When Target port transitions personality from one to another (NVMe <-->
FCP), there could be some overlap of the two where one layer is going down
while the other layer is coming up. This overlap can cause temporary I/O
error. Detect those errors/transitions and recover from them. Triggers
session tear down and allow relogin to re-drive the connection under the
following conditions:
- NVMe command error
- On PRLO + N2N (rida format 2)
Link: https://lore.kernel.org/r/20210817051315.2477-11-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For target port that register itself as both FCP + NVMe, initiator driver
will try to login one mode at a time. If the last mode did not succeed,
then driver will try the other mode.
When error is encountered, current code only flip to other mode one time
(NVMe->FCP) and remain on the last mode. Driver wrongly assumed target
port does not support PRLI NVMe, instead it was not ready to receive PRLI.
This patch will alternate back and forth on every PRLI failure until login
retry count has depleted or it is succeeded.
Link: https://lore.kernel.org/r/20210817051315.2477-10-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The abort callback gets called only when it gets posted to firmware. The
refcounting is done properly in the callback. On internal errors, the
callback is not invoked leading to a hung I/O. Fix this by having separate
error code when command gets returned from firmware.
Link: https://lore.kernel.org/r/20210817051315.2477-9-njavali@marvell.com
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently driver saves the personality type (FCP|NVMe) at the start of
first discovery of the remote device. If the remote device personality do
change over time, then qla driver needs to present that to user to decide.
Link: https://lore.kernel.org/r/20210817051315.2477-8-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For initiator mode, always do secure login when authentication app started.
Also remove redundant flags to indicate secure connection.
Link: https://lore.kernel.org/r/20210817051315.2477-7-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For EDIF + N2N to work, firmware 9.8 or later is required. The driver will
pause after PLOGI to allow app to authenticate. Once authentication
completes, app will tell driver to do PRLI.
Link: https://lore.kernel.org/r/20210817051315.2477-6-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The following hung task call trace was seen:
[ 1230.183294] INFO: task qla2xxx_wq:523 blocked for more than 120 seconds.
[ 1230.197749] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[ 1230.205585] qla2xxx_wq D 0 523 2 0x80004000
[ 1230.205636] Workqueue: qla2xxx_wq qlt_free_session_done [qla2xxx]
[ 1230.205639] Call Trace:
[ 1230.208100] __schedule+0x2c4/0x700
[ 1230.211607] schedule+0x38/0xa0
[ 1230.214769] schedule_timeout+0x246/0x2f0
[ 1230.222651] wait_for_completion+0x97/0x100
[ 1230.226921] qlt_free_session_done+0x6a0/0x6f0 [qla2xxx]
[ 1230.232254] process_one_work+0x1a7/0x360
...when device side port resets were done.
Abort threads were getting out without processing due to the "deleted"
flag check. The delete thread, meanwhile, could not proceed with a
logout (that would have cleared out pending requests) as the logout IOCB
work was not progressing. It appears like the hung qlt_free_session_done()
thread is causing the ha->wq works on hold. The qlt_free_session_done()
was hung waiting for nvme_fc_unregister_remoteport() + localport_delete cb
to be complete, which would only happen when all I/Os are released.
Fix this by allowing abort to progress until device delete is completely
done. This should make the qlt_free_session_done() proceed without hang and
thus clear up the deadlock.
Link: https://lore.kernel.org/r/20210817051315.2477-5-njavali@marvell.com
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
edif_enabled is prematurely turned on if hardware is capable of handling
the feature. However, firmware also needs to support EDIF before enabling
this bit.
Link: https://lore.kernel.org/r/20210817051315.2477-4-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Reject inflight AUTH ELS if driver is going through session recovery.
Link: https://lore.kernel.org/r/20210817051315.2477-3-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When firmware indicates session has been torn down via UPDATE SA IOCB or
ELS Passthrough IOCB, the driver needs to also tear down the session.
Link: https://lore.kernel.org/r/20210817051315.2477-2-njavali@marvell.com
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The sd_spinup_disk() function logs what is happening. Unfortunately this
output stops if the media was marked as removed in the meantime. Add a
print for this case too.
Link: https://lore.kernel.org/r/CWXP265MB26803209FD08A64222EEEA02C4FD9@CWXP265MB2680.GBRP265.PROD.OUTLOOK.COM
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Christian Loehle <cloehle@hyperstone.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The initial device scan might take some time, and there really is no need
to wait for it during probe(). So return immediately from scsi_scan_host()
during probe() and avoid any udev stalls during booting.
Link: https://lore.kernel.org/r/20210817075306.11315-1-mwilck@suse.com
Acked-by: Tyrel Datwyler <tyreld@linux.ibm.com>
Signed-off-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin Wilck <mwilck@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
TCM fails to pass the following tests in libiscsi:
SCSI.ExtendedCopy.DescrType
SCSI.ExtendedCopy.DescrLimits
SCSI.ExtendedCopy.ParamHdr
SCSI.ExtendedCopy.ValidSegDescr
SCSI.ExtendedCopy.ValidTgtDescr
The xcopy code always returns the same NOT READY sense key for all detected
errors. Change the sense key for invalid requests to ILLEGAL REQUEST, and
for aborted transfers to COPY ABORTED.
Link: https://lore.kernel.org/r/20210803145410.80147-3-s.samoylenko@yadro.com
Fixes: d877d7275b ("target: Fix a deadlock between the XCOPY code and iSCSI session shutdown")
Reviewed-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Reviewed-by: Konstantin Shelekhin <k.shelekhin@yadro.com>
Signed-off-by: Sergey Samoylenko <s.samoylenko@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently, backend drivers can fail I/O with SAM_STAT_CHECK_CONDITION which
gets us TCM_LOGICAL_UNIT_COMMUNICATION_FAILURE.
Add a new helper that allows backend drivers to fail with specific sense
codes.
This is based on a patch from Mike Christie <michael.christie@oracle.com>.
Cc: Mike Christie <michael.christie@oracle.com>
Link: https://lore.kernel.org/r/20210803145410.80147-2-s.samoylenko@yadro.com
Reviewed-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Sergey Samoylenko <s.samoylenko@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The return value of transport_kmap_data_sg() is assigned to the variable
buf:
buf = transport_kmap_data_sg(cmd);
And then it is checked:
if (!buf) {
This indicates that buf can be NULL. However, it is dereferenced in the
following statements:
if (!(buf[3] & 0x80))
buf[3] |= 0x80;
if (!(buf[2] & 0x80))
buf[2] |= 0x80;
To fix these possible null-pointer dereferences, dereference buf and call
transport_kunmap_data_sg() only when buf is not NULL.
Link: https://lore.kernel.org/r/20210810040414.248167-1-islituo@gmail.com
Reported-by: TOTE Robot <oslab@tsinghua.edu.cn>
Reviewed-by: Bodo Stroesser <bostroesser@gmail.com>
Signed-off-by: Tuo Li <islituo@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
It is never read, so get rid of it.
Link: https://lore.kernel.org/r/1628862553-179450-4-git-send-email-john.garry@huawei.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
It is never read. Setting it and the request tag seems dodgy anyway.
Link: https://lore.kernel.org/r/1628862553-179450-3-git-send-email-john.garry@huawei.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Use scsi_cmd_to_rq(cmd)->tag instead.
Link: https://lore.kernel.org/r/1628862553-179450-2-git-send-email-john.garry@huawei.com
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: John Garry <john.garry@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This driver has some left over "return 1" on failure style code mixed with
"return negative error codes" style code. The caller doesn't care so we
should just convert everything to return negative error codes.
Then there was a problem that there were two variables used to store error
codes which just resulted in confusion. If qedf_alloc_bdq() returned a
negative error code, we accidentally returned success instead of
propagating the error code. So get rid of the "rc" variable and use
"status" every where.
Also remove the "status = 0" initialization so that these sorts of bugs
will be detected by the compiler in the future.
Link: https://lore.kernel.org/r/20210810085023.GA23998@kili
Fixes: 61d8658b4a ("scsi: qedf: Add QLogic FastLinQ offload FCoE driver framework.")
Acked-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This function had some left over code that returned 1 on error instead
negative error codes. Convert everything to use negative error codes. The
caller treats all non-zero returns the same so this does not affect run
time.
A couple places set "rc" instead of "status" so those error paths ended up
returning success by mistake. Get rid of the "rc" variable and use
"status" everywhere.
Remove the bogus "status = 0" initialization, as a future proofing measure
so the compiler will warn about uninitialized error codes.
Link: https://lore.kernel.org/r/20210810084753.GD23810@kili
Fixes: ace7f46ba5 ("scsi: qedi: Add QLogic FastLinQ offload iSCSI driver framework.")
Acked-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Return -EINVAL on failure instead of success.
Link: https://lore.kernel.org/r/20210810084613.GB23810@kili
Fixes: a91aaae024 ("scsi: smartpqi: allow for larger raid maps")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Avoid allocating firmware dump and only allocate a single queue for a kexec
kernel.
Link: https://lore.kernel.org/r/20210810043720.1137-12-njavali@marvell.com
Cc: stable@vger.kernel.org
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Suppress logging of retryable errors. These can still be seen if extended
logging is enabled.
Link: https://lore.kernel.org/r/20210810043720.1137-11-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When user creates multiple NPIVs, the switch capabilities field is checked
before a vport is allowed to be created. This field is being toggled if a
switch scan is in progress. This creates erroneous reject of vport create.
Link: https://lore.kernel.org/r/20210810043720.1137-10-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Over time, fcport->port_type became a flag field. The flags within this
field were not defined properly. This caused external tools to read wrong
info.
Link: https://lore.kernel.org/r/20210810043720.1137-8-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
To be consistent with other OS drivers, register OS name and version in
FDMI-1 fabric registration.
Link: https://lore.kernel.org/r/20210810043720.1137-6-njavali@marvell.com
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Arun Easi <aeasi@marvell.com>
Signed-off-by: Nilesh Javali <njavali@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Driver fastpath employs doorbells to indicate to the device that work is
available. Each doorbell translates to a message sent to the device over
PCI. These messages are queued by the doorbell queue HW block, and handled
by the HW.
If a sufficient amount of CPU cores are sending messages at a sufficient
rate, the queue can overflow, and messages can be dropped. There are many
entities in the driver which can send doorbell messages. When overflow
happens, a fatal HW attention is indicated, and the Doorbell HW block stops
accepting new doorbell messages until recovery procedure is done.
When overflow occurs, all doorbells are dropped. Since doorbells are
aggregatives, if more doorbells are sent nothing has to be done. But if
the "last" doorbell is dropped, the doorbelling entity doesn’t know this
happened, and may wait forever for the device to perform the action. The
doorbell recovery mechanism addresses just that - it sends the last
doorbell of every entity.
[mkp: fix missing brackets reported by Guenter Roeck]
Link: https://lore.kernel.org/r/20210804221412.5048-1-smalin@marvell.com
Co-developed-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Manish Rangankar <mrangankar@marvell.com>
Signed-off-by: Shai Malin <smalin@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Since all scsi_cmnd.request users are gone, remove the request pointer
from struct scsi_cmnd.
Link: https://lore.kernel.org/r/20210809230355.8186-53-bvanassche@acm.org
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Prepare for removal of the request pointer by using scsi_cmd_to_rq()
instead. This patch does not change any functionality.
Link: https://lore.kernel.org/r/20210809230355.8186-52-bvanassche@acm.org
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Prepare for removal of the request pointer by using scsi_cmd_to_rq()
instead. This patch does not change any functionality.
Link: https://lore.kernel.org/r/20210809230355.8186-51-bvanassche@acm.org
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Prepare for removal of the request pointer by using scsi_cmd_to_rq()
instead. This patch does not change any functionality.
Link: https://lore.kernel.org/r/20210809230355.8186-50-bvanassche@acm.org
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Prepare for removal of the request pointer by using scsi_cmd_to_rq()
instead. This patch does not change any functionality.
Link: https://lore.kernel.org/r/20210809230355.8186-49-bvanassche@acm.org
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
