Wu [Fri, 8 Jul 2022 18:46:55 +0000 (13:46 -0500)]
scsi: smartpqi: Add controller fw version to console log

Print controller firmware version to OS message log during driver
initialization or after OFA.

Link: https://lore.kernel.org/r/165730601536.177165.17698744242908911822.stgit@brunhilda
Reviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Mike McGowen <mike.mcgowen@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Gilbert Wu <Gilbert.Wu@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: smartpqi: Shorten drive visibility after removal
Mike McGowen [Fri, 8 Jul 2022 18:46:50 +0000 (13:46 -0500)]
scsi: smartpqi: Shorten drive visibility after removal

Check the response code returned from the LUN reset task management
function and if it indicates the LUN is not valid, do not retry.

Reduce rescan worker delay to 5 seconds for the event handler only.

The removal of a drive from the OS could have been delayed up to 30 seconds
after being physically pulled.

The driver was retrying a LUN reset 3 times even though the return code
indiciated the LUN was no longer valid. There was a 10 second delay between
each retry. Additionally, the rescan worker was scheduled to run 10 seconds
after the driver received the event.

Link: https://lore.kernel.org/r/165730601025.177165.9416869335174437006.stgit@brunhilda
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Kevin Barnett <kevin.barnett@microchip.com>
Signed-off-by: Mike McGowen <Mike.McGowen@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: fnic: Refactor code in fnic probe to initialize SCSI layer
Karan Tilak Kumar [Thu, 7 Jul 2022 20:51:55 +0000 (13:51 -0700)]
scsi: fnic: Refactor code in fnic probe to initialize SCSI layer

Refactor code from fnic probe into a different function so that
scsi layer initialization code is grouped together.
Also, add log messages for better debugging.

Link: https://lore.kernel.org/r/20220707205155.692688-1-kartilak@cisco.com
Co-developed-by: Gian Carlo Boffa <gcboffa@cisco.com>
Signed-off-by: Gian Carlo Boffa <gcboffa@cisco.com>
Co-developed-by: Arulprabhu Ponnusamy <arulponn@cisco.com>
Signed-off-by: Arulprabhu Ponnusamy <arulponn@cisco.com>
Signed-off-by: Karan Tilak Kumar <kartilak@cisco.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: gvp11.c: Fix DMA mask calculation error
Michael Schmitz [Wed, 13 Jul 2022 07:49:13 +0000 (19:49 +1200)]
scsi: gvp11.c: Fix DMA mask calculation error

DMA masks given in the Zorro ID table don't contain the 2 byte alignment
quirk seen in the GVP11_XFER_MASK macro from gvp11.h so no need to account
for that.

DMA masks passed to dma_set_mask_and_coherent() must be 64 bit, add the
missing cast in the TO_DMA_MASK macro used to convert driver DMA masks to
DMA API masks.

Link: https://lore.kernel.org/r/6d1d88ee-1cf6-c735-1e6d-bafd2096e322@gmail.com
Link: https://lore.kernel.org/r/20220713074913.7873-1-schmitzmic@gmail.com
Fixes: 158da6bcae7a ("scsi: gvp11: Convert m68k WD33C93 drivers to DMA API")
CC: linux-scsi@vger.kernel.org
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: core: Remove UIC_HIBERN8_ENTER_RETRIES
Keoseong Park [Fri, 8 Jul 2022 05:20:06 +0000 (14:20 +0900)]
scsi: ufs: core: Remove UIC_HIBERN8_ENTER_RETRIES

Commit 4db7a2360597 ("scsi: ufs: Fix concurrency of error handler and other
error recovery paths") removed all callers of UIC_HIBERN8_ENTER_RETRIES.
Hence also remove the macro itself.

Link: https://lore.kernel.org/r/20220708052006epcms2p2d1129dbf95fd77f46906200ccb0a9ccd@epcms2p2
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Keoseong Park <keosung.park@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-exynos: Change ufs phy control sequence
Chanho Park [Wed, 6 Jul 2022 02:02:55 +0000 (11:02 +0900)]
scsi: ufs: ufs-exynos: Change ufs phy control sequence

Since commit 1599069a62c6 ("phy: core: Warn when phy_power_on is called
before phy_init"), the following warning has been reported:

phy_power_on was called before phy_init

To address this, we need to remove phy_power_on from exynos_ufs_phy_init()
and move it after phy_init. phy_power_off and phy_exit are also necessary
in exynos_ufs_remove().

Link: https://lore.kernel.org/r/20220706020255.151177-4-chanho61.park@samsung.com
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Chanho Park <chanho61.park@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: zfcp: Drop redundant "the" in the comments
Jiang Jian [Wed, 6 Jul 2022 15:59:40 +0000 (17:59 +0200)]
scsi: zfcp: Drop redundant "the" in the comments

There is an unexpected word "the" in the comments that needs to be dropped.

file: ./drivers/s390/scsi/zfcp_diag.h
line: 5
* Definitions for handling diagnostics in the the zfcp device driver.
changed to
* Definitions for handling diagnostics in the zfcp device driver.

Link: https://lore.kernel.org/r/20220621114207.106405-1-jiangjian@cdjrlc.com
Link: https://lore.kernel.org/r/4c02fefa19ab46f0f163990bde3ce10bd9c7caf1.1657122360.git.bblock@linux.ibm.com
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: zfcp: Declare zfcp_sdev_attrs as static
Julian Wiedmann [Wed, 6 Jul 2022 15:59:39 +0000 (17:59 +0200)]
scsi: zfcp: Declare zfcp_sdev_attrs as static

These attributes are now only accessed through the
zfcp_sysfs_sdev_attr_group.

Link: https://lore.kernel.org/r/0791b9149ebfa39e6b8eab093113cd2527dbf3d3.1657122360.git.bblock@linux.ibm.com
Fixes: d8d7cf3f7d07 ("scsi: zfcp: Switch to attribute groups")
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: Skip last hci reset to get valid register values
Junwoo Lee [Tue, 5 Jul 2022 08:35:38 +0000 (17:35 +0900)]
scsi: ufs: Skip last hci reset to get valid register values

Once the host link startup fails 3 times, all host registers are reset to
default values except in ufshcd_hba_enable().

The ufs host controller is disabled and enabled in ufshcd_hba_enable().
Consequently we need to skip last hci reset to get valid host register
values.

e.g.
[    1.898026] [2:  kworker/u16:2:  211] ufs: link startup failed 1
[    1.898133] [2:  kworker/u16:2:  211] host_regs: 000000001383ff1f 00000000 00000300 00000000
[    1.898141] [2:  kworker/u16:2:  211] host_regs: 0000001000000106 000001ce 00000000 00000000
[    1.898148] [2:  kworker/u16:2:  211] host_regs: 0000002000000000 00000470 00000000 00000000
[    1.898155] [2:  kworker/u16:2:  211] host_regs: 0000003000000008 00000003 00000000 00000000
[    1.898163] [2:  kworker/u16:2:  211] host_regs: 0000004000000000 00000000 00000000 00000000
[    1.898171] [2:  kworker/u16:2:  211] host_regs: 0000005000000000 00000000 00000000 00000000
[    1.898177] [2:  kworker/u16:2:  211] host_regs: 0000006000000000 00000000 00000000 00000000
[    1.898186] [2:  kworker/u16:2:  211] host_regs: 0000007000000000 00000000 00000000 00000000
[    1.898194] [2:  kworker/u16:2:  211] host_regs: 0000008000000000 00000000 00000000 00000000
[    1.898201] [2:  kworker/u16:2:  211] host_regs: 0000009000000000 00000000 00000000 00000000

Link: https://lore.kernel.org/r/20220705083538.15143-1-sh043.lee@samsung.com
Reviewed-by: Avri Altman <avri.altman@wdc.com>
Signed-off-by: Junwoo Lee <junwoo80.lee@samsung.com>
Signed-off-by: Seunghui Lee <sh043.lee@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-pci: Enable WriteBooster capability on ADL
Daniil Lunev [Tue, 5 Jul 2022 06:53:26 +0000 (16:53 +1000)]
scsi: ufs: ufs-pci: Enable WriteBooster capability on ADL

Sets the WriteBooster capability flag when ADL's UFS controller is used.

Link: https://lore.kernel.org/r/20220705165316.v2.1.Ib5ebec952d9a59f5c69c89b694777f517d22466d@changeid
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Daniil Lunev <dlunev@chromium.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Copyright updates for 14.2.0.5 patches
James Smart [Fri, 1 Jul 2022 21:14:25 +0000 (14:14 -0700)]
scsi: lpfc: Copyright updates for 14.2.0.5 patches

Update copyrights to 2022 for files modified in the 14.2.0.5 patch set.

Link: https://lore.kernel.org/r/20220701211425.2708-13-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Update lpfc version to 14.2.0.5
James Smart [Fri, 1 Jul 2022 21:14:24 +0000 (14:14 -0700)]
scsi: lpfc: Update lpfc version to 14.2.0.5

Update lpfc version to 14.2.0.5

Link: https://lore.kernel.org/r/20220701211425.2708-12-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Remove Menlo/Hornet related code
James Smart [Fri, 1 Jul 2022 21:14:23 +0000 (14:14 -0700)]
scsi: lpfc: Remove Menlo/Hornet related code

The Menlo/Hornet adapter was never released to the field. As such, driver
code specific to the adapter is unnecessary and should be removed.

Link: https://lore.kernel.org/r/20220701211425.2708-11-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Refactor lpfc_nvmet_prep_abort_wqe() into lpfc_sli_prep_abort_xri()
James Smart [Fri, 1 Jul 2022 21:14:22 +0000 (14:14 -0700)]
scsi: lpfc: Refactor lpfc_nvmet_prep_abort_wqe() into lpfc_sli_prep_abort_xri()

lpfc_nvmet_prep_abort_wqe() has a lot of common code with
lpfc_sli_prep_abort_xri().

Delete lpfc_nvmet_prep_abort_wqe() as the wqe can be filled out using the
generic lpfc_sli_prep_abort_xri routine(). Add the wqec option to
lpfc_sli_prep_abort_xri() for lpfc_nvmet_prep_abort_wqe().

Link: https://lore.kernel.org/r/20220701211425.2708-10-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Revert RSCN_MEMENTO workaround for misbehaved configuration
James Smart [Fri, 1 Jul 2022 21:14:21 +0000 (14:14 -0700)]
scsi: lpfc: Revert RSCN_MEMENTO workaround for misbehaved configuration

The RSCN_MEMENTO logic was to workaround a target that does not register
both FCP and NVMe FC4 types at the same time.  This caused the
configuration to not produce a second RSCN for the NVMe FC4 type
registration in a timely manner.  The intention of the RSCN_MEMENTO flag
was to always signal to try NVMe PRLI.

However, there are other FCP-only target arrays in correctly behaved
configurations that reject the NVMe PRLI followed by a LOGO leading to
never rediscovering the target after an issue_lip (as LOGO causes a repeat
of PLOGI/PRLIs).

Revert the RSCN_MEMENTO patch as it is causing correctly behaved configs to
fail while it exists only to succeed on a misbehaved config.

Link: https://lore.kernel.org/r/20220701211425.2708-9-jsmart2021@gmail.com
Fixes: 1045592fc968 ("scsi: lpfc: Introduce FC_RSCN_MEMENTO flag for tracking post RSCN completion")
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Fix lost NVMe paths during LIF bounce stress test
James Smart [Fri, 1 Jul 2022 21:14:20 +0000 (14:14 -0700)]
scsi: lpfc: Fix lost NVMe paths during LIF bounce stress test

During a target link bounce test, the driver sees a mismatch between the
NPortId and the WWPN on the node structures (ndlps) involved.  When this
occurs, the driver "swaps" the ndlp and new_ndlp node parameters to restore
WWPN/DID uniqueness in the fc_nodes list per vport.  However, the driver
neglected to swap the nlp_fc4_type in the ndlp passed to
lpfc_plogi_confirm_nport causing a failure to recover the NVMe PLOGI/PRLI
and ultimately the NVMe paths.

Correct confirm_nport to preserve the fc4 types from the new-ndlp when the
data is moved over ot the ndlp structure.

Link: https://lore.kernel.org/r/20220701211425.2708-8-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Fix attempted FA-PWWN usage after feature disable
James Smart [Fri, 1 Jul 2022 21:14:19 +0000 (14:14 -0700)]
scsi: lpfc: Fix attempted FA-PWWN usage after feature disable

Disabling FA-PWWN should be effective after port reset, but in some cases
it was found to be impossible to clear FA-PWWN usage without a driver
reload.

Clean up FA-PWWN flag management to make enable and disable of the feature
more robust.

Link: https://lore.kernel.org/r/20220701211425.2708-7-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Fix possible memory leak when failing to issue CMF WQE
James Smart [Fri, 1 Jul 2022 21:14:18 +0000 (14:14 -0700)]
scsi: lpfc: Fix possible memory leak when failing to issue CMF WQE

There is no corresponding free routine if lpfc_sli4_issue_wqe fails to
issue the CMF WQE in lpfc_issue_cmf_sync_wqe.

If ret_val is non-zero, then free the iocbq request structure.

Link: https://lore.kernel.org/r/20220701211425.2708-6-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Remove extra atomic_inc on cmd_pending in queuecommand after VMID
James Smart [Fri, 1 Jul 2022 21:14:17 +0000 (14:14 -0700)]
scsi: lpfc: Remove extra atomic_inc on cmd_pending in queuecommand after VMID

VMID introduced an extra increment of cmd_pending, causing double-counting
of the I/O. The normal increment ios performed in lpfc_get_scsi_buf.

Link: https://lore.kernel.org/r/20220701211425.2708-5-jsmart2021@gmail.com
Fixes: 33c79741deaf ("scsi: lpfc: vmid: Introduce VMID in I/O path")
Cc: <stable@vger.kernel.org> # v5.14+
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Set PU field when providing D_ID in XMIT_ELS_RSP64_CX iocb
James Smart [Fri, 1 Jul 2022 21:14:16 +0000 (14:14 -0700)]
scsi: lpfc: Set PU field when providing D_ID in XMIT_ELS_RSP64_CX iocb

When providing a D_ID in XMIT_ELS_RSP64_CX iocb the PU field should
be set to 3 to describe the parameter being passed to firmware.

Link: https://lore.kernel.org/r/20220701211425.2708-4-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Prevent buffer overflow crashes in debugfs with malformed user input
James Smart [Fri, 1 Jul 2022 21:14:15 +0000 (14:14 -0700)]
scsi: lpfc: Prevent buffer overflow crashes in debugfs with malformed user input

Malformed user input to debugfs results in buffer overflow crashes.  Adapt
input string lengths to fit within internal buffers, leaving space for NULL
terminators.

Link: https://lore.kernel.org/r/20220701211425.2708-3-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: lpfc: Fix uninitialized cqe field in lpfc_nvme_cancel_iocb()
James Smart [Fri, 1 Jul 2022 21:14:14 +0000 (14:14 -0700)]
scsi: lpfc: Fix uninitialized cqe field in lpfc_nvme_cancel_iocb()

In lpfc_nvme_cancel_iocb(), a cqe is created locally from stack storage.
The code didn't initialize the total_data_placed word, inheriting stack
content.

Initialize the total_data_placed word.

Link: https://lore.kernel.org/r/20220701211425.2708-2-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoMerge branch '5.19/scsi-fixes' into 5.20/scsi-staging
Martin K. Petersen [Thu, 7 Jul 2022 21:20:43 +0000 (17:20 -0400)]
Merge branch '5.19/scsi-fixes' into 5.20/scsi-staging

Bring in fixes to resolve a merge conflict in the lpfc driver update.

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: sd: Rework asynchronous resume support
Bart Van Assche [Thu, 30 Jun 2022 19:57:03 +0000 (12:57 -0700)]
scsi: sd: Rework asynchronous resume support

For some technologies, e.g. an ATA bus, resuming can take multiple
seconds. Waiting for resume to finish can cause a very noticeable delay.
Hence this commit that restores the behavior from before "scsi: core: pm:
Rely on the device driver core for async power management" for most SCSI
devices.

This commit introduces a behavior change: if the START command fails, do
not consider this as a SCSI disk resume failure.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=215880
Link: https://lore.kernel.org/r/20220630195703.10155-3-bvanassche@acm.org
Fixes: a19a93e4c6a9 ("scsi: core: pm: Rely on the device driver core for async power management")
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: John Garry <john.garry@huawei.com>
Cc: ericspero@icloud.com
Cc: jason600.groome@gmail.com
Tested-by: jason600.groome@gmail.com
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: core: Move the definition of SCSI_QUEUE_DELAY
Bart Van Assche [Thu, 30 Jun 2022 19:57:02 +0000 (12:57 -0700)]
scsi: core: Move the definition of SCSI_QUEUE_DELAY

Move the definition of SCSI_QUEUE_DELAY to just above the function that
uses it.

Link: https://lore.kernel.org/r/20220630195703.10155-2-bvanassche@acm.org
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: John Garry <john.garry@huawei.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: megaraid_sas: Clean up some inconsistent indenting
Jiapeng Chong [Thu, 30 Jun 2022 07:41:52 +0000 (15:41 +0800)]
scsi: megaraid_sas: Clean up some inconsistent indenting

This was found by coccicheck:

drivers/scsi/megaraid/megaraid_sas_base.c:3950 process_fw_state_change_wq() warn: inconsistent indenting.

Link: https://lore.kernel.org/r/20220630074152.29171-1-jiapeng.chong@linux.alibaba.com
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: gvp11: Convert m68k WD33C93 drivers to DMA API
Michael Schmitz [Thu, 30 Jun 2022 03:33:02 +0000 (15:33 +1200)]
scsi: gvp11: Convert m68k WD33C93 drivers to DMA API

Use dma_map_single() for gvp11 driver (leave bounce buffer logic
unchanged).

Use dma_set_mask_and_coherent() to avoid explicit cache flushes.

Compile-tested only.

CC: linux-scsi@vger.kernel.org
Link: https://lore.kernel.org/r/6d1d88ee-1cf6-c735-1e6d-bafd2096e322@gmail.com
Link: https://lore.kernel.org/r/20220630033302.3183-4-schmitzmic@gmail.com
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
--

Changes from v1:

Arnd Bergmann:
- reorder bounce buffer copy and dma mapping

2 years agoscsi: a2091: Convert m68k WD33C93 drivers to DMA API
Michael Schmitz [Thu, 30 Jun 2022 03:33:01 +0000 (15:33 +1200)]
scsi: a2091: Convert m68k WD33C93 drivers to DMA API

Use dma_map_single() for a2091 driver (leave bounce buffer logic
unchanged).

Use dma_set_mask_and_coherent() to avoid explicit cache flushes.

Compile-tested only.

CC: linux-scsi@vger.kernel.org
Link: https://lore.kernel.org/r/6d1d88ee-1cf6-c735-1e6d-bafd2096e322@gmail.com
Link: https://lore.kernel.org/r/20220630033302.3183-3-schmitzmic@gmail.com
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
--

Changes from v1:

Arnd Bergmann:
- reorder mapping and bounce buffer copy

2 years agoscsi: a3000: Convert m68k WD33C93 drivers to DMA API
Michael Schmitz [Thu, 30 Jun 2022 03:33:00 +0000 (15:33 +1200)]
scsi: a3000: Convert m68k WD33C93 drivers to DMA API

Use dma_map_single() for a3000 driver (leave bounce buffer logic
unchanged).

Use dma_set_mask_and_coherent() to avoid explicit cache flushes.

Compile-tested only.

CC: linux-scsi@vger.kernel.org
Link: https://lore.kernel.org/r/6d1d88ee-1cf6-c735-1e6d-bafd2096e322@gmail.com
Link: https://lore.kernel.org/r/20220630033302.3183-2-schmitzmic@gmail.com
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
--

Changes from v1:

- restore bounce buffer allocation (dropped in v1)

Arnd Bergmann:
- reorder dma mapping and bounce buffer copy

2 years agoscsi: core: Shorten long warning messages
Li Zhijian [Thu, 30 Jun 2022 02:38:11 +0000 (02:38 +0000)]
scsi: core: Shorten long warning messages

sdev_printk() will only accept messages up to 128 bytes.

Shorten strings exceeding 128 bytes avoid printing an incomplete sentence
like:

  [  475.156955] sd 9:0:0:0: Warning! Received an indication that the LUN assignments on this target have changed. The Linux SCSI layer does not automatical

Link: https://lore.kernel.org/r/20220630024516.1571209-1-lizhijian@fujitsu.com
Suggested-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Li Zhijian <lizhijian@fujitsu.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Detect UNMAP support post configuration
Mike Christie [Tue, 28 Jun 2022 20:02:30 +0000 (15:02 -0500)]
scsi: target: Detect UNMAP support post configuration

On our backend we can do something similar to LIO where we can enable and
disable UNMAP support on the fly. In the SCSI/block layer we can detect
this by just doing a rescan. However, LIO cannot detect this change because
we only check during the initial configuration. This patch allows UNMAP
detection to also happen when the user tries to turn it on.

Link: https://lore.kernel.org/r/20220628200230.15052-6-michael.christie@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Add file configure_unmap callout
Mike Christie [Tue, 28 Jun 2022 20:02:29 +0000 (15:02 -0500)]
scsi: target: Add file configure_unmap callout

Move file's UNMAP setup code to a configure_unmap callout.

Link: https://lore.kernel.org/r/20220628200230.15052-5-michael.christie@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Add iblock configure_unmap callout
Mike Christie [Tue, 28 Jun 2022 20:02:28 +0000 (15:02 -0500)]
scsi: target: Add iblock configure_unmap callout

Move iblock's UNMAP setup code to a configure_unmap callout.

Link: https://lore.kernel.org/r/20220628200230.15052-4-michael.christie@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Add callout to configure UNMAP settings
Mike Christie [Tue, 28 Jun 2022 20:02:27 +0000 (15:02 -0500)]
scsi: target: Add callout to configure UNMAP settings

Add a callout to configure a backend's UNMAP settings. This will be used to
allow userspace to configure UNMAP after the initial device setup, similar
to how we can set up the other attributes post device configuration.

Link: https://lore.kernel.org/r/20220628200230.15052-3-michael.christie@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Remove incorrect zero blocks WRITE_SAME check
Mike Christie [Tue, 28 Jun 2022 20:02:26 +0000 (15:02 -0500)]
scsi: target: Remove incorrect zero blocks WRITE_SAME check

We use WSNZ=1 so if we get a WRITE_SAME with zero logical blocks we are
supposed to fail it. We do this check and failure in target_core_sbc.c
before calling into the backend, so we can remove the incorrect check in
target_core_file.

Link: https://lore.kernel.org/r/20220628200230.15052-2-michael.christie@oracle.com
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: mpi3mr: Increase cmd_per_lun to 128
Sreekanth Reddy [Tue, 28 Jun 2022 07:48:48 +0000 (13:18 +0530)]
scsi: mpi3mr: Increase cmd_per_lun to 128

Increase cmd_per_lun to 128.

Link: https://lore.kernel.org/r/20220628074848.5036-3-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: mpi3mr: Enable shared host tagset
Sreekanth Reddy [Tue, 28 Jun 2022 07:48:47 +0000 (13:18 +0530)]
scsi: mpi3mr: Enable shared host tagset

Enable shared host tagset to make sure that total outstanding I/O count can
not exceed controller's can_queue setting.

Link: https://lore.kernel.org/r/20220628074848.5036-2-sreekanth.reddy@broadcom.com
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: qla2xxx: Check correct variable in qla24xx_async_gffid()
Dan Carpenter [Wed, 22 Jun 2022 06:21:55 +0000 (09:21 +0300)]
scsi: qla2xxx: Check correct variable in qla24xx_async_gffid()

There is a copy and paste bug here.  It should check ".rsp" instead of
".req".  The error message is copy and pasted as well so update that too.

Link: https://lore.kernel.org/r/YrK1A/t3L6HKnswO@kili
Fixes: 9c40c36e75ff ("scsi: qla2xxx: edif: Reduce Initiator-Initiator thrashing")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-qcom: Remove unneeded code
ChanWoo Lee [Mon, 27 Jun 2022 23:55:45 +0000 (08:55 +0900)]
scsi: ufs: ufs-qcom: Remove unneeded code

Checks information about tx_lanes, but is not used.

Since commit 1e1e465c6d23 ("scsi/ufs: qcom: Remove ufs_qcom_phy_*() calls
from host"), tx_lanes is deprecated.

As a result, ufs_qcom_link_startup_notify -> POST_CHANGE action does
nothing.  If it is not going to be updated, it can be removed.

Link: https://lore.kernel.org/r/20220627235545.16943-1-cw9316.lee@samsung.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: ChanWoo Lee <cw9316.lee@samsung.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Fix invalid access to vccqx
Alice Chao [Thu, 23 Jun 2022 03:50:52 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Fix invalid access to vccqx

NULL pointer access issue was found for the regulator released
by ufs_mtk_vreg_fix_vccq(). Simply fix this issue by clearing
the released vreg pointer in ufs_hba struct.

Link: https://lore.kernel.org/r/20220623035052.18802-9-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Alice Chao <alice.chao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support performance boosting
Peter Wang [Thu, 23 Jun 2022 03:50:51 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Support performance boosting

Add pm-qos request to support performance boosting in MediaTek UFS
platforms.

At the same time, adjust the order of function calls to be symmetric during
the low-power control flow.

Link: https://lore.kernel.org/r/20220623035052.18802-8-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support host power control
Po-Wen Kao [Thu, 23 Jun 2022 03:50:50 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Support host power control

Add interfaces for controlling the host power to optimize the power
consumption in MediaTek UFS platforms.

Link: https://lore.kernel.org/r/20220623035052.18802-7-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Po-Wen Kao <powen.kao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Disable reset confirm feature by UniPro
Po-Wen Kao [Thu, 23 Jun 2022 03:50:49 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Disable reset confirm feature by UniPro

In MediaTek UFS platforms, UniPro will not return reset confirm if it is in
POWERDOWN state thus hang issue may happen while disabling UFSHCI. Simply
disable this feature before UniPro leaves POWERDOWN state.

Link: https://lore.kernel.org/r/20220623035052.18802-6-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Po-Wen Kao <powen.kao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Add stage information for ref-clk control
Peter Wang [Thu, 23 Jun 2022 03:50:48 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Add stage information for ref-clk control

Add "PRE_CHANGE" and "POST_CHANGE" information for ref-clk control to
precisely configure the low-power state of the parent of ref-clk.

Link: https://lore.kernel.org/r/20220623035052.18802-5-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Prevent host hang by setting CLK_CG early
Peter Wang [Thu, 23 Jun 2022 03:50:47 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Prevent host hang by setting CLK_CG early

Some UFSHCI hosts in MediaTek UFS platform need workaround to prevent host
hang issue by setting CLK_CG bit before host is enabled.

This operation shall have no side effect on those platforms which do not
support this bit.

Link: https://lore.kernel.org/r/20220623035052.18802-4-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Always add delays for VCC operations
Peter Wang [Thu, 23 Jun 2022 03:50:46 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Always add delays for VCC operations

MediaTek decides to always add delays before and after VCC is turned-off.

Link: https://lore.kernel.org/r/20220623035052.18802-3-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Fix build warnings
Stanley Chu [Thu, 23 Jun 2022 03:50:45 +0000 (11:50 +0800)]
scsi: ufs: ufs-mediatek: Fix build warnings

Fix build warnings:

1.
../drivers/ufs/host/ufs-mediatek.c:1375:5: error: no previous prototype for function 'ufs_mtk_system_suspend' [-Werror,-Wmissing-prototypes]
int ufs_mtk_system_suspend(struct device *dev)
    ^
../drivers/ufs/host/ufs-mediatek.c:1375:1: note: declare 'static' if the function is not intended to be used outside of this translation unit
int ufs_mtk_system_suspend(struct device *dev)
^
static

2.
../drivers/ufs/host/ufs-mediatek.c:702:50: error: format specifies type 'unsigned long' but the argument has type 'int' [-Werror,-Wformat]
                snprintf(vcc_name, MAX_VCC_NAME, "vcc-ufs%lu", ver);
                                                         ~~~   ^~~
                                                         %d

Link: https://lore.kernel.org/r/20220623035052.18802-2-stanley.chu@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: dpt_i2o: Remove obsolete driver
Arnd Bergmann [Fri, 24 Jun 2022 15:52:25 +0000 (17:52 +0200)]
scsi: dpt_i2o: Remove obsolete driver

The dpt_i2o driver was fixed to stop using virt_to_bus() in 2008, but it
still has a stale reference in an error handling code path that could never
work. I submitted a patch to fix this reference earlier, but Hannes
Reinecke suggested that removing the driver may be just as good here.

The i2o driver layer was removed in 2015 with commit 4a72a7af462d
("staging: remove i2o subsystem"), but the even older dpt_i2o scsi driver
stayed around.

The last non-cleanup patches I could find were from Miquel van Smoorenburg
and Mark Salyzyn back in 2008, they might know if there is any chance of
the hardware still being used anywhere.

Link: https://lore.kernel.org/linux-scsi/CAK8P3a1XfwkTOV7qOs1fTxf4vthNBRXKNu8A5V7TWnHT081NGA@mail.gmail.com/T/
Link: https://lore.kernel.org/r/20220624155226.2889613-3-arnd@kernel.org
Cc: Miquel van Smoorenburg <mikevs@xs4all.net>
Cc: Mark Salyzyn <salyzyn@android.com>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: BusLogic: Remove bus_to_virt()
Arnd Bergmann [Fri, 24 Jun 2022 15:52:24 +0000 (17:52 +0200)]
scsi: BusLogic: Remove bus_to_virt()

The BusLogic driver is the last remaining driver that relies on the
deprecated bus_to_virt() function, which in turn only works on a few
architectures, and is incompatible with both swiotlb and iommu support.

Before commit 391e2f25601e ("[SCSI] BusLogic: Port driver to 64-bit."), the
driver had a dependency on x86-32, presumably because of this
problem. However, the change introduced another bug that made it still
impossible to use the driver on any 64-bit machine.

This was in turn fixed in commit 56f396146af2 ("scsi: BusLogic: Fix 64-bit
system enumeration error for Buslogic"), 8 years later, which shows that
there are not a lot of users.

Maciej is still using the driver on 32-bit hardware, and Khalid mentioned
that the driver works with the device emulation used in VirtualBox and
VMware. Both of those only emulate it for Windows 2000 and older operating
systems that did not ship with the better LSI logic driver.

Do a minimum fix that searches through the list of descriptors to find one
that matches the bus address. This is clearly as inefficient as was
indicated in the code comment about the lack of a bus_to_virt()
replacement. A better fix would likely involve changing out the entire
descriptor allocation for a simpler one, but that would be much more
invasive.

Link: https://lore.kernel.org/r/20220624155226.2889613-2-arnd@kernel.org
Cc: Maciej W. Rozycki <macro@orcam.me.uk>
Cc: Matt Wang <wwentao@vmware.com>
Tested-by: Khalid Aziz <khalid@gonehiking.org>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Acked-by: Khalid Aziz <khalid@gonehiking.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: fcoe: Remove redundant assignment to variable 'wlen'
Colin Ian King [Thu, 23 Jun 2022 16:47:10 +0000 (17:47 +0100)]
scsi: fcoe: Remove redundant assignment to variable 'wlen'

Variable wlen is being assigned a value that is never read, it is being
re-assigned with a different value later on. The assignment is redundant
and can be removed.

Cleans up clang scan build warning:
drivers/scsi/fcoe/fcoe.c:1491:2: warning: Value stored to 'wlen'
is never read [deadcode.DeadStores]

Link: https://lore.kernel.org/r/20220623164710.76831-1-colin.i.king@gmail.com
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: hisi_sas: Limit max hw sectors for v3 HW
John Garry [Thu, 23 Jun 2022 12:41:59 +0000 (20:41 +0800)]
scsi: hisi_sas: Limit max hw sectors for v3 HW

If the controller is behind an IOMMU then the IOMMU IOVA caching range can
affect performance, as discussed in [0].

Limit the max HW sectors to not exceed this limit. We need to hardcode the
value until a proper DMA mapping API is available.

[0] https://lore.kernel.org/linux-iommu/20210129092120.1482-1-thunder.leizhen@huawei.com/

Link: https://lore.kernel.org/r/1655988119-223714-1-git-send-email-john.garry@huawei.com
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufshcd: Constify pointed data
Krzysztof Kozlowski [Thu, 23 Jun 2022 10:24:32 +0000 (12:24 +0200)]
scsi: ufs: ufshcd: Constify pointed data

For code safety, constify arrays and pointers to data which is not
modified.

Link: https://lore.kernel.org/r/20220623102432.108059-4-krzysztof.kozlowski@linaro.org
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufshcd-pltfrm: Constify pointed data
Krzysztof Kozlowski [Thu, 23 Jun 2022 10:24:31 +0000 (12:24 +0200)]
scsi: ufs: ufshcd-pltfrm: Constify pointed data

For code safety, constify pointers to data which is not modified.

Link: https://lore.kernel.org/r/20220623102432.108059-3-krzysztof.kozlowski@linaro.org
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: MAINTAINERS: Update cxgb3i and cxgb4i maintainer
Varun Prakash [Sun, 12 Jun 2022 12:13:40 +0000 (17:43 +0530)]
scsi: MAINTAINERS: Update cxgb3i and cxgb4i maintainer

Link: https://lore.kernel.org/r/20220612121340.6746-1-varun@chelsio.com
Signed-off-by: Varun Prakash <varun@chelsio.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: mpt3sas: Remove flush_scheduled_work() call
Tetsuo Handa [Thu, 9 Jun 2022 13:26:48 +0000 (22:26 +0900)]
scsi: mpt3sas: Remove flush_scheduled_work() call

It seems to me that mpt3sas driver is using dedicated workqueues and is not
calling schedule{,_delayed}_work{,_on}(). Then, there will be no work to
flush using flush_scheduled_work().

Link: https://lore.kernel.org/r/f3b97c7c-1094-4e46-20d8-4321716d6f3f@I-love.SAKURA.ne.jp
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: trace: Print driver_tag and scheduler_tag in SCSI trace
Changyuan Lyu [Tue, 21 Jun 2022 18:11:25 +0000 (18:11 +0000)]
scsi: trace: Print driver_tag and scheduler_tag in SCSI trace

Trace events like scsi_dispatch_cmd_start and scsi_dispatch_cmd_done are
useful for tracking a command throughout its lifetime. But for some ATA
passthrough commands, the information printed in current logs is not enough
to identify and match them. For example, if two threads send SMART cmd to
the same disk at the same time, their trace logs may look the same, which
makes it hard to match scsi_dispatch_cmd_done and scsi_dispatch_cmd_start.

Printing tags can help us solve the problem.  Further, if a command failed
for some reason and then is retried, its driver_tag will change. So
scheduler_tag is also included such that we can track the retries of a
command.

Link: https://lore.kernel.org/r/20220621181125.3211399-1-changyuanl@google.com
Reviewed-by: Vishakha Channapattan <vishakhavc@google.com>
Reviewed-by: Jolly Shah <jollys@google.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Changyuan Lyu <changyuanl@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: sym53c8xx_2: Remove redundant "with"
Jiang Jian [Tue, 21 Jun 2022 16:26:31 +0000 (00:26 +0800)]
scsi: sym53c8xx_2: Remove redundant "with"

Remove redundant "with" in comment

Link: https://lore.kernel.org/r/20220621162631.25353-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: snic: Remove redundant "on"
Jiang Jian [Tue, 21 Jun 2022 15:01:36 +0000 (23:01 +0800)]
scsi: snic: Remove redundant "on"

Remove redundant "on" in comment.

Link: https://lore.kernel.org/r/20220621150136.9264-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: core: Fix typos in error messages
Wolfram Sang [Tue, 21 Jun 2022 14:46:53 +0000 (16:46 +0200)]
scsi: ufs: core: Fix typos in error messages

Should be "retries" instead of "retires".

Link: https://lore.kernel.org/r/20220621144653.62630-1-wsa+renesas@sang-engineering.com
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: aacraid: Remove redundant "the"
Jiang Jian [Tue, 21 Jun 2022 14:23:46 +0000 (22:23 +0800)]
scsi: aacraid: Remove redundant "the"

Remove redundant "the" in comment.

Link: https://lore.kernel.org/r/20220621142346.6429-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: target: Remove duplicate "the"
Jiang Jian [Tue, 21 Jun 2022 14:20:12 +0000 (22:20 +0800)]
scsi: target: Remove duplicate "the"

Remove redundant "the" in comment".

Link: https://lore.kernel.org/r/20220621142012.4622-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: message: fusion: Drop redundant "the"
Jiang Jian [Tue, 21 Jun 2022 10:31:19 +0000 (18:31 +0800)]
scsi: message: fusion: Drop redundant "the"

Redundant "the" in the comments can be dropped.

Link: https://lore.kernel.org/r/20220621103119.91492-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: hisi_sas: Align comments
Jiang Jian [Tue, 21 Jun 2022 07:24:05 +0000 (15:24 +0800)]
scsi: hisi_sas: Align comments

Properly align comment lines in slot_index_alloc_quirk_v2_hw().

Link: https://lore.kernel.org/r/20220621072405.34394-1-jiangjian@cdjrlc.com
Signed-off-by: Jiang Jian <jiangjian@cdjrlc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: mpt3sas: Fix whitespace and spelling mistake
Zhang Jiaming [Fri, 17 Jun 2022 10:11:03 +0000 (18:11 +0800)]
scsi: mpt3sas: Fix whitespace and spelling mistake

There is a spelling mistake in _base_sas_ioc_info(). Change 'cant' to
'can't'.

Also fix up whitespace.

Link: https://lore.kernel.org/r/20220617101103.3162-1-jiaming@nfschina.com
Signed-off-by: Zhang Jiaming <jiaming@nfschina.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: mpt3sas: Fix typo in comment
Ren Yu [Fri, 17 Jun 2022 08:15:57 +0000 (16:15 +0800)]
scsi: mpt3sas: Fix typo in comment

Spelling mistake in comment: non-succesfull -> non-successful.

Link: https://lore.kernel.org/r/20220617081557.9009-1-renyu@nfschina.com
Signed-off-by: Ren Yu <renyu@nfschina.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: core: bsg: Remove usage of the deprecated ida_simple_xxx() API
Bo Liu [Fri, 17 Jun 2022 02:04:30 +0000 (22:04 -0400)]
scsi: core: bsg: Remove usage of the deprecated ida_simple_xxx() API

Use ida_alloc_xxx()/ida_free() instead of
ida_simple_get()/ida_simple_remove().

The latter is deprecated and more verbose.

Link: https://lore.kernel.org/r/20220617020430.2300-1-liubo03@inspur.com
Signed-off-by: Bo Liu <liubo03@inspur.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: libiscsi: Improve conn_send_pdu API
Mike Christie [Thu, 16 Jun 2022 22:45:57 +0000 (17:45 -0500)]
scsi: libiscsi: Improve conn_send_pdu API

The conn_send_pdu API is evil in that it returns a pointer to an
iscsi_task, but that task might have been freed already so you can't touch
it. This patch splits the task allocation and transmission, so functions
like iscsi_send_nopout() can access the task before its sent and do
whatever bookkeeping is needed before it is sent.

Link: https://lore.kernel.org/r/20220616224557.115234-10-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Try to avoid taking back_lock in xmit path
Mike Christie [Thu, 16 Jun 2022 22:45:56 +0000 (17:45 -0500)]
scsi: iscsi: Try to avoid taking back_lock in xmit path

We need the back lock when freeing a task, so we hold it when calling
__iscsi_put_task() from the completion path to make it easier and to avoid
having to retake it in that path. For iscsi_put_task() we just grabbed it
while also doing the decrement on the refcount but it's only really needed
if the refcount is zero and we free the task. This modifies
iscsi_put_task() to just take the lock when needed then has the xmit path
use it. Normally we will then not take the back lock from the xmit path. It
will only be rare cases where the network is so fast that we get a response
right after we send the header/data.

Link: https://lore.kernel.org/r/20220616224557.115234-9-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Remove iscsi_get_task back_lock requirement
Mike Christie [Thu, 16 Jun 2022 22:45:55 +0000 (17:45 -0500)]
scsi: iscsi: Remove iscsi_get_task back_lock requirement

We currently require that the back_lock is held when calling the functions
that manipulate the iscsi_task refcount. The only reason for this is to
handle races where we are handling SCSI-ml EH callbacks and the cmd is
completing at the same time the normal completion path is running, and we
can't return from the EH callback until the driver has stopped accessing
the cmd. Holding the back_lock while also accessing the task->state made it
simple to check that a cmd is completing and also get/put a refcount at the
same time, and at the time we were not as concerned about performance.

The problem is that we don't want to take the back_lock from the xmit path
for normal I/O since it causes contention with the completion path if the
user has chosen to try and split those paths on different CPUs (in this
case abusing the CPUs and ignoring caching improves perf for some uses).

Begins to remove the back_lock requirement for iscsi_get/put_task by
removing the requirement for the get path. Instead of always holding the
back_lock we detect if something has done the last put and is about to call
iscsi_free_task(). A subsequent commit will then allow iSCSI code to do the
last put on a task and only grab the back_lock if the refcount is now zero
and it's going to call iscsi_free_task().

Link: https://lore.kernel.org/r/20220616224557.115234-8-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Remove unneeded task state check
Mike Christie [Thu, 16 Jun 2022 22:45:54 +0000 (17:45 -0500)]
scsi: iscsi: Remove unneeded task state check

Commit 5923d64b7ab6 ("scsi: libiscsi: Drop taskqueuelock") added an extra
task->state because for commit 6f8830f5bbab ("scsi: libiscsi: add lock
around task lists to fix list corruption regression") we didn't know why we
ended up with cmds on the list and thought it might have been a bad target
sending a response while we were still sending the cmd. We were never able
to get a target to send us a response early, because it turns out the bug
was just a race in libiscsi/libiscsi_tcp where:

 1. iscsi_tcp_r2t_rsp() queues a r2t to tcp_task->r2tqueue.

 2. iscsi_tcp_task_xmit() runs iscsi_tcp_get_curr_r2t() and sees we have a
    r2t. It dequeues it and iscsi_tcp_task_xmit() starts to process it.

 3. iscsi_tcp_r2t_rsp() runs iscsi_requeue_task() and puts the task on the
    requeue list.

 4. iscsi_tcp_task_xmit() sends the data for r2t. This is the final chunk
    of data, so the cmd is done.

 5. target sends the response.

 6. On a different CPU from #3, iscsi_complete_task() processes the
    response.  Since there was no common lock for the list, the lists/tasks
    pointers are not fully in sync, so could end up with list corruption.

Since it was just a race on our side, remove the extra check and fix up the
comments.

Link: https://lore.kernel.org/r/20220616224557.115234-7-michael.christie@oracle.com
Reviewed-by: Wu Bo <wubo40@huawei.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi_tcp: Drop target_alloc use
Mike Christie [Thu, 16 Jun 2022 22:45:53 +0000 (17:45 -0500)]
scsi: iscsi_tcp: Drop target_alloc use

For software iSCSI, we do a session per host so there is no need to set the
target's can_queue since it's the same as the host one. Setting it just
results in extra atomic checks in the main I/O path.

Link: https://lore.kernel.org/r/20220616224557.115234-6-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Wu Bo <wubo40@huawei.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi_tcp: Tell net when there's more data
Mike Christie [Thu, 16 Jun 2022 22:45:52 +0000 (17:45 -0500)]
scsi: iscsi_tcp: Tell net when there's more data

If we have more data, set the MSG_SENDPAGE_NOTLAST in case we go down the
sendpage path.

Link: https://lore.kernel.org/r/20220616224557.115234-5-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Run recv path from workqueue
Mike Christie [Thu, 16 Jun 2022 22:45:51 +0000 (17:45 -0500)]
scsi: iscsi: Run recv path from workqueue

We don't always want to run the recv path from the network softirq because
when we have to have multiple sessions sharing the same CPUs, some sessions
can eat up the NAPI softirq budget and affect other sessions or users.

Allow us to queue the recv handling to the iscsi workqueue so we can have
the scheduler/wq code try to balance the work and CPU use across all
sessions' worker threads.

Note: It wasn't the original intent of the change but a nice side effect is
that for some workloads/configs we get a nice performance boost. For a
simple read heavy test:

  fio --direct=1 --filename=/dev/dm-0  --rw=randread --bs=256K
    --ioengine=libaio --iodepth=128 --numjobs=4

where the iscsi threads, fio jobs, and rps_cpus share CPUs we see a 32%
throughput boost. We also see increases for small I/O IOPs tests but it's
not as high.

Link: https://lore.kernel.org/r/20220616224557.115234-4-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Add recv workqueue helpers
Mike Christie [Thu, 16 Jun 2022 22:45:50 +0000 (17:45 -0500)]
scsi: iscsi: Add recv workqueue helpers

Add helpers to allow the drivers to run their recv paths from libiscsi's
workqueue.

Link: https://lore.kernel.org/r/20220616224557.115234-3-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Rename iscsi_conn_queue_work()
Mike Christie [Thu, 16 Jun 2022 22:45:49 +0000 (17:45 -0500)]
scsi: iscsi: Rename iscsi_conn_queue_work()

Rename iscsi_conn_queue_work() to iscsi_conn_queue_xmit() to reflect that
it handles queueing of xmits only.

Link: https://lore.kernel.org/r/20220616224557.115234-2-michael.christie@oracle.com
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Wu Bo <wubo40@huawei.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Fix session removal on shutdown
Mike Christie [Thu, 16 Jun 2022 22:27:38 +0000 (17:27 -0500)]
scsi: iscsi: Fix session removal on shutdown

When the system is shutting down, iscsid is not running so we will not get
a response to the ISCSI_ERR_INVALID_HOST error event. The system shutdown
will then hang waiting on userspace to remove the session.

This has libiscsi force the destruction of the session from the kernel when
iscsi_host_remove() is called from a driver's shutdown callout.

This fixes a regression added in qedi boot with commit d1f2ce77638d ("scsi:
qedi: Fix host removal with running sessions") which made qedi use the
common session removal function that waits on userspace instead of rolling
its own kernel based removal.

Link: https://lore.kernel.org/r/20220616222738.5722-7-michael.christie@oracle.com
Fixes: d1f2ce77638d ("scsi: qedi: Fix host removal with running sessions")
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: qedi: Use QEDI_MODE_NORMAL for error handling
Mike Christie [Thu, 16 Jun 2022 22:27:37 +0000 (17:27 -0500)]
scsi: qedi: Use QEDI_MODE_NORMAL for error handling

When handling errors that lead to host removal use QEDI_MODE_NORMAL. There
is currently no difference in behavior between NORMAL and SHUTDOWN, but in
a subsequent commit we will want to know when we are called from the
pci_driver shutdown callout vs remove/err_handler so we know when userspace
is up.

Link: https://lore.kernel.org/r/20220616222738.5722-6-michael.christie@oracle.com
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Add helper to remove a session from the kernel
Mike Christie [Thu, 16 Jun 2022 22:27:36 +0000 (17:27 -0500)]
scsi: iscsi: Add helper to remove a session from the kernel

During qedi shutdown we need to stop the iSCSI layer from sending new nops
as pings and from responding to target ones and make sure there is no
running connection cleanups. Commit d1f2ce77638d ("scsi: qedi: Fix host
removal with running sessions") converted the driver to use the libicsi
helper to drive session removal, so the above issues could be handled. The
problem is that during system shutdown iscsid will not be running so when
we try to remove the root session we will hang waiting for userspace to
reply.

Add a helper that will drive the destruction of sessions like these during
system shutdown.

Link: https://lore.kernel.org/r/20220616222738.5722-5-michael.christie@oracle.com
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Clean up bound endpoints during shutdown
Mike Christie [Thu, 16 Jun 2022 22:27:35 +0000 (17:27 -0500)]
scsi: iscsi: Clean up bound endpoints during shutdown

In the next patch we allow drivers to drive session removal during
shutdown. In this case iscsid will not be running, so we need to detect
bound endpoints and disconnect them. This moves the bound ep check so we
now always check.

Link: https://lore.kernel.org/r/20220616222738.5722-4-michael.christie@oracle.com
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Allow iscsi_if_stop_conn() to be called from kernel
Mike Christie [Thu, 16 Jun 2022 22:27:34 +0000 (17:27 -0500)]
scsi: iscsi: Allow iscsi_if_stop_conn() to be called from kernel

iscsi_if_stop_conn() is only called from the userspace interface but in a
subsequent commit we will want to call it from the kernel interface to
allow drivers like qedi to remove sessions from inside the kernel during
shutdown. This removes the iscsi_uevent code from iscsi_if_stop_conn() so we
can call it in a new helper.

Link: https://lore.kernel.org/r/20220616222738.5722-3-michael.christie@oracle.com
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Fix HW conn removal use after free
Mike Christie [Thu, 16 Jun 2022 22:27:33 +0000 (17:27 -0500)]
scsi: iscsi: Fix HW conn removal use after free

If qla4xxx doesn't remove the connection before the session, the iSCSI
class tries to remove the connection for it. We were doing a
iscsi_put_conn() in the iter function which is not needed and will result
in a use after free because iscsi_remove_conn() will free the connection.

Link: https://lore.kernel.org/r/20220616222738.5722-2-michael.christie@oracle.com
Tested-by: Nilesh Javali <njavali@marvell.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Fix build error and type mismatch
Ren Zhijie [Sun, 19 Jun 2022 11:54:32 +0000 (19:54 +0800)]
scsi: ufs: ufs-mediatek: Fix build error and type mismatch

If CONFIG_PM_SLEEP is not set.

make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu-, will fail:

drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_vreg_fix_vcc’:
drivers/ufs/host/ufs-mediatek.c:688:46: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 4 has type ‘long unsigned int’ [-Wformat=]
    snprintf(vcc_name, MAX_VCC_NAME, "vcc-opt%u", res.a1);
                                             ~^   ~~~~~~
                                             %lu
drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_system_suspend’:
drivers/ufs/host/ufs-mediatek.c:1371:8: error: implicit declaration of function ‘ufshcd_system_suspend’; did you mean ‘ufs_mtk_system_suspend’? [-Werror=implicit-function-declaration]
  ret = ufshcd_system_suspend(dev);
        ^~~~~~~~~~~~~~~~~~~~~
        ufs_mtk_system_suspend
drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_system_resume’:
drivers/ufs/host/ufs-mediatek.c:1386:9: error: implicit declaration of function ‘ufshcd_system_resume’; did you mean ‘ufs_mtk_system_resume’? [-Werror=implicit-function-declaration]
  return ufshcd_system_resume(dev);
         ^~~~~~~~~~~~~~~~~~~~
         ufs_mtk_system_resume
cc1: some warnings being treated as errors

The declaration of func "ufshcd_system_suspend()" depends on
CONFIG_PM_SLEEP, so the function wrapper ufs_mtk_system_suspend() should
wrapped by CONFIG_PM_SLEEP too.

Link: https://lore.kernel.org/r/20220619115432.205504-1-renzhijie2@huawei.com
Fixes: 3fd23b8dfb54 ("scsi: ufs: ufs-mediatek: Fix the timing of configuring device regulators")
Reported-by: Hulk Robot <hulkci@huawei.com>
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Ren Zhijie <renzhijie2@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: iscsi: Make iscsi_unregister_transport() return void
Max Gurtovoy [Thu, 16 Jun 2022 08:02:10 +0000 (11:02 +0300)]
scsi: iscsi: Make iscsi_unregister_transport() return void

This function always returns 0. We can make it return void to simplify the
code. Also, no caller ever checks the return value of this function.

Link: https://lore.kernel.org/r/20220616080210.18531-1-mgurtovoy@nvidia.com
Reviewed-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Disable unused VCCQx power rails
Stanley Chu [Thu, 16 Jun 2022 05:37:25 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Disable unused VCCQx power rails

Some MediaTek UFS platforms support different VCCQx power rails, for
example, both 1.2v and 1.8v VCCQx, in a single kernel image.

To optimize the system power consumption, provide a way to disable and
release the unused power rail during the device probing.

Link: https://lore.kernel.org/r/20220616053725.5681-12-stanley.chu@mediatek.com
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support multiple VCC sources
Stanley Chu [Thu, 16 Jun 2022 05:37:24 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Support multiple VCC sources

Support multiple VCC sources in MediaTek UFS platforms.

Two options are provided and distinguished by specific device tree
attributes as below examples,

[Option 1: By numbering]
mediatek,ufs-vcc-by-num;
vcc-opt1-supply = <&mt6373_vbuck4_ufs>;
vcc-opt2-supply = <&mt6363_vemc>;

[Option 2: By UFS version]
mediatek,ufs-vcc-by-ver;
vcc-ufs3-supply = <&mt6373_vbuck4_ufs>;

Link: https://lore.kernel.org/r/20220616053725.5681-11-stanley.chu@mediatek.com
Signed-off-by: Alice Chao <alice.chao@mediatek.com>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: Export regulator functions
Stanley Chu [Thu, 16 Jun 2022 05:37:23 +0000 (13:37 +0800)]
scsi: ufs: Export regulator functions

Export below regulator functions to allow vendors to
customize regulator configuration in their own platforms.

int ufshcd_populate_vreg(struct device *dev, const char *name,
                         struct ufs_vreg **out_vreg);
int ufshcd_get_vreg(struct device *dev, struct ufs_vreg *vreg);

Link: https://lore.kernel.org/r/20220616053725.5681-10-stanley.chu@mediatek.com
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support low-power mode for parents of VCCQx
Stanley Chu [Thu, 16 Jun 2022 05:37:22 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Support low-power mode for parents of VCCQx

Provide the facility to configure parents of VCCQx power rails as low-power
or full-power mode in MediaTek UFS platforms.

Link: https://lore.kernel.org/r/20220616053725.5681-9-stanley.chu@mediatek.com
Signed-off-by: Alice Chao <alice.chao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support flexible parameters for SMC calls
Stanley Chu [Thu, 16 Jun 2022 05:37:21 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Support flexible parameters for SMC calls

Provide flexible number of parameters for UFS SMC calls to be used in a
subsequent commit.

Link: https://lore.kernel.org/r/20220616053725.5681-8-stanley.chu@mediatek.com
Signed-off-by: Alice Chao <alice.chao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Support low-power mode for VCCQ
Peter Wang [Thu, 16 Jun 2022 05:37:20 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Support low-power mode for VCCQ

Allow VCCQ to enter low-power mode, and also remove the restriction of VCC
because VCCQ/VCCQ2 can be changed to low-power mode even if VCC stays on
while the device is not in active power mode.

Link: https://lore.kernel.org/r/20220616053725.5681-7-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Prevent device regulators setting LPM incorrectly
Po-Wen Kao [Thu, 16 Jun 2022 05:37:19 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Prevent device regulators setting LPM incorrectly

Device regulatrs are allowed to enter low-power mode if neither device is
not in active mode, nor VCC does not keep on.

Fix this by adding conditions before LPM decision.

Link: https://lore.kernel.org/r/20220616053725.5681-6-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Po-Wen Kao <powen.kao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Fix the timing of configuring device regulators
Po-Wen Kao [Thu, 16 Jun 2022 05:37:18 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Fix the timing of configuring device regulators

Currently the LPM configurations of device regulators may not work since
VCC is not disabled yet while ufs_mtk_vreg_set_lpm() is executed.

Fix this by changing the timing of invoking ufs_mtk_vreg_set_lpm().

Link: https://lore.kernel.org/r/20220616053725.5681-5-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Po-Wen Kao <powen.kao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: ufs-mediatek: Introduce workaround for power mode change
CC Chou [Thu, 16 Jun 2022 05:37:17 +0000 (13:37 +0800)]
scsi: ufs: ufs-mediatek: Introduce workaround for power mode change

Some MediaTek SoC chips need special flow for power mode change, especially
for chips supporting HS-G5.

Enable the workaround by setting the host-specific capability.

Link: https://lore.kernel.org/r/20220616053725.5681-4-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: CC Chou <cc.chou@mediatek.com>
Signed-off-by: Eddie Huang <eddie.huang@mediatek.com>
Signed-off-by: Dennis Yu <tun-yu.yu@mediatek.com>
Signed-off-by: Peter Wang <peter.want@medaitek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
2 years agoscsi: ufs: Fix ADAPT logic for HS-G5
Stanley Chu [Thu, 16 Jun 2022 05:37:16 +0000 (13:37 +0800)]
scsi: ufs: Fix ADAPT logic for HS-G5

ADAPT now is added not only for HS Gear4 mode but also higher gears.  Fix
the logic for higher gears.

Link: https://lore.kernel.org/r/20220616053725.5681-3-stanley.chu@mediatek.com
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>