On a dual controller setup with multipath enabled, some MEDIUM ERRORs
caused both paths to be failed, thus I/O got queued/blocked since the
'queue_if_no_path' feature is enabled by default on IPR controllers.
This example disabled 'queue_if_no_path' so the I/O failure is seen at
the sg_dd program. Notice that after the sg_dd test-case, both paths
are in 'failed' state, and both path/priority groups are in 'enabled'
state (not 'active') -- which would block I/O with 'queue_if_no_path'.
# sg_dd if=/dev/dm-2 bs=4096 count=1 dio=1 verbose=4 blk_sgio=0
<...>
read(unix): count=4096, res=-1
sg_dd: reading, skip=0 : Input/output error
<...>
# dmesg
[...] sd 2:2:16:0: [sds] FAILED Result: hostbyte=DID_OK driverbyte=DRIVER_SENSE
[...] sd 2:2:16:0: [sds] Sense Key : Medium Error [current]
[...] sd 2:2:16:0: [sds] Add. Sense: Unrecovered read error - recommend rewrite the data
[...] sd 2:2:16:0: [sds] CDB: Read(10) 28 00 00 00 00 00 00 00 20 00
[...] blk_update_request: I/O error, dev sds, sector 0
[...] device-mapper: multipath: Failing path 65:32.
<...>
[...] device-mapper: multipath: Failing path 65:224.
# multipath -l
1IBM_IPR-0_59C2AE0000001F80 dm-2 IBM ,IPR-0 59C2AE00
size=5.2T features='0' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=0 status=enabled
| `- 2:2:16:0 sds 65:32 failed undef running
`-+- policy='service-time 0' prio=0 status=enabled
`- 1:2:7:0 sdae 65:224 failed undef running
This is not the desired behavior. The dm-multipath explicitly checks
for the MEDIUM ERROR case (and a few others) so not to fail the path
(e.g., I/O to other sectors could potentially happen without problems).
See dm-mpath.c :: do_end_io_bio() -> noretry_error() !->! fail_path().
The problem trace is:
1) ipr_scsi_done() // SENSE KEY/CHECK CONDITION detected, go to..
2) ipr_erp_start() // ipr_is_gscsi() and masked_ioasc OK, go to..
3) ipr_gen_sense() // masked_ioasc is IPR_IOASC_MED_DO_NOT_REALLOC,
// so set DID_PASSTHROUGH.
4) scsi_decide_disposition() // check for DID_PASSTHROUGH and return
// early on, faking a DID_OK.. *instead*
// of reaching scsi_check_sense().
// Had it reached the latter, that would
// set host_byte to DID_MEDIUM_ERROR.
5) scsi_finish_command()
6) scsi_io_completion()
7) __scsi_error_from_host_byte() // That would be converted to -ENODATA
<...>
8) dm_softirq_done()
9) multipath_end_io()
10) do_end_io()
11) noretry_error() // And that is checked in dm-mpath :: noretry_error()
// which would cause fail_path() not to be called.
With this patch applied, the I/O is failed but the paths are not. This
multipath device continues accepting more I/O requests without blocking.
(and notice the different host byte/driver byte handling per SCSI layer).
# dmesg
[...] sd 2:2:7:0: [sdaf] Done: SUCCESS Result: hostbyte=0x13 driverbyte=DRIVER_OK
[...] sd 2:2:7:0: [sdaf] CDB: Read(10) 28 00 00 00 00 00 00 00 40 00
[...] sd 2:2:7:0: [sdaf] Sense Key : Medium Error [current]
[...] sd 2:2:7:0: [sdaf] Add. Sense: Unrecovered read error - recommend rewrite the data
[...] blk_update_request: critical medium error, dev sdaf, sector 0
[...] blk_update_request: critical medium error, dev dm-6, sector 0
[...] sd 2:2:7:0: [sdaf] Done: SUCCESS Result: hostbyte=0x13 driverbyte=DRIVER_OK
[...] sd 2:2:7:0: [sdaf] CDB: Read(10) 28 00 00 00 00 00 00 00 10 00
[...] sd 2:2:7:0: [sdaf] Sense Key : Medium Error [current]
[...] sd 2:2:7:0: [sdaf] Add. Sense: Unrecovered read error - recommend rewrite the data
[...] blk_update_request: critical medium error, dev sdaf, sector 0
[...] blk_update_request: critical medium error, dev dm-6, sector 0
[...] Buffer I/O error on dev dm-6, logical block 0, async page read
# multipath -l 1IBM_IPR-0_59C2AE0000001F80
1IBM_IPR-0_59C2AE0000001F80 dm-6 IBM ,IPR-0 59C2AE00
size=5.2T features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
|-+- policy='service-time 0' prio=0 status=active
| `- 2:2:7:0 sdaf 65:240 active undef running
`-+- policy='service-time 0' prio=0 status=enabled
`- 1:2:7:0 sdh 8:112 active undef running
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch fixes a hang that can occur in ATA EH with ipr. With ipr's
usage of libata, commands should never end up on ap->eh_done_q. The
timeout function we use for ipr, even for SATA devices, is
scsi_times_out, so ATA_QCFLAG_EH_SCHEDULED never gets set for ipr and EH
is driven completely by ipr and SCSI. The SCSI EH thread ends up calling
ipr's eh_device_reset_handler, which then calls
ata_std_error_handler. This ends up calling ipr_sata_reset, which issues
a reset to the device. This should result in all pending commands
getting failed back and having ata_qc_complete called for them, which
should end up clearing ATA_QCFLAG_FAILED as qc->flags gets zeroed in
ata_qc_free. This ensures that when we end up in ata_eh_finish, we
don't do anything more with the command.
On adapters that only support a single interrupt and when running with
two MSI-X vectors or less, the adapter firmware guarantees that
responses to all outstanding commands are sent back prior to sending the
response to the SATA reset command. On newer adapters supporting
multiple HRRQs, however, this can no longer be guaranteed, since the
command responses and reset response may be processed on different
HRRQs.
If ipr returns from ipr_sata_reset before the outstanding command was
returned, this sends us down the path of __ata_eh_qc_complete which then
moves the associated scsi_cmd from the work_q in
scsi_eh_bus_device_reset to ap->eh_done_q, which then will sit there
forever and we will be wedged.
This patch fixes this up by ensuring that any outstanding commands are
flushed before returning from eh_device_reset_handler for a SATA device.
Reported-by: David Jeffery <djeffery@redhat.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch closes up some potential race conditions observed in the
error handling paths in ipr while debugging an issue resulting in a hang
with SATA error handling. These patches ensure we are holding the
correct lock when adding and removing commands from the free and pending
queues in some error scenarios.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This fixes a race condition in the error handlomg paths of ipr. While a
command is outstanding to the adapter, it is placed on a pending queue
for the hrrq it is associated with, while holding the HRRQ lock. When a
command is completed, it is removed from the pending queue, under HRRQ
lock, and placed on a local list. This list is then iterated through
without any locks and each command's done function is invoked, inside of
which, the command gets returned to the free list while grabbing the
HRRQ lock. This fixes two race conditions when commands have been
removed from the pending list but have not yet been added to the free
list. Both of these changes fix race conditions that could result in
returning success from eh_abort_handler and then later calling scsi_done
for the same request.
The first race condition is in ipr_cancel_op. It looks through each
pending queue to see if the command to be aborted is still outstanding
or not. Rather than looking on the pending queue, reverse the logic to
check to look for commands that are NOT on the free queue. The second
race condition can occur when in ipr_wait_for_ops where we are waiting
for responses for commands we've aborted.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Removes some code in __ipr_eh_dev_reset which was modifying the ipr_cmd
done function. This should have already been setup at command allocation
time and if its since been changed, it means we are in the ipr_erp*
functions and need to wait for them to complete and don't want to
override that here.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Following a command abort or device reset, ipr's EH handlers wait for
the commands getting aborted to get sent back from the adapter prior to
returning from the EH handler. This fixes up some cases where the
completion handler was not getting called, which would have resulted in
the EH thread waiting until it timed out, greatly extending EH time.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Fix typos and add the following to the scripts/spelling.txt:
intialization||initialization
The "inintialization" in drivers/acpi/spcr.c is a different pattern but
I fixed it as well in this commit.
Link: http://lkml.kernel.org/r/1481573103-11329-16-git-send-email-yamada.masahiro@socionext.com
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
LSIs must be ack'ed with an MMIO otherwise they remain asserted
forever. This is controlled by the "clear_isr" flag.
While we set that flag properly when deciding initially whether to use
LSIs or MSIs, we fail to set it if we first chose MSIs, the test fails,
then fallback to LSIs.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Switch the ipr driver to use pci_alloc_irq_vectors. We need to two
calls to pci_alloc_irq_vectors as ipr only supports multiple MSI-X
vectors, but not multiple MSI vectors.
Otherwise this cleans up a lot of cruft and allows to use a common
request_irq loop for irq types, which happens to only iterate over a
single line in the non MSI-X case.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This update includes the usual round of major driver updates (hpsa, be2iscsi,
hisi_sas, zfcp, cxlflash). There's a new incarnation of hpsa called smartpqi
for which a driver is added, there's some cleanup work of the ibm vscsi target
and updates to libfc, plus a whole host of minor fixes and updates and finally
the removal of several ISA drivers which seem not to have been used for years.
Signed-off-by: James Bottomley <jejb@linux.vnet.ibm.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2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=cyTO
-----END PGP SIGNATURE-----
Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI updates from James Bottomley:
"This update includes the usual round of major driver updates (hpsa,
be2iscsi, hisi_sas, zfcp, cxlflash). There's a new incarnation of hpsa
called smartpqi for which a driver is added, there's some cleanup work
of the ibm vscsi target and updates to libfc, plus a whole host of
minor fixes and updates and finally the removal of several ISA drivers
which seem not to have been used for years"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (173 commits)
scsi: mvsas: Mark symbols static where possible
scsi: pm8001: Mark symbols static where possible
scsi: arcmsr: Simplify user_len checking
scsi: fcoe: fix off by one in eth2fc_speed()
scsi: dtc: remove from tree
scsi: t128: remove from tree
scsi: pas16: remove from tree
scsi: u14-34f: remove from tree
scsi: ultrastor: remove from tree
scsi: in2000: remove from tree
scsi: wd7000: remove from tree
scsi: scsi_dh_alua: Fix memory leak in alua_rtpg()
scsi: lpfc: Mark symbols static where possible
scsi: hpsa: correct call to hpsa_do_reset
scsi: ufs: Get a TM service response from the correct offset
scsi: ibmvfc: Fix I/O hang when port is not mapped
scsi: megaraid_sas: clean function declarations in megaraid_sas_base.c up
scsi: ipr: Remove redundant messages at adapter init time
scsi: ipr: Don't log unnecessary 9084 error details
scsi: smartpqi: raid bypass lba calculation fix
...
Whenever multiple HRRQs are enabled, which is the default setting now,
we end up seeing the following message logged prior to initialization of
each HRRQ:
Starting IOA initialization sequence
This results in 16 of these messages on most adapters, which serves
little purpose. Change to just log this once.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A 9084 error gets logged by the ipr adapter when adapter raw mode gets
enabled. A bunch of unformatted hex data also gets logged for this
error, which is of little use, so let's avoid logging it by default in
order to avoid the log getting polluted with useless data.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This patch implements functions for pushing HCAM (host controlled
asynchronous messages) error buffers to userspace through sysfs
attributes. Reads to the "async_err_log" attribute will result in a
single HCAM buffer being copied to userspace; one can process the next
HCAM buffer by writing any string to the same attribute.
A new list was added to the ioa_cfg structure to store the HCAM buffers
for later reporting. We also send a KOBJ_CHANGE event whenever a new
HCAM buffer is made available to userspace.
Signed-off-by: Heitor Ricardo Alves de Siqueira <halves@linux.vnet.ibm.com>
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Commit b195d5e2bf ("ipr: Wait to do async scan until scsi host is
initialized") fixed async scan for ipr, but broke sync scan for ipr.
This fixes sync scan back up.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reported-and-tested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Fix to return error code -ENOMEM from the workqueue alloc error handling
case instead of 0, as done elsewhere in this function.
Signed-off-by: Wei Yongjun <weiyj.lk@gmail.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When performing an async scan, make sure the kthread doing scanning
doesn't start before the scsi host is fully initialized.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Increase MSIX vectors from 2 to 16 by default.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add the appropriate definitions and table entries for new adapters
support.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If we fall back to using LSI on the Croc or Crocodile chip we need to
clear the interrupt so we don't hang the system.
Cc: <stable@vger.kernel.org>
Tested-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Commit d63c7dd5bc ("ipr: Fix out-of-bounds null overwrite") removed
the end of line handling when storing the update_fw sysfs attribute.
This changed the userpace API because it started refusing writes
terminated by a line feed, which broke the update tools we already have.
This patch re-adds that handling, so both a write terminated by a line
feed or not can make it through with the update.
Fixes: d63c7dd5bc ("ipr: Fix out-of-bounds null overwrite")
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Cc: Insu Yun <wuninsu@gmail.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
- Remove usage of ib_query_device and instead store attributes in
ib_device struct
- Move iopoll out of block and into lib, rename to irqpoll, and use
in several places in the rdma stack as our new completion queue
polling library mechanism. Update the other block drivers that
already used iopoll to use the new mechanism too.
- Replace the per-entry GID table locks with a single GID table lock
- IPoIB multicast cleanup
- Cleanups to the IB MR facility
- Add support for 64bit extended IB counters
- Fix for netlink oops while parsing RDMA nl messages
- RoCEv2 support for the core IB code
- mlx4 RoCEv2 support
- mlx5 RoCEv2 support
- Cross Channel support for mlx5
- Timestamp support for mlx5
- Atomic support for mlx5
- Raw QP support for mlx5
- MAINTAINERS update for mlx4/mlx5
- Misc ocrdma, qib, nes, usNIC, cxgb3, cxgb4, mlx4, mlx5 updates
- Add support for remote invalidate to the iSER driver (pushed through the
RDMA tree due to dependencies, acknowledged by nab)
- Update to NFSoRDMA (pushed through the RDMA tree due to dependencies,
acknowledged by Bruce)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQIcBAABAgAGBQJWoSygAAoJELgmozMOVy/dDjsP/2vbTda2MvQfkfkGEZBQdJSg
095RN0gQgCJdg78lAl8yuaK8r4VN/7uefpDtFdudH1I/Pei7X0wxN9R1UzFNG4KR
AD53lz92IVPs15328SbPR2kvNWISR9aBFQo3rlElq3Grqlp0EMn2Ou1vtu87rekF
aMllxr8Nl0uZhP+eWusOsYpJUUtwirLgRnrAyfqo2UxZh/TMIroT0TCx1KXjVcAg
dhDARiZAdu3OgSc6OsWqmH+DELEq6dFVA5F+DDBGAb8bFZqlJc7cuMHWInwNsNXT
so4bnEQ835alTbsdYtqs5DUNS8heJTAJP4Uz0ehkTh/uNCcvnKeUTw1c2P/lXI1k
7s33gMM+0FXj0swMBw0kKwAF2d9Hhus9UAN7NwjBuOyHcjGRd5q7SAnfWkvKx000
s9jVW19slb2I38gB58nhjOh8s+vXUArgxnV1+kTia1+bJSR5swvVoWRicRXdF0vh
TvLX/BjbSIU73g1TnnLNYoBTV3ybFKQ6bVdQW7fzSTDs54dsI1vvdHXi3bYZCpnL
HVwQTZRfEzkvb0AdKbcvf8p/TlaAHem3ODqtO1eHvO4if1QJBSn+SptTEeJVYYdK
n4B3l/dMoBH4JXJUmEHB9jwAvYOpv/YLAFIvdL7NFwbqGNsC3nfXFcmkVORB1W3B
KEMcM2we4bz+uyKMjEAD
=5oO7
-----END PGP SIGNATURE-----
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma
Pull rdma updates from Doug Ledford:
"Initial roundup of 4.5 merge window patches
- Remove usage of ib_query_device and instead store attributes in
ib_device struct
- Move iopoll out of block and into lib, rename to irqpoll, and use
in several places in the rdma stack as our new completion queue
polling library mechanism. Update the other block drivers that
already used iopoll to use the new mechanism too.
- Replace the per-entry GID table locks with a single GID table lock
- IPoIB multicast cleanup
- Cleanups to the IB MR facility
- Add support for 64bit extended IB counters
- Fix for netlink oops while parsing RDMA nl messages
- RoCEv2 support for the core IB code
- mlx4 RoCEv2 support
- mlx5 RoCEv2 support
- Cross Channel support for mlx5
- Timestamp support for mlx5
- Atomic support for mlx5
- Raw QP support for mlx5
- MAINTAINERS update for mlx4/mlx5
- Misc ocrdma, qib, nes, usNIC, cxgb3, cxgb4, mlx4, mlx5 updates
- Add support for remote invalidate to the iSER driver (pushed
through the RDMA tree due to dependencies, acknowledged by nab)
- Update to NFSoRDMA (pushed through the RDMA tree due to
dependencies, acknowledged by Bruce)"
* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dledford/rdma: (169 commits)
IB/mlx5: Unify CQ create flags check
IB/mlx5: Expose Raw Packet QP to user space consumers
{IB, net}/mlx5: Move the modify QP operation table to mlx5_ib
IB/mlx5: Support setting Ethernet priority for Raw Packet QPs
IB/mlx5: Add Raw Packet QP query functionality
IB/mlx5: Add create and destroy functionality for Raw Packet QP
IB/mlx5: Refactor mlx5_ib_qp to accommodate other QP types
IB/mlx5: Allocate a Transport Domain for each ucontext
net/mlx5_core: Warn on unsupported events of QP/RQ/SQ
net/mlx5_core: Add RQ and SQ event handling
net/mlx5_core: Export transport objects
IB/mlx5: Expose CQE version to user-space
IB/mlx5: Add CQE version 1 support to user QPs and SRQs
IB/mlx5: Fix data validation in mlx5_ib_alloc_ucontext
IB/sa: Fix netlink local service GFP crash
IB/srpt: Remove redundant wc array
IB/qib: Improve ipoib UD performance
IB/mlx4: Advertise RoCE v2 support
IB/mlx4: Create and use another QP1 for RoCEv2
IB/mlx4: Enable send of RoCE QP1 packets with IP/UDP headers
...
Return value of snprintf is not bound by size value, 2nd argument.
(https://www.kernel.org/doc/htmldocs/kernel-api/API-snprintf.html).
Return value is number of printed chars, can be larger than 2nd
argument. Therefore, it can write null byte out of bounds ofbuffer.
Since snprintf puts null, it does not need to put additional null byte.
Signed-off-by: Insu Yun <wuninsu@gmail.com>
Reviewed-by: Shane Seymour <shane.seymour@hpe.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
There is no good reason to keep them apart, and this makes using the API
a bit simpler.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
There is no good reason to start out disabled - drivers can control if
the poll instance can be scheduled by simply not scheduling it yet.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
The new name is irq_poll as iopoll is already taken. Better suggestions
welcome.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
This patch changes the !blk-mq path to the same defaults as the blk-mq
I/O path by always enabling block tagging, and always using host wide
tags. We've had blk-mq available for a few releases so bugs with
this mode should have been ironed out, and this ensures we get better
coverage of over tagging setup over different configs.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Some new adapters require a special Configure Cache Parameters command
to enable the adapter write cache, so send this during the adapter
initialization if the adapter requires it.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add an IOA Inquiry command for Page 0xC4 during IOA initialization to
collect cache capabilities, particularly to check if Sync IOA Write
Cache is supported.
Inquiry will happen right after Cap Inquiry on page 0xD0; and will
execute only if the "Supported Pages" field in Inquiry Page 0x0 shows
support for Page 0xC4. Otherwise, assume Sync IOA Write Cache is
not supported.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
According to the IPR specification, Inhibit Underlength Checking bit
must be disabled when issuing commands to vsets. Enabling it in this
case might cause SCSI commands to fail with an Illegal Request, so make
sure we keep this bit cleared when resource is a vset.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add a holding pattern prior to collecting dump data, to wait for the IOA
indication that the Mailbox register is stable and won't change without
an explicit reset. This ensures we'll be collecting meaningful dump
data, even when dumping right after an adapter reset.
In the event of a timeout, we still force the dump, since a partial dump
still might be useful.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Remove unnecessary check that disabled SIS pipe commands for SIS-32
devices. This change was sufficient to enable raw mode and send SIS
pipe commands for a 57B3 device.
Fixes: f8ee25d7d2 ("ipr: AF DASD raw mode implementation in ipr driver")
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Disable underlength error verification based on count of bytes actually
transferred for AF DASD devices when SIS pipe mode is enabled.
This avoids unexpected underlength errors when issuing some commands in
raw mode.
Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Some misc fixes for endianness checking with sparse so sparse with
endian checking now runs clean. Fixes a minor bug in the process
which was uncovered by sparse which would result in unnecessary
error recovery for check conditions.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
On LE system, users see the wrong device_id attribute. This patch
does necessary byte swapping for device_id attribute and works on
both of LE and BE systems.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Fixes another signed / unsigned array indexing bug in the ipr driver.
Currently, when hrrq_index wraps, it becomes a negative number. We
do the modulo, but still have a negative number, so we end up indexing
backwards in the array. Given where the hrrq array is located in memory,
we probably won't actually reference memory we don't own, but nonetheless
ipr is still looking at data within struct ipr_ioa_cfg and interpreting it as
struct ipr_hrr_queue data, so bad things could certainly happen.
Each ipr adapter has anywhere from 1 to 16 HRRQs. By default, we use 2 on new
adapters. Let's take an example:
Assume ioa_cfg->hrrq_index=0x7fffffffe and ioa_cfg->hrrq_num=4:
The atomic_add_return will then return -1. We mod this with 3 and get -2, add
one and get -1 for an array index.
On adapters which support more than a single HRRQ, we dedicate HRRQ to adapter
initialization and error interrupts so that we can optimize the other queues
for fast path I/O. So all normal I/O uses HRRQ 1-15. So we want to spread the
I/O requests across those HRRQs.
With the default module parameter settings, this bug won't hit, only when
someone sets the ipr.number_of_msix parameter to a value larger than 3 is when
bad things start to happen.
Cc: <stable@vger.kernel.org>
Tested-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
When ipr's internal driver trace was changed to an atomic, a signed/unsigned
bug slipped in which results in us indexing backwards in our memory buffer
writing on memory that does not belong to us. This patch fixes this by removing
the modulo and instead just mask off the low bits.
Cc: <stable@vger.kernel.org>
Tested-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Make sure we have the host lock held when calling scsi_report_bus_reset. Fixes
a crash seen as the __devices list in the scsi host was changing as we were
iterating through it.
Cc: <stable@vger.kernel.org>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
This is the usual grab bag of driver updates (lpfc, qla2xxx, storvsc, aacraid,
ipr) plus an assortment of minor updates. There's also a major update to
aic1542 which moves the driver into this millenium.
Signed-off-by: James Bottomley <JBottomley@Odin.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2
iQEcBAABAgAGBQJVL/DEAAoJEDeqqVYsXL0MOwgIALPlgI0aMAtX5wLxzPMLB/2j
fhNlsB9XZ6TeYIqE7syOY7geVJqsbACMGmDhGHs5Gt6jkTnwix/G49x3T1PXBODZ
frz8GgNB6iGSqfCp+YbhJkTNHdudDIy2LrQ92EzNMb2+x0v6KTYTSq2dekgrC1zK
8GUZ9bEzuxEGaBx9TK/Sy6H8QpvMtqqJig2eCL189U3JMMU3okWtSGya708u5Whh
knbUgraMxFWNs+oHJHFclVYvekP+61i/TVyacQEM4KLDsmlxsLn49eRdiGMY6rpX
LgDIvMjggQhbY2WcCXzetF7tsFFl0joJp1wFK1fUn9YN5e+J3MRWYVBDt8FMPX8=
=OBny
-----END PGP SIGNATURE-----
Merge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI updates from James Bottomley:
"This is the usual grab bag of driver updates (lpfc, qla2xxx, storvsc,
aacraid, ipr) plus an assortment of minor updates. There's also a
major update to aic1542 which moves the driver into this millenium"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (106 commits)
change SCSI Maintainer email
sd, mmc, virtio_blk, string_helpers: fix block size units
ufs: add support to allow non standard behaviours (quirks)
ufs-qcom: save controller revision info in internal structure
qla2xxx: Update driver version to 8.07.00.18-k
qla2xxx: Restore physical port WWPN only, when port down detected for FA-WWPN port.
qla2xxx: Fix virtual port configuration, when switch port is disabled/enabled.
qla2xxx: Prevent multiple firmware dump collection for ISP27XX.
qla2xxx: Disable Interrupt handshake for ISP27XX.
qla2xxx: Add debugging info for MBX timeout.
qla2xxx: Add serdes read/write support for ISP27XX
qla2xxx: Add udev notification to save fw dump for ISP27XX
qla2xxx: Add message for sucessful FW dump collected for ISP27XX.
qla2xxx: Add support to load firmware from file for ISP 26XX/27XX.
qla2xxx: Fix beacon blink for ISP27XX.
qla2xxx: Increase the wait time for firmware to be ready for P3P.
qla2xxx: Fix crash due to wrong casting of reg for ISP27XX.
qla2xxx: Fix warnings reported by static checker.
lpfc: Update version to 10.5.0.0 for upstream patch set
lpfc: Update copyright to 2015
...
This patch implements raw mode support for AF DASD in ipr driver
which allows for tools to send commands directly to physical
devices which are members of RAID arrays when enabled in the firmware.
[jejb: fix up whitespace]
Signed-off-by: Wen Xiong<wenxiong@linux.vnet.ibm.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Daniel Kreling <kreling@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Re-enable write same support for ipr RAID adapters.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Daniel Kreling <kreling@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Fixes a possible oops during adapter initialization in some
memory allocation failure error paths scenarios.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Daniel Kreling <kreling@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
The pci_set_pcie_reset_state has changed semantics to not be callable
from interrupt context, so change ipr's usage of the API to comply with
this change by ensuring this occurs from a workqueue.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Daniel Kreling <kreling@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
Currently when performing a reboot with an ipr adapter,
the adapter gets shutdown completely, flushing all write
cache, as well as performing a full hardware reset of the card
during the shutdown phase of the old kernel. This ensures
the adapter is in a fully quiesced state across the reboot.
There are scenarios, however, such as when performing
kexec, where this full adapter shutdown is not required
and not desired, since it can make the reboot process take
noticeably longer.
This patch adds a module parameter to allow for skipping the
full shutdown during reboot. Rather than performing a full
adapter shutdown and reset, we simply cancel any outstanding
error buffers, place the adapter into a state where it has no
memory of any DMA addresses from the old kernel, then disable
the device. This significantly speeds up kexec boot, particularly
in configurations with multiple ipr adapters.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Reviewed-by: Daniel Kreling <kreling@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>
SAS controller has its own tag allocation, which doesn't directly match to ATA
tag, so SAS and SATA have different code path for ata tags. Originally we use
port->scsi_host (98bd4be1) to destinguish SAS controller, but libsas set
->scsi_host too, so we can't use it for the destinguish, we add a new flag for
this purpose.
Without this patch, the following oops can happen because scsi-mq uses
a host-wide tag map shared among all devices with some integer tag
values >= ATA_MAX_QUEUE. These unexpectedly high tag values cause
__ata_qc_from_tag() to return NULL, which is then dereferenced in
ata_qc_new_init().
BUG: unable to handle kernel NULL pointer dereference at 0000000000000058
IP: [<ffffffff804fd46e>] ata_qc_new_init+0x3e/0x120
PGD 32adf0067 PUD 32adf1067 PMD 0
Oops: 0002 [#1] SMP DEBUG_PAGEALLOC
Modules linked in: iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi igb
i2c_algo_bit ptp pps_core pm80xx libsas scsi_transport_sas sg coretemp
eeprom w83795 i2c_i801
CPU: 4 PID: 1450 Comm: cydiskbench Not tainted 4.0.0-rc3 #1
Hardware name: Supermicro X8DTH-i/6/iF/6F/X8DTH, BIOS 2.1b 05/04/12
task: ffff8800ba86d500 ti: ffff88032a064000 task.ti: ffff88032a064000
RIP: 0010:[<ffffffff804fd46e>] [<ffffffff804fd46e>] ata_qc_new_init+0x3e/0x120
RSP: 0018:ffff88032a067858 EFLAGS: 00010046
RAX: 0000000000000000 RBX: ffff8800ba0d2230 RCX: 000000000000002a
RDX: ffffffff80505ae0 RSI: 0000000000000020 RDI: ffff8800ba0d2230
RBP: ffff88032a067868 R08: 0000000000000201 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000000 R12: ffff8800ba0d0000
R13: ffff8800ba0d2230 R14: ffffffff80505ae0 R15: ffff8800ba0d0000
FS: 0000000041223950(0063) GS:ffff88033e480000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b
CR2: 0000000000000058 CR3: 000000032a0a3000 CR4: 00000000000006e0
Stack:
ffff880329eee758 ffff880329eee758 ffff88032a0678a8 ffffffff80502dad
ffff8800ba167978 ffff880329eee758 ffff88032bf9c520 ffff8800ba167978
ffff88032bf9c520 ffff88032bf9a290 ffff88032a0678b8 ffffffff80506909
Call Trace:
[<ffffffff80502dad>] ata_scsi_translate+0x3d/0x1b0
[<ffffffff80506909>] ata_sas_queuecmd+0x149/0x2a0
[<ffffffffa0046650>] sas_queuecommand+0xa0/0x1f0 [libsas]
[<ffffffff804ea544>] scsi_dispatch_cmd+0xd4/0x1a0
[<ffffffff804eb50f>] scsi_queue_rq+0x66f/0x7f0
[<ffffffff803e5098>] __blk_mq_run_hw_queue+0x208/0x3f0
[<ffffffff803e54b8>] blk_mq_run_hw_queue+0x88/0xc0
[<ffffffff803e5c74>] blk_mq_insert_request+0xc4/0x130
[<ffffffff803e0b63>] blk_execute_rq_nowait+0x73/0x160
[<ffffffffa0023fca>] sg_common_write+0x3da/0x720 [sg]
[<ffffffffa0025100>] sg_new_write+0x250/0x360 [sg]
[<ffffffffa0025feb>] sg_write+0x13b/0x450 [sg]
[<ffffffff8032ec91>] vfs_write+0xd1/0x1b0
[<ffffffff8032ee54>] SyS_write+0x54/0xc0
[<ffffffff80689932>] system_call_fastpath+0x12/0x17
tj: updated description.
Fixes: 12cb5ce101 ("libata: use blk taging")
Reported-and-tested-by: Tony Battersby <tonyb@cybernetics.com>
Signed-off-by: Shaohua Li <shli@fb.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Fixes a race condition in abort handling that was injected
when multiple interrupt support was added. When only a single
interrupt is present, the adapter guarantees it will send
responses for aborted commands prior to the response for the
abort command itself. With multiple interrupts, these responses
generally come back on different interrupts, so we need to
ensure the abort thread waits until the aborted command is
complete so we don't perform a double completion. This race
condition was being hit frequently in environments which
were triggering command timeouts, which was resulting in
a double completion causing a kernel oops.
Cc: <stable@vger.kernel.org>
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Tested-by: Wendy Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Set the scsi_level correctly for disk arrays such
that things like the rotational field get set
properly by sd.c.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong<wenxiong@linux.vnet.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
