Command allocation is the thing that takes the longest in the main i/o
path, so check for controller lockup immediately after this to prevent
submitting commands to locked up controller as much as possible.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
In the code that translates logical drive LBAs to physical
drive LBAs if we overflow the raid map disk data array we
will get the wrong answers. We do not expect that to happen,
but best to be on the safe side and guard against it anyway.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acking controller events on controllers that do not support
it can cause such controllers to lock up.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Joe Handzik <joseph.t.handzik@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
In set_encrypt_ioaccel2() and in hpsa_scsi_ioaccel_raid_map
there were BUG_ONs that looked like this:
BUG_ON(!(dev->offload_config && dev->offload_enabled));
But, In hpsa_ack_ctlr_events() we have this,
/* Stop sending new RAID offload reqs via the IO accelerator */
scsi_block_requests(h->scsi_host);
for (i = 0; i < h->ndevices; i++)
h->dev[i]->offload_enabled = 0;
hpsa_drain_accel_commands(h);
So, we set offload_enabled = 0 for all drives, then do this
drain_accel_commands, so that means accel commands could still
be in flight, ie. perhaps having just been submitted into
hpsa_scsi_ioaccel_raid_map concurrent with ->offload_enabled
having just been set to zero.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Performance enhancement. Remove spin_locks from the driver.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Empirically, this improves performance slightly (~2% max IOPS) by
allowing cmd_alloc to remember where it left off searching for
free commands between calls instead of always starting its search
at command 0.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This means changing the allocator to reference count commands.
The reference count is now the authoritative indicator of whether a
command is allocated or not. The h->cmd_pool_bits bitmap is now
only a heuristic hint to speed up the allocation process, it is no
longer the authoritative record of allocated commands.
Since we changed the command allocator to use reference counting
as the authoritative indicator of whether a command is allocated,
fail_all_outstanding_cmds needs to use the reference count not
h->cmd_pool_bits for this purpose.
Fix hpsa_drain_accel_commands to use the reference count as the
authoritative indicator of whether a command is allocated instead of
the h->cmd_pool_bits bitmap.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
When using the ioaccel submission methods, requests destined for RAID volumes
are sometimes diverted to physical devices. The OS has no or limited
knowledge of these physical devices, so it is up to the driver to avoid
pushing the device too hard. It is better to honor the physical device queue
limit rather than making the device spew zillions of TASK SET FULL responses.
This is so that hpsa based devices support /sys/block/sdNN/device/queue_type
of simple, which lets the SCSI midlayer automatically adjust the queue_depth
based on TASK SET FULL and GOOD status.
Adjust the queue depth for a new device after it is created based on the
maximum queue depths of the physical devices that constitute the
device. This drops the maximum queue depth from .can_queue of 1024 to
something like 174 for single-drive RAID-0, 348 for two-drive RAID-1, etc.
It also adjusts for the ratio of data to parity drives.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Instead of kicking the commands all the way back to the mid
layer, use a work queue. This enables having a mechanism for
the driver to be able to resubmit the commands down the "normal"
raid path without turning off the ioaccel feature entirely
whenever an error is encountered on the ioaccel path, and
prevent excessive rescanning of devices.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Factor out the bottom part of the queuecommand function
which is the part that builds commands for submitting down
the "normal' RAID stack path of a Smart Array.
Need to factor this out to improve how commands that
were initially sent down one of the "ioaccellerated"
paths but which have some sort of error condition are
retried down the "normal" path.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The original reasoning behind doing this was faulty. An error
of some sort would be encountered, accelerated i/o would be
disabled for that logical drive, the command would be kicked
back out to the SCSI midlayer for a retry, and since i/o accelerator
mode was disabled, it would get retried down the RAID path.
However, something needs to turn ioaccellerator mode back on,
and this rescan request was what did that. However, it was racy,
and extremely bad for performance to rescan all devices, so,
don't do that.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
By not doing maintaining a list of queued commands, we can eliminate some spin
locking in the main i/o path and gain significant improvement in IOPS. Remove
the queuing code and the code that calls it; remove now-unused interrupt code;
remove DIRECT_LOOKUP_BIT.
Now that the passthru commands share the same command pool as
the main i/o path, and the total size of the pool is less than
or equal to the number of commands that will fit in the hardware
fifo, there is no need to check to see if we are exceeding the
hardware fifo's depth.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We have commands reserved for internal use.
This is laying the groundwork for removing the internal
queue of commands from the driver so that the locks that
protect that queue may be removed.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We need to reserve some commands for device rescans,
aborts, and the pass through ioctls, etc. so we cannot
give them all to the scsi mid layer.
This is in preparation for removing cmd_special_alloc and
cmd_special_free so that we can stop queuing commands internally
in the driver so that we can remove the locks thta protect the
queue that we will no longer have.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
If hpsa_allocate_cmd_pool failed, we were calling two functions unnecessarily:
hpsa_free_sg_chain_blocks(h);
hpsa_free_cmd_pool(h);
This didn't cause any problem, as those functions can tolerate being called
when what they free hasn't been allocated (relevant pointers would be NULL)
but it is potentially confusing.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Partial allocation failure wasn't handled correctly
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Return the actual error code instead of a generic error code.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Make the function name more descriptive. We use more than
one interrupt.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Enhance error reporting.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Cleanup comments to be more specific. Make messages more
informational.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Encapsulate the conditional predicate which tests for legacy controllers
in a separate function and rework the code comments.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
There is nothing worrisome about the "Waiting for controller to
respond to no-op" print, so use dev_info rather than dev_warn.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
If the board ID lookup function fails, return the return
code rather than return -ENODEV.
The only board ID failure reason right now is -ENODEV,
so this just provides more informative prints in kdump
and adapts to future changes.
Tested with error injection while booting with
reset_devices
on the kernel command line:
[ 62.804324] injecting error in inj_hpsa_lookup_board_id: 1 11
[ 62.804423] hpsa 0000:04:00.0: Board ID not found
(the pci probe layer does not print an additional
message if -ENODEV is the reason)
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Return the real reason for kdump_hard_reset failure rather
than change them all to -ENODEV.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The queue depth printed at startup is in decimal, so
shouldn't have a 0x prefix.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
In MSI and MSI-X mode, where hpsa asks for more than one interrupt,
hpsa_request_irqs forgets if the first request_irq call failed
if later ones succeed.
It needs to exit the loop on any failure rather than continue,
freeing all irqs that were requested until that point.
Also, it needs to clear out the q numbers up to MAX_REPLY_QUEUES.
The same is true for the general hpsa_free_irqs function.
Tested with error injection of -ENOSYS on the 4th call:
[ 9.277691] injecting error in inj_request_irq: 1 4
[ 9.277780] hpsa 0000:02:00.0: failed to get irq 35 for hpsa1
[ 10.711623] scsi host1: Error handler scsi_eh_1 exiting
[ 10.739170] hpsa: probe of 0000:02:00.0 failed with error -38
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Remove unused variable in hpsa_free_cmd_pool.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Fabian Frederick <fabf@skynet.be>
Acked-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Change the function names to have hpsa prefix.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
HP now uses RAID-6 rather than RAID-ADG (Advanced Data Guarding)
as the marketing name for our implementation of RAID-6.
The driver considers RAID-1 and RAID-1+0 to be the same level, and
considers RAID-1ADM and RAID-1+0ADM to be the same level. Parenthesis
can be used to reflect the optional +0 portion of both those RAID levels.
Rename: RAID-ADG to RAID-6
RAID-1(1+0) to RAID-1(+0)
RAID-1(ADM) to RAID-1(+0)ADM
Also, add another const after the pointer type as suggested
by checkpatch.pl so the array is:
static const char * const raid_label[]
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We change drive queue depths to match drive reported queue depths.
The name of the SML function was changed from scsi_adjust_queue_depth
changed to scsi_change_queue_depth.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Change how SA controllers are reset by changing PCI power levels.
The hpsa driver was finding the PCI_PM_CTRL_STATE_MASK offset
then reading/writing a bitmask to change the power state. There
are kernel functions that do the same operations. Better to use
the kernel functions.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Sometimes when the card is restarted it may cause -
"irq 16: nobody cared (try booting with the "irqpoll" option)"
that is likely caused so, that the card, after the hard reset
finishes, pulls on the irq. Disabling the ints before or after
the hpsa_kdump_hard_reset_controller fixes it.
At this point we can't know in which state the card is,
so using SA5_INTR_OFF + SA5_REPLY_INTR_MASK_OFFSET defines directly,
instead of the function the drivers provides, seems to be apropriate.
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
There is a potential memory leak in hpsa_kdump_hard_reset_controller.
Reviewed-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Correct endiness issues reported by sparse. SA controllers are
little endian. This patch ensures endiness correctness.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Drop the now unused reason argument from the ->change_queue_depth method.
Also add a return value to scsi_adjust_queue_depth, and rename it to
scsi_change_queue_depth now that it can be used as the default
->change_queue_depth implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Reviewed-by: Hannes Reinecke <hare@suse.de>
We won't ever queue more commands than the host allows. Instead of
letting drivers either reject or ignore this case handle it in
common code. Note that various driver use internal constant or
variables that are assigned to both shost->can_queue and checked
in ->change_queue_depth - I did remove those checks as well.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Reviewed-by: Hannes Reinecke <hare@suse.de>
It is already using atomic test_and_set_bit to do the
allocation.
There is some microscopic chance of starvation, but it is
so microscopic that it should never happen in reality.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
If the kernel is booted with the reset_device parameter, which
is done for kdump, then the driver needs to call pci_set_master
after pci_enable_device to reenable bus mastering (since
the preceding pci_disable_device call disables bus mastering).
Also, place that after pci_request_regions both in the
kdump code and the normal pci_init code.
Remove the comment summarizing what pci_set_master
does, with the incomplete commentary on the impact of
pci_disable_device.
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
There isn't anything in hpsa that requires the host lock to be held
during queuecommand.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Reviewed-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We were printing a lot of useless information before ultimately
just passing things up to the SCSI mid layer. Just let the
midlayer handle it without LLD chatter.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Reviewed-by: Joe Handzik <joseph.t.handzik@hp.com>
Reviewed-by: Scott Teel <scott.teel@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Use atomics for commands_outstanding instead of protecting with spin locks.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Reviewed-by: Joe Handzik <joseph.t.handzik@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Using bit fields for hardware command fields isn't portable and
relies on assumptions about how the compiler lays out the bits.
We can fix this in the driver's internal command structure, but the
ioctl interface we can't change because it is part of the
userland ABI.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Webb Scales <webb.scales@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The hardware needs little endian scatter gather addresses and
lengths but we were not bothering to convert from cpu byte
order as we should have been. On Intel, this is all just
a bunch of no-ops macros, but it makes the code endian-clean(er).
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Robert Elliott <elliott@hp.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We were allocating roughly double the amount of memory
we should be due to ReportLUNdata and ExtendedReportLUNdata
containing a non-zero sized array but adding extra memory
to allocate as if the array were zero sized.
Track the logical and physical sizes separately.
Allocate the memory based on the specific data
structure sizes.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Webb Scales <webb.scales@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
In the case of LUN data changing, the driver will
auto rescan and so it's not even true that "action" is
"required".
Remove "action required" phrases from warning messages and
replace with description phrases.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Reviewed-by: Joe Handzik <joseph.t.handzik@hp.com>
Reviewed-by: Webb Scales <webb.scales@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Correct the size calculation of the chained SG block
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Webb Scales <webbnh@hp.com>
Reviewed-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Reviewed-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Fix a couple of pci id table mistakes:
Subdevice ID 0x3323 missing from product[] table
(another name for HP Smart Storage 1210m)
Bogus 0x1925 subdevice id removed from hpsa_pci_device_id[] (no such thing.)
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
RAID-1ADM is unusable with dev_warn called on every command.
Signed-off-by: Robert Elliott <elliott@hp.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Stephen M. Cameron <stephenmcameron@gmail.com>
Reviewed-by: Webb Scales <webbnh@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Clean up issues reported when running sparse.
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Webb Scales <webb.scales@hp.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>