This patch adds support for an alternate I/O path in the scsi midlayer
which uses the blk-mq infrastructure instead of the legacy request code.

Use of blk-mq is fully transparent to drivers, although for now a host
template field is provided to opt out of blk-mq usage in case any unforseen
incompatibilities arise.

In general replacing the legacy request code with blk-mq is a simple and
mostly mechanical transformation.  The biggest exception is the new code
that deals with the fact the I/O submissions in blk-mq must happen from
process context, which slightly complicates the I/O completion handler.
The second biggest differences is that blk-mq is build around the concept
of preallocated requests that also include driver specific data, which
in SCSI context means the scsi_cmnd structure.  This completely avoids
dynamic memory allocations for the fast path through I/O submission.

Due the preallocated requests the MQ code path exclusively uses the
host-wide shared tag allocator instead of a per-LUN one.  This only
affects drivers actually using the block layer provided tag allocator
instead of their own.  Unlike the old path blk-mq always provides a tag,
although drivers don't have to use it.

For now the blk-mq path is disable by defauly and must be enabled using
the "use_blk_mq" module parameter.  Once the remaining work in the block
layer to make blk-mq more suitable for slow devices is complete I hope
to make it the default and eventually even remove the old code path.

Based on the earlier scsi-mq prototype by Nicholas Bellinger.

Thanks to Bart Van Assche and Robert Elliot for testing, benchmarking and
various sugestions and code contributions.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Blk-mq drivers usually preallocate their S/G list as part of the request,
but if we want to support the very large S/G lists currently supported by
the SCSI code that would tie up a lot of memory in the preallocated request
pool.  Add support to the scatterlist code so that it can initialize a
S/G list that uses a preallocated first chunks and dynamically allocated
additional chunks.  That way the scsi-mq code can preallocate a first
page worth of S/G entries as part of the request, and dynamically extend
the S/G list when needed.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Replace the calls to the various blk_end_request variants with opencode
equivalents.  Blk-mq is using a model that gives the driver control
between the bio updates and the actual completion, and making the old
code follow that same model allows us to keep the code more similar for
both paths.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

This saves us an atomic operation for each I/O submission and completion
for the usual case where the driver doesn't set a per-target can_queue
value.  Only a few iscsi hardware offload drivers set the per-target
can_queue value at the moment.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Seems like these counters are missing any sort of synchronization for
updates, as a over 10 year old comment from me noted.  Fix this by
using atomic counters, and while we're at it also make sure they are
in the same cacheline as the _busy counters and not needlessly stored
to in every I/O completion.

With the new model the _busy counters can temporarily go negative,
so all the readers are updated to check for > 0 values.  Longer
term every successful I/O completion will reset the counters to zero,
so the temporarily negative values will not cause any harm.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Avoid taking the queue_lock to check the per-device queue limit.  Instead
we do an atomic_inc_return early on to grab our slot in the queue,
and if necessary decrement it after finishing all checks.

Unlike the host and target busy counters this doesn't allow us to avoid the
queue_lock in the request_fn due to the way the interface works, but it'll
allow us to prepare for using the blk-mq code, which doesn't use the
queue_lock at all, and it at least avoids a queue_lock round trip in
scsi_device_unbusy, which is still important given how busy the queue_lock
is.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Avoid taking the host-wide host_lock to check the per-host queue limit.
Instead we do an atomic_inc_return early on to grab our slot in the queue,
and if necessary decrement it after finishing all checks.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Avoid taking the host-wide host_lock to check the per-target queue limit.
Instead we do an atomic_inc_return early on to grab our slot in the queue,
and if necessary decrement it after finishing all checks.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Prepare for not taking a host-wide lock in the dispatch path by pushing
the lock down into the places that actually need it.  Note that this
patch is just a preparation step, as it will actually increase lock
roundtrips and thus decrease performance on its own.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

The blk-mq code path will set this to a different function, so make the
code simpler by setting it up in a legacy-request specific place.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Make sure we only have the logic for requeing commands in one place.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Factor out a helper to set the _blocked values, which we'll reuse for the
blk-mq code path.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NWebb Scales <webbnh@hp.com>
Acked-by: NJens Axboe <axboe@kernel.dk>
Tested-by: NBart Van Assche <bvanassche@acm.org>
Tested-by: NRobert Elliott <elliott@hp.com>

Factor out command setup code that will be shared with the blk-mq code path.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NWebb Scales <webbnh@hp.com>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>

The data direction fiel in the SCSI command is derived only from the block
request structure.  Move setting it up into common code instead of
duplicating it in the ULDs.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>

We should call the device handler prep_fn for all TYPE_FS requests,
not just simple read/write calls that are handled by the disk driver.

Restructure the common I/O code to call the prep_fn handler and zero
out the CDB, and just leave the call to scsi_init_io to the ULDs.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>

scsi_init_io should only be called for requests that transfer data,
so move the assert that a request has segments from the callers into
scsi_init_io.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>

During IO with fabric faults, one generally sees several "Unhandled error
code" messages in the syslog as shown below:

sd 4:0:6:2: [sdbw] Unhandled error code
sd 4:0:6:2: [sdbw] Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK
sd 4:0:6:2: [sdbw] CDB: Read(10): 28 00 00 00 00 00 00 00 08 00
end_request: I/O error, dev sdbw, sector 0

This comes from scsi_io_completion (in scsi_lib.c) while handling error
codes other than DID_RESET or not deferred sense keys i.e. this is
actually handled by the SCSI mid layer. But what gets displayed here is
"Unhandled error code" which is quite misleading as it indicates
something that is not addressed by the mid layer.

The description string is based on the sense key and sometimes on the
additional sense code;
since the ACTION_FAIL case always prints the sense key and the
additional sense code, this patch removes the description string
completely because it does not add useful information.
Signed-off-by: NMaurizio Lombardi <mlombard@redhat.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NPaolo Bonzini <pbonzini@redhat.com>
Reviewed-by: NHannes Reinecke <hare@suse.de>

Using dev_printk variants prefixes the logging message with
the originating device, which makes debugging easier.
Signed-off-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NChristoph Hellwig <hch@lst.de>

Flush commands don't transfer data and thus need to be special cased
in the I/O completion handler so that we can propagate errors to
the block layer and filesystem.
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>
Reported-by: NSteven Haber <steven@qumulo.com>
Tested-by: NSteven Haber <steven@qumulo.com>
Reviewed-by: NMartin K. Petersen <martin.petersen@oracle.com>
Cc: stable@vger.kernel.org
Signed-off-by: NChristoph Hellwig <hch@lst.de>

With the optimizations around not clearing the full request at alloc
time, we are leaving some of the needed init for REQ_TYPE_BLOCK_PC
up to the user allocating the request.

Add a blk_rq_set_block_pc() that sets the command type to
REQ_TYPE_BLOCK_PC, and properly initializes the members associated
with this type of request. Update callers to use this function instead
of manipulating rq->cmd_type directly.

Includes fixes from Christoph Hellwig <hch@lst.de> for my half-assed
attempt.
Signed-off-by: NJens Axboe <axboe@fb.com>

Instead of letting the ULD play games with the prep_fn move back to
the model of a central prep_fn with a callback to the ULD.  This
already cleans up and shortens the code by itself, and will be required
to properly support blk-mq in the SCSI midlayer.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NNicholas Bellinger <nab@linux-iscsi.org>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>
Reviewed-by: NHannes Reinecke <hare@suse.de>

By folding scsi_end_request into its only caller we can significantly clean
up the completion logic.  We can use simple goto labels now to only have
a single place to finish or requeue command there instead of the previous
convoluted logic.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NNicholas Bellinger <nab@linux-iscsi.org>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>
Reviewed-by: NHannes Reinecke <hare@suse.de>

Instead of trying to guess when we have a BIDI buffer in scsi_release_buffers
add a function to explicitly free the BIDI ressoures in the one place that
handles them.  This avoids needing a special __scsi_release_buffers for the
case where we already have freed the request as well.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>

We're seeing a case where the contents of scmd->result isn't being reset after
a SCSI command encounters an error, is resubmitted, times out and then gets
handled.  The error handler acts on the stale result of the previous error
instead of the timeout.  Fix this by properly zeroing the scmd->status before
the command is resubmitted.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Patch

commit 04796336
Author: Christoph Hellwig <hch@infradead.org>
Date:   Thu Feb 20 14:20:55 2014 -0800

    [SCSI] do not manipulate device reference counts in scsi_get/put_command

Introduced a use after free:I in the kill case of scsi_prep_return we have to
release our device reference, but we do this trying to reference the just
freed command.  Use the local sdev pointer instead.

Fixes: 04796336Reported-by: NJoe Lawrence <joe.lawrence@stratus.com>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Patch

commit 04796336
Author: Christoph Hellwig <hch@infradead.org>
Date:   Thu Feb 20 14:20:55 2014 -0800

    [SCSI] do not manipulate device reference counts in scsi_get/put_command

Introduced a use after free: when scsi_init_io fails we have to release our
device reference, but we do this trying to reference the just freed command.
Add a local scsi_device pointer to fix this.

Fixes: 04796336Reported-by: NSander Eikelenboom <linux@eikelenboom.it>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

This was used in the olden days, back when onions were proper
yellow. Basically it mapped to the current buffer to be
transferred. With highmem being added more than a decade ago,
most drivers map pages out of a bio, and rq->buffer isn't
pointing at anything valid.

Convert old style drivers to just use bio_data().

For the discard payload use case, just reference the page
in the bio.
Signed-off-by: NJens Axboe <axboe@fb.com>

cmd_flags in struct request is now 64 bits wide but the scsi_execute
functions truncated arguments passed to int leading to errors. Make sure
the flags parameters are u64.
Signed-off-by: NMartin K. Petersen <martin.petersen@oracle.com>
Cc: Jens Axboe <axboe@fb.com>
CC: Jan Kara <jack@suse.cz>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: NJens Axboe <axboe@fb.com>

The queue parameter is never used, just get rid of it.
Signed-off-by: NJens Axboe <axboe@fb.com>

Avoid a spurious device get/put pair by cleaning up scsi_requeue_command
and folding scsi_unprep_request into it.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Eliminate a get_device() / put_device() pair from scsi_next_command().
Both are atomic operations hence removing these slightly improves
performance.

[hch: slight changes due to different context]
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

SCSI devices may only be removed by calling scsi_remove_device().
That function must invoke blk_cleanup_queue() before the final put
of sdev->sdev_gendev. Since blk_cleanup_queue() waits for the
block queue to drain and then tears it down, scsi_request_fn cannot
be active anymore after blk_cleanup_queue() has returned and hence
the get_device()/put_device() pair in scsi_request_fn is unnecessary.
Signed-off-by: NBart Van Assche <bvanassche@acm.org>
Reviewed-by: NTejun Heo <tj@kernel.org>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Reviewed-by: NMike Christie <michaelc@cs.wisc.edu>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Many callers won't need this and we can optimize them away.  In addition
the handling in the __-prefixed variants was inconsistant to start with.

Based on an earlier patch from Bart Van Assche.

[jejb: fix kerneldoc probelm picked up by Fengguang Wu]
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

If we don't have starved devices we don't need to take the host lock
to iterate over them.  Also split the function up to be more clear.
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Reviewed-by: NHannes Reinecke <hare@suse.de>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

Currently, scsi error handling in scsi_io_completion() tries to
unconditionally requeue scsi command when device keeps some error state.
For example, UNIT_ATTENTION causes infinite retry with
action == ACTION_RETRY.
This is because retryable errors are thought to be temporary and the scsi
device will soon recover from those errors. Normally, such retry policy is
appropriate because the device will soon recover from temporary error state.

But there is no guarantee that device is able to recover from error state
immediately. Some hardware error can prevent device from recovering.

This patch adds timeout in scsi_io_completion() to avoid infinite command
retry in scsi_io_completion(). Once scsi command retry time is longer than
this timeout, the command is treated as failure.
Signed-off-by: NEiichi Tsukata <eiichi.tsukata.xh@hitachi.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>

We must use a 64-bit for this, otherwise overflowed bits get lost, and
that can result in a lower than intended value set.

Fixes: 8e0cb8a1 ("ARM: 7797/1: mmc: Use dma_max_pfn(dev) helper for bounce_limit calculations")
Fixes: 7d35496d ("ARM: 7796/1: scsi: Use dma_max_pfn(dev) helper for bounce_limit calculations")
Tested-Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NUlf Hansson <ulf.hansson@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

DMA bounce limit is the maximum direct DMA'able memory beyond which
bounce buffers has to be used to perform dma operations. SCSI driver
relies on dma_mask but its calculation is based on max_*pfn which
don't have uniform meaning across architectures. So make use of
dma_max_pfn() which is expected to return the DMAable maximum pfn
value across architectures.

Cc: linux-scsi@vger.kernel.org
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Generate a uevent when the following Unit Attention ASC/ASCQ
codes are received:

    2A/01  MODE PARAMETERS CHANGED
    2A/09  CAPACITY DATA HAS CHANGED
    38/07  THIN PROVISIONING SOFT THRESHOLD REACHED
    3F/03  INQUIRY DATA HAS CHANGED
    3F/0E  REPORTED LUNS DATA HAS CHANGED

Log kernel messages when the following Unit Attention ASC/ASCQ
codes are received that are not as specific as those above:

    2A/xx  PARAMETERS CHANGED
    3F/xx  TARGET OPERATING CONDITIONS HAVE CHANGED

Added logic to set expecting_lun_change for other LUNs on the target
after REPORTED LUNS DATA HAS CHANGED is received, so that duplicate
uevents are not generated, and clear expecting_lun_change when a
REPORT LUNS command completes, in accordance with the SPC-3
specification regarding reporting of the 3F 0E ASC/ASCQ UA.

[jejb: remove SPC3 test in scsi_report_lun_change and some docbook fixes and
       unused variable fix, both reported by Fengguang Wu]
Signed-off-by: NEwan D. Milne <emilne@redhat.com>
Signed-off-by: NJames Bottomley <JBottomley@Parallels.com>