| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
Pull SCSI updates from James Bottomley:
"Updates to the usual drivers (ufs, qla2xx, target, lpfc, smartpqi,
mpi3mr).
The main driver change that might cause issues on down the road is the
conversion of some of our oldest surviving drivers to the DMA API
(should only affect m68k).
The only major core change is the rework of async resume; the rest are
either completely trivial or for updating deprecated APIs"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (195 commits)
scsi: target: Remove XDWRITEREAD emulated support
scsi: megaraid: Remove the static variable initialisation
scsi: ch: Do not initialise statics to 0
scsi: ufs: core: Fix spelling mistake "Cannnot" -> "Cannot"
scsi: target: iscsi: Do not require target authentication
scsi: target: iscsi: Allow AuthMethod=None
scsi: target: iscsi: Support base64 in CHAP
scsi: target: iscsi: Add support for extended CDB AHS
scsi: ufs: dt-bindings: Add SC8280XP binding
scsi: target: iscsi: Fix clang -Wformat warnings
scsi: ufs: core: Read device property for ref clock
scsi: libsas: Resume SAS host for phy reset or enable via sysfs
scsi: hisi_sas: Modify v3 HW SATA completion error processing
scsi: hisi_sas: Relocate DMA unmap of SMP task
scsi: hisi_sas: Remove unnecessary variable to hold DMA map elements
scsi: hisi_sas: Call hisi_sas_slave_configure() from slave_configure_v3_hw()
scsi: mpi3mr: Delete a stray tab
scsi: mpi3mr: Unlock on error path
scsi: mpi3mr: Reduce VD queue depth on detecting throttling
scsi: mpi3mr: Resource Based Metering
...
|
|
Negotiated link rate needs to be updated to 'Disabled' when phy is stopped.
Link: https://lore.kernel.org/r/20220708205026.969161-1-changyuanl@google.com
Reviewed-by: Igor Pylypiv <ipylypiv@google.com>
Signed-off-by: Changyuan Lyu <changyuanl@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In commit 05c6c029a44d ("scsi: pm80xx: Increase number of supported
queues"), support for 80xx chip was improved by enabling multiple HW
queues.
In this, like other SCSI MQ HBA drivers at the time, the HW queues were not
exposed to upper layer, and instead the driver managed the queues
internally.
However, this management duplicates blk-mq code. In addition, the HW queue
management is sub-optimal for a system where the number of CPUs exceeds the
HW queues - this is because queues are selected in a round-robin fashion,
when it would be better to make adjacent CPUs submit on the same queue. And
finally, the affinity of the completion queue interrupts is not set to
mirror the cpu<->HQ queue mapping, which is suboptimal.
As such, for when MSIX is supported, expose HW queues to upper layer. We
always use queue index #0 for "internal" commands, i.e. anything which does
not come from the block layer, so omit this from the affinity spreading.
Link: https://lore.kernel.org/r/1654879602-33497-5-git-send-email-john.garry@huawei.com
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The current code is buggy in that the tags are set up after they are needed
in pm80xx_chip_init() -> pm80xx_set_sas_protocol_timer_config(). The tag
depth is earlier read in pm80xx_chip_init() -> read_main_config_table().
Add a post init callback to do the pm80xx work which needs to be done after
reading the tags. I don't see a better way to do this.
Link: https://lore.kernel.org/r/1654879602-33497-3-git-send-email-john.garry@huawei.com
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Executing driver on servers with more than 32 CPUs were faced with command
timeouts. This is because we were not geting completions for commands
submitted on IQ32 - IQ63.
Set E64Q bit to enable upper inbound and outbound queues 32 to 63 in the
MPI main configuration table.
Added 500ms delay after successful MPI initialization as mentioned in
controller datasheet.
Link: https://lore.kernel.org/r/20220411064603.668448-3-Ajish.Koshy@microchip.com
Fixes: 05c6c029a44d ("scsi: pm80xx: Increase number of supported queues")
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
When upper inbound and outbound queues 32-63 are enabled, we see upper
vectors 32-63 in interrupt service routine. We need corresponding registers
to handle masking and unmasking of these upper interrupts.
To achieve this, we use registers MSGU_ODMR_U(0x34) to mask and
MSGU_ODMR_CLR_U(0x3C) to unmask the interrupts. In these registers bit 0-31
represents interrupt vectors 32-63.
Link: https://lore.kernel.org/r/20220411064603.668448-2-Ajish.Koshy@microchip.com
Fixes: 05c6c029a44d ("scsi: pm80xx: Increase number of supported queues")
Reviewed-by: John Garry <john.garry@huawei.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Firmware expects host driver to clear scratchpad rsvd 0 register after
non-fatal error is found.
This is done when firmware raises fatal error interrupt and indicates
non-fatal error. At this point firmware updates scratchpad rsvd 0 register
with non-fatal error value. Here host has to clear the register after
reading it during non-fatal errors.
Rename:
- MSGU_HOST_SCRATCH_PAD_6 to MSGU_SCRATCH_PAD_RSVD_0
- MSGU_HOST_SCRATCH_PAD_7 to MSGU_SCRATCH_PAD_RSVD_1
Link: https://lore.kernel.org/r/20220222092618.108198-1-Ajish.Koshy@microchip.com
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Both pm8001_send_abort_all() and pm80xx_send_abort_all() are called only
for a non null device with the NCQ_READ_LOG_FLAG set, so remove the device
check on entry of these functions. Furthermore, setting the
NCQ_ABORT_ALL_FLAG device id flag and clearing the NCQ_READ_LOG_FLAG is
always done before calling these functions. Move these operations inside
the functions.
Link: https://lore.kernel.org/r/20220220031810.738362-31-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The task argument of the pm8001_ccb_task_free() function can be inferred
from the ccb argument ccb_task field. So there is no need to have this
argument. Likewise, the ccb_index argument is always equal to the ccb tag
field and is not needed either. Remove both arguments and update all call
sites. The pm8001_ccb_task_free_done() helper is also modified to match
this change.
Link: https://lore.kernel.org/r/20220220031810.738362-30-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
There is no need to pass a pointer to a struct inbound_queue_table to
pm8001_mpi_build_cmd(). Passing the start index in the inbound queue table
of the adapter is enough. This change allows avoiding the declaration of a
struct inbound_queue_table pointer (circularQ variables) in many functions,
simplifying the code.
While at it, blank lines are added i(e.g. after local variable
declarations) to make the code more readable.
Link: https://lore.kernel.org/r/20220220031810.738362-28-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Introduce the pm8001_ccb_alloc() and pm8001_ccb_free() helpers to replace
the typical code patterns:
res = pm8001_tag_alloc(pm8001_ha, &ccb_tag);
if (res)
...
ccb = &pm8001_ha->ccb_info[ccb_tag];
ccb->device = pm8001_ha_dev;
ccb->ccb_tag = ccb_tag;
ccb->task = task;
ccb->n_elem = 0;
and
ccb->task = NULL;
ccb->ccb_tag = PM8001_INVALID_TAG;
pm8001_tag_free(pm8001_ha, tag);
With the simpler function calls:
ccb = pm8001_ccb_alloc(pm8001_ha, pm8001_ha_dev, task);
if (!ccb)
...
and
pm8001_ccb_free(pm8001_ha, ccb);
The pm8001_ccb_alloc() helper ensures that all fields of the ccb info
structure for the newly allocated tag are all initialized, except the
buf_prd field. The pm8001_ccb_free() helper clears the initialized fields
and the ccb tag to ensure that iteration over the adapter ccb_info array
detects ccbs that are in use.
All call site of the pm8001_tag_alloc() function that use a ccb info
associated with an allocated tag are converted to use the new helpers.
Link: https://lore.kernel.org/r/20220220031810.738362-27-damien.lemoal@opensource.wdc.com
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In pm8001_chip_set_dev_state_req(), pm8001_chip_fw_flash_update_req(),
pm80xx_chip_phy_ctl_req() and pm8001_chip_reg_dev_req() add missing calls
to pm8001_tag_free() to free the allocated tag when pm8001_mpi_build_cmd()
fails.
Similarly, in pm8001_exec_internal_task_abort(), if the chip ->task_abort
method fails, the tag allocated for the abort request task must be
freed. Add the missing call to pm8001_tag_free().
Link: https://lore.kernel.org/r/20220220031810.738362-22-damien.lemoal@opensource.wdc.com
Reviewed-by: John Garry <john.garry@huawei.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The function pm8001_tag_alloc() determines free tags using the function
find_first_zero_bit() which can return 0 when the first bit of the bitmap
being inspected is 0. As such, tag 0 is a valid tag value that should not
be dismissed as invalid. Fix the functions pm8001_work_fn(),
mpi_sata_completion(), pm8001_mpi_task_abort_resp() and
pm8001_open_reject_retry() to not dismiss 0 tags as invalid.
The value 0xffffffff is used for invalid tags for unused ccb information
structures. Add the macro definition PM8001_INVALID_TAG to define this
value.
Link: https://lore.kernel.org/r/20220220031810.738362-20-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In mpi_set_phy_profile_req() and in pm8001_set_phy_profile_single(), if
pm8001_tag_alloc() fails to allocate a new tag, a warning message is issued
but the uninitialized tag variable is still used to build a command. Avoid
this by returning early in case of tag allocation failure.
Also make sure to always return the error code returned by
pm8001_tag_alloc() when this function fails instead of an arbitrary value.
Link: https://lore.kernel.org/r/20220220031810.738362-18-damien.lemoal@opensource.wdc.com
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In pm80xx_send_abort_all(), the n_elem field of the ccb used is not
initialized to 0. This missing initialization sometimes lead to the task
completion path seeing the ccb with a non-zero n_elem resulting in the
execution of invalid dma_unmap_sg() calls in pm8001_ccb_task_free(),
causing a crash such as:
[ 197.676341] RIP: 0010:iommu_dma_unmap_sg+0x6d/0x280
[ 197.700204] RSP: 0018:ffff889bbcf89c88 EFLAGS: 00010012
[ 197.705485] RAX: dffffc0000000000 RBX: 0000000000000000 RCX: ffffffff83d0bda0
[ 197.712687] RDX: 0000000000000002 RSI: 0000000000000000 RDI: ffff88810dffc0d0
[ 197.719887] RBP: 0000000000000000 R08: 0000000000000000 R09: ffff8881c790098b
[ 197.727089] R10: ffffed1038f20131 R11: 0000000000000001 R12: 0000000000000000
[ 197.734296] R13: ffff88810dffc0d0 R14: 0000000000000010 R15: 0000000000000000
[ 197.741493] FS: 0000000000000000(0000) GS:ffff889bbcf80000(0000) knlGS:0000000000000000
[ 197.749659] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 197.755459] CR2: 00007f16c1b42734 CR3: 0000000004814000 CR4: 0000000000350ee0
[ 197.762656] Call Trace:
[ 197.765127] <IRQ>
[ 197.767162] pm8001_ccb_task_free+0x5f1/0x820 [pm80xx]
[ 197.772364] ? do_raw_spin_unlock+0x54/0x220
[ 197.776680] pm8001_mpi_task_abort_resp+0x2ce/0x4f0 [pm80xx]
[ 197.782406] process_oq+0xe85/0x7890 [pm80xx]
[ 197.786817] ? lock_acquire+0x194/0x490
[ 197.790697] ? handle_irq_event+0x10e/0x1b0
[ 197.794920] ? mpi_sata_completion+0x2d70/0x2d70 [pm80xx]
[ 197.800378] ? __wake_up_bit+0x100/0x100
[ 197.804340] ? lock_is_held_type+0x98/0x110
[ 197.808565] pm80xx_chip_isr+0x94/0x130 [pm80xx]
[ 197.813243] tasklet_action_common.constprop.0+0x24b/0x2f0
[ 197.818785] __do_softirq+0x1b5/0x82d
[ 197.822485] ? do_raw_spin_unlock+0x54/0x220
[ 197.826799] __irq_exit_rcu+0x17e/0x1e0
[ 197.830678] irq_exit_rcu+0xa/0x20
[ 197.834114] common_interrupt+0x78/0x90
[ 197.840051] </IRQ>
[ 197.844236] <TASK>
[ 197.848397] asm_common_interrupt+0x1e/0x40
Avoid this issue by always initializing the ccb n_elem field to 0 in
pm8001_send_abort_all(), pm8001_send_read_log() and
pm80xx_send_abort_all().
Link: https://lore.kernel.org/r/20220220031810.738362-17-damien.lemoal@opensource.wdc.com
Fixes: c6b9ef5779c3 ("[SCSI] pm80xx: NCQ error handling changes")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
NCQ NON DATA is an NCQ command with the DMA_NONE DMA direction and so a
register-device-to-host-FIS response is expected for it.
However, for an IO_SUCCESS case, mpi_sata_completion() expects a
set-device-bits-FIS for any ata task with an use_ncq field true, which
includes NCQ NON DATA commands.
Fix this to correctly treat NCQ NON DATA commands as non-data by also
testing for the DMA_NONE DMA direction.
Link: https://lore.kernel.org/r/20220220031810.738362-16-damien.lemoal@opensource.wdc.com
Fixes: dbf9bfe61571 ("[SCSI] pm8001: add SAS/SATA HBA driver")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In the pm8001_chip_sata_req() and pm80xx_chip_sata_req() functions, all
tasks with a DMA direction of DMA_NONE (no data transfer) are initialized
using the ATAP value 0x04. However, NCQ NON DATA commands, while being
DMA_NONE commands are NCQ commands and need to be initialized using the
value 0x07 for ATAP, similarly to other NCQ commands.
Make sure that NCQ NON DATA command tasks are initialized similarly to
other NCQ commands by also testing the task "use_ncq" field in addition to
the DMA direction. While at it, reorganize the code into a chain of if -
else if - else to avoid useless affectations and debug messages.
Link: https://lore.kernel.org/r/20220220031810.738362-15-damien.lemoal@opensource.wdc.com
Fixes: dbf9bfe61571 ("[SCSI] pm8001: add SAS/SATA HBA driver")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
In mpi_set_phy_profile_req() and pm8001_set_phy_profile_single(), use
cpu_to_le32() to initialize the ppc_phyid field of struct
set_phy_profile_req. Furthermore, fix the definition of the reserved field
of this structure to be an array of __le32, to match the use of
cpu_to_le32() when assigning values. These changes remove several sparse
type warnings.
Link: https://lore.kernel.org/r/20220220031810.738362-13-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Make sure that the __le32 fields of struct sata_cmd are manipulated after
applying the correct endian conversion. That is, use cpu_to_le32() for
assigning values and le32_to_cpu() for consulting a field value. In
particular, make sure that the calculations for the 4G boundary check are
done using CPU endianness and *not* little endian values. With these fixes,
many sparse warnings are removed.
While at it, fix some code identation and add blank lines after variable
declarations and in some other places to make this code more readable.
Link: https://lore.kernel.org/r/20220220031810.738362-12-damien.lemoal@opensource.wdc.com
Fixes: 0ecdf00ba6e5 ("[SCSI] pm80xx: 4G boundary fix.")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Make sure that the __le32 fields of struct ssp_ini_io_start_req are
manipulated after applying the correct endian conversion. That is, use
cpu_to_le32() for assigning values and le32_to_cpu() for consulting a field
value. In particular, make sure that the calculations for the 4G boundary
check are done using CPU endianness and *not* little endian values. With
these fixes, many sparse warnings are removed.
While at it, add blank lines after variable declarations and in some other
places to make this code more readable.
Link: https://lore.kernel.org/r/20220220031810.738362-11-damien.lemoal@opensource.wdc.com
Fixes: 0ecdf00ba6e5 ("[SCSI] pm80xx: 4G boundary fix.")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
All fields of the kek_mgmt_req structure have the type __le32. So make sure
to use cpu_to_le32() to initialize them. This suppresses the sparse
warning:
warning: incorrect type in assignment (different base types)
expected restricted __le32 [addressable] [assigned] [usertype] new_curidx_ksop
got int
Link: https://lore.kernel.org/r/20220220031810.738362-10-damien.lemoal@opensource.wdc.com
Fixes: f5860992db55 ("[SCSI] pm80xx: Added SPCv/ve specific hardware functionalities and relevant changes in common files")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
All fields of the SASProtocolTimerConfig structure have the __le32 type.
As such, use cpu_to_le32() to initialize them. This change suppresses many
sparse warnings:
warning: incorrect type in assignment (different base types)
expected restricted __le32 [addressable] [usertype] pageCode
got int
Note that the check to limit the value of the STP_IDLE_TMO field is removed
as this field is initialized using the fixed (and small) value defined by
the STP_IDLE_TIME macro.
The pm8001_dbg() calls printing the values of the SASProtocolTimerConfig
structure fileds are changed to use le32_to_cpu() to present the values in
human readable form.
Link: https://lore.kernel.org/r/20220220031810.738362-9-damien.lemoal@opensource.wdc.com
Fixes: a6cb3d012b98 ("[SCSI] pm80xx: thermal, sas controller config and error handling update")
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The fields of the set_ctrl_cfg_req structure have the __le32 type, so use
cpu_to_le32() to assign them. This removes the sparse warnings:
warning: incorrect type in assignment (different base types)
expected restricted __le32
got unsigned int
Link: https://lore.kernel.org/r/20220220031810.738362-8-damien.lemoal@opensource.wdc.com
Fixes: 842784e0d15b ("pm80xx: Update For Thermal Page Code")
Fixes: f5860992db55 ("[SCSI] pm80xx: Added SPCv/ve specific hardware functionalities and relevant changes in common files")
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The declaration of the local variable destination1 in pm80xx_pci_mem_copy()
as a pointer to a u32 results in the sparse warning:
warning: incorrect type in assignment (different base types)
expected unsigned int [usertype]
got restricted __le32 [usertype]
Furthermore, the destination" argument of pm80xx_pci_mem_copy() is wrongly
declared with the const attribute.
Fix both problems by changing the type of the "destination" argument to
"__le32 *" and use this argument directly inside the pm80xx_pci_mem_copy()
function, thus removing the need for the destination1 local variable.
Link: https://lore.kernel.org/r/20220220031810.738362-6-damien.lemoal@opensource.wdc.com
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Since the sata_cmd struct is zeroed out before its fields are initialized,
there is no need for using "|=" to initialize the ncqtag_atap_dir_m
field. Using a standard assignment removes the sparse warning:
warning: invalid assignment: |=
Also, since the ncqtag_atap_dir_m field has type __le32, use cpu_to_le32()
to generate the assigned value.
Link: https://lore.kernel.org/r/20220220031810.738362-5-damien.lemoal@opensource.wdc.com
Fixes: c6b9ef5779c3 ("[SCSI] pm80xx: NCQ error handling changes")
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Pull 5.17 fixes branch into 5.18 tree to resolve a few pm8001 driver
merge conflicts.
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
This flag is now only ever set, so delete it.
This also avoids a use-after-free in the pm8001 queue path, as reported in
the following:
https://lore.kernel.org/linux-scsi/c3cb7228-254e-9584-182b-007ac5e6fe0a@huawei.com/T/#m28c94c6d3ff582ec4a9fa54819180740e8bd4cfb
https://lore.kernel.org/linux-scsi/0cc0c435-b4f2-9c76-258d-865ba50a29dd@huawei.com/
[mkp: checkpatch + two SAS_TASK_AT_INITIATOR references]
Link: https://lore.kernel.org/r/1644489804-85730-3-git-send-email-john.garry@huawei.com
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Currently a use-after-free may occur if a sas_task is aborted by the upper
layer before we handle the I/O completion in mpi_ssp_completion() or
mpi_sata_completion().
In this case, the following are the two steps in handling those I/O
completions:
- Call complete() to inform the upper layer handler of completion of
the I/O.
- Release driver resources associated with the sas_task in
pm8001_ccb_task_free() call.
When complete() is called, the upper layer may free the sas_task. As such,
we should not touch the associated sas_task afterwards, but we do so in the
pm8001_ccb_task_free() call.
Fix by swapping the complete() and pm8001_ccb_task_free() calls ordering.
Link: https://lore.kernel.org/r/1643289172-165636-4-git-send-email-john.garry@huawei.com
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
make W=1 complains of an undescribed function parameter:
drivers/scsi/pm8001/pm80xx_hwi.c:3938: warning: Function parameter or member 'circularQ' not described in 'process_one_iomb'
Fix it.
Link: https://lore.kernel.org/r/1643289172-165636-2-git-send-email-john.garry@huawei.com
Reported-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Current code handles completions for SATA devices in mpi_sata_completion()
and mpi_sata_event().
However, at the time when any SATA event happens, for almost all the event
types, the command is still in the target. It is therefore incorrect to
complete the task in sata_event().
There are some events for which we get sata_completions, some need recovery
procedure and others abort. All the tasks must be completed via
sata_completion() path.
Removed the task done related code from sata_events(). For tasks where we
don't get completions, let top layer call abort() to abort the command post
timeout.
Link: https://lore.kernel.org/r/20220124082255.86223-1-Ajish.Koshy@microchip.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Co-developed-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
According to the comment in check_fw_ready() we should not check the
IOP1_READY field in register SCRATCH_PAD_1 for 8008 or 8009 controllers.
However we check this very field in process_oq() for processing the highest
index interrupt vector. The highest interrupt vector is checked as the FW
is programmed to signal fatal errors through this irq.
Change that function to not check IOP1_READY for those mentioned
controllers, but do check ILA_READY in both cases.
The reason I assume that this was not hit earlier was because we always
allocated 64 MSI(X), and just did not pass the vector index check in
process_oq(), i.e. the handler never ran for vector index 63.
Link: https://lore.kernel.org/r/1642508105-95432-1-git-send-email-john.garry@huawei.com
Tested-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Pull more SCSI updates from James Bottomley:
"This series is all the stragglers that didn't quite make the first
merge window pull. It's mostly minor updates and bug fixes of merge
window code"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: nsp_cs: Check of ioremap return value
scsi: ufs: ufs-mediatek: Fix error checking in ufs_mtk_init_va09_pwr_ctrl()
scsi: ufs: Modify Tactive time setting conditions
scsi: efct: Remove useless DMA-32 fallback configuration
scsi: message: fusion: mptctl: Use dma_alloc_coherent()
scsi: message: fusion: mptsas: Use dma_alloc_coherent()
scsi: message: fusion: Use dma_alloc_coherent() in mptsas_exp_repmanufacture_info()
scsi: message: fusion: mptbase: Use dma_alloc_coherent()
scsi: message: fusion: Use dma_alloc_coherent() in mpt_alloc_fw_memory()
scsi: message: fusion: Remove usage of the deprecated "pci-dma-compat.h" API
scsi: megaraid: Avoid mismatched storage type sizes
scsi: hisi_sas: Remove unused variable and check in hisi_sas_send_ata_reset_each_phy()
scsi: aic79xx: Remove redundant error variable
scsi: pm80xx: Port reset timeout error handling correction
scsi: mpi3mr: Fix formatting problems in some kernel-doc comments
scsi: mpi3mr: Fix some spelling mistakes
scsi: mpt3sas: Update persistent trigger pages from sysfs interface
scsi: core: Fix scsi_mode_select() interface
scsi: aacraid: Fix spelling of "its"
scsi: qedf: Fix potential dereference of NULL pointer
|
|
Pull SCSI updates from James Bottomley:
"This series consists of the usual driver updates (ufs, pm80xx, lpfc,
mpi3mr, mpt3sas, hisi_sas, libsas) and minor updates and bug fixes.
The most impactful change is likely the switch from GFP_DMA to
GFP_KERNEL in a bunch of drivers, but even that shouldn't affect too
many people"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (121 commits)
scsi: mpi3mr: Bump driver version to 8.0.0.61.0
scsi: mpi3mr: Fixes around reply request queues
scsi: mpi3mr: Enhanced Task Management Support Reply handling
scsi: mpi3mr: Use TM response codes from MPI3 headers
scsi: mpi3mr: Add io_uring interface support in I/O-polled mode
scsi: mpi3mr: Print cable mngnt and temp threshold events
scsi: mpi3mr: Support Prepare for Reset event
scsi: mpi3mr: Add Event acknowledgment logic
scsi: mpi3mr: Gracefully handle online FW update operation
scsi: mpi3mr: Detect async reset that occurred in firmware
scsi: mpi3mr: Add IOC reinit function
scsi: mpi3mr: Handle offline FW activation in graceful manner
scsi: mpi3mr: Code refactor of IOC init - part2
scsi: mpi3mr: Code refactor of IOC init - part1
scsi: mpi3mr: Fault IOC when internal command gets timeout
scsi: mpi3mr: Display IOC firmware package version
scsi: mpi3mr: Handle unaligned PLL in unmap cmnds
scsi: mpi3mr: Increase internal cmnds timeout to 60s
scsi: mpi3mr: Do access status validation before adding devices
scsi: mpi3mr: Add support for PCIe Managed Switch SES device
...
|
|
Error handling steps were not in sequence as per the programmers
manual. Expected sequence:
- PHY_DOWN (PORT_IN_RESET)
- PORT_RESET_TIMER_TMO
- Host aborts pending I/Os
- Host deregister the device
- Host sends HW_EVENT_PHY_DOWN ACK
Previously we were sending HW_EVENT_PHY_DOWN ACK first and then deregister
the device. Fix this to use the expected sequence.
Link: https://lore.kernel.org/r/20211228111753.10802-1-Ajish.Koshy@microchip.com
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
The driver supports a "direct" mode of operation, where the SMP req frame
is directly copied into the command payload (and vice-versa for the SMP
resp).
To get at the SMP req frame data in the scatterlist the driver uses
phys_to_virt() on the DMA mapped memory dma_addr_t . This is broken, and
subsequently crashes as follows when an IOMMU is enabled:
Unable to handle kernel paging request at virtual address
ffff0000fcebfb00
...
pc : pm80xx_chip_smp_req+0x2d0/0x3d0
lr : pm80xx_chip_smp_req+0xac/0x3d0
pm80xx_chip_smp_req+0x2d0/0x3d0
pm8001_task_exec.constprop.0+0x368/0x520
pm8001_queue_command+0x1c/0x30
smp_execute_task_sg+0xdc/0x204
sas_discover_expander.part.0+0xac/0x6cc
sas_discover_root_expander+0x8c/0x150
sas_discover_domain+0x3ac/0x6a0
process_one_work+0x1d0/0x354
worker_thread+0x13c/0x470
kthread+0x17c/0x190
ret_from_fork+0x10/0x20
Code: 371806e1 910006d6 6b16033f 54000249 (38766b05)
---[ end trace b91d59aaee98ea2d ]---
note: kworker/u192:0[7] exited with preempt_count 1
Instead use kmap_atomic().
--
Difference to v1:
- use kmap_atomic() in both locations
Difference to v2:
- add whitespace around arithmetic (Damien)
Link: https://lore.kernel.org/r/1639390248-213603-1-git-send-email-john.garry@huawei.com
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Tracepoints for tracking controller and ATA commands issued and completed.
Link: https://lore.kernel.org/r/20211115215750.131696-2-changyuanl@google.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Co-developed-by: Akshat Jain <akshatzen@google.com>
Signed-off-by: Akshat Jain <akshatzen@google.com>
Signed-off-by: Changyuan Lyu <changyuanl@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Address-of operator cannot return NULL.
Link: https://lore.kernel.org/r/20211101232825.2350233-3-ipylypiv@google.com
Reviewed-by: Vishakha Channapattan <vishakhavc@google.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Phy ID is located in the least significant byte of the 4-byte field.
mpi_phy_stop_resp() already applies such mask.
Link: https://lore.kernel.org/r/20211101232825.2350233-2-ipylypiv@google.com
Reviewed-by: Vishakha Channapattan <vishakhavc@google.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
This is a follow up cleanup to the commit 924a3541eab0 ("scsi: libsas:
aic94xx: hisi_sas: mvsas: pm8001: Use dev_is_expander()")
Link: https://lore.kernel.org/r/20210929025807.646589-1-ipylypiv@google.com
Reviewed-by: Vishakha Channapattan <vishakhavc@google.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Commit 1f02beff224e ("scsi: pm80xx: Remove global lock from outbound queue
processing") introduced a lock per outbound queue. Prior to that change the
driver was using a global lock for all outbound queues.
While processing the I/O responses and events the driver takes the outbound
queue spinlock and is supposed to release it in pm8001_ccb_task_free_done()
before calling command done(). Since the older code was using a global
lock, pm8001_ccb_task_free_done() was releasing the global spin lock. The
change that split the lock per outbound queue did not consider this and
pm8001_ccb_task_free_done() was still releasing the global lock.
Link: https://lore.kernel.org/r/20210906170404.5682-3-Ajish.Koshy@microchip.com
Fixes: 1f02beff224e ("scsi: pm80xx: Remove global lock from outbound queue processing")
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
During phyup event, the firmware provides the phy_id and port_id and driver
is supposed to use these during device handle registration. Previously the
driver was using the port id value from libsas during device handle
registration. Since id can be different from the one assigned by firmware,
this can lead to wrong device registration and drives not showing up.
Use firmware assigned port id during device registration.
Link: https://lore.kernel.org/r/20210906170404.5682-2-Ajish.Koshy@microchip.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Ajish Koshy <Ajish.Koshy@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Fix kernel-doc warnings then test again, wash, rinse, find more, then
repeat more/again.
Also fix spellos, some grammar, and some punctuation.
../drivers/scsi/pm8001/pm8001_ctl.c:557: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
** pm8001_ctl_fatal_log_show - fatal error logging
../drivers/scsi/pm8001/pm8001_ctl.c:577: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
** non_fatal_log_show - non fatal error logging
../drivers/scsi/pm8001/pm8001_ctl.c:622: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
** pm8001_ctl_gsm_log_show - gsm dump collection
Link: https://lore.kernel.org/r/20210708165723.8594-1-rdunlap@infradead.org
Cc: Jack Wang <jinpu.wang@cloud.ionos.com>
Cc: linux-scsi@vger.kernel.org
Cc: "James E.J. Bottomley" <jejb@linux.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Pull SCSI updates from James Bottomley:
"This series consists of the usual driver updates (ufs, ibmvfc,
megaraid_sas, lpfc, elx, mpi3mr, qedi, iscsi, storvsc, mpt3sas) with
elx and mpi3mr being new drivers.
The major core change is a rework to drop the status byte handling
macros and the old bit shifted definitions and the rest of the updates
are minor fixes"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (287 commits)
scsi: aha1740: Avoid over-read of sense buffer
scsi: arcmsr: Avoid over-read of sense buffer
scsi: ips: Avoid over-read of sense buffer
scsi: ufs: ufs-mediatek: Add missing of_node_put() in ufs_mtk_probe()
scsi: elx: libefc: Fix IRQ restore in efc_domain_dispatch_frame()
scsi: elx: libefc: Fix less than zero comparison of a unsigned int
scsi: elx: efct: Fix pointer error checking in debugfs init
scsi: elx: efct: Fix is_originator return code type
scsi: elx: efct: Fix link error for _bad_cmpxchg
scsi: elx: efct: Eliminate unnecessary boolean check in efct_hw_command_cancel()
scsi: elx: efct: Do not use id uninitialized in efct_lio_setup_session()
scsi: elx: efct: Fix error handling in efct_hw_init()
scsi: elx: efct: Remove redundant initialization of variable lun
scsi: elx: efct: Fix spelling mistake "Unexected" -> "Unexpected"
scsi: lpfc: Fix build error in lpfc_scsi.c
scsi: target: iscsi: Remove redundant continue statement
scsi: qla4xxx: Remove redundant continue statement
scsi: ppa: Switch to use module_parport_driver()
scsi: imm: Switch to use module_parport_driver()
scsi: mpt3sas: Fix error return value in _scsih_expander_add()
...
|
|
This patch prepares for converting SAM status codes into an enum. Without
this patch converting SAM status codes into an enumeration type would
trigger complaints about enum type mismatches for the SAS code.
Link: https://lore.kernel.org/r/20210524025457.11299-2-bvanassche@acm.org
Cc: Hannes Reinecke <hare@suse.com>
Cc: Artur Paszkiewicz <artur.paszkiewicz@intel.com>
Cc: Jason Yan <yanaijie@huawei.com>
Reviewed-by: John Garry <john.garry@huawei.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
When driver is loaded after rmmod some drives are not showing up during
discovery.
SATA drives are directly attached to the controller connected phys. During
device discovery, the IDENTIFY command (qc timeout (cmd 0xec)) is timing out
during revalidation. This will trigger abort from host side and controller
successfully aborts the command and returns success. Post this successful
abort response ATA library decides to mark the disk as NODEV.
To overcome this, inside pm8001_scan_start() after phy_start() call, add get
start response and wait for few milliseconds to trigger next phy start.
This millisecond delay will give sufficient time for the controller state
machine to accept next phy start.
Link: https://lore.kernel.org/r/20210505120103.24497-1-ajish.koshy@microchip.com
Signed-off-by: Ajish Koshy <ajish.koshy@microchip.com>
Signed-off-by: Viswas G <viswas.g@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Introduce spin lock for outbound queue. With this, driver need not acquire
HBA global lock for outbound queue processing.
Link: https://lore.kernel.org/r/20210415103352.3580-9-Viswas.G@microchip.com
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Ruksar Devadi <Ruksar.devadi@microchip.com>
Signed-off-by: Ashokkumar N <Ashokkumar.N@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
Producer index(PI) outbound queue and consumer index(CI) for Outbound queue
are in DMA memory. During resume(), the stale PI and CI Values will lead to
unexpected behavior. These values should be reset to 0 during driver
reinitialization.
Link: https://lore.kernel.org/r/20210415103352.3580-8-Viswas.G@microchip.com
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Ruksar Devadi <Ruksar.devadi@microchip.com>
Signed-off-by: Ashokkumar N <Ashokkumar.N@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
When controller runs into fatal error, I/Os get stuck with no response,
handler event is defined to complete the pending I/Os (SAS task and
internal task) and also perform the cleanup for the drives.
Link: https://lore.kernel.org/r/20210415103352.3580-7-Viswas.G@microchip.com
Acked-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Ruksar Devadi <Ruksar.devadi@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Signed-off-by: Ashokkumar N <Ashokkumar.N@microchip.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
checkpatch reports the following:
ERROR: space prohibited before that ',' (ctx:WxW)
+int pm8001_mpi_general_event(struct pm8001_hba_info *pm8001_ha , void *piomb);
Remove unnecessary whitespace.
Link: https://lore.kernel.org/r/1617886593-36421-2-git-send-email-luojiaxing@huawei.com
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Luo Jiaxing <luojiaxing@huawei.com>
Signed-off-by: Jianqin Xie <xiejianqin@hisilicon.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
|
mpi_uninit_check() is not being called in an atomic context. The only
caller of mpi_uninit_check() is pm80xx_chip_soft_rst().
Callers of pm80xx_chip_soft_rst():
- pm8001_ioctl_soft_reset()
- pm8001_pci_probe()
- pm8001_pci_remove()
- pm8001_pci_suspend()
- pm8001_pci_resume()
There was a similar fix for mpi_init_check() in commit
d71023af4bec ("scsi: pm80xx: Do not busy wait in MPI init check")
Link: https://lore.kernel.org/r/20210406180534.1924345-3-ipylypiv@google.com
Reviewed-by: Vishakha Channapattan <vishakhavc@google.com>
Acked-by: Jack Wang <jinpu.wang@ionos.com>
Signed-off-by: Igor Pylypiv <ipylypiv@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
|
Linus Torvalds
Changyuan Lyu
John Garry
Ajish Koshy
Damien Le Moal
Martin K. Petersen
Igor Pylypiv
Randy Dunlap
Bart Van Assche
Viswas G
Ruksar Devadi
Luo Jiaxing