Commits · e9b1108316b9b5beee03f731c7c9c7c874e537fa · Kirill Smelkov / linux

17 Nov, 2020 15 commits

scsi: lpfc: Fix refcounting around SCSI and NVMe transport APIs · e9b11083

James Smart authored Nov 15, 2020

Due to bug history and code review, the node reference counting approach in
the driver isn't implemented consistently with how the scsi and nvme
transport perform registrations and unregistrations and their callbacks.
This resulted in many bad/stale node pointers.

Reword the driver so that reference handling is performed as follows:

- The initial node reference is taken on structure allocation

- Take a reference on any add/register call to the transport

- Remove a reference on any delete/unregister call to the transport

- After the node has fully removed from both the SCSI and NVMEe transports
(dev_loss_callbacks have called back) call the discovery engine
DEVICE_RM event which will remove the final reference and release the
node structure.

- Alter dev_loss handling when a vport or base port is unloading.

- Remove the put_node handling - no longer needed.

- Rewrite the vport_delete handling on reference counts. Part of this
effort was driven from the FDISC not registering with the transport and
disrupting the model for node reference counting.

- Deleted lpfc_nlp_remove. Pushed it's remaining ops into
lpfc_nlp_release.

- Several other small code cleanups.

Link: https://lore.kernel.org/r/20201115192646.12977-5-james.smart@broadcom.comCo-developed-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

e9b11083

scsi: lpfc: Fix removal of SCSI transport device get and put on dev structure · 95f0ef8a

James Smart authored Nov 15, 2020

The lpfc driver is calling get_device and put_device on scsi_fc_transport
device structure. When this code was removed, the driver triggered an oops
in "scsi_is_host_dev" when the first SCSI target was unregistered from the
transport.

The reason the calls were necessary is that the driver is calling
scsi_remove_host too early, before the target rports are unregistered and
the scsi devices disconnected from the scsi_host. The fc_host was torn
down during fc_remove_host.

Fix by moving the lpfc_pci_remove_one_s3/s4 calls to scsi_remove_host to
after the nodes are cleaned up. Remove the get_device and put_device calls
and the supporting code.

Link: https://lore.kernel.org/r/20201115192646.12977-4-james.smart@broadcom.comCo-developed-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

95f0ef8a

scsi: lpfc: Rework locations of ndlp reference taking · 4430f7fd

James Smart authored Nov 15, 2020

Now that the driver has gone to a normal ref interface (with no odd logic)
the discovery logic needs to be updated to reworked so that it properly
takes references when it should and give them up when it should.

Rework the driver for the following get/put model:

- Move gets to just before an I/O is issued. Add gets for places where an
I/O was issued without one.

- Ensure that failures from lpfc_nlp_get() are handled by the driver.

- Check and fix the placement of lpfc_nlp_puts relative to io completions.
Note: some of these paths may not release the reference on the exact io
completion as the reference is held as the code takes another step in
the discovery thread and which may cause another io to be issued.

- Rearrange some code for error processing and calling lpfc_nlp_put.

- Fix some places of incorrect reference freeing that was causing the
premature releasing of the structure.

- Nvmet plogi handling performs unreg_rpi's. The reference counts were
unbalanced resulting in premature node removal. In some cases this
caused loss of node discovery. Corrected the reftaking around nvmet
plogis.

Nodes that experience devloss now get released from the node list now that
there is a proper reference taking.

Link: https://lore.kernel.org/r/20201115192646.12977-3-james.smart@broadcom.comCo-developed-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

4430f7fd

scsi: lpfc: Rework remote port ref counting and node freeing · 307e3380

James Smart authored Nov 15, 2020

When a remote port is disconnected and disappears, its node structure
(ndlp) stays allocated and on a vport node list. While on the list it can
be matched, thus requires validation checks on state to be added in
numerous code paths. If the node comes back, its possible for there to be
multiple node structures for the same device on the vport node list. There
is no reason to keep the node structure around after it is no longer in
existence, and the current implementation creates problems for itself
(multiple nodes) and lots of unnecessary code for state validation.

Additionally, the reference taking on the node structure didn't follow the
normal model used by the kernel kref api. It included lots of odd logic to
match state with reference count. The combination of this odd logic plus
the way it was implicitly used in the discovery engine made its reference
taking implementation suspect and extremely hard to follow.

Change the driver such that the reference taking routines are now normal
ref increments/decrements and callout on refcount=0.

With this in place, the rework can be done such that the node structure is
fully removed and deallocated when the remote port no longer exists and all
references are removed. This removal logic, and the basic ref counting are
intrically tied, thus in a single patch.

Link: https://lore.kernel.org/r/20201115192646.12977-2-james.smart@broadcom.comCo-developed-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

307e3380

scsi: be2iscsi: Mark beiscsi_attrs with static keyword · 4ab2990a

Zou Wei authored Nov 14, 2020

Fix the following sparse warning:

./be_main.c:167:25: warning: symbol 'beiscsi_attrs' was not declared. Should it be static?

Link: https://lore.kernel.org/r/1605339474-22329-1-git-send-email-zou_wei@huawei.comSigned-off-by: Zou Wei <zou_wei@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

4ab2990a

scsi: ufs-qcom: Only select QCOM_SCM if SCSI_UFS_CRYPTO · 6ac63216

Eric Biggers authored Nov 13, 2020

QCOM_SCM is only needed to make the qcom_scm_*() calls in ufs-qcom-ice.c,
which is only compiled when SCSI_UFS_CRYPTO=y. So don't unnecessarily
enable QCOM_SCM when SCSI_UFS_CRYPTO=n.

Link: https://lore.kernel.org/r/20201114004754.235378-1-ebiggers@kernel.orgSigned-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

6ac63216

scsi: hpsa: Strip out a bunch of set but unused variables · 1fc65919

Lee Jones authored Nov 12, 2020

Fixes the following W=1 kernel build warning(s):

drivers/scsi/hpsa.c: In function ‘hpsa_volume_offline’:
drivers/scsi/hpsa.c:3885:5: warning: variable ‘scsi_status’ set but not used [-Wunused-but-set-variable]
drivers/scsi/hpsa.c:3884:6: warning: variable ‘cmd_status’ set but not used [-Wunused-but-set-variable]
drivers/scsi/hpsa.c: In function ‘hpsa_update_scsi_devices’:
drivers/scsi/hpsa.c:4354:9: warning: variable ‘n_ext_target_devs’ set but not used [-Wunused-but-set-variable]
drivers/scsi/hpsa.c: In function ‘hpsa_scatter_gather’:
drivers/scsi/hpsa.c:4583:36: warning: variable ‘last_sg’ set but not used [-Wunused-but-set-variable]
drivers/scsi/hpsa.c: In function ‘hpsa_init_one’:
drivers/scsi/hpsa.c:8639:6: warning: variable ‘dac’ set but not used [-Wunused-but-set-variable]
drivers/scsi/hpsa.c: In function ‘hpsa_enter_performant_mode’:
drivers/scsi/hpsa.c:9300:7: warning: variable ‘rc’ set but not used [-Wunused-but-set-variable]

Link: https://lore.kernel.org/r/20201112101929.GC1997862@dell
Cc: Don Brace <don.brace@microchip.com>
Cc: Bugfixes to <esc.storagedev@microsemi.com>
Cc: storagedev@microchip.com
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1fc65919

scsi: ufs: Add retry flow for failed HBA enabling · 6081b12c

Stanley Chu authored Nov 12, 2020

Once HBA enabling has failed, add retry mechanism and allow vendors to
apply specific tweaks before the next retry. For example, vendors can do
vendor-specific host reset flow in variant function
"ufshcd_vops_hce_enable_notify()".

Link: https://lore.kernel.org/r/20201112054537.22494-1-stanley.chu@mediatek.comReviewed-by: Can Guo <cang@codeaurora.org>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

6081b12c

scsi: smartpqi: Update version to 1.2.16-012 · 5443bdc4

Don Brace authored Nov 11, 2020

Link: https://lore.kernel.org/r/160512629093.2359.13675060282143622110.stgit@brunhildaReviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Reviewed-by: Gerry Morong <gerry.morong@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

5443bdc4

scsi: smartpqi: Correct pqi_sas_smp_handler busy condition · 408bdd7e

Don Brace authored Nov 11, 2020

Remove unbalanced call to pqi_ctrl_unbusy.

Link: https://lore.kernel.org/r/160512628513.2359.17193493825283879603.stgit@brunhildaReviewed-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: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

408bdd7e

scsi: smartpqi: Correct driver removal with HBA disks · 1bdf6e93

Don Brace authored Nov 11, 2020

Correct rmmod hangs when using HBA disks with write cache enabled.

Do not set controller flag "in_shutdown" during rmmod. SCSI SYNCHRONIZE
CACHE(10) and SCSI SYNCHRONIZE CACHE(16) requests were blocked with
SCSI_MLQUEUE_HOST_BUSY.

Link: https://lore.kernel.org/r/160512627928.2359.10698615071827614781.stgit@brunhildaReviewed-by: Scott Benesh <scott.benesh@microchip.com>
Reviewed-by: Scott Teel <scott.teel@microchip.com>
Signed-off-by: Don Brace <don.brace@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

1bdf6e93

scsi: pm8001: Remove unused variable 'value' · a364a3ea

Lee Jones authored Nov 16, 2020

Hasn't been used since 2009.

Fixes the following W=1 kernel build warning(s):

 drivers/scsi/pm8001/pm8001_hwi.c: In function ‘mpi_set_phys_g3_with_ssc’:
 drivers/scsi/pm8001/pm8001_hwi.c:415:6: warning: variable ‘value’ set but not used [-Wunused-but-set-variable]

Link: https://lore.kernel.org/r/20201116104119.816527-1-lee.jones@linaro.org
Cc: Jack Wang <jinpu.wang@cloud.ionos.com>
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

a364a3ea

scsi: core: Fix -Wformat for scsi_host · 88392820

Nick Desaulniers authored Nov 07, 2020

Clang is more aggressive about -Wformat warnings when the format flag
specifies a type smaller than the parameter. Turns out, struct Scsi_Host's
member can_queue is actually an int. Fixes:

[-Wformat]
shost_rd_attr(can_queue, "%hd\n");
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
                          %d

Link: https://github.com/ClangBuiltLinux/linux/issues/378
Link: https://lore.kernel.org/r/20201107081132.2629071-1-ndesaulniers@google.comReviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
warning: format specifies type 'short' but the argument has type 'int'

88392820

scsi: core: Replace while-loop by for-loop in scsi_vpd_lun_id() · 16d6317e

Martin Wilck authored Oct 29, 2020

This makes the code slightly more readable.

Link: https://lore.kernel.org/r/20201029170846.14786-2-mwilck@suse.comReviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

16d6317e

scsi: core: Fix VPD LUN ID designator priorities · 2e4209b3

Martin Wilck authored Oct 29, 2020

The current implementation of scsi_vpd_lun_id() uses the designator length
as an implicit measure of priority. This works most of the time, but not
always. For example, some Hitachi storage arrays return this in VPD 0x83:

VPD INQUIRY: Device Identification page
  Designation descriptor number 1, descriptor length: 24
    designator_type: T10 vendor identification,  code_set: ASCII
    associated with the Addressed logical unit
      vendor id: HITACHI
      vendor specific: 5030C3502025
  Designation descriptor number 2, descriptor length: 6
    designator_type: vendor specific [0x0],  code_set: Binary
    associated with the Target port
      vendor specific: 08 03
  Designation descriptor number 3, descriptor length: 20
    designator_type: NAA,  code_set: Binary
    associated with the Addressed logical unit
      NAA 6, IEEE Company_id: 0x60e8
      Vendor Specific Identifier: 0x7c35000
      Vendor Specific Identifier Extension: 0x30c35000002025
      [0x60060e8007c350000030c35000002025]

The current code would use the first descriptor because it's longer than
the NAA descriptor. But this is wrong, the kernel is supposed to prefer NAA
descriptors over T10 vendor ID. Designator length should only be used to
compare designators of the same type.

This patch addresses the issue by separating designator priority and
length.

Link: https://lore.kernel.org/r/20201029170846.14786-1-mwilck@suse.com
Fixes: 9983bed3 ("scsi: Add scsi_vpd_lun_id()")
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

2e4209b3

11 Nov, 2020 25 commits

scsi: bnx2fc: Fix comparison to bool warning · cb2b4e8f

Kaixu Xia authored Nov 08, 2020

Fix the following coccicheck warning:

./drivers/scsi/bnx2fc/bnx2fc_fcoe.c:2089:5-23: WARNING: Comparison to bool
./drivers/scsi/bnx2fc/bnx2fc_fcoe.c:2187:5-23: WARNING: Comparison to bool

Link: https://lore.kernel.org/r/1604767920-8361-1-git-send-email-kaixuxia@tencent.comReported-by: Tosk Robot <tencent_os_robot@tencent.com>
Acked-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Kaixu Xia <kaixuxia@tencent.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

cb2b4e8f

scsi: isci: Don't use PCI helper functions · 44ff81f2

Vaibhav Gupta authored Nov 07, 2020

PCI helper functions such as pci_enable/disable_device(),
pci_save/restore_state(), pci_set_power_state(), etc. were used by the
legacy framework to perform standard operations related to PCI PM.

This driver is using the generic framework and thus calls for those
functions should be dropped as those tasks are now performed by the PCI
core.

Link: https://lore.kernel.org/r/20201107100420.149521-1-vaibhavgupta40@gmail.comAcked-by: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Signed-off-by: Vaibhav Gupta <vaibhavgupta40@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

44ff81f2

scsi: ufs-mediatek: Add HS-G4 support · 638e6271

Stanley Chu authored Oct 29, 2020

Provide HS-G4 support in MediaTek UFS platforms.

To support HS-G4, introduce mechanism to get the MediaTek UFS controller
version. With such information, driver can make right decision to apply
different configurations in different controllers.

Link: https://lore.kernel.org/r/20201029115750.24391-7-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

638e6271

scsi: ufs: Add enums for UniPro version higher than 1.6 · 801909ac

Stanley Chu authored Oct 29, 2020

Some vendors need newer UniPro version to decide if some features can be
enabled or not.

Simply add missing enums for the latest UniPro versions.

Link: https://lore.kernel.org/r/20201029115750.24391-6-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

801909ac

scsi: ufs-mediatek: Support option to disable auto-hibern8 · bf8fab86

Stanley Chu authored Oct 29, 2020

Support an option to allow users to disable auto-hibern8 feature.

Instead, enable hibern8-during-clk-gating feature to keep similar power
consumption.

Link: https://lore.kernel.org/r/20201029115750.24391-5-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

bf8fab86

scsi: ufs-mediatek: Decouple features from platform bindings · ac8c2459

Stanley Chu authored Oct 29, 2020

UFS proprietary features (including features introduced later) in MediaTek
UFS platforms have complicated combinations among different platforms.

To ease code readability and maintenance, decouple all proprietary features
from platform bindings. Each feature would be enabled only if specific
property string is defined in device tree node.

Link: https://lore.kernel.org/r/20201029115750.24391-4-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ac8c2459

scsi: ufs-mediatek: Support VA09 regulator operations · cf137b3e

Stanley Chu authored Oct 29, 2020

Some MediaTek UFS platforms need to control VA09 power
specifically. Provide such control according to the device tree binding.

Link: https://lore.kernel.org/r/20201029115750.24391-3-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

cf137b3e

scsi: ufs-mediatek: Assign arguments with correct type · a004147a

Stanley Chu authored Oct 29, 2020

In ufs_mtk_unipro_set_lpm(), use specific unsigned values as the argument
to invoke ufshcd_dme_set().

At the same time, change the name of ufs_mtk_unipro_set_pm() to
ufs_mtk_unipro_set_lpm() to align the naming convention in MediaTek UFS
driver.

Link: https://lore.kernel.org/r/20201029115750.24391-2-stanley.chu@mediatek.comSigned-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

a004147a

scsi: core: Return BLK_STS_AGAIN for ALUA transitioning · 0d882320

Hannes Reinecke authored Sep 30, 2020

Whenever we encounter a sense code of ALUA transitioning in
scsi_io_completion() it means that the SCSI midlayer ran out of retries
trying to wait for ALUA transitioning.  In these cases we should be passing
up the error, but signalling that the I/O might be retried, preferably on
another path.  So return BLK_STS_AGAIN in these cases.

[mkp: typo + fallthrough]

Link: https://lore.kernel.org/r/20200930080256.90964-5-hare@suse.deReviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

0d882320

scsi: scsi_dh_alua: Set 'transitioning' state on Unit Attention · fb908850

Hannes Reinecke authored Sep 30, 2020

We should be setting the 'transitioning' ALUA state once we get a Unit
Attention indicating the array is in transitioning. There are arrays which
cannot respond to an RTPG while in transitioning, and others have issues
correctly reporting the state. So better to set the state during Unit
Attention handling and wait for TUR / RTPG to run its course.

Link: https://lore.kernel.org/r/20200930080256.90964-4-hare@suse.deReviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

fb908850

scsi: scsi_dh_alua: Return BLK_STS_AGAIN for ALUA transitioning state · 268940b8

Hannes Reinecke authored Sep 30, 2020

When the ALUA state indicates transitioning we should not retry the command
immediately, but rather complete the command with BLK_STS_AGAIN to signal
the completion handler that it might be retried. This allows multipathing
to redirect the command to another path if possible, and avoid stalls
during lengthy transitioning times.

Link: https://lore.kernel.org/r/20200930080256.90964-3-hare@suse.deReviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

268940b8

scsi: block: Return status code in blk_mq_end_request() · e21ee5a6

Hannes Reinecke authored Sep 30, 2020

blk_mq_end_request() will use the block status returned from queue_rq() as
argument, except in one instance in blk_mq_dispatch_rq_list(), where the
generic BLK_STS_IOERR is used.

Link: https://lore.kernel.org/r/20200930080256.90964-2-hare@suse.deReviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

e21ee5a6

scsi: MAINTAINERS: Make Bodo target_core_user maintainer · e9326b5a

Mike Christie authored Nov 04, 2020

Bodo knows the code better than me now, has time to review patches and is
excellent at it, and has lots of ideas for how to make the driver better,
so this patch adds him as the maintainer.

There was no entry in there already. Andy had posted on the list here:

https://www.spinics.net/lists/target-devel/msg14690.html

when it got transitioned to me. I added an entry because several companies
used it and I thought it would be easy for them to find Bodo.

Link: https://lore.kernel.org/r/1604521666-16573-1-git-send-email-michael.christie@oracle.com
Cc: Bodo Stroesser <bostroesser@gmail.com>
Signed-off-by: Mike Christie <michael.christie@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>