Shuah Khan [Wed, 20 Feb 2013 17:24:34 +0000 (11:24 -0600)]
[SCSI] hpsa: Check for dma_mapping_error in hpsa_map_one
Signed-off-by: Shuah Khan <shuah.khan@hp.com> Signed-off-by: Stephen M. Cameron <scameron@beardog.cce.hp.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Dan Carpenter [Mon, 11 Feb 2013 19:03:18 +0000 (22:03 +0300)]
[SCSI] dc395x: uninitialized variable in device_alloc()
This bug was introduced back in bitkeeper days in 2003. We use
"dcb->dev_mode" before it has been initialized.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Acked-by: Oliver Neukum <oliver@neukum.org> Cc: <stable@vger.kernel.org> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
The range checking from fe542396 was bad. We would still end up walking
beyond the array as Type 3 is defined to be 4 in the protection
bitmask. Instead use ARRAY_SIZE() for the range check.
Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Brian King [Thu, 31 Jan 2013 05:45:08 +0000 (23:45 -0600)]
[SCSI] ipr: Fix oops while resetting an ipr adapter
When resetting an ipr adapter, we use scsi_block_requests to
block any new commands from scsi core, and then unblock after
the reset. When hotplug removing an adapter, we shut it down
and go through this same code, but we've seen issues with
scsi_unblock_requests running after the adapter's memory has
been freed. There is really no need to block/unblock when
the adapter is being removed, so this patch skips the
block/unblock and will immediately fail any commands that
happen to make it to queuecommand while the adapter is
being shutdown.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Hiral Patel [Wed, 13 Feb 2013 01:01:02 +0000 (17:01 -0800)]
[SCSI] fnic: Fnic Trace Utility
Fnic Trace utility is a tracing functionality built directly into fnic driver
to trace events. The benefit that trace buffer brings to fnic driver is the
ability to see what it happening inside the fnic driver. It also provides the
capability to trace every IO event inside fnic driver to debug panics, hangs
and potentially IO corruption issues. This feature makes it easy to find
problems in fnic driver and it also helps in tracking down strange bugs in a
more manageable way. Trace buffer is shared across all fnic instances for
this implementation.
Signed-off-by: Hiral Patel <hiralpat@cisco.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Hiral Patel [Wed, 13 Feb 2013 01:01:01 +0000 (17:01 -0800)]
[SCSI] fnic: New debug flags and debug log messages
Added new fnic debug flags for identifying IO state at every stage of IO while
debugging and also added more log messages for better debugging capability.
Hiral Patel [Wed, 13 Feb 2013 01:01:00 +0000 (17:01 -0800)]
[SCSI] fnic: fnic driver may hit BUG_ON on device reset
The issue was observed when LUN Reset is issued through IOCTL or sg_reset
utility.
fnic driver issues LUN RESET to firmware. On successful completion of device
reset, driver cleans up all the pending IOs that were issued prior to device
reset. These pending IOs are expected to be in ABTS_PENDING state. This works
fine, when the device reset operation resulted from midlayer, but not when
device reset was triggered from IOCTL path as the pending IOs were not in
ABTS_PENDING state. execution path hits panic if the pending IO is not in
ABTS_PENDING state.
Changes:
The fix replaces BUG_ON check in fnic_clean_pending_aborts() with marking
pending IOs as ABTS_PENDING if they were not in ABTS_PENDING state and skips
if they were already in ABTS_PENDING state. An extra check is added to validate
the abort status of the commands after a delay of 2 * E_D_TOV using a
helper function. The helper function returns 1 if it finds any pending IO in
ABTS_PENDING state, belong to the LUN on which device reset was issued else 0.
With this, device reset operation returns success only if the helper funciton
returns 0, otherwise it returns failure.
Other changes:
- Removed code in fnic_clean_pending_aborts() that returns failure if it finds
io_req NULL, instead of returning failure added code to continue with next io
- Added device reset flags for debugging in fnic_terminate_rport_io,
fnic_rport_exch_reset, and fnic_clean_pending_aborts
Hiral Patel [Wed, 13 Feb 2013 01:00:59 +0000 (17:00 -0800)]
[SCSI] fnic: Fix SGEs limit
Driver allows IOs with more SGEs than max SGEs supported by Palo. The current
max SGEs supported by the fnic driver is 1024. The current register settings
on Palo supports a max of 256 only. Palo would return any IO with more than
256 SGEs with an error indicating INVALID_SGLS. Fnic driver should limit the
max supported SGLs in the driver to 256 to avoid this error.
Hiral Patel [Wed, 13 Feb 2013 01:00:58 +0000 (17:00 -0800)]
[SCSI] fnic: fixing issues in device and firmware reset code
1. Handling overlapped firmware resets
This fix serialize multiple firmware resets to avoid situation where fnic
device fails to come up for link up event, when firmware resets are issued
back to back. If there are overlapped firmware resets are issued,
the firmware reset operation checks whether there is any firmware reset in
progress, if so it polls for its completion in a loop with 100ms delay.
2. Handling device reset timeout
fnic_device_reset code has been modified to handle Device reset timeout:
- Issue terminate on device reset timeout.
- Introduced flags field (one of the scratch fields in scsi_cmnd).
With this, device reset request would have DEVICE_RESET flag set for other
routines to determine the type of the request.
Also modified fnic_terminate_rport_io, fnic_rport_exch_rset, completion
routines to handle SCSI commands with DEVICE_RESET flag.
3. LUN/Device Reset hangs when issued through IOCTL using utilities like
sg_reset.
Each SCSI command is associated with a valid tag, fnic uses this tag to
retrieve associated scsi command on completion. the LUN/Device Reset issued
through IOCTL resulting into a SCSI command that is not associated with a
valid tag. So fnic fails to retrieve associated scsi command on completion,
which causes hang. This fix allocates tag, associates it with the
scsi command and frees the tag, when the operation completed.
4. Preventing IOs during firmware reset.
Current fnic implementation allows IO submissions during firmware reset.
This fix synchronizes IO submissions and firmware reset operations.
It ensures that IOs issued to fnic prior to reset will be issued to the
firmware before firmware reset.
Sakthivel K [Mon, 4 Feb 2013 06:40:02 +0000 (12:10 +0530)]
[SCSI] pm80xx: fix for memory region free
All memory regions are allocated based on variables total_len
and alignment but free was based on element_size.
Signed-off-by: Sakthivel K <DL.Team.PMC.SPC.Linux.open.source@pmcs.com> Signed-off-by: Anand Kumar S <DL.Team.PMC.SPC.Linux.open.source@pmcs.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
adam radford [Sat, 9 Feb 2013 23:29:20 +0000 (15:29 -0800)]
[SCSI] megaraid_sas: Dont load DevHandle unless FastPath enabled
The following patch for megaraid_sas will fix an issue where the
driver should not be loading the DevHandle unless FastPath is enabled.
If FastPath was not enabled, this means the hardware raid map
validation failed for some reason, or the map was corrupted, which
could mean the DevHandle could be invalid.
Signed-off-by: Adam Radford <aradford@gmail.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Chad Dupuis [Fri, 8 Feb 2013 06:58:04 +0000 (01:58 -0500)]
[SCSI] qla2xxx: Wait for IDC complete event to finish loopback operation.
Wait for the IDC complete AEN before returning the loopback operation back to
the application to make sure the port is put back into normal operations.
Signed-off-by: Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Chad Dupuis [Fri, 8 Feb 2013 06:58:01 +0000 (01:58 -0500)]
[SCSI] qla2xxx: Don't process RSCNs for a vport on the same physical adapter.
Currently,the driver is processes RSCNs for each new NPIV ports that is created.
Processing the RSCN includes a name server query to see what has changed at the
name server side. The name server query is performed by the physical port and
each virtual port on the physical adapter (since the RSCN is passed to each
virtual port for processing). As the number of virtual ports being created
increases, this causes a lot of traffic and busies the firmware.
Processing the RSCN for a virtual port we already have a priori knowledge of is
not necessary so check the 24-bit fabric ID of the RSCN entry and skip
processing it if the RSCN is for a virtual port we already know about.
Signed-off-by: Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Andrew Vasquez [Fri, 8 Feb 2013 06:57:58 +0000 (01:57 -0500)]
[SCSI] qla2xxx: Correct list-iteration bug in Report-ID Acquisition codes.
Code in qla24xx_report_id_acquisition() incorrectly assumed that
upon completion of list iteration (with no match), the 'pos'
(vp) variable passed to list_for_each_entry() would be set to
NULL. In this context, if the firmware were to return an
unrecognized vp_idx, the follow-on assignments to vp-members
could result in corruption of the structure.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Paul Bolle [Fri, 8 Feb 2013 06:57:55 +0000 (01:57 -0500)]
[SCSI] qla2xxx: silence two GCC warnings.
Compiling qla_gs.o (part of the qla2xxx module) triggers two GCC
warnings:
drivers/scsi/qla2xxx/qla_gs.c: In function ‘qla2x00_fdmi_rhba’:
drivers/scsi/qla2xxx/qla_gs.c:1339:7: warning: array subscript is above array bounds [-Warray-bounds]
drivers/scsi/qla2xxx/qla_gs.c: In function ‘qla2x00_fdmi_register’:
drivers/scsi/qla2xxx/qla_gs.c:1663:15: warning: array subscript is above array bounds [-Warray-bounds]
It seems that the sequence of a strcpy followed by a strlen confuses GCC
when it is keeping track of array bounds here. (It is not clear to me
which array triggers this warning and by how much GCC thinks the
subscript is above its bounds. Neither is it clear to me why comparable
code in these two functions doesn't trigger this warning.)
An easy way to silence these warnings is to use preprocessor macros and
strncpy, as that apparently gives GCC enough information to keep track
of array bounds.
Signed-off-by: Paul Bolle <pebolle@tiscali.nl> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Chad Dupuis [Fri, 8 Feb 2013 06:57:52 +0000 (01:57 -0500)]
[SCSI] qla2xxx: Move loopback mode reset after chip reset check.
If we need to do a chip reset because of a serious loopback error don't try to
reset the loopback mode on the port as the mailbox command will timeout.
Signed-off-by: Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Chad Dupuis [Fri, 8 Feb 2013 06:57:51 +0000 (01:57 -0500)]
[SCSI] qla2xxx: Extra loopback error handling for ISP83xx.
Add the following error handling for loopback diagnostic mode with ISP83xx:
1. If we do not receive an MBA_DCBX_COMPLETE after our initial set port
configuration command, try to reset the port back into normal operation.
If that fails, take a FCoE dump and then reset the chip.
2. After completing the loopback diagnostic operation, if the reset of the port
back into normal operation fails then reset the port so we take a FCoE dump
and then reset the chip.
3. When we receive an IDC notification and the requested operation is loopback
extend the loop down timer so the link does not appear to down for an
extended period of time.
[jejb: fix checkpatch issue] Signed-off-by: Chad Dupuis <chad.dupuis@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Andrew Vasquez [Fri, 8 Feb 2013 06:57:42 +0000 (01:57 -0500)]
[SCSI] qla2xxx: Correct race in loop_state assignment during reset handling.
There's a subtle race in the loop/bus-reset handling whereby a
VHA's loop-state can get incorrectly set to 'down' after the
loop-reset and firmware's completion of link re-negotiation. The
original code incorrectly assumes that firmware AENs would arrive
only after mailbox-command execution to initiate the link-flap.
Here's a good case with the old code (AENs arrive after
mailbox-command completion):
qla2xxx [0000:03:00.1]-8012:91: BUS RESET ISSUED nexus=91:0:4.
qla2xxx [0000:03:00.1]-287d:91: FCPort state transitioned from ONLINE to LOST - portid=010100.
qla2xxx [0000:03:00.1]-580e:91: Asynchronous P2P MODE received.
qla2xxx [0000:03:00.1]-287d:91: FCPort state transitioned from ONLINE to LOST - portid=010400.
qla2xxx [0000:03:00.1]-802b:91: BUS RESET SUCCEEDED nexus=91:0:4.
qla2xxx [0000:03:00.1]-480b:91: Reset marker scheduled.
qla2xxx [0000:03:00.1]-5812:91: Port database changed ffff 0006 0000.
qla2xxx [0000:03:00.1]-505f:91: Link is operational (4 Gbps).
qla2xxx [0000:03:00.1]-480c:91: Reset marker end.
qla2xxx [0000:03:00.1]-480f:91: Loop resync scheduled.
qla2xxx [0000:03:00.1]-8837:91: F/W Ready - OK.
qla2xxx [0000:03:00.1]-883a:91: fw_state=3 (7, 0, 0, 0) curr time=170b8f315.
qla2xxx [0000:03:00.1]-280e:91: HBA in F P2P topology.
qla2xxx [0000:03:00.1]-2812:91: qla2x00_configure_hba success
qla2xxx [0000:03:00.1]-2814:91: Configure loop -- dpc flags = 0x5260.
notice how the 'Port database changed' (8014) arrived after the
bus-reset handler completed 'BUS RESET SUCCEEDED'.
Now, here's a failing case with the old code (AENs arrive before
mailbox-command completion):
qla2xxx [0000:03:00.1]-8012:91: BUS RESET ISSUED nexus=91:0:0.
qla2xxx [0000:03:00.1]-580e:91: Asynchronous P2P MODE received.
qla2xxx [0000:03:00.1]-287d:91: FCPort state transitioned from ONLINE to LOST - portid=010100.
qla2xxx [0000:03:00.1]-287d:91: FCPort state transitioned from ONLINE to LOST - portid=010400.
qla2xxx [0000:03:00.1]-4800:91: DPC handler sleeping.
qla2xxx [0000:03:00.1]-5812:91: Port database changed ffff 0006 0000.
qla2xxx [0000:03:00.1]-505f:91: Link is operational (4 Gbps).
qla2xxx [0000:03:00.1]-802b:91: BUS RESET SUCCEEDED nexus=91:0:0.
qla2xxx [0000:03:00.1]-480b:91: Reset marker scheduled.
qla2xxx [0000:03:00.1]-480c:91: Reset marker end.
qla2xxx [0000:03:00.1]-480f:91: Loop resync scheduled.
qla2xxx [0000:03:00.1]-8837:91: F/W Ready - OK.
qla2xxx [0000:03:00.1]-883a:91: fw_state=3 (7, 0, 0, 0) curr time=170be9eb2.
qla2xxx [0000:03:00.1]-280e:91: HBA in F P2P topology.
qla2xxx [0000:03:00.1]-2812:91: qla2x00_configure_hba success
qla2xxx [0000:03:00.1]-2814:91: Configure loop -- dpc flags = 0x5260.
qla2xxx [0000:03:00.1]-281e:91: Needs RSCN update and loop transition.
qla2xxx [0000:03:00.1]-286a:91: qla2x00_configure_loop *** FAILED ***.
qla2xxx [0000:03:00.1]-4810:91: Loop resync end.
qla2xxx [0000:03:00.1]-4800:91: DPC handler sleeping.
This race would ultimately lead to devices go unexpectedly
offline until another link-flap or chip-reset would cause driver
re-discovery to take place.
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com> Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Xiangliang Yu [Tue, 29 Jan 2013 16:25:53 +0000 (00:25 +0800)]
[SCSI] mvsas: fixed timeout issue when removing module
Root cause is libsas will clear asd_sas_port phy_mask value in sas_port_deform
after triggering destruct workqueue, but the workqueue will send sync cmd and
still need phy_mask value. Now, mvsas using asd_sas_phy setting instead of
asd_sas_port setting.
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Julia Lawall [Mon, 21 Jan 2013 13:02:49 +0000 (14:02 +0100)]
[SCSI] bnx2fc: adjust duplicate test
Delete successive tests to the same location. The code tested the result
of a previous allocation, that itself was already tested. It is changed to
test the result of the most recent allocation.
A simplified version of the semantic match that finds this problem is as
follows: (http://coccinelle.lip6.fr/)
// <smpl>
@s exists@
local idexpression y;
expression x,e;
@@
Manish Rangankar [Mon, 21 Jan 2013 04:51:03 +0000 (23:51 -0500)]
[SCSI] qla4xxx: wait for boot target login response during probe.
Boot from SAN fails in some cases where boot target login response
is slow. In this case, probe finishes early and triggers init, which
doesn't find remote disk to mount root which leads to BFS failure.
Poornima Vonti [Mon, 21 Jan 2013 04:51:01 +0000 (23:51 -0500)]
[SCSI] qla4xxx: Re-register IRQ handler while retrying initialize of adapter
Problem:
If initialization of adapter fails for some reason, then the interrupt handlers
are released. The interrupt handlers are not registered again when initialization
of adapter is retried.
Solution:
Re-register the interrupt handler when adapter initialization is retried.
Karen Higgins [Mon, 21 Jan 2013 04:51:00 +0000 (23:51 -0500)]
[SCSI] qla4xxx: Throttle active IOCBs to firmware limits
Problem:
QLA4xxx firmware may assert if given more IOCBs than it can handle.
Solution:
The driver was updated to throttle the number of active IOCBs based
on the total number of IOCB buffers received from get_firmware_status
mbx_sts[2].
Nilesh Javali [Mon, 21 Jan 2013 04:50:56 +0000 (23:50 -0500)]
[SCSI] qla4xxx: Add spurious interrupt messages under debug level 2
The spurious interrupt messages are seen when interrupt lines are
shared between different adapters or modules. Example, interrupt
lines are shared between qlcnic and qla4xxx and USB modules, then the
console is flooded with flurry of spurious interrupt messages
which are expected in such scenario.
Hence put these messages under debug level 2 for INTx interrupt mode.
[SCSI] cxgb4i: Remove the scsi host device when removing device
When doing a hotplug remove of a cxgb4 device, there are still dandling
symlinks at /sys/class/scsi_host/hostX to the removed PCI device. The
upper layer device may also try to send data, which may crash the
system.
The DETACH message from the lower level driver is sent to the ULD when
the device is removed, when the scsi host should be removed from the
system, avoiding any problems.
After this patch, there are no more dangling symlinks and many attempts
to crash the system while there is SCSI activity and removing the device
have failed. Adding the device back again works as expected, with the
scsi hosts showing up again.
Based on a patch by Karen Xie.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@linux.vnet.ibm.com> Cc: Karen Xie <kxie@chelsio.com> Reviewed-by: Mike Christie <michaelc@cs.wisc.edu> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
This patch implements blk iopoll in ipr driver for performance improvement.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
This patch reduces lock contention while implementing distributed
completion processing.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
[SCSI] ipr: Add support for MSI-X and distributed completion
The new generation IBM SAS Controllers will support MSI-X interrupts and
Distributed Completion Processing features. This patch add these support
in ipr device driver.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Brian King [Fri, 11 Jan 2013 23:43:49 +0000 (17:43 -0600)]
[SCSI] ipr: Resource path error logging cleanup
The resource path as displayed by the ipr driver is the
location string identifying a location on the SAS fabric.
This patch adds the SCSI host number such that error logs
can be more easily correlated in multiple adapter configurations.
Signed-off-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
[SCSI] ipr: Add sereral new CCIN definitions for new adapters support
Add the appropriate definitions and table entries for new adapter support.
Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com> Acked-by: Brian King <brking@linux.vnet.ibm.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Jesper Juhl [Wed, 26 Dec 2012 20:31:51 +0000 (21:31 +0100)]
[SCSI] csiostor: Don't leak mem or fail to release firmware in csio_hw_flash_config()
If kzalloc() or csio_hw_check_fwconfig() fail we may leave the
csio_hw_flash_config() function without freeing allocated memory or
firmware. This should take care of the leaks.
[SCSI] bnx2fc: Map the doorbell register between offload and enable requests
We used to map doorbell register after FW enable request is complete.
This causes a race condition when unsolicited event is received, and FW
sends a CQE for it. Since the doorbell is not mapped, driver does not
arm CQ, which means FW will not notify the driver for further CQ
completions. To resolve this, map the doorbell between offload and
enable, so that driver is ready to receive the unsolicited packets and
arm the CQ as soon as FW enable is performed.
Signed-off-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
[SCSI] bnx2fc: Move offload/upload wait logic into a function
Moved the 'waiting for upload/offload completion' logic to respective
functions as it is called in multiple places. Also this is done in
preperation for the next patch.
Signed-off-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
Software FCoE target always advertises RETRY bit even when there are no
tape LUNs behind the target. This causes the driver to enable FW support
for sequence level error recovery and perform REC/SRR. This patch
arrests the behavior by not enabling SLER feature for this target.
Signed-off-by: Bhanu Prakash Gollapudi <bprakash@broadcom.com> Signed-off-by: James Bottomley <JBottomley@Parallels.com>
projects / linux-beck.git / log