| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The spec recommends that for transfer length larger than the max-single-cmd
attribute (bMAX_DATA_SIZE_FOR_HPB_SINGLE_CMD) it is possible to couple
pre-requests with the HPB-READ command. Being a recommendation, using
pre-requests can be perceived merely as a means of optimization. A common
practice was to send pre-requests for chunks within some interval, and
leave the READ10 untouched if larger.
Now that the pre-request flows have been removed, all the commands are
single commands. Properly handle this attribute and do not send HPB-READ
for transfer lengths larger than max-single-cmd.
[mkp: resolve conflict]
Fixes: 09d9e4d04187 ("scsi: ufs: ufshpb: Remove HPB2.0 flows")
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Daejun Park <[email protected]>
Signed-off-by: Avri Altman <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
The partial UFS revert in 5.15 is needed for some additional fixes in
the 5.16 SCSI tree. Merge the fixes branch.
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
The changes to issue the abort from the scmd->abort_work instead of the EH
thread introduced a problem if eh_deadline is used. If aborting the
command(s) is successful, and there are never any scmds added to the
shost->eh_cmd_q, there is no code path which will reset the ->last_reset
value back to zero.
The effect of this is that after a successful abort with no EH thread
activity, a subsequent timeout, perhaps a long time later, might
immediately be considered past a user-set eh_deadline time, and the host
will be reset with no attempt at recovery.
Fix this by resetting ->last_reset back to zero in scmd_eh_abort_handler()
if it is determined that the EH thread will not run to do this.
Thanks to Gopinath Marappan for investigating this problem.
Link: https://lore.kernel.org/r/[email protected]
Fixes: e494f6a72839 ("[SCSI] improved eh timeout handler")
Cc: [email protected]
Signed-off-by: Ewan D. Milne <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
When the value of error is printed, it will always be 0. We should print
the correct error code when scsi_bsg_register_queue() fails.
Link: https://lore.kernel.org/r/[email protected]
Fixes: ead09dd3aed5 ("scsi: bsg: Simplify device registration")
Cc: Jens Axboe <[email protected]>
Cc: Christoph Hellwig <[email protected]>
Reviewed-by: Christoph Hellwig <[email protected]>
Signed-off-by: Jackie Liu <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
The Host Performance Buffer feature allows UFS read commands to carry the
physical media addresses along with the LBAs, thus allowing less internal
L2P-table switches in the device. HPB1.0 allowed a single LBA, while
HPB2.0 increases this capacity up to 255 blocks.
Carrying more than a single record, the read operation is no longer purely
of type "read" but a "hybrid" command: Writing the physical address to the
device in one operation and reading back the required payload in another.
The JEDEC HPB spec defines two commands for this operation:
HPB-WRITE-BUFFER (0x2) to write the physical addresses to device, and
HPB-READ to read the payload.
With the current HPB design the UFS driver has no alternative but to divide
the READ request into 2 separate commands: HPB-WRITE-BUFFER and HPB-READ.
This causes a great deal of aggravation to the block layer guys who
demanded that we completely revert the entire HPB driver regardless of the
huge amount of corporate effort already invested in it.
As a compromise, remove only the pieces that implement the 2.0
specification. This is done as a matter of urgency for the final 5.15
release.
Link: https://lore.kernel.org/r/[email protected]
Tested-by: Avri Altman <[email protected]>
Tested-by: Bean Huo <[email protected]>
Reviewed-by: Bart Van Assche <[email protected]>
Reviewed-by: Bean Huo <[email protected]>
Co-developed-by: James Bottomley <[email protected]>
Signed-off-by: James Bottomley <[email protected]>
Signed-off-by: Avri Altman <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Testing revealed a problem with how the reference tag was handled for
a WRITE_INSERT operation. The SCSI_PROT_REF_CHECK flag is not set when
the controller is asked to generate the protection information
(i.e. not DIX). And as a result the initial reference tag would not be
set in the WRITE_INSERT case.
Separate handling of the REF_CHECK and REF_INCREMENT flags to align
with both the DIX spec and the MPI implementation.
Link: https://lore.kernel.org/r/[email protected]
Fixes: b3e2c72af1d5 ("scsi: mpt3sas: Use the proper SCSI midlayer interfaces for PI")
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alexey Dobriyan <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Introduce virtual host driver for ExynosAuto v9 UFS mHCI.
The VH (Virtual Host) only supports data transfer functions so most of
physical features must be disabled by setting the following quirks:
- UFSHCD_QUIRK_BROKEN_UIC_CMD
- UFSHCD_QUIRK_SKIP_PH_CONFIGURATION
Before initialization, the VH must wait until PH is ready. This wait is
currently implemented via polling.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
The UFS controller of the ExynosAuto v9 SoC supports a multi-host interface
for I/O virtualization. In general, we're using para-virtualized driver to
support a block device by several virtual machines.
Multi-host functionality extends the host controller by providing register
interfaces that can be used by each VM's UFS drivers respectively. This way
we can provide direct access to the UFS device for multiple VMs similar to
PCIe SR-IOV.
We divide this M-HCI as PH (Physical Host) and VHs (Virtual Host). The PH
supports all UFSHCI functions (all SAPs) like a conventional UFSHCI but the
VH only supports data transfer functions. Thus, except UTP_CMD_SAP and
UTP_TMPSAP, the PH should handle all the physical features.
Provide an initial implementation of PH part. M-HCI can support up to four
interfaces (1 for a PH and 3 for VHs) but this patch initially supports
only 1 PH and 1 VH. For this, we uses TASK_TAG[7:5] field so TASK_TAG[4:0]
for 32 doorbell will be supported. After the PH is initiated, this will
send a ready message to VHs through a mailbox register. The message handler
is not fully implemented yet such as supporting reset / abort cases.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Reviewed-by: Inki Dae <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Add support for the ExynosAuto v9 SoC. This requires controlling the UFS IP
shareability register via syscon and regmap. The offset of the register
can be different according to the UFS instance and SoC specific offset
value. As a result, we need to get the offset value from DT property.
Unlike exynos7, this implementation has a different M-PHY setting which
must be configured via exynosauto_ufs_pre_link.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Add driver-specific pre/post_hce_enable callbacks to execute extra
initializations before and after hce_enable_notify callback.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
To leverage the initialization code for the other variant of the exynos-ufs
driver, factor out the assignment part.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
To skip exynos_ufs_config_phy_*_attr settings for exynos-ufs variant,
provide EXYNOS_UFS_OPT_SKIP_CONFIG_PHY_ATTR as an opts flag.
For the ExynosAuto v9 SoC's controller, M-Phy timing setting is not
required and most of vendor-specific configuration will be performed in the
pre_link callback function.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
By default, ufs_hba_exynos_ops will be used. Add support for a custom
version of ufs_hba_variant_ops because some variants of exynos-ufs will use
only few callbacks.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Add setup_clocks callback to control/gate clocks by ufshcd. To avoid
calling before initialization, check whether UFS is on or not and call it
initially from pre_link callback.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: Kiwoong Kim <[email protected]>
Cc: Krzysztof Kozlowski <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
This patch adds REFCLKOUT_STOP control to CLK_STOP_MASK. This permits
enabling/disabling reference clock out control for the UFS device.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
The compatible field of exynos_ufs_drv_data is not necessary because
of_device_id already has it. Thus, we don't need it anymore and we can get
drv_data by device_get_match_data.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
To support 167MHz PCLK, we need to adjust the maximum value.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Alim Akhtar <[email protected]>
Reviewed-by: Inki Dae <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Samsung ExynosAuto v9 SoC virtual hosts do not support device management.
Add a quirk to skip the physical host interface configuration part that
cannot be performed in the virtual host.
Link: https://lore.kernel.org/r/[email protected]
Cc: James E.J. Bottomley <[email protected]>
Cc: Martin K. Petersen <[email protected]>
Cc: Bart Van Assche <[email protected]>
Suggested-by: Alim Akhtar <[email protected]>
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: jongmin jeong <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Samsung ExynosAuto v9 SoC has two types of host controller interface to
support the virtualization of UFS Device. One is the physical host (PH)
that is the same as conventional UFSHCI, and the other is the virtual host
(VH) that supports data transfer function only.
In this configuration the virtual host does not support UIC commands. Add a
quirk to return 0 when the UIC command send function is called.
Link: https://lore.kernel.org/r/[email protected]
Cc: Alim Akhtar <[email protected]>
Cc: James E.J. Bottomley <[email protected]>
Cc: Martin K. Petersen <[email protected]>
Cc: Bart Van Assche <[email protected]>
Reviewed-by: Alim Akhtar <[email protected]>
Signed-off-by: jongmin jeong <[email protected]>
Signed-off-by: Chanho Park <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
v4.17 commit 86b87cde0b55 ("scsi: core: host template attribute groups")
introduced explicit sysfs_create_groups() in scsi_sysfs_add_sdev()
and sysfs_remove_groups() in __scsi_remove_device(), both for sdev_gendev,
based on a new field const struct attribute_group **sdev_groups
of struct scsi_host_template.
Commit 92c4b58b15c5 ("scsi: core: Register sysfs attributes earlier")
removed above explicit (de)registration of scsi_device attribute groups.
It also converted all scsi_device attributes and attribute_groups to
end up in a new field const struct attribute_group *gendev_attr_groups[6]
of struct scsi_device. However, that new field was not used anywhere.
Surprisingly, this only caused missing LLDD specific scsi_device sysfs
attributes. Whereas, scsi core attributes from scsi_sdev_attr_groups
did continue to exist because of scsi_dev_type.groups.
We separate scsi core attibutes from LLDD specific attributes.
Hence, we keep the initializing assignment scsi_dev_type =
{ .groups = scsi_sdev_attr_groups, } as this takes care of core
attributes. Without the separation, it would cause attribute double
registration due to scsi_dev_type.groups and sdev_gendev.groups.
Julian suggested to assign the sdev_groups pointer of the
scsi_host_template directly to the groups pointer of sdev_gendev.
This way we can delete the container scsi_device.gendev_attr_groups
and the loop copying each entry from hostt->sdev_groups to
sdev->gendev_attr_groups.
Alternative approaches ruled out:
Assigning gendev_attr_groups to sdev_dev has no visible effect.
Assigning sdev->gendev_attr_groups to scsi_dev_type.groups
caused scsi_device of all scsi host types to get LLDD specific
attributes of the LLDD for which the last sdev alloc happened to occur,
as that overwrote scsi_dev_type.groups,
e.g. scsi_debug had zfcp-specific scsi_device attributes.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 92c4b58b15c5 ("scsi: core: Register sysfs attributes earlier")
Suggested-by: Julian Wiedmann <[email protected]>
Reviewed-by: Bart Van Assche <[email protected]>
Signed-off-by: Steffen Maier <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Various EDIF bsgs did not properly fill out the reply_payload_rcv_len
field. This causes app to parse empty data in the return payload.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 7ebb336e45ef ("scsi: qla2xxx: edif: Add start + stop bsgs")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
db_flags field is a bit field. Replace value check with bit flag check.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Currently, firmware limits ELS payload to FC frame size/2112. This patch
adjusts memory buffer size to be able to handle max ELS payload.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 84318a9f01ce ("scsi: qla2xxx: edif: Add send, receive, and accept for auth_els")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
On ipsec start by remote port, target port may use RSCN to trigger
initiator to relogin. If driver is already in the process of a relogin,
then ignore the RSCN and allow the current relogin to continue. This
reduces thrashing of the connection.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Modify trace messages for additional debugability.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
This patch is per review comment by Hannes Reinecke from previous
submission to replace list_for_each_safe with list_for_each_entry_safe.
Link: https://lore.kernel.org/r/[email protected]
Reviewed-by: Hannes Reinecke <[email protected]>
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
On session down, driver will flush all stale messages and doorbell
events. This prevents authentication application from having to process
stale data.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 4de067e5df12 ("scsi: qla2xxx: edif: Add N2N support for EDIF")
Reviewed-by: Himanshu Madhani <[email protected]>
Co-developed-by: Karunakara Merugu <[email protected]>
Signed-off-by: Karunakara Merugu <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Current driver does unnecessary pause for each session to get to certain
state before allowing the app start call to return. In larger environment,
this introduces a long delay. Originally the delay was meant to
synchronize app and driver. However, the with current implementation the
two sides use various events to synchronize their state.
The same is applied to the authentication failure call.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 4de067e5df12 ("scsi: qla2xxx: edif: Add N2N support for EDIF")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
On app start, all sessions need to be reset to see if secure connection can
be made. Fix the broken check which prevents that process.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 4de067e5df12 ("scsi: qla2xxx: edif: Add N2N support for EDIF")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
When user uses issue_lip to do link bounce, driver sends additional target
reset to remote device before resetting the link. The target reset would
affect other paths with active I/Os. This patch will remove the unnecessary
target reset.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 5854771e314e ("[SCSI] qla2xxx: Add ISPFX00 specific bus reset routine")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Current code does list element deletion and addition in and out of lock
protection. This patch moves deletion behind lock.
list_add double add: new=ffff9130b5eb89f8, prev=ffff9130b5eb89f8,
next=ffff9130c6a715f0.
------------[ cut here ]------------
kernel BUG at lib/list_debug.c:31!
invalid opcode: 0000 [#1] SMP PTI
CPU: 1 PID: 182395 Comm: kworker/1:37 Kdump: loaded Tainted: G W OE
--------- - - 4.18.0-193.el8.x86_64 #1
Hardware name: HP ProLiant DL160 Gen8, BIOS J03 02/10/2014
Workqueue: qla2xxx_wq qla2x00_iocb_work_fn [qla2xxx]
RIP: 0010:__list_add_valid+0x41/0x50
Code: 85 94 00 00 00 48 39 c7 74 0b 48 39 d7 74 06 b8 01 00 00 00 c3 48 89 f2
4c 89 c1 48 89 fe 48 c7 c7 60 83 ad 97 e8 4d bd ce ff <0f> 0b 0f 1f 00 66 2e
0f 1f 84 00 00 00 00 00 48 8b 07 48 8b 57 08
RSP: 0018:ffffaba306f47d68 EFLAGS: 00010046
RAX: 0000000000000058 RBX: ffff9130b5eb8800 RCX: 0000000000000006
RDX: 0000000000000000 RSI: 0000000000000096 RDI: ffff9130b7456a00
RBP: ffff9130c6a70a58 R08: 000000000008d7be R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000001 R12: ffff9130c6a715f0
R13: ffff9130b5eb8824 R14: ffff9130b5eb89f8 R15: ffff9130b5eb89f8
FS: 0000000000000000(0000) GS:ffff9130b7440000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007efcaaef11a0 CR3: 000000005200a002 CR4: 00000000000606e0
Call Trace:
qla24xx_async_gnl+0x113/0x3c0 [qla2xxx]
? qla2x00_iocb_work_fn+0x53/0x80 [qla2xxx]
? process_one_work+0x1a7/0x3b0
? worker_thread+0x30/0x390
? create_worker+0x1a0/0x1a0
? kthread+0x112/0x130
Link: https://lore.kernel.org/r/[email protected]
Fixes: 726b85487067 ("qla2xxx: Add framework for async fabric discovery")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
For RSCN of type "Area, Domain, or Fabric", which indicate a portion or
entire fabric was disturbed, current driver does not set the scan_need flag
to indicate a session was affected by the disturbance. This in turn can
lead to I/O timeout and delay of relogin. Hence initiate relogin in the
event of fabric disturbance.
Link: https://lore.kernel.org/r/[email protected]
Fixes: 1560bafdff9e ("scsi: qla2xxx: Use complete switch scan for RSCN events")
Reviewed-by: Himanshu Madhani <[email protected]>
Signed-off-by: Quinn Tran <[email protected]>
Signed-off-by: Nilesh Javali <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
'sli4->ext[i].use_map' is a bitmap. Use 'bitmap_zalloc()' to simplify code,
improve the semantic and avoid some open-coded arithmetic in allocator
arguments.
Also change the corresponding 'kfree()' into 'bitmap_free()' to keep
consistency.
Link: https://lore.kernel.org/r/2a0a83949fb896a0a236dcca94dfdc8486d489f5.1635104793.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Replace two cpu_to_le32() calls by a single cpu_to_le64() call.
Additionally, issue a warning if the length of an scatter gather list
element exceeds what is allowed by the UFSHCI specification.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Before modifying struct ufshcd_sg_entry, add a compile-time structure size
check.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Casting an int explicitly to u16 when passed as an argument to a function
is not necessary.
Since prd_table and ucd_prdt_ptr both have type struct ufshcd_sg_entry *,
remove the casts from assignments of these two to each other.
This patch does not change any functionality.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Make it easier to test the impact of the UFS error handler on software that
submits SCSI commands to the UFS driver.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Introduce an array for debugfs attributes to make it easier to add new
debugfs attributes. Change the value of the inode.i_private pointer for
debugfs attributes from a pointer to the HBA data structure to a pointer to
the attribute description for the stats attribute. Store the HBA pointer in
the private data of the parent inode instead.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Make it possible to call ufshcd_schedule_eh_work() from other source files
than ufshcd.c. Additionally, convert a source code comment into a
lockdep_assert_held() call.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Kernel logs are hard to comprehend without information about what the UFS
error handler is doing. Hence this patch that logs information about error
handler activity.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Introduce an enumeration type for the overall command status to allow the
compiler to perform more static type checking.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Make the descriptions above data structures that come from the UFS
specification match the terminology from that specification. This makes it
easier to find these data structures while reading the UFS specification.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
successfully"
Commit 73dc3c4ac703 ("scsi: ufs: Retry aborted SCSI commands instead of
completing these successfully") is not necessary. If a SCSI command is
aborted successfully the UFS controller has not modified the command status
and the command status still has the value assigned by
ufshcd_prepare_req_desc_hdr(), namely OCS_INVALID_COMMAND_STATUS. The
function ufshcd_transfer_rsp_status() requeues commands that have an
invalid command status. Hence revert commit 73dc3c4ac703.
Link: https://lore.kernel.org/r/[email protected]
Acked-by: Avri Altman <[email protected]>
Signed-off-by: Bart Van Assche <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Currently TMF commands are removed from de_device.dev_tmf_list at the very
end of se_cmd lifecycle. However, se_lun unlinks from se_cmd upon a command
status (response) being queued in transport layer. This means that LUN and
backend device can be deleted in the meantime and a panic will occur:
target_tmr_work()
cmd->se_tfo->queue_tm_rsp(cmd); // send abort_rsp to a wire
transport_lun_remove_cmd(cmd) // unlink se_cmd from se_lun
- // - // - // -
<<<--- lun remove
<<<--- core backend device remove
- // - // - // -
qlt_handle_abts_completion()
tfo->free_mcmd()
transport_generic_free_cmd()
target_put_sess_cmd()
core_tmr_release_req() {
if (dev) { // backend device, can not be null
spin_lock_irqsave(&dev->se_tmr_lock, flags); //<<<--- CRASH
Call Trace:
NIP [c000000000e1683c] _raw_spin_lock_irqsave+0x2c/0xc0
LR [c00800000e433338] core_tmr_release_req+0x40/0xa0 [target_core_mod]
Call Trace:
(unreliable)
0x0
target_put_sess_cmd+0x2a0/0x370 [target_core_mod]
transport_generic_free_cmd+0x6c/0x1b0 [target_core_mod]
tcm_qla2xxx_complete_mcmd+0x28/0x50 [tcm_qla2xxx]
process_one_work+0x2c4/0x5c0
worker_thread+0x88/0x690
For the iSCSI protocol this is easily reproduced:
- Send some SCSI sommand
- Send Abort of that command over iSCSI
- Remove LUN on target
- Send next iSCSI command to acknowledge the Abort_Response
- Target panics
There is no need to keep the command in tmr_list until response completion,
so move the removal from tmr_list from the response completion to the
response queueing when the LUN is unlinked. Move the removal from state
list too as it is a subject to the same race condition.
Link: https://lore.kernel.org/r/[email protected]
Fixes: c66ac9db8d4a ("[SCSI] target: Add LIO target core v4.0.0-rc6")
Reviewed-by: Roman Bolshakov <[email protected]>
Reviewed-by: Mike Christie <[email protected]>
Signed-off-by: Dmitry Bogdanov <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
Update lpfc version to 14.0.0.3.
Link: https://lore.kernel.org/r/[email protected]
Co-developed-by: Justin Tee <[email protected]>
Signed-off-by: Justin Tee <[email protected]>
Signed-off-by: James Smart <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
A link bounce to a slow fabric may observe FDISC response delays lasting
longer than devloss tmo. Current logic decrements the final fabric node
kref during a devloss tmo event. This results in a NULL ptr dereference
crash if the FDISC completes for that fabric node after devloss tmo.
Fix by adding the NLP_IN_RECOV_POST_DEV_LOSS flag, which is set when
devloss tmo triggers and we've noticed that fabric node recovery has
already started or finished in between the time lpfc_dev_loss_tmo_callbk
queues lpfc_dev_loss_tmo_handler. If fabric node recovery succeeds, then
the driver reverses the devloss tmo marked kref put with a kref get. If
fabric node recovery fails, then the final kref put relies on the ELS
timing out or the REG_LOGIN cmpl routine.
Link: https://lore.kernel.org/r/[email protected]
Co-developed-by: Justin Tee <[email protected]>
Signed-off-by: Justin Tee <[email protected]>
Signed-off-by: James Smart <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
If an FC link down transition while PLOGIs are outstanding to fabric well
known addresses, outstanding ABTS requests may result in a NULL pointer
dereference. Driver unload requests may hang with repeated "2878" log
messages.
The Link down processing results in ABTS requests for outstanding ELS
requests. The Abort WQEs are sent for the ELSs before the driver had set
the link state to down. Thus the driver is sending the Abort with the
expectation that an ABTS will be sent on the wire. The Abort request is
stalled waiting for the link to come up. In some conditions the driver may
auto-complete the ELSs thus if the link does come up, the Abort completions
may reference an invalid structure.
Fix by ensuring that Abort set the flag to avoid link traffic if issued due
to conditions where the link failed.
Link: https://lore.kernel.org/r/[email protected]
Co-developed-by: Justin Tee <[email protected]>
Signed-off-by: Justin Tee <[email protected]>
Signed-off-by: James Smart <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|
|
A remote nport can stop responding to PLOGI beyond the ELS I/O timeout
under some fault conditions. When this happens, the non-response triggers
a dev_loss_tmo event from the transport which causes the driver to abort
the PLOGI and stop any retries. This was due to a policy in the ELS
completion handler whenever an ELS was terminated due to driver request.
Revise the ELS completion path to detect PLOGIs that were aborted and
allow retries.
Link: https://lore.kernel.org/r/[email protected]
Co-developed-by: Justin Tee <[email protected]>
Signed-off-by: Justin Tee <[email protected]>
Signed-off-by: James Smart <[email protected]>
Signed-off-by: Martin K. Petersen <[email protected]>
|