For every IO, memory of PAGE size is allocated for handling NVMe native
PRPS. And in addition to that for every IO (chains need per IO * chain
buffer size, e.g. 38 * 128byte) amount of memory is allocated for chain
buffers.
However, at any point of time; the IO request can be for NVMe target
device (where PRP's page is used for framing PRP's) or can be for SCSI
target device (where chain buffers are used for framing chain
SGE's). This patch modifies the driver to reuse same pre-allocated PRP
page buffers as a chain buffer for IO's targeted for SCSI target
devices. No need to allocate separate buffers for chain SGE's buffers.
Suppose if the number of chain buffers need for IO doesn't fit in the
PRP Page size then driver maintain's separate buffers for those extra
chain buffers that exceeds the PRP page size. For example consider PRP
page size as 4K and chain buffer size as 128 bytes, then number of chain
buffers that can fit in PRP page is 4096/128 => 32. if the number of
chain buffer need per IO exceeds 32; for example consider number of
chains need per IO is 36 then for remaining 4 chain buffer's driver
allocates them individual.
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Introduces Chain lookup table/tracker and implements accessing chain
buffer using smid. Removed link list based access of chain buffer which
requires lock and allocated as many chains needed.
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Instead of allocating RDPQ array (This stores the address's of each RDPQ
pools) at run time, now it will be allocated once during driver load
time and same will be reused during host reset operation also (instead
of allocating & freeing this buffer on the fly during every host reset
operation) and then freed during driver unload.
Signed-off-by: Chaitra P B <chaitra.basappa@broadcom.com>
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch adds driver changes for supporting the Unified Fabric Port
(UFP). This is a new paritioning mode wherein MFW provides the set of
parameters to be used by the device such as traffic class, outer-vlan
tag value, priority type etc. Drivers receives this info via notifications
from mfw and configures the hardware accordingly.
Signed-off-by: Sudarsana Reddy Kalluru <Sudarsana.Kalluru@cavium.com>
Signed-off-by: Ariel Elior <ariel.elior@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We can rely on the dma-mapping code to handle any DMA limits that is
bigger than the ISA DMA mask for us (either using an iommu or swiotlb),
so remove setting the block layer bounce limit for anything but the
unchecked_isa_dma case, or the bouncing for highmem pages.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Trivial fix to spelling mistake in module parameter description text
[mkp: applied by hand]
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Trivial fix to spelling mistake in module parameter description text
[mkp: applied by hand]
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The sanity check on u->in_connection_align_insertion_frequency is being
performed twice and hence the first check can be removed since it is
redundant. Cleans up cppcheck warning:
drivers/scsi/ibmvscsi/ibmvscsi.c:1711: (warning) Identical inner 'if'
condition is always true.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
commit b60710ec7d ("scsi: aacraid: enable sending of TMFs from
aac_hba_send()") allows aac_hba_send() to send scsi commands, and TMF
requests, but the existing code only updates the iu_type for scsi
commands. For TMF requests we are sending an unknown iu_type to
firmware, which causes a fault.
Include iu_type prior to determining the validity of the command
Reported-by: Noah Misner <nmisner@us.ibm.com>
Fixes: b60710ec7d ("aacraid: enable sending of TMFs from aac_hba_send()")
Fixes: 423400e64d ("aacraid: Include HBA direct interface")
Tested-by: Noah Misner <nmisner@us.ibm.com>
cc: stable@vger.kernel.org
Signed-off-by: Dave Carroll <david.carroll@microsemi.com>
Reviewed-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Reviewed-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The vmw_pvscsi driver returns DID_ABORT for commands aborted internally
by the adapter, leading to the filesystem going read-only. Change the
result to DID_BUS_BUSY, causing the kernel to retry the command.
Signed-off-by: Jim Gill <jgill@vmware.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
All three instance of ->smp_handler deal with highmem backed requests
just fine.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
do_gettimeofday() is deprecated since it will stop working in 2038 on
32-bit platforms, leading to incorrect times passed to the firmware.
On 64-bit platforms the current code appears to be fine, as the
calculation passes an 8-bit century number into the firmware that can
represent times long in the future (possibly until 25599).
Using ktime_get_real_seconds() to get a 64-bit seconds value and
time64_to_tm() to convert it into the firmware format greatly simplifies
the ips timekeeping code, makes 32-bit and 64-bit behave the same way
here, and gets us closer to removing the deprecated interfaces.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
do_gettimeofday() is deprecated because of the y2038 overflow. Here, we
use the result to pass into a 32-bit field in the firmware, which still
risks an overflow, but if the firmware is written to expect unsigned
values, it can at least last until y2106, and there is not much we can
do about it.
This changes do_gettimeofday() to ktime_get_real_seconds(), which at
least simplifies the code a bit, and avoids the deprecated
interface. I'm adding a comment about the overflow to document what
happens.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In the case when the phy_mask is bitwise anded with the phy_index bit is
zero the continue statement currently jumps to the next iteration of the
while loop and phy_index is never actually incremented, potentially
causing an infinite loop if phy_index is less than SCI_MAX_PHS. Fix this
by turning the while loop into a for loop.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
new_tape_buffer() is never called in atomic context. new_tape_buffer()
is only called by st_probe(), which is only set as ".probe" in struct
scsi_driver.
Despite never getting called from atomic context, new_tape_buffer()
calls kzalloc() with GFP_ATOMIC, which does not sleep for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL, which
can sleep and improve the possibility of sucessful allocation.
This is found by a static analysis tool named DCNS written by myself.
And I also manually check it.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
st_probe() is never called in atomic context. st_probe() is only set as
".probe" in struct scsi_driver.
Despite never getting called from atomic context, st_probe() calls
kzalloc() with GFP_ATOMIC, which does not sleep for allocation.
GFP_ATOMIC is not necessary and can be replaced with GFP_KERNEL, which
can sleep and improve the possibility of sucessful allocation.
This is found by a static analysis tool named DCNS written by myself.
And I also manually check it.
Signed-off-by: Jia-Ju Bai <baijiaju1990@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
On Fujitsu ETERNUS systems, sense code ABORTED COMMAND with ASC/Q C1/01
is used to indicate temporary condition where the storage-internal path
to a target is switched from one controller to another. SCSI commands
that return with this error code must be retried unconditionally
(i.e. without the "maybe_retry" logic in scsi_decide_disposition);
otherwise dm-multipath might initiate a failover from a healthy path
e.g. for REQ_FAILFAST_DEV commands.
Introduce a new blist flag for this case.
[mkp: applied by hand]
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
EMC Symmetrix returns 'internal target error' for a variety of
conditions, most of which will be transient. So we should always retry
it, even with failfast set. Otherwise we'd get spurious path flaps with
multipath.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Warn if a device (or the user) sets blist flags which are unknown
or have been removed. This should enable us to reuse freed blist
bits in later releases.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Space for SCSI blist flags is gradually running out. Change the type to
__u64 and fix a checkpatch complaint about symbolic mode flags in
scsi_devinfo.c.
Make checkpatch happy by replacing simple_strtoul() with kstrtoull().
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Use the just introduced const_ilog2() macro to avoid sparse errors.
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This is a best effort for estimating on how busy the ring buffer is for
that channel, based on available buffer to write in percentage. It is
still possible that at the time of actual ring buffer write, the space
may not be available due to other processes may be writing at the time.
Selecting a channel based on how full it is can reduce the possibility
that a ring buffer write will fail, and avoid the situation a channel is
over busy.
Now it's possible that storvsc can use a smaller ring buffer size
(e.g. 40k bytes) to take advantage of cache locality.
Signed-off-by: Long Li <longli@microsoft.com>
Reviewed-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Since SCSI scanning occurs asynchronously, since sd_revalidate_disk() is
called from sd_probe_async() and since sd_revalidate_disk() calls
sd_zbc_read_zones() it can happen that sd_zbc_read_zones() is called
concurrently with blkdev_report_zones() and/or blkdev_reset_zones(). That can
cause these functions to fail with -EIO because sd_zbc_read_zones() e.g. sets
q->nr_zones to zero before restoring it to the actual value, even if no drive
characteristics have changed. Avoid that this can happen by making the
following changes:
- Protect the code that updates zone information with blk_queue_enter()
and blk_queue_exit().
- Modify sd_zbc_setup_seq_zones_bitmap() and sd_zbc_setup() such that
these functions do not modify struct scsi_disk before all zone
information has been obtained.
Note: since commit 055f6e18e0 ("block: Make q_usage_counter also track
legacy requests"; kernel v4.15) the request queue freezing mechanism also
affects legacy request queues.
Fixes: 89d9475610 ("sd: Implement support for ZBC devices")
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Cc: stable@vger.kernel.org # v4.16
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
scsi_io_completion() translates the sense key ILLEGAL REQUEST / ASC 0x21 into
ACTION_FAIL. That means that setting cmd->allowed to zero in sd_zbc_complete()
for this sense code / ASC combination is not necessary. Hence remove the code
that resets cmd->allowed from sd_zbc_complete().
Signed-off-by: Bart Van Assche <bart.vanassche@wdc.com>
Cc: Damien Le Moal <damien.lemoal@wdc.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.com>
Reviewed-by: Damien Le Moal <damien.lemoal@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The default already is to never bounce, so the call is a no-op.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The default already is to never bounce, so the call is a no-op.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
New combined SCSI driver for all ESP based Zorro SCSI boards for m68k Amiga.
Code largely based on board specific parts of the old drivers (blz1230.c,
blz2060.c, cyberstorm.c, cyberstormII.c, fastlane.c which were removed after
the 2.6 kernel series for lack of maintenance) with contributions by Tuomas
Vainikka (TCQ bug tests and workaround) and Finn Thain (TCQ bugfix by use of
PIO in extended message in transfer).
New Kconfig option and Makefile entries for new Amiga Zorro ESP SCSI driver
included in this patch.
Use DMA transfers wherever possible, with board-specific DMA set-up functions
copied from the old driver code. Three byte reselection messages do appear to
cause DMA timeouts. So wire up a PIO transfer routine for these
instead. esp_reselect_with_tag explicitly sets
esp->cmd_block_dma as target address for the message bytes but PIO
requires a virtual address. Substiute kernel virtual address
esp->cmd_block in PIO transfer call if DMA address is esp->cmd_block_dma
and phase is message in.
PIO code taken from mac_esp.c where the reselection timeout issue was debugged
and fixed first, with minor macro and function rename.
Signed-off-by: Michael Schmitz <schmitzmic@gmail.com>
Reviewed-by: Finn Thain <fthain@telegraphics.com.au>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Tested-by: Christian T. Steigies <cts@debian.org>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A drive being sanitized will return NOT READY / ASC 0x4 / ASCQ
0x1b ("LOGICAL UNIT NOT READY. SANITIZE IN PROGRESS").
Prevent spinning up the drive until this condition clears.
[mkp: tweaked commit message]
Signed-off-by: Mahesh Rajashekhara <mahesh.rajashekhara@microsemi.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add UFS Protocol Information Units(upiu) trace events for ufs driver,
used to trace various ufs transaction types- command, task-management
and device management.
The trace-point format is generic and can be easily adapted to trace
other upius if needed.
Currently tracing ufs transaction of type 'device management', which
this patch introduce, cannot be obtained from any other trace.
Device management transactions are used for communication with the
device such as reading and writing descriptor or attributes etc.
Signed-off-by: Ohad Sharabi <ohad.sharabi@sandisk.com>
Reviewed-by: Stanislav Nijnikov <stanislav.nijnikov@wdc.com>
Reviewed-by: Bart Van Assche <bart.vanassche@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A patch titled: "[PATCH v2] scsi_debug: implement IMMED bit" introduced
long delays to the Start stop unit (SSU) and Synchronize cache (SC)
commands when the IMMED bit is clear. This patch makes those delays
more realistic. It causes SSU to only delay when the start stop state is
changed; SC only delays when there's been a write since the previous
SC. It also reduced the SC delay from 1 second to 50 milliseconds.
Signed-off-by: Douglas Gilbert <dgilbert@interlog.com>
Tested-by: Ming Lei <ming.lei@redhat.com>
Reported-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Instead of always multicasting responses, send a unicast netlink message
directed at the correct pid. This will be needed if we ever want to
support multiple userspace processes interacting with the kernel over
iSCSI netlink simultaneously. Limitations can currently be seen if you
attempt to run multiple iscsistart commands in parallel.
We've fixed up the userspace issues in iscsistart that prevented
multiple instances from running, so now attempts to speed up booting by
bringing up multiple iscsi sessions at once in the initramfs are just
running into misrouted responses that this fixes.
Signed-off-by: Chris Leech <cleech@redhat.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The revision field is currently unused by the devinfo pattern matching
code. Combine two blacklist entries into one.
$ egrep "Generic.*Storage-SMC" /proc/scsi/device_info
'Generic' 'USB Storage-SMC' 0x402
'Generic' 'USB Storage-SMC' 0x402
[mkp: tweaked commit desc]
Cc: Hannes Reinecke <hare@suse.de>
Cc: Martin K. Petersen <martin.petersen@oracle.com>
Cc: James E.J. Bottomley <jejb@linux.vnet.ibm.com>
Cc: SCSI ML <linux-scsi@vger.kernel.org>
Signed-off-by: Xose Vazquez Perez <xose.vazquez@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Remote port disappearance/reappearances would cause a series of RSCN
events to be delivered to the driver. During the resulting GID_FT
handling, the driver clears the fc4 settings on the remote port, which
makes it skip registration. As such, the nvme associations eventually
fail and return io errors to the applications.
Correct by not clearng the nlp_fc4_types for all nodes in
lpfc_issue_gidft. Instead, when the GID_FT response is handled, clear
the nlp_fc4_types of FCP and NVME prior to evaluating the fc4_type
returned by the GID_FT response. This approach leaves "skipped" nodes
with their nlp_fc4_types intacted.
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Points referencing local port structures didn't accommodate cases where
the localport may not be registered yet.
Add NULL pointer checks to logic.
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
On tests adding and removing a remote port, calls to nvme_info would
eventually show fewer target ports discovered than were present in the
san. Additionally, the following error messages were seen:
6031 RemotePort Registration failed err: -116, DID x471301
There is a race condition that exists between the driver and the nvme
transport on remote port unregister vs the confirmed deletion. It's
possible that the driver may rediscover the remote port and reregister
the remote port before a prior unregister delete callback was made (as
it rebinded to the prior remoteport structure). However, the driver was
coded to expect the callback before seeing the remote port again thus a
new registration. The logic results in the driver having an invalid
remoteport pointer set.
Correct by tracking when waiting for the delete callback. In cases where
the ndlp remoteport pointer is updated, it is only cleared when the wait
has not been superceded by a prior registration.
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
During target-side port faults, the driver would not recover all target
port logins. This resulted in a loss of nvme device discovery.
The driver is coded to wait for all GID_FT requests to complete before
restarting discovery. A fault is seen where the outstanding GIT_FT
counts are not properly decremented, thus discovery would never
start. Another fault was found in the clearing of the gidft_inp counter
that would be skipped in this condition. And a third fault found with
lpfc_nvme_register_port that would remove a reverence on the ndlp which
then allows a node swap on a port address change to prematurely remove
the reference and release the ndlp.
The following changes are made:
- Correct the decrementing of the outstanding GID_FT counters.
- In RSCN handling, no longer zero the counter before calling to issue
another GID_FT.
- No longer remove the reference on the dlp when the ndlp->nrport value
is not yet null.
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The patch to enlarge WQ/CQ creation keys off of an adapter response that
indicates support for the larger values. Older adapters return an
incorrect response and are limited in size. Thus the adapters fail the
WQ creation steps.
Augment the WQ sizing checks with a check on the older adapter types and
limit them to the restricted sizes.
Fixes: c176ffa084 ("scsi: lpfc: Increase CQ and WQ sizes for SCSI")
Signed-off-by: Dick Kennedy <dick.kennedy@broadcom.com>
Signed-off-by: James Smart <james.smart@broadcom.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
