iSCSI session destruction can be arbitrarily slow, since it might require
network operations and serialization inside the SCSI layer. This patch
adds a new user event to trigger the destruction work asynchronously,
releasing the rx_queue_mutex as soon as the operation is queued and before
it is performed. This change allows other operations to run in other
sessions in the meantime, removing one of the major iSCSI bottlenecks for
us.
To prevent the session from being used after the destruction request, we
remove it immediately from the sesslist. This simplifies the locking
required during the asynchronous removal.
Link: https://lore.kernel.org/r/20200227195945.761719-1-krisman@collabora.com
Co-developed-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Co-developed-by: Khazhismel Kumykov <khazhy@google.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Signed-off-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Signed-off-by: Frank Mayhar <fmayhar@google.com>
Signed-off-by: Khazhismel Kumykov <khazhy@google.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When an SVC node goes down as part of a node reboot, its WWPNs are moved to
the remaining node. When the node is back online, its WWPNs are moved
back. The result is that the WWPN moves from one NPort_ID to another, then
back again. The ibmvfc driver was forcing the old port to be removed, but
not sending an implicit logout. When the WWPN showed up at the new
location, the PLOGI failed as there was already a login established for the
old scsi id. The patch below fixes this by ensuring we always send an
implicit logout for any scsi id associated with an rport prior to calling
fc_remote_port_delete.
Link: https://lore.kernel.org/r/1582767943-16611-1-git-send-email-brking@linux.vnet.ibm.com
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The current codebase makes use of the zero-length array language extension
to the C90 standard, but the preferred mechanism to declare variable-length
types such as these ones is a flexible array member[1][2], introduced in
C99:
struct foo {
int stuff;
struct boo array[];
};
By making use of the mechanism above, we will get a compiler warning in
case the flexible array does not occur last in the structure, which will
help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.
Also, notice that, dynamic memory allocations won't be affected by this
change:
"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]
This issue was found with the help of Coccinelle.
[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 7649773293 ("cxgb3/l2t: Fix undefined behaviour")
Link: https://lore.kernel.org/r/20200224161406.GA21454@embeddedor
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Satish Kharat <satishkh@cisco.com>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Use scsi_host_block() and scsi_host_unblock() instead of
scsi_block_requests()/scsi_unblock_requests() to block and unblock I/O.
This has the advantage that the block layer will stop sending I/O to the
adapter instead of having the SCSI midlayer requeueing I/O internally.
Link: https://lore.kernel.org/r/20200228075318.91255-10-hare@suse.de
Reviewed-by: Christoph Hellwig <hch@lst.de>
Acked-by: Balsundar P < Balsundar.P@microchip.com>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Rename the badly named function into adpt_i2o_scsi_complete(), and make it
a void function as the return value is never used. This also fixes a
potential use-after-free as the return value might be evaluated from the
command result after the command has been freed.
Link: https://lore.kernel.org/r/20200228075318.91255-2-hare@suse.de
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For N2N, fc_port struct is created during report id acquisition. At
later time, the loop resync (fabric, n2n, loop) would trigger the rest
of the login using the created fc_port struct. The loop resync logic
can trigger another fc_port allocation if the 1st allocation was not
able to execute. This patch prevents the 2nd allocation trigger.
Link: https://lore.kernel.org/r/20200226224022.24518-15-hmadhani@marvell.com
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Quinn Tran <qutran@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
commit e4e3a2ce95 ("scsi: qla2xxx: Add ability to autodetect SFP
type") takes a heavy handed approach to BPM (Buffer Plus Management)
enablement:
1) During hardware initialization, if an LR-capable transceiver is
recognized, the driver schedules a disruptive post-initialization
chip-reset (ISP-ABORT) to allow the BPM settings to be sent to the
firmware. This chip-reset will result in (short-term) path-loss to
all fc-rports and their attached SCSI devices.
2) LR-detection is triggered during any link-up event, resulting in a
refresh and potential chip-reset
Based on firmware-team guidance, upon LR-capable transceiver
recognition, the driver's hardware initialization code will now
re-execute firmware with the new BPM settings, then continue on with
driver initialization. To address the second issue, the driver
performs LR-capable detection upon the driver receiving a
transceiver-insertion asynchronous event from firmware. No short-term
path loss is needed with this new semantic.
Link: https://lore.kernel.org/r/20200226224022.24518-10-hmadhani@marvell.com
Signed-off-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Andrew Vasquez <andrewv@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch fixes the following sparse warnings:
drivers/scsi/qla2xxx/qla_mbx.c:120:21: warning: restricted pci_channel_state_t degrades to integer
drivers/scsi/qla2xxx/qla_mbx.c:120:37: warning: restricted pci_channel_state_t degrades to integer
>From include/linux/pci.h:
enum pci_channel_state {
/* I/O channel is in normal state */
pci_channel_io_normal = (__force pci_channel_state_t) 1,
/* I/O to channel is blocked */
pci_channel_io_frozen = (__force pci_channel_state_t) 2,
/* PCI card is dead */
pci_channel_io_perm_failure = (__force pci_channel_state_t) 3,
};
Link: https://lore.kernel.org/r/20200220043441.20504-4-bvanassche@acm.org
Cc: Roman Bolshakov <r.bolshakov@yadro.com>
Cc: Martin Wilck <mwilck@suse.com>
Cc: Quinn Tran <qutran@marvell.com>
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Acked-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Since the SCSI core does not reuse the tag of the SCSI command that is
being aborted by .eh_abort() before .eh_abort() has finished it is not
necessary to check from inside that callback whether or not the SCSI
command has already completed. Instead, rely on the firmware to return an
error code when attempting to abort a command that has already
completed. Additionally, rely on the firmware to return an error code when
attempting to abort an already aborted command.
In qla2x00_abort_srb(), use blk_mq_request_started() instead of
sp->completed and sp->aborted.
Link: https://lore.kernel.org/r/20200220043441.20504-2-bvanassche@acm.org
Cc: Martin Wilck <mwilck@suse.com>
Cc: Quinn Tran <qutran@marvell.com>
Reviewed-by: Daniel Wagner <dwagner@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Acked-by: Himanshu Madhani <hmadhani@marvell.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
