Calvin Owens [Fri, 14 Aug 2015 01:48:10 +0000 (18:48 -0700)]
mpt2sas: Refcount fw_events and fix unsafe list usage
The fw_event_work struct is concurrently referenced at shutdown, so
add a refcount to protect it, and refactor the code to use it.
Additionally, refactor _scsih_fw_event_cleanup_queue() such that it
no longer iterates over the list without holding the lock, since
_firmware_event_work() concurrently deletes items from the list.
Calvin Owens [Fri, 14 Aug 2015 01:48:09 +0000 (18:48 -0700)]
mpt2sas: Refcount sas_device objects and fix unsafe list usage
These objects can be referenced concurrently throughout the driver, we
need a way to make sure threads can't delete them out from under each
other. This patch adds the refcount, and refactors the code to use it.
Additionally, we cannot iterate over the sas_device_list without
holding the lock, or we risk corrupting random memory if items are
added or deleted as we iterate. This patch refactors _scsih_probe_sas()
to use the sas_device_list in a safe way.
scsi_dh: don't allow to detach device handlers at runtime
The I/O submission and completion paths call into the device handler
without any synchronization agains detachment. So disallow detaching
device handlers at runtime.
scsi_dh: add a common helper to get a scsi_device from a request_queue
And cleanup the various messy opencoded versions of this. Note that this
moves the sdev_state checks outside the queue_lock coverage, but as
we don't hold the lock over the activation they are only advisory anyway.
Add a ->handler and a ->handler_data field to struct scsi_device and kill
this indirection. Also move struct scsi_device_handler to scsi_dh.h so that
changes to it don't require rebuilding every SCSI LLDD.
Stop building scsi_dh as a separate module and integrate it fully into the
core SCSI code with explicit callouts at bus scan time. For now the
callouts are placed at the same point as the old bus notifiers were called,
but in the future we will be able to look at ALUA INQUIRY data earlier on.
Note that this also means that the device handler modules need to be loaded
by the time we scan the bus. The next patches will add support for
autoloading device handlers at bus scan time to make sure they are always
loaded if they are enabled in the kernel config.
dm-mpath, scsi_dh: don't let dm detach device handlers
While allowing dm-mpath to attach device handlers is a functionality we need
for backwards compatibility reason there is no reason to reference count
them and detach them if dm-mpath stops using the device for some reason.
If the device handler works for the given device it can just stay attached,
and we can take the retain_hw_handler codepath.
Hannes Reinecke [Mon, 6 Jul 2015 11:07:58 +0000 (13:07 +0200)]
aic94xx: Skip reading user settings if flash is not found
If no user settings are found it's pointless trying to
read them from flash. So skip that step.
This also fixes a compilation warning about uninitialized variables in
aic94xx.
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.
Sreekanth Reddy [Tue, 30 Jun 2015 06:55:05 +0000 (12:25 +0530)]
mpt3sas: When device is blocked followed by unblock fails, unfreeze the I/Os
Issue: When the disks are getting discovered and assigned device
handles by the kernel, a device block followed by an unblock
(due to broadcast primitives) issued by the driver is
interspersed by the kernel changing the state of the device.
Therefore the unblock by the driver results in a no operation
within the kernel API.
To fix this one, the below patch checks the return of the unblock API
and performs a block followed by an unblock to unfreeze the block
layer's I/O queue. Sufficient checks and prints are also added in the
driver to identify this condition caused by the kernel.
Sreekanth Reddy [Tue, 30 Jun 2015 06:55:04 +0000 (12:25 +0530)]
mpt3sas: Call dma_mapping_error() API after mapping an address with dma_map_single() API
Added dma_mapping_error() API after mapping an address with dma_map_single()
API. Otherwise when CONFIG_DMA_API_DEBUG is enabled in the kernel, then it
complains about mpt3sas driver not calling dma_mapping_error after mapping an
address with dma_map_single
Sreekanth Reddy [Tue, 30 Jun 2015 06:55:03 +0000 (12:25 +0530)]
mpt3sas: Use alloc_ordered_workqueue() API instead of create_singlethread_workqueue() API
Created a thread using alloc_ordered_workqueue() API in order to process
the works from firmware Work-queue sequentially instead of
create_singlethread_workqueue() API.
Sreekanth Reddy [Tue, 30 Jun 2015 06:55:01 +0000 (12:25 +0530)]
mpt3sas: Return host busy error status to SML when DMA mapping of scatter gather list fails for a SCSI command
scsi_dma_map API will return a negative value (i.e. -ENOMEM)
if DMA mapping of sg lists fails and zero if the sg list in the
SCSI cmd is NULL. But drivers doesn't handled sg list DMA mapping
failure case properly.
So, Updated the code to return host busy error status to SCSI MID Layer(SML),
when DMA mapping of scatter gather list fails for a SCSI command.
So that SML will retry this SCSI cmd after some time.
Sreekanth Reddy [Tue, 30 Jun 2015 06:55:00 +0000 (12:25 +0530)]
mpt3sas: Complete the SCSI command with DID_RESET status for log_info value 0x0x32010081
For any SCSI command, if the driver receives
IOC status = SCSI_IOC_TERMINATED and log info = 0x32010081 then
that command will be completed with DID_RESET host status.
The definition of this log info value is
"Virtual IO has failed and has to be retried".
Firmware will provide this log info value with IOC Status
"SCSI_IOC_TERMINATED", whenever a drive (with is a part of a volume)
is pulled and pushed back within some minimal delay.
With this log info value, firmware informs the driver to retry the
failed IO command infinite times, so to provide some time for the
firmware to discover the reinserted drive successfully instated of
just retrying failed command for five times(doesn't giving enough
time for firmware to complete the drive discovery) and failing the
IO permanently even though drive came back successfully.
Sreekanth Reddy [Tue, 30 Jun 2015 06:54:52 +0000 (12:24 +0530)]
mpt3sas: Provides the physical location of sas drives
This Patch will provide more details of the devices such as slot number,
enclosure logical id, enclosure level & connector name in the following
scenarios,
- When end device is added in the topology,
- When the end device is removed from the setup,
- When the SCSI mid layer issues TASK ABORT/ DEVICE RESET/ TARGET RESET during
error handling,
- When any command to the device fails with Sense key Hardware error or Medium
error or Unit Attention,
- When firmware returns device error or device not ready status for the end
device,
- When a Predicted fault is detected on an end device.
This information can be used by the user to identify the location of the
desired drive in the topology.
Driver will get these information by reading the sas device page0.
Sreekanth Reddy [Tue, 30 Jun 2015 06:54:49 +0000 (12:24 +0530)]
mpt3sas: Don't block the drive when drive addition under the control of SML
During hot-plugging of a disk(having a flaky link), the disk addition
stops and any further disk addition or removal doesn't happen on that
controller.
This is because, when driver receives DELAY_NOT_RESPONDING event for a disk
while it is undergoing addition at the SCSI Transport layer, the driver
would block the I/O to that disk resulting in a deadlock. i.e the disk
addition work couldn't be completed at the SCSI Transport Layer as it
can't send any I/Os (such as Inquiry, Report LUNs etc) to the disk as
I/Os are blocked to this drive. Also any subsequent device removal
(TARGET_NOT_RESPONDING) or link update(RC_PHY_CHANGED) event couldn't be
processed as they are in the queue to get processed after disk addition
event.
Description of Change:
Don't block the drive when drive addition is under the control of SML.
So that SML won't be blocked of issuing the device dicovery commands
(such as Inquiry, Report LUNs etc).
Sreekanth Reddy [Tue, 30 Jun 2015 06:54:48 +0000 (12:24 +0530)]
mpt3sas: Get IOC_FACTS information using handshake protocol only after HBA card gets into READY or Operational state.
Driver initialization fails if driver tries to send IOC facts request message
when the IOC is in reset or in a fault state.
This patch will make sure that
1.Driver to send IOC facts request message only if HBA is in operational or
ready state.
2.If IOC is in fault state, a diagnostic reset would be issued.
3.If IOC is in reset state then driver will wait for 10 seconds to exit out
of reset state. If the HBA continues to be in reset state, then the HBA
wouldn't be claimed by the driver.
Sreekanth Reddy [Tue, 30 Jun 2015 06:54:47 +0000 (12:24 +0530)]
mpt3sas: Added Combined Reply Queue feature to extend up-to 96 MSIX vector support
In this patch, increased the number of MSIX vector support for SAS3 C0 HBAs to
up-to 96.
Following are changes that are done in this patch
1. This feature is enabled only for SAS3 C0 and higher revision cards and also
only when reply post free queue count is greater than 8.
2. To support this feature 12 SupplementalReplyPostHostIndex system interfaces
are used. MSI-X index numbered from 0 to 7 use the first
SupplementalReplyPostHostIndex system interface to update its corresponding
ReplyPostHostIndex values, MSI-X index numbered from 8 to 15 will use the
second SupplementalReplyPostHostIndex system interface and so on. These 12
SuppementalReplyPostHostIndex system interfaces address are saved in the array
replyPostRegisterIndex[].
3. As each SupplementalReplyPostHostIndex register supports 8 MSI-X
vectors. So MSIxIndex field in these register must contain a value between 0
and 7.
4. After processing the reply descriptors from a reply post free queues then
update the new reply post host index value in ReplyPostHostIndex field and
(msix_index mod 8) value in MSIxIndex field of SupplementalReplyPostHostIndex
register. The Address of this SupplementalReplyPostHostIndex register is
retrived from (msix_index/8)th entry of replyPostRegisterIndex[] array.
Keith Mange [Thu, 13 Aug 2015 15:43:46 +0000 (08:43 -0700)]
storvsc: Rather than look for sets of specific protocol versions, make decisions based on ranges.
Rather than look for sets of specific protocol versions,
make decisions based on ranges. This will be safer and require fewer changes
going forward as we add more storage protocol versions.
Matthew R. Ochs [Wed, 26 Aug 2015 23:36:12 +0000 (18:36 -0500)]
cxlflash: Remove unused variable from queuecommand
The queuecommand routine has a local dev pointer used for the
dev_* prints. The two prints that currently exist are tucked
under a debug define and thus can be left out. Use the actual
location instead of a local to avoid this warning.
This patch is intended to be applied after the "CXL Flash Error
Recovery and Superpipe" series.
Dan Carpenter [Fri, 14 Aug 2015 20:35:25 +0000 (23:35 +0300)]
cxlflash: off by one bug in cxlflash_show_port_status()
The > should be >= or we read one element past the end of the array.
Fixes: c21e0bbfc485 ('cxlflash: Base support for IBM CXL Flash Adapter') Signed-off-by: Dan Carpenter <[email protected]> Acked-by: Matthew R. Ochs <[email protected]> Signed-off-by: James Bottomley <[email protected]>
Matthew R. Ochs [Fri, 14 Aug 2015 02:47:53 +0000 (21:47 -0500)]
cxlflash: Virtual LUN support
Add support for physical LUN segmentation (virtual LUNs) to device
driver supporting the IBM CXL Flash adapter. This patch allows user
space applications to virtually segment a physical LUN into N virtual
LUNs, taking advantage of the translation features provided by this
adapter.
Matthew R. Ochs [Fri, 14 Aug 2015 02:47:43 +0000 (21:47 -0500)]
cxlflash: Superpipe support
Add superpipe supporting infrastructure to device driver for the IBM CXL
Flash adapter. This patch allows userspace applications to take advantage
of the accelerated I/O features that this adapter provides and bypass the
traditional filesystem stack.
Matthew R. Ochs [Fri, 14 Aug 2015 02:47:34 +0000 (21:47 -0500)]
cxlflash: Base error recovery support
Introduce support for enhanced I/O error handling.
A device state is added to track 3 possible states of the device:
Normal - the device is operating normally and is fully operational
Limbo - the device is in a reset/recovery scenario and its operational
status is paused
Failed/terminating - the device has either failed to be reset/recovered
or is being terminated (removed); it is no longer
operational
All operations are allowed when the device is operating normally. When the
device transitions to limbo state, I/O must be paused. To help accomplish
this, a wait queue is introduced where existing and new threads can wait
until the device is no longer in limbo. When coming out of limbo, threads
need to check the state and error out gracefully when encountering the
failed state. When the device transitions to the failed/terminating state,
normal operations are no longer allowed. Only specially designated
operations related to graceful cleanup are permitted.
Arun Easi [Tue, 4 Aug 2015 17:38:02 +0000 (13:38 -0400)]
qla2xxx: Fix missing device login retries.
On certain conditions, login failures will just invoke
qla2x00_mark_device_lost() with the intend to do login again;
but if login_retry has been set already, that would fail to set the
relogin needed flag which is required to wakeup the DPC to retry.
Chad Dupuis [Tue, 4 Aug 2015 17:37:57 +0000 (13:37 -0400)]
qla2xxx: Do not reset adapter if SRB handle is in range.
If an SRB is NULL but the handle is in range just drop the
command instead of also resetting the adapter. If the handle
is in range then the command was valid at some point and may
have been aborted. Resetting the adapter can lead to extended
recovery times in this case.
Hiral Patel [Tue, 4 Aug 2015 17:37:56 +0000 (13:37 -0400)]
qla2xxx: Do not crash system for sp ref count zero
Aovid crashing the system in the scenario where firmware
just completes the command and it can not find the command
during abort mailbox processing. This scenario can lead to
sp reference counter being zero. Instead of crashing the
system, use WARN_ON to print warning in log file.
scsi_error: should not get sense for timeout IO in scsi error handler
scsi_error: should not get sense for timeout IO in scsi error handler
When an IO timeout occurs, the IO will be aborted in
scsi_abort_command() and SCSI_EH_ABORT_SCHEDULED will be set. Because
of that, the SCSI_EH_CANCEL_CMD will be clear in scsi_eh_scmd_add().
So when scsi error handler starts, it will get sense for this
timeout IO and the scmd of the IO request will be reused. In that
case, the scmd may be double released when racing with io_done(),
which will result in crash.
SO SCSI_EH_ABORT_SCHEDULED should also be checked when getting sense.
The bug maybe reproduced when the link between host and disk is
unstable.
Viswas G [Tue, 11 Aug 2015 09:36:30 +0000 (15:06 +0530)]
pm80xx: Add PORT RECOVERY TIMEOUT support
PORT RECOVERY TIMEOUT is the maximum time between the controller's
detection of the PHY down until the receipt of the ID_Frame (from the
same remote SAS port). If the time expires before the ID_FRAME is
received, the port is considered INVALID and can be removed. The
IOP_EVENT_PORT_RECOVERY_TIMER_TMO event is reported following the
IOP_EVENT_ PHY_DOWN event when the PHY/port does not recover after
Port Recovery Time.
Viswas G [Tue, 11 Aug 2015 09:36:29 +0000 (15:06 +0530)]
pm80xx: Remove unnecessary phy disconnect while link error
If the link error happens, we don't need to disconnect the phy,
which will remove the drive. Instead acknowledging the controller
and logging the error will be enough.
Viswas G [Tue, 11 Aug 2015 09:36:28 +0000 (15:06 +0530)]
pm80xx: Fix for Incorrect DMA Unmapping of SG List
In pm8001_ccb_task_free(), the dma unmapping is done based on
ccb->n_elem value. This should be initialized to zero in the
task_abort(). Otherwise, pm8001_ccb_task_free() will try for
dma_unmap_sg() which is invalid for task abort and can lead to
kernel crash.
qla2xxx: Return the fabric command state for non-task management requests
Invoking get_cmd_state for qla2xxx always returns 0. Instead change it
to return the actual fabric state from qla_tgt_cmd. This will help with
debugging.
A regression was introduced into the hpsa driver a while back so
non-zero LUNs of multi-LUN devices may no longer be presented via
a SAS based Smart Array. I have not done a bisection to discover
the change that caused it.
The CISS firmware specification (available on sourceforge)
defines an 8 byte lunid that describes devices that the Smart
Array can see/present to the system. The current code in the hpsa
driver attempts to find matches for non-zero LUNs with LUN 0 for
a bus/target by zeroing out byte 4 of the lunid and find a match.
This method is sufficient for SCSI based Smart Arrays because
byte 5 is always 0. For SAS based Smart arrays byte 5 of the
lunid contains the path number for a multipath device and
either one or two bits (the documentation does not define how
many bits are used but it appears it may be one only) that
indicate if the given path number in byte 5 must always be
used to access that device. Byte 5 may not always be zero.
The following are lunids (spaces added for clarity) for a
MSL2024 single drive library connected via a H241 Smart Array:
In the 4th byte (counting from 0) you can see that the tape
is LUN 0 and the changer is LUN 1. The 0x80 set in the 5th byte
for the tape drive means the driver should force access to
path 0 (the library in this case was connected to one path only
anyway).
After the changes we can see the following in the dmesg output:
scsi 0:3:0:0: RAID HP H241 1.18 \
PQ: 0 ANSI: 5
scsi 0:2:0:0: Sequential-Access HP Ultrium 6-SCSI 354W \
PQ: 0 ANSI: 6
scsi 0:2:0:1: Medium Changer HP MSL G3 Series 8.70 \
PQ: 0 ANSI: 5
Showing that the changer is correctly identified as LUN 1 of
bus 2 target 0. Before the change the changer device is not seen.
mpt2sas, mpt3sas: Abort initialization if no memory I/O resources detected
Driver crashes if the BIOS do not set up at least one
memory I/O resource. This failure can happen if the device is too
slow to respond during POST and is missed by the BIOS, but Linux
then detects the device later in the boot process.
The "return QLA_SUCCESS" statement just above the "fw_load_failed"
label cannot be reached, hence remove it. Additionally remove the
"else" keyword since the code block below the if-statement ends
with a return statement.
This patch removes 21 casts between an __iomem pointer type and
another data type but also introduces five new casts (see also
the casts with "__force"). Although this patch does not change
any functionality, IMHO the code with __force casts needs further
review.
Whether htonl() or __constant_htonl() is used, if the argument
is a constant the conversion happens at compile time. Hence leave
out the __constant_ prefix for this and other endianness
conversion functions. This improves source code readability.
qla2xxx: Replace two macros with an inline function
Replace the QLA82XX_ADDR_IN_RANGE() and QLA8044_ADDR_IN_RANGE() macros
with the inline function addr_in_range(). This avoids that the compiler
reports the following warning when building with W=1: comparison of
unsigned expression >= 0 is always true.
Destroy st_index_idr on module exit, reclaiming the allocated memory.
This was detected by the following semantic patch (written by Luis Rodriguez
<[email protected]>)
<SmPL>
@ defines_module_init @
declarer name module_init, module_exit;
declarer name DEFINE_IDR;
identifier init;
@@
projects / linux.git / log