This patch will add support for non-secure Aero adapter PCI IDs. Driver
will throw an error message when a non-secure type controller is
detected. Purpose of this interface is to avoid interacting with any
firmware which is not secured/signed by Broadcom. Any tampering on Firmware
component will be detected by hardware and it will be communicated to the
driver to avoid any further interaction with that component.
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Aero adapters provides Atomic Request Descriptor as an alternative method
for posting an entry onto a request queue. The posting of an Atomic Request
Descriptor is an atomic operation, providing a safe mechanism for multiple
processors on the host to post requests without synchronization. This
Atomic Request Descriptor format is identical to first 32 bits of Default
Request Descriptor and uses only 32 bits.
If Aero adapters support Atomic descriptor, driver should use it for
posting IOs and DCMDs to firmware.
Signed-off-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Chandrakanth Patil <chandrakanth.patil@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Added the logic for collecting IOP log respective to event log size.
Signed-off-by: Deepak Ukey <deepak.ukey@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Reviewed-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Added support to read event log size from MPI configuration table and
export through sysfs.
Signed-off-by: Deepak Ukey <deepak.ukey@microchip.com>
Signed-off-by: Viswas G <Viswas.G@microchip.com>
Reviewed-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Enable msix load balance only when combined reply queue mode is disabled on
the SAS3 and above generation HBA devices.
Earlier msix load balance used to enable if the number of online cpus is
greater than the number of MSI-X vectors enabled on the HBA. Combined reply
queue mode will be disabled only on those HBA which works in shared
resources mode. I.e. on SAS3 HBAs it will be <= 8 and on SAS35 HBA devices
it will be <= 16.
- Before this patch if system has 256 logical CPUs and HBA exposes 128
MSI-X vectors, driver will enable msix load balance.
- After this patch if system has 256 logical CPUs and HBA exposes 128
MSI-X vectors, driver will disable msix load balance.
- After this patch if system has 256 logical CPUs and HBA exposes 16 MSI-X
vectors (due to combined reply queue mode being off in HW), driver will
enable msix load balance.
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Even though 'smp_affinity_enable' module parameter is enabled, if the
number of online CPUs is bigger than the number of msix vectors enabled on
that HBA, then smp affinity settings should be disabled only for this HBA.
But currently the smp affinity setting is disabled globally and hence smp
affinity will be disabled for subsequent HBAs even though number of msix
vectors enabled for this HBA matches the number of online CPU.
To fix this, define a per HBA variable smp_affinity_enable. Initially this
variable is initialized with smp_affinity_enable module parameter value. If
this HBA has less number of msix vectors configured when compared to number
of online cpus, then only this HBA's variable smp_affinity_enable is set to
zero.
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When enabling high iops queues, the driver should use the HBA's configured
PCIe link speed instead of looking for the maximum link speed.
I.e. enable high iops queues only if Aero/Sea HBA's configured PCIe link
speed is set to 16GT/s.
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Currently default perf_mode is set to 'balanced' on Intel architecture
machines and on other machines default perf_mode is set to 'latency' mode.
This CPU architecture check is removed and the default perf_mode mode is
set to 'balanced' mode on all machines.
User can choose the required performance mode using perf_mode module
parameter.
Signed-off-by: Sreekanth Reddy <sreekanth.reddy@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The ufs-tool stable release v1.0 is available at:
https://github.com/westerndigitalcorporation/ufs-tool
Feedback and bug reports, as always, are welcomed.
Signed-off-by: Arthur Simchaev <Arthur.Simchaev@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If cb_arg alloc failed, we can't release the struct orig_io_req refcount
before we take its refcount. As Saurav said, move the srr_err label down
to avoid unnecessary refcount release and nullptr free.
Signed-off-by: Lin Yi <teroincn@163.com>
Acked-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If cb_arg alloc failed, we can't release the struct orig_io_req refcount
before we take its refcount. As Saurav said, move the rec_err label down
to avoid unnecessary refcount release and nullptr free.
Signed-off-by: Lin Yi <teroincn@163.com>
Acked-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Update the driver version to 2.12.10.
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
- Reduce the sg_tablesize to 255.
- Reduce the MAX BDs firmware can handle to 255.
- Return IO to ML if BD goes more then 255 after split.
- Correct the size of each BD split to 0xffff.
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If firmware sends either cleanup or abort completion, it means other won't
be sent. Clean out flags for other as well.
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Separate out abort and cleanup flag and completion, to have better
understaning of what is getting processed.
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In certain tests where the SCSI error handler issues an abort that is
already outstanding, we will cleanup the command so that the SCSI error
handler can proceed. In some of these cases we were seeing a command
mismatch:
kernel: scsi host2: bnx2fc: xid:0x42b eh_abort - refcnt = 2
kernel: bnx2fc: eh_abort: io_req (xid = 0x42b) already in abts processing
kernel: scsi host2: bnx2fc: xid:0x42b Entered bnx2fc_initiate_cleanup
kernel: scsi host2: bnx2fc: xid:0x42b CLEANUP io_req xid = 0x80b
kernel: scsi host2: bnx2fc: xid:0x80b cq_compl- cleanup resp rcvd
kernel: scsi host2: bnx2fc: xid:0x42b complete - rx_state = 9
kernel: scsi host2: bnx2fc: xid:0x42b Entered process_cleanup_compl refcnt = 2, cmd_type = 1
kernel: scsi host2: bnx2fc: xid:0x42b scsi_done. err_code = 0x7
kernel: scsi host2: bnx2fc: xid:0x42b sc=ffff8807f93dfb80, result=0x7, retries=0, allowed=5
kernel: ------------[ cut here ]------------
kernel: WARNING: at /root/rpmbuild/BUILD/netxtreme2-7.14.43/obj/default/bnx2fc-2.12.1/driver/bnx2fc_io.c:1347 bnx2fc_eh_abort+0x56f/0x680 [bnx2fc]()
kernel: xid=0x42b refcount=-1
kernel: Modules linked in:
kernel: nls_utf8 isofs sr_mod cdrom tcp_lp dm_round_robin xt_CHECKSUM iptable_mangle ipt_MASQUERADE nf_nat_masquerade_ipv4 iptable_nat nf_nat_ipv4 nf_nat nf_conntrack_ipv4 nf_defrag_ipv4 xt_conntrack nf_conntrack ipt_REJECT nf_reject_ipv4 tun bridge ebtable_filter ebtables fuse ip6table_filter ip6_tables iptable_filter bnx2fc(OE) cnic(OE) uio fcoe libfcoe 8021q libfc garp mrp scsi_transport_fc stp llc scsi_tgt vfat fat dm_service_time intel_powerclamp coretemp intel_rapl iosf_mbi kvm_intel kvm irqbypass crc32_pclmul ghash_clmulni_intel aesni_intel lrw gf128mul glue_helper ablk_helper cryptd ses enclosure ipmi_ssif i2c_core hpilo hpwdt wmi sg ipmi_devintf pcspkr ipmi_si ipmi_msghandler shpchp acpi_power_meter dm_multipath nfsd auth_rpcgss nfs_acl lockd grace sunrpc ip_tables xfs sd_mod crc_t10dif
kernel: crct10dif_generic bnx2x(OE) crct10dif_pclmul crct10dif_common crc32c_intel mdio ptp pps_core libcrc32c smartpqi scsi_transport_sas fjes uas usb_storage dm_mirror dm_region_hash dm_log dm_mod
kernel: CPU: 9 PID: 2012 Comm: scsi_eh_2 Tainted: G W OE ------------ 3.10.0-514.el7.x86_64 #1
kernel: Hardware name: HPE Synergy 480 Gen10/Synergy 480 Gen10 Compute Module, BIOS I42 03/21/2018
kernel: ffff8807f25a3d98 0000000015e7fa0c ffff8807f25a3d50 ffffffff81685eac
kernel: ffff8807f25a3d88 ffffffff81085820 ffff8807f8e39000 ffff880801ff7468
kernel: ffff880801ff7610 0000000000002002 ffff8807f8e39014 ffff8807f25a3df0
kernel: Call Trace:
kernel: [<ffffffff81685eac>] dump_stack+0x19/0x1b
kernel: [<ffffffff81085820>] warn_slowpath_common+0x70/0xb0
kernel: [<ffffffff810858bc>] warn_slowpath_fmt+0x5c/0x80
kernel: [<ffffffff8168d842>] ? _raw_spin_lock_bh+0x12/0x50
kernel: [<ffffffffa0549e6f>] bnx2fc_eh_abort+0x56f/0x680 [bnx2fc]
kernel: [<ffffffff814570af>] scsi_error_handler+0x59f/0x8b0
kernel: [<ffffffff81456b10>] ? scsi_eh_get_sense+0x250/0x250
kernel: [<ffffffff810b052f>] kthread+0xcf/0xe0
kernel: [<ffffffff810b0460>] ? kthread_create_on_node+0x140/0x140
kernel: [<ffffffff81696418>] ret_from_fork+0x58/0x90
kernel: [<ffffffff810b0460>] ? kthread_create_on_node+0x140/0x140
kernel: ---[ end trace 42deb88f2032b111 ]---
The reason that there was a mismatch is that the SCSI command is actual
returned from the cleanup handler. In previous testing, the type of
cleanup notification we'd get from the CQE did not trigger the code that
returned the SCSI command. To overcome the previous behavior we would put
a reference in bnx2fc_abts_cleanup() to account for the SCSI command.
However, in cases where the SCSI command is actually off, we end up with an
extra put.
The fix for this is to only take the extra put in bnx2fc_abts_cleanup if
the completion for the cleanup times out.
Signed-off-by: Chad Dupuis <cdupuis@marvell.com>
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
For bnx2fc, the source FCoE MAC is stored in the fcoe_port struct in the
data_src_mac field. Currently this is set in fcoe_ctlr_recv_flogi which
ends up setting it by simply using fc_fcoe_set_mac() which only uses the
default FCF-MAC. We still want to store the source FCoE MAC in
port->data_src_mac but we want to snoop the FLOGI response payload so as to
set it in the following method:
1. If a granted_mac is found, use that.
2. If not granted_mac is there but there is a FCF-MAP from the FCF then
create the MAC from the FCF-MAP and the destination ID from the frame.
3. If there is no FCF-MAP the use the spec. default FCF-MAP and the
destination ID from the frame.
Signed-off-by: Chad Dupuis <cdupuis@marvell.com>
Signed-off-by: Saurav Kashyap <skashyap@marvell.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In the case of UPIU/DME request execution failed in UFS device,
ufs_bsg_request() will complete the failed bsg job by calling
bsg_job_done(). Meanwhile, it returns this error status to blk-mq layer,
then triggers blk-mq completing this request again, this will cause the
following panic.
Call trace:
ll_sc___cmpxchg_case_acq_32+0x4/0x20
complete+0x28/0x70
blk_end_sync_rq+0x24/0x30
blk_mq_end_request+0xb8/0x118
bsg_job_put+0x4c/0x58
bsg_complete+0x20/0x30
blk_done_softirq+0xb4/0xe8
do_softirq+0x154/0x3f0
run_ksoftirqd+0x4c/0x68
smpboot_thread_fn+0x22c/0x268
kthread+0x130/0x138
ret_from_fork+0x10/0x1c
Code: f84107fe d65f03c0 d503201f f9800011 (885ffc10)
---[ end trace d92825bff6326e66 ]---
Kernel panic - not syncing: Fatal exception in interrupt
This patch is to fix this issue. The solution is to complete the ufs-bsg
job only if no error happened.
[mkp: commit description tweak]
Fixes: df032bf27a (scsi: ufs: Add a bsg endpoint that supports UPIUs)
Signed-off-by: Bean Huo <beanhuo@micron.com>
Reviewed-by: Avri Altman <Avri.Altman@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Correct dev_dbg to dev_err, so as to print out the error information in
case of DME command failed.
Signed-off-by: Bean Huo <beanhuo@micron.com>
Reviewed-by: Avri Altman <Avri.Altman@wdc.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The 'affinity_hint_set' is not used any longer since commit
0d9f0a52c8 ("virtio_scsi: use virtio IRQ affinity").
Signed-off-by: Dongli Zhang <dongli.zhang@oracle.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Use existing macros. No functional change.
[mkp: typo]
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Acked-by: Suganath Prabu <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Support is easier with all driver parameters visible in sysfs. Also I've
replaced a constant with an octal permission.
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Acked-by: Suganath Prabu <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
New Qualcomm AArch64 based laptops are now available which use UFS as their
primary data storage medium. These devices are supplied with ACPI support
out of the box. This patch ensures the Qualcomm UFS driver will be bound
when the "QCOM24A5" H/W device is advertised as present.
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
One of the more common cases of allocation size calculations is finding the
size of a structure that has a zero-sized array at the end, along with
memory for some number of elements for that array. For example:
struct MR_PD_CFG_SEQ_NUM_SYNC {
...
struct MR_PD_CFG_SEQ seq[1];
} __packed;
Make use of the struct_size() helper instead of an open-coded version in
order to avoid any potential type mistakes.
So, replace the following form:
sizeof(struct MR_PD_CFG_SEQ_NUM_SYNC) + (sizeof(struct MR_PD_CFG_SEQ) * (MAX_PHYSICAL_DEVICES - 1))
with:
struct_size(pd_sync, seq, MAX_PHYSICAL_DEVICES - 1)
This code was detected with the help of Coccinelle.
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Acked-by: Sumit Saxena <sumit.saxena@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A system bus error during a PDMA send operation can result in bytes being
lost. Theoretically that could cause the target to remain in DATA OUT phase
and the initiator (expecting a phase change) would time-out waiting for the
Last Byte Sent flag. Should that happen, fail the transfer so the core
driver will stop using PDMA with this target.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add support for Apple's custom "SCSI DMA" chip. This patch doesn't make use
of its DMA capability. Just the PDMA capability is sufficient to improve
sequential read throughput by a factor of 5.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: Joshua Thompson <funaho@jurai.org>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A system bus error during a PDMA transfer can mess up the calculation of
the transfer residual (the PDMA handshaking hardware lacks a byte
counter). This results in data corruption.
The algorithm in this patch anticipates a bus error by starting each
transfer with a MOVE.B instruction. If a bus error is caught the transfer
will be retried. If a bus error is caught later in the transfer (for a
MOVE.W instruction) the transfer gets failed and subsequent requests for
that target will use PIO instead of PDMA.
This avoids the "!REQ and !ACK" error so the severity level of that message
is reduced to KERN_DEBUG.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: stable@vger.kernel.org # v4.14+
Fixes: 3a0f64bfa9 ("mac_scsi: Fix pseudo DMA implementation")
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Reported-by: Chris Jones <chris@martin-jones.com>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Some targets introduce delays when handshaking the response to certain
commands. For example, a disk may send a 96-byte response to an INQUIRY
command (or a 24-byte response to a MODE SENSE command) too slowly.
Apparently the first 12 or 14 bytes are handshaked okay but then the system
bus error timeout is reached while transferring the next word.
Since the scsi bus phase hasn't changed, the driver then sets the target
borken flag to prevent further PDMA transfers. The driver also logs the
warning, "switching to slow handshake".
Raise the PDMA threshold to 512 bytes so that PIO transfers will be used
for these commands. This default is sufficiently low that PDMA will still
be used for READ and WRITE commands.
The existing threshold (16 bytes) was chosen more or less at random.
However, best performance requires the threshold to be as low as possible.
Those systems that don't need the PIO workaround at all may benefit from
mac_scsi.setup_use_pdma=1
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: stable@vger.kernel.org # v4.14+
Fixes: 3a0f64bfa9 ("mac_scsi: Fix pseudo DMA implementation")
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A PDMA error is handled in the core driver by setting the device's 'borken'
flag and aborting the command. Unfortunately, do_abort() is not
dependable. Perform a SCSI bus reset instead, to make sure that the command
fails and gets retried.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: stable@vger.kernel.org # v4.20+
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2")
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The reselection interrupt gets disabled during selection and must be
re-enabled when hostdata->connected becomes NULL. If it isn't re-enabled a
disconnected command may time-out or the target may wedge the bus while
trying to reselect the host. This can happen after a command is aborted.
Fix this by enabling the reselection interrupt in NCR5380_main() after
calls to NCR5380_select() and NCR5380_information_transfer() return.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: stable@vger.kernel.org # v4.9+
Fixes: 8b00c3d5d4 ("ncr5380: Implement new eh_abort_handler")
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
This reverts commit 4822827a69.
The purpose of that commit was to suppress a timeout warning message which
appeared to be caused by target latency. But suppressing the warning is
undesirable as the warning may indicate a messed up transfer count.
Another problem with that commit is that 15 ms is too long to keep
interrupts disabled as interrupt latency can cause system clock drift and
other problems.
Cc: Michael Schmitz <schmitzmic@gmail.com>
Cc: stable@vger.kernel.org
Fixes: 4822827a69 ("scsi: ncr5380: Increase register polling limit")
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Tested-by: Michael Schmitz <schmitzmic@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Host reset oopses because it calls wd719x_chip_init, which calls
request_firmware, under a spinlock. Stop the RISC first, then flush active
SCBs under a spinlock. Finally call wd719x_chip_init unlocked.
Also found and fixed more bugs during tests:
Affected active SCBs were not flushed during abort, bus and device
reset. This caused problems in a following host reset (hang or oops).
Device and bus reset failed under load because the result of the reset
command is WD719X_SUE_TERM or WD719X_SUE_RESET. Don't treat these codes as
error in wd719x_wait_done.
wd719x_direct_cmd for RESET/ABORT commands didn't work properly, causing
timeouts. Looks like it was caused by the WD719X_DISABLE_INT bit. Not
setting it for RESET/ABORT commands seems to fix the probem. Also lower
the log level of the corresponding "direct command completed" message to
debug.
Unfortunately, my documentation is missing some pages, including page
67 (SPIDER67.gif) about resets :(
Reported-by: Hariprasad Kelam <hariprasad.kelam@gmail.com>
Signed-off-by: Ondrej Zary <linux@zary.sk>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The previous patch guarantees that srp_queuecommand() does not get
invoked while reconnecting occurs. Hence remove the code from
srp_queuecommand() that prevents command queueing while reconnecting.
This patch avoids that the following can appear in the kernel log:
BUG: sleeping function called from invalid context at kernel/locking/mutex.c:747
in_atomic(): 1, irqs_disabled(): 0, pid: 5600, name: scsi_eh_9
1 lock held by scsi_eh_9/5600:
#0: (rcu_read_lock){....}, at: [<00000000cbb798c7>] __blk_mq_run_hw_queue+0xf1/0x1e0
Preemption disabled at:
[<00000000139badf2>] __blk_mq_delay_run_hw_queue+0x78/0xf0
CPU: 9 PID: 5600 Comm: scsi_eh_9 Tainted: G W 4.15.0-rc4-dbg+ #1
Hardware name: Dell Inc. PowerEdge R720/0VWT90, BIOS 2.5.4 01/22/2016
Call Trace:
dump_stack+0x67/0x99
___might_sleep+0x16a/0x250 [ib_srp]
__mutex_lock+0x46/0x9d0
srp_queuecommand+0x356/0x420 [ib_srp]
scsi_dispatch_cmd+0xf6/0x3f0
scsi_queue_rq+0x4a8/0x5f0
blk_mq_dispatch_rq_list+0x73/0x440
blk_mq_sched_dispatch_requests+0x109/0x1a0
__blk_mq_run_hw_queue+0x131/0x1e0
__blk_mq_delay_run_hw_queue+0x9a/0xf0
blk_mq_run_hw_queue+0xc0/0x1e0
blk_mq_start_hw_queues+0x2c/0x40
scsi_run_queue+0x18e/0x2d0
scsi_run_host_queues+0x22/0x40
scsi_error_handler+0x18d/0x5f0
kthread+0x11c/0x140
ret_from_fork+0x24/0x30
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Laurence Oberman <loberman@redhat.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Leon Romanovsky <leonro@mellanox.com>
Cc: Doug Ledford <dledford@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Several SCSI transport and LLD drivers surround code that does not
tolerate concurrent calls of .queuecommand() with scsi_target_block() /
scsi_target_unblock(). These last two functions use
blk_mq_quiesce_queue() / blk_mq_unquiesce_queue() for scsi-mq request
queues to prevent concurrent .queuecommand() calls. However, that is
not sufficient to prevent .queuecommand() calls from scsi_send_eh_cmnd().
Hence surround the .queuecommand() call from the SCSI error handler with
code that avoids that .queuecommand() gets called in the blocked state.
Note: converting the .queuecommand() call in scsi_send_eh_cmnd() into
code that calls blk_get_request() + blk_execute_rq() is not an option
since scsi_send_eh_cmnd() must be able to make forward progress even
if all requests have been allocated.
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The ability to modify the SCSI device state was introduced by commit
638127e579a4 ("[PATCH] Fix error handler offline behaviour"; v2.6.12). That
same commit introduced the following device states:
{ SDEV_CREATED, "created" },
{ SDEV_RUNNING, "running" },
{ SDEV_CANCEL, "cancel" },
{ SDEV_DEL, "deleted" },
{ SDEV_QUIESCE, "quiesce" },
{ SDEV_OFFLINE, "offline" },
The SDEV_BLOCK state was introduced later to avoid that an FC cable pull
would immediately result in an I/O error (commit 1094e682310e; "[PATCH]
suspending I/Os to a device"; v2.6.12). That same patch introduced the
ability to set the SDEV_BLOCK state from user space. I'm not sure whether
that ability was introduced on purpose or accidentally.
Since there is agreement that only writing "running" or "offline" into
the SCSI sysfs device state attribute makes sense, restrict sysfs writes
to these values.
This patch makes sure that SDEV_BLOCK is only used for its original
purpose, namely to allow transport drivers and LLDs to block further
.queuecommand() calls while transport layer or adapter recovery is in
progress.
Note: a web search for "/sys/class/scsi_device" AND "device/state"
revealed several storage configuration guides. The instructions I found
in these guides tell users to write the value "running" or "offline" in
the SCSI device state sysfs attribute and no other values.
[mkp: typo]
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ming Lei <ming.lei@redhat.com>
Cc: Hannes Reinecke <hare@suse.de>
Cc: Johannes Thumshirn <jthumshirn@suse.de>
Cc: James Smart <james.smart@broadcom.com>
Cc: Ewan D. Milne <emilne@redhat.com>
Cc: Laurence Oberman <loberman@redhat.com>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
IEEE_8021QAZ_APP_SEL_STREAM is a valid selector for iSCSI connections, so
add code to use IEEE_8021QAZ_APP_SEL_STREAM selector to get priority mask.
Signed-off-by: Varun Prakash <varun@chelsio.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Use 'kasprintf()' instead of:
- snprintf(NULL, 0...
- kmalloc(...
- snprintf(...
This is less verbose and saves 7 bytes (i.e. the space for '/(null)') if
'udev->dev_config' is NULL.
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Mike Christie <mchristi@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Adding functionality to allow the SCSI queue depth to be changed by
utilizing the "scsi_change_queue_depth" function.
[mkp: checkpatch]
Signed-off-by: Branden Bonaby <brandonbonaby94@gmail.com>
Reviewed-by: Michael Kelley <mikelley@microsoft.com>
Reviewed-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In preparation to enabling -Wimplicit-fallthrough, mark switch cases where
we are expecting to fall through.
This patch fixes the following warning:
drivers/scsi/mpt3sas/mpt3sas_base.c: In function _base_update_ioc_page1_inlinewith_perf_mode :
drivers/scsi/mpt3sas/mpt3sas_base.c:4510:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
if (ioc->high_iops_queues) {
^
drivers/scsi/mpt3sas/mpt3sas_base.c:4530:2: note: here
case MPT_PERF_MODE_LATENCY:
^~~~
Warning level 3 was used: -Wimplicit-fallthrough=3
This patch is part of the ongoing efforts to enable -Wimplicit-fallthrough.
Fixes: 30cb97023f38 ("scsi: mpt3sas: Introduce perf_mode module parameter")
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Many times in libsas, and in LLDDs which use libsas, the check for an
expander device is re-implemented or open coded.
Use dev_is_expander() instead. We rename this from
sas_dev_type_is_expander() to not spill so many lines in referencing.
Signed-off-by: John Garry <john.garry@huawei.com>
Reviewed-by: Jason Yan <yanaijie@huawei.com>
Reviewed-by: Jack Wang <jinpu.wang@cloud.ionos.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Update driver version from 28.100.00.00 to 29.100.00.00
This is equivalent to Phase 10 OOB driver.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
1. Introduce module parameter perf_mode for only Aero/Sea generation HBAs.
2. Update IOC page1 fields according to performance mode.
Below are the performance modes that can be enabled with module parameter
perf_mode:
0: Balanced - Few high iops reply queues will be enabled. Interrupt
coalescing will be enabled only for these high iops reply descriptor
queues.
1: Iops - Interrupt coalescing will be enabled on all reply queues.
Coalescing timeout is set to 0x20.This is default value for Aero.
2: Latency - Interrupt coalescing will be enabled on all reply queues.
Coalescing timeout is set to 0xA. This is a legacy behavior similar to
Ventura & Invader HBA series.
Default perf mode set by driver will be balanced mode if the following
conditions are met:
- CPU vendor = Intel;
- Aero controller working in 16GT/s pcie speed
Performance mode will be set to latency mode for all other cases.
4k Random Read IO performance numbers on 24 SAS SSD drives for above three
permormance modes. Performance data is from Intel Skylake and HGST SS300
(drive model SDLL1DLR400GCCA1).
IOPs:
-----------------------------------------------------------------------
|perf_mode | qd = 1 | qd = 64 | note |
|-------------|--------|---------|-------------------------------------
|balanced | 259K | 3061k | Provides max performance numbers |
| | | | both on lower QD workload & |
| | | | also on higher QD workload |
|-------------|--------|---------|-------------------------------------
|iops | 220K | 3100k | Provides max performance numbers |
| | | | only on higher QD workload. |
|-------------|--------|---------|-------------------------------------
|latency | 246k | 2226k | Provides good performance numbers |
| | | | only on lower QD worklaod. |
-----------------------------------------------------------------------
Avarage Latency:
-----------------------------------------------------
|perf_mode | qd = 1 | qd = 64 |
|-------------|--------------|----------------------|
|balanced | 92.05 usec | 501.12 usec |
|-------------|--------------|----------------------|
|iops | 108.40 usec | 498.10 usec |
|-------------|--------------|----------------------|
|latency | 97.10 usec | 689.26 usec |
-----------------------------------------------------
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Enable interrupt coalescing only on high iops queues.
In ioc config page 1, offset 0x14 (ProductSpecific field) is used to
determine interrupt coalescing enabled/disabled on per reply descriptor
post queue group(8) basis. If 31st bit is zero, then interrupt coalescing
is enabled for all reply descriptor post queues. If 31st bit is set to one,
then user can enable/disable interrupt coalescing on per reply descriptor
post queue group(8) basis. So to enable interrupt coalescing only on first
reply descriptor post queue group (i.e. on high iops queues), set bit 0 and
31.
This configuration should reset during driver unload or shutdown to the
default settings. For this, the driver takes copy of default ioc page 1 and
copies back the default or unmodified ioc page1 during unload and
shutdown. This means that on next driver load (e.g. if older version driver
is loaded by user), current modified changes on ioc page1 won't take
effect.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
High iops queues are mapped to non-managed irqs. Set affinity of
non-managed irqs to local numa node. Low latency queues are mapped to
managed irqs.
Driver reserves some reply queues for max iops (through
pci_alloc_irq_vectors_affinity and .pre_vectors interface). The rest of
queues are for low latency.
Based on io workload in io submission path, driver will decide which group
of reply queues (either high iops queues or low latency queues) to be
used. High iops queues will be mapped to local numa node of controller and
low latency queues will be mapped to cpus across numa nodes. In general,
high iops and low latency queues should fit into 128 reply queues
which is the max number of reply queues supported by Aero/Sea.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
In the IO submission path _base_get_msix_index is called twice. Initially
while getting the smid and subsequently while posting the request
descriptor (RD).
Refactor code to query msix index only while posting the request
descriptor. Save determined msix index in msix_io field.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The driver will use round-robin method for io submission in batches within
the high iops queues when the number of in-flight ios on the target device
is larger than 8. Otherwise the driver will use low latency reply queues.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Code refactoring.
In function _base_get_msix_index, add scmd as second argument. This change
is made in preparation for the next patch where we introduce a new function
to get the MSI-X index for high iops queues.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Aero controllers support balanced performance mode through the ability to
configure queues with different properties.
Reply queues with interrupt coalescing enabled are called "high iops reply
queues" and reply queues with interrupt coalescing disabled are called "low
latency reply queues".
The driver configures a combination of high iops and low latency reply
queues if:
- HBA is an AERO controller;
- MSI-X vectors supported by the HBA is 128;
- Total CPU count in the system more than high iops queue count;
- Driver is loaded with default max_msix_vectors module parameter; and
- System booted in non-kdump mode.
Signed-off-by: Suganath Prabu S <suganath-prabu.subramani@broadcom.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
