memory allocated by kmem_cache_alloc() should be freed using
kmem_cache_free(), not kfree().
Fixes: ad669505c4 ("scsi: target/core: Make sure that target_wait_for_sess_cmds() waits long enough")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Reviewed-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Initialise the t10_wwn vendor, model and revision defaults when a device is
allocated instead of when it's enabled. This ensures that custom vendor or
model strings set prior to enablement are not later overwritten with
default values.
The TRANSPORT_FLAG_PASSTHROUGH conditional can be dropped for the following
reasons:
- target_core_pscsi overwrites the defaults in the
pscsi_configure_device() callback.
+ the contents is then only used for ConfigFS via
$pscsi_dev/statistics/scsi_lu/vend, etc.
- target_core_user doesn't touch the defaults, nor are they used for
anything outside of ConfigFS.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The vendor_id attribute will allow for the modification of the T10 Vendor
Identification string returned in inquiry responses. Its value can be
viewed and modified via the ConfigFS path at:
target/core/$backstore/$name/wwn/vendor_id
"LIO-ORG" remains the default value, which is set when the backstore device
is enabled.
[mkp: corrected VPD page number]
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In preparation for supporting user provided vendor strings, add an extra
byte to the vendor, model and revision arrays in struct t10_wwn. This
ensures that the full INQUIRY data can be carried in the arrays along with
a null-terminator.
Change a number of array readers and writers so that they account for
explicit null-termination:
- The pscsi_set_inquiry_info() and emulate_model_alias_store() codepaths
don't currently explicitly null-terminate; fix this.
- Existing t10_wwn field dumps use for-loops which step over
null-terminators for right-padding.
+ Use printf with width specifiers instead.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
spc5r17.pdf specifies:
4.3.1 ASCII data field requirements
ASCII data fields shall contain only ASCII printable characters (i.e.,
code values 20h to 7Eh) and may be terminated with one or more ASCII null
(00h) characters. ASCII data fields described as being left-aligned
shall have any unused bytes at the end of the field (i.e., highest
offset) and the unused bytes shall be filled with ASCII space characters
(20h).
LIO currently space-pads the T10 VENDOR IDENTIFICATION and PRODUCT
IDENTIFICATION fields in the standard INQUIRY data. However, the PRODUCT
REVISION LEVEL field in the standard INQUIRY data as well as the T10 VENDOR
IDENTIFICATION field in the INQUIRY Device Identification VPD Page are
zero-terminated/zero-padded.
Fix this inconsistency by using space-padding for all of the above fields.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bryant G. Ly <bly@catalogicsoftware.com>
Reviewed-by: Lee Duncan <lduncan@suse.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Roman Bolshakov <r.bolshakov@yadro.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Due to the "make ABORT and LUN RESET handling synchronous" patch, cmd->work
is only modified from the regular command execution path and no longer
asynchronously by the code that executes task management functions. Since
the regular command execution code is sequential per command, no locking is
required to manipulate cmd->work. Hence stop protecting cmd->work
manipulations with locking.
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Instead of invoking target driver callback functions from the context that
handles an abort or LUN RESET task management function, only set the abort
flag from that context and perform the actual abort handling from the
context of the regular command processing flow. This approach has the
advantage that the task management code becomes much easier to read and to
verify since the number of potential race conditions against the command
processing flow is strongly reduced.
This patch has been tested by running the following two shell commands
concurrently for about ten minutes for both the iSCSI and the SRP target
drivers ($dev is an initiator device node connected with storage provided
by the target driver under test):
* fio with data verification enabled on a filesystem mounted on top of
$dev.
* while true; do sg_reset -d $dev; echo -n .; sleep .1; done
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The TASK ABORTED STATUS (TAS) bit is defined as follows in SAM:
"TASK_ABORTED: this status shall be returned if a command is aborted by a
command or task management function on another I_T nexus and the control
mode page TAS bit is set to one". TAS handling is spread over the target
core and the iSCSI target driver. If a LUN RESET is received, the target
core will send the TASK_ABORTED response for all commands for which such a
response has to be sent. If an ABORT TASK is received, only the iSCSI
target driver will send the TASK_ABORTED response for the commands for
which that response has to be sent. That is a bug since all target drivers
have to honor the TAS bit. Fix this by moving the code that handles TAS
from the iSCSI target driver into the target core. Additionally, if a
command has been aborted, instead of sending the TASK_ABORTED status from
the context that processes the SCSI command send it from the context of the
ABORT TMF. The core_tmr_abort_task() change in this patch causes the
CMD_T_TAS flag to be set if a TASK_ABORTED status has to be sent back to
the initiator that submitted the command. If that flag has been set
transport_cmd_finish_abort() will send the TASK_ABORTED response.
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Instead of allowing the code that aborts a SCSI command to finish before
all iSCSI data frames have been received, make that code wait until all
iSCSI data frames have been received. Introduce a new member variable in
the target driver template to communicate that information from the iSCSI
target driver to the target core. This change allows to leave out the check
whether or not it is already safe to send the TASK_ABORTED reply from
transport_send_task_abort().
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A quote from SAM-5: "The order in which task management requests are
processed is not specified by the SCSI architecture model. The SCSI
architecture model does not require in-order delivery of such task
management requests or processing by the task manager in the order
received. To guarantee the processing order of task management requests
referencing sent to a specific logical unit, an application client should
not have more than one such task management request pending to that logical
unit." This means that it is safe to use the system workqueues instead of
tmr_wq for processing TMFs. An intended side effect of this patch is that
it enables concurrent processing of TMFs.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A session must only be released after all code that accesses the session
structure has finished. Make sure that this is the case by introducing a
new command counter per session that is only decremented after the
.release_cmd() callback has finished. This patch fixes the following crash:
BUG: KASAN: use-after-free in do_raw_spin_lock+0x1c/0x130
Read of size 4 at addr ffff8801534b16e4 by task rmdir/14805
CPU: 16 PID: 14805 Comm: rmdir Not tainted 4.18.0-rc2-dbg+ #5
Call Trace:
dump_stack+0xa4/0xf5
print_address_description+0x6f/0x270
kasan_report+0x241/0x360
__asan_load4+0x78/0x80
do_raw_spin_lock+0x1c/0x130
_raw_spin_lock_irqsave+0x52/0x60
srpt_set_ch_state+0x27/0x70 [ib_srpt]
srpt_disconnect_ch+0x1b/0xc0 [ib_srpt]
srpt_close_session+0xa8/0x260 [ib_srpt]
target_shutdown_sessions+0x170/0x180 [target_core_mod]
core_tpg_del_initiator_node_acl+0xf3/0x200 [target_core_mod]
target_fabric_nacl_base_release+0x25/0x30 [target_core_mod]
config_item_release+0x9c/0x110 [configfs]
config_item_put+0x26/0x30 [configfs]
configfs_rmdir+0x3b8/0x510 [configfs]
vfs_rmdir+0xb3/0x1e0
do_rmdir+0x262/0x2c0
do_syscall_64+0x77/0x230
entry_SYSCALL_64_after_hwframe+0x49/0xbe
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Since transport_clear_lun_ref() already waits until the percpu-refcount
.release() method is called, it is not necessary to wait first until
percpu_ref_kill_and_confirm() has finished transitioning the refcount into
atomic mode. Remove the code that waits for percpu_ref_kill_and_confirm()
to complete and also the completion object that is used by that code. This
patch does not change the behavior of the SCSI target code.
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: David Disseldorp <ddiss@suse.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch does not change any functionality but avoids that sparse
complains about the queue_cmd_ring() function and its callers.
Fixes: 6fd0ce7972 ("tcmu: prep queue_cmd_ring to be used by unmap wq")
Reviewed-by: David Disseldorp <ddiss@suse.de>
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Even if we have no waiters on any of the sbitmap_queue wait states, we
still have to loop every entry to check. We do this for every IO, so
the cost adds up.
Shift a bit of the cost to the slow path, when we actually have waiters.
Wrap prepare_to_wait_exclusive() and finish_wait(), so we can maintain
an internal count of how many are currently active. Then we can simply
check this count in sbq_wake_ptr() and not have to loop if we don't
have any sleepers.
Convert the two users of sbitmap with waiting, blk-mq-tag and iSCSI.
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
iscsi_target_mod is the only LIO fabric where fabric_ops.name differs from
the fabric_ops.fabric_name string. fabric_ops.name is used when matching
target/$fabric ConfigFS create paths, so rename it .fabric_alias and
fallback to target/$fabric vs .fabric_name comparison if .fabric_alias
isn't initialised. iscsi_target_mod is the only fabric module to set
.fabric_alias . All other fabric modules rely on .fabric_name matching and
can drop the duplicate string.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
All fabrics return a const string. In all cases *except* iSCSI the
get_fabric_name() string matches fabric_ops.name.
Both fabric_ops.get_fabric_name() and fabric_ops.name are user-facing, with
the former being used for PR/ALUA state and the latter for ConfigFS
(config/target/$name), so we unfortunately need to keep both strings around
for now. Replace the useless .get_fabric_name() accessor function with a
const string fabric_name member variable.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
On write, the pi_prot_format configfs attribute invokes the device
format_prot() callback if present. Read dumps the contents of
se_dev_attrib.pi_prot_format which is always zero. Make the configfs
attribute write-only, and drop the always zero se_dev_attrib.pi_prot_format
storage.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If the fw supports returning VIN/VIVLD in FW_VI_CMD save it
in port_info structure else retrieve these from viid and save
them in port_info structure. Do the same for smt_idx from
FW_VI_MAC_CMD
Signed-off-by: Santosh Rastapur <santosh@chelsio.com>
Signed-off-by: Ganesh Goudar <ganeshgr@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The new emulate_pr backstore attribute allows for Persistent Reservation
and SCSI2 RESERVE/RELEASE support to be completely disabled. This can be
useful for scenarios such as:
- Ensuring ATS (Compare & Write) usage on recent VMware ESXi initiators.
- Allowing clustered (e.g. tcm-user) backends to block such requests,
avoiding the multi-node reservation state propagation.
When explicitly disabled, PR and RESERVE/RELEASE requests receive Invalid
Command Operation Code response sense data.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Reviewed-by: Mike Christie <mchristi@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Merge in -rc3 to resolve a few conflicts, but also to get a few
important fixes that have gone into mainline since the block
4.21 branch was forked off (most notably the SCSI queue issue,
which is both a conflict AND needed fix).
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now there's no difference between blk_put_request() and
__blk_put_request() anymore, get rid of the underscore version and
convert the few callers.
Reviewed-by: Hannes Reinecke <hare@suse.com>
Tested-by: Ming Lei <ming.lei@redhat.com>
Reviewed-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull more SCSI updates from James Bottomley:
"This is a set of minor small (and safe changes) that didn't make the
initial pull request plus some bug fixes"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: mvsas: Remove set but not used variable 'id'
scsi: qla2xxx: Remove two arguments from qlafx00_error_entry()
scsi: qla2xxx: Make sure that qlafx00_ioctl_iosb_entry() initializes 'res'
scsi: qla2xxx: Remove a set-but-not-used variable
scsi: qla2xxx: Make qla2x00_sysfs_write_nvram() easier to analyze
scsi: qla2xxx: Declare local functions 'static'
scsi: qla2xxx: Improve several kernel-doc headers
scsi: qla2xxx: Modify fall-through annotations
scsi: 3w-sas: 3w-9xxx: Use unsigned char for cdb
scsi: mvsas: Use dma_pool_zalloc
scsi: target: Don't request modules that aren't even built
scsi: target: Set response length for REPORT TARGET PORT GROUPS
Pull AFS updates from Al Viro:
"AFS series, with some iov_iter bits included"
* 'work.afs' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (26 commits)
missing bits of "iov_iter: Separate type from direction and use accessor functions"
afs: Probe multiple fileservers simultaneously
afs: Fix callback handling
afs: Eliminate the address pointer from the address list cursor
afs: Allow dumping of server cursor on operation failure
afs: Implement YFS support in the fs client
afs: Expand data structure fields to support YFS
afs: Get the target vnode in afs_rmdir() and get a callback on it
afs: Calc callback expiry in op reply delivery
afs: Fix FS.FetchStatus delivery from updating wrong vnode
afs: Implement the YFS cache manager service
afs: Remove callback details from afs_callback_break struct
afs: Commit the status on a new file/dir/symlink
afs: Increase to 64-bit volume ID and 96-bit vnode ID for YFS
afs: Don't invoke the server to read data beyond EOF
afs: Add a couple of tracepoints to log I/O errors
afs: Handle EIO from delivery function
afs: Fix TTL on VL server and address lists
afs: Implement VL server rotation
afs: Improve FS server rotation error handling
...
Pull SCSI updates from James Bottomley:
"This is mostly updates of the usual drivers: UFS, esp_scsi, NCR5380,
qla2xxx, lpfc, libsas, hisi_sas.
In addition there's a set of mostly small updates to the target
subsystem a set of conversions to the generic DMA API, which do have
some potential for issues in the older drivers but we'll handle those
as case by case fixes.
A new myrs driver for the DAC960/mylex raid controllers to replace the
block based DAC960 which is also being removed by Jens in this merge
window.
Plus the usual slew of trivial changes"
[ "myrs" stands for "MYlex Raid Scsi". Obviously. Silly of me to even
wonder. There's also a "myrb" driver, where the 'b' stands for
'block'. Truly, somebody has got mad naming skillz. - Linus ]
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (237 commits)
scsi: myrs: Fix the processor absent message in processor_show()
scsi: myrs: Fix a logical vs bitwise bug
scsi: hisi_sas: Fix NULL pointer dereference
scsi: myrs: fix build failure on 32 bit
scsi: fnic: replace gross legacy tag hack with blk-mq hack
scsi: mesh: switch to generic DMA API
scsi: ips: switch to generic DMA API
scsi: smartpqi: fully convert to the generic DMA API
scsi: vmw_pscsi: switch to generic DMA API
scsi: snic: switch to generic DMA API
scsi: qla4xxx: fully convert to the generic DMA API
scsi: qla2xxx: fully convert to the generic DMA API
scsi: qla1280: switch to generic DMA API
scsi: qedi: fully convert to the generic DMA API
scsi: qedf: fully convert to the generic DMA API
scsi: pm8001: switch to generic DMA API
scsi: nsp32: switch to generic DMA API
scsi: mvsas: fully convert to the generic DMA API
scsi: mvumi: switch to generic DMA API
scsi: mpt3sas: switch to generic DMA API
...
If, for example, I don't enable CONFIG_TCM_PSCSI, then every time I load
the target subsystem, I get an annoying
Unable to load target_core_pscsi
kernel log message. Instead let's only request_module() on things if
that code is enabled.
Signed-off-by: Roland Dreier <roland@purestorage.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In the iov_iter struct, separate the iterator type from the iterator
direction and use accessor functions to access them in most places.
Convert a bunch of places to use switch-statements to access them rather
then chains of bitwise-AND statements. This makes it easier to add further
iterator types. Also, this can be more efficient as to implement a switch
of small contiguous integers, the compiler can use ~50% fewer compare
instructions than it has to use bitwise-and instructions.
Further, cease passing the iterator type into the iterator setup function.
The iterator function can set that itself. Only the direction is required.
Signed-off-by: David Howells <dhowells@redhat.com>
COMPARE AND WRITE command execution starts with a call of
sbc_compare_and_write(). That function locks the caw_sem member in the
backend device data structure and submits a read request to the backend
driver. Upon successful completion of the read compare_and_write_callback()
gets called. That last function compares the data that has been read. If it
matches transport_complete_callback is set to compare_and_write_post and a
write request is submitted. compare_and_write_post() submits a write request
to the backend driver.
XDWRITEREAD command execution starts with sbc_execute_rw() submitting a
read to the backend device. Upon successful completion of the read the
xdreadwrite_callback() gets called. That function xors the data that has
been read with the data in the data-out buffer and stores the result in
the data-in buffer.
Call transport_complete_callback() not only if COMPARE AND WRITE fails but
also if XDWRITEREAD fails. This makes the code more systematic. Make sure
that the callback functions handle (cmd, false, NULL) argument triples fine.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The purpose of sg_alloc_table() is to allocate and initialize an
sg-list. Use that function instead of open-coding it. This patch will
make it easier to share code for caching sg-list allocations between the
SCSI and NVMe target cores.
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Cc: Nicholas Bellinger <nab@linux-iscsi.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
sess_err_stats are currently filled on NOP ping timeout, but not Data-Out
timeout. Stash details of Data-Out timeouts using a
ISCSI_SESS_ERR_CXN_TIMEOUT value for last_sess_failure_type.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Replace existing nested code blocks with helper function calls.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Events resulting in connection outages like this should be logged as
errors. Include the I_T Nexus in the message to aid path identification.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Data-Out timeouts resulting in connection outages should be logged as
errors. Include the I_T Nexus in the message to aid path identification.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Move the ISCSI_IQN_LEN definition up, so that it can be used in more
places instead of a hardcoded value.
Signed-off-by: David Disseldorp <ddiss@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
With the addition of commit 00d909a107 ("scsi: target: Make the session
shutdown code also wait for commands that are being aborted") in v4.19-rc, it
incorrectly assumes no signals will be pending for task_struct executing the
normal session shutdown and I/O quiesce code-path.
For example, iscsi-target and iser-target issue SIGINT to all kthreads as part
of session shutdown. This has been the behaviour since day one.
As-is when signals are pending with se_cmds active in se_sess->sess_cmd_list,
wait_event_interruptible_lock_irq_timeout() returns a negative number and
immediately kills the machine because of the do while (ret <= 0) loop that was
added in commit 00d909a107 to spin while backend I/O is taking any amount of
extended time (say 30 seconds) to complete.
Here's what it looks like in action with debug plus delayed backend I/O
completion:
[ 4951.909951] se_sess: 000000003e7e08fa before target_wait_for_sess_cmds
[ 4951.914600] target_wait_for_sess_cmds: signal_pending: 1
[ 4951.918015] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 0
[ 4951.921639] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 1
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 2
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 3
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 4
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 5
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 6
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 7
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 8
[ 4951.921944] wait_event_interruptible_lock_irq_timeout ret: -512 signal_pending: 1 loop count: 9
... followed by the usual RCU CPU stalls and deadlock.
There was never a case pre commit 00d909a107 where
wait_for_complete(&se_cmd->cmd_wait_comp) was able to be interrupted, so to
address this for v4.19+ moving forward go ahead and use
wait_event_lock_irq_timeout() instead so new code works with all fabric
drivers.
Also for commit 00d909a107, fix a minor regression in
target_release_cmd_kref() to only wake_up the new se_sess->cmd_list_wq only
when shutdown has actually been triggered via se_sess->sess_tearing_down.
Fixes: 00d909a107 ("scsi: target: Make the session shutdown code also wait for commands that are being aborted")
Cc: <stable@vger.kernel.org> # v4.19+
Cc: Bart Van Assche <bvanassche@acm.org>
Cc: Mike Christie <mchristi@redhat.com>
Cc: Hannes Reinecke <hare@suse.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Sagi Grimberg <sagi@grimberg.me>
Cc: Bryant G. Ly <bryantly@linux.vnet.ibm.com>
Tested-by: Nicholas Bellinger <nab@linux-iscsi.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Reviewed-by: Bryant G. Ly <bly@catalogicsoftware.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
csk leak can happen if a new TCP connection gets established after
cxgbit_accept_np() returns, to fix this leak free remaining csk in
cxgbit_free_np().
Signed-off-by: Varun Prakash <varun@chelsio.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When T10 PI is enabled on a backing device for the iblock backstore, the PI
SGL for the entire command is attached to the first bio only. This works fine
if the command is covered by a single bio, but can result in ref tag errors in
the client for the other bios in a multi-bio command, e.g.
[ 47.631236] sda: ref tag error at location 2048 (rcvd 0)
[ 47.637658] sda: ref tag error at location 4096 (rcvd 0)
[ 47.644228] sda: ref tag error at location 6144 (rcvd 0)
The command will be split into multiple bios if the number of data SG elements
exceeds BIO_MAX_PAGES (see iblock_get_bio()).
The bios may later be split again in the block layer on the host after
iblock_submit_bios(), depending on the queue limits of the backing device.
The block and SCSI layers will pass through the whole PI SGL down to the LLDD
however that first bio is split up, but the LLDD may only use the portion that
corresponds to the data length (depends on the LLDD, tested with scsi_debug).
Split the PI SGL across the bios in the command, so each bio's
bio_integrity_payload contains the protection information for the data in the
bio. Use an sg_mapping_iter to keep track of where we are in PI SGL, so we
know where to start with the next bio.
Signed-off-by: Greg Edwards <gedwards@ddn.com>
Reviewed-by: Mike Christie <mchristi@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
