In Current code, Req Q 0, RespQ 0 & hardware_lock are the main resources
for sending and process completion of Target IO. These resources are now
referenced behind a new qpair/"struct qla_qpair base_qpair". Main path
IO handle will access those resources via the qpair pointer in
preparation for Target MQ.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Merge active/outstanding cmd arrays from target side and initiator side
together in prepration for Target Multi Queue support.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Previously, the ql2xexchoffld module parameter was used to control the
max number of exchanges to be offload onto host memory.
Module parameter ql_dm_tgt_ex_pct was used to control the percentage of
exchanges allocated to the Target side.
With this patch, module parameter ql_dm_tgt_ex_pct is no longer used to
control exchanges for the driver. New module parameter ql2xiniexchg is
added to control exchanges between target mode and initiator mode.
With the updated module parameters, users can control the exact number
of exchanges for either Initiator or Target. The exchange offload
feature will be automatically enabled when the total number of exchanges
exceeds 2048 limit.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Hung task timeouts can result if a qlogic board breaks unexpectedly
while running I/O. These tasks become hung because command srb reference
counts are not going to zero, hence the affected srbs and commands do
not get freed. This fix accounts for this extra reference in the srbs in
the case of a board failure.
Fixes: a465537ad1 ("qla2xxx: Disable the adapter and skip error recovery in case of register disconnect")
Signed-off-by: Bill Kuzeja <william.kuzeja@stratus.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Fixes following signature in the stack trace:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000374
IP: [<ffffffffa06ec8eb>] qla2x00_sp_free_dma+0xeb/0x2a0 [qla2xxx]
Cc: <stable@vger.kernel.org> # v4.10+
Signed-off-by: Joe Carnuccio <joe.carnuccio@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When pci_enable_device() or pci_enable_device_mem() fail in
qla2x00_probe_one() we bail out but do a call to
pci_disable_device(). This causes the dev_WARN_ON() in
pci_disable_device() to trigger, as the device wasn't enabled
previously.
So instead of taking the 'probe_out' error path we can directly return
*iff* one of the pci_enable_device() calls fails.
Additionally rename the 'probe_out' goto label's name to the more
descriptive 'disable_device'.
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Fixes: e315cd28b9 ("[SCSI] qla2xxx: Code changes for qla data structure refactoring")
Cc: <stable@vger.kernel.org>
Reviewed-by: Bart Van Assche <bart.vanassche@sandisk.com>
Reviewed-by: Giridhar Malavali <giridhar.malavali@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull SCSI fixes from James Bottomley:
"This is seven small fixes which are all for user visible issues that
fortunately only occur in rare circumstances.
The most serious is the sr one in which QEMU can cause us to read
beyond the end of a buffer (I don't think it's exploitable, but just
in case).
The next is the sd capacity fix which means all non 512 byte sector
drives greater than 2TB fail to be correctly sized.
The rest are either in new drivers (qedf) or on error legs"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: ipr: do not set DID_PASSTHROUGH on CHECK CONDITION
scsi: aacraid: fix PCI error recovery path
scsi: sd: Fix capacity calculation with 32-bit sector_t
scsi: qla2xxx: Add fix to read correct register value for ISP82xx.
scsi: qedf: Fix crash due to unsolicited FIP VLAN response.
scsi: sr: Sanity check returned mode data
scsi: sd: Consider max_xfer_blocks if opt_xfer_blocks is unusable
There are several local or function parameter pointers that are being
assigned NULL after a kfree where and these have no effect and hence can
be removed.
Fixes various cppcheck warnings:
"Assignment of function parameter has no effect outside the
function. Did you forget dereferencing it"
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Add fix to read correct register value for ISP82xx, during check for
register disconnect.ISP82xx has different base register.
Fixes: a465537ad1 ("qla2xxx: Disable the adapter and skip error recovery in case of register disconnect")
Signed-off-by: Sawan Chandak <sawan.chandak@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull SCSI fixes from James Bottomley:
"Thirteen small fixes: The hopefully final effort to get the lpfc nvme
kconfig problems sorted, there's one important sg fix (user can induce
read after end of buffer) and one minor enhancement (adding an extra
PCI ID to qedi). The rest are a set of minor fixes, which mostly occur
as user visible in error legs or on specific devices"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: ufs: remove the duplicated checking for supporting clkscaling
scsi: lpfc: fix building without debugfs support
scsi: lpfc: Fix PT2PT PRLI reject
scsi: hpsa: fix volume offline state
scsi: libsas: fix ata xfer length
scsi: scsi_dh_alua: Warn if the first argument of alua_rtpg_queue() is NULL
scsi: scsi_dh_alua: Ensure that alua_activate() calls the completion function
scsi: scsi_dh_alua: Check scsi_device_get() return value
scsi: sg: check length passed to SG_NEXT_CMD_LEN
scsi: ufshcd-platform: remove the useless cast in ERR_PTR/IS_ERR
scsi: qedi: Add PCI device-ID for QL41xxx adapters.
scsi: aacraid: Fix potential null access
scsi: qla2xxx: Fix crash in qla2xxx_eh_abort on bad ptr
Current driver wait for FW to be in the ready state before
processing in-coming commands. For Arbitrated Loop or
Point-to- Point (not switch), FW Ready state can take a while.
FW will transition to ready state after all Nports have been
logged in. In the mean time, certain initiators have completed
the login and starts IO. Driver needs to start processing all
queues if FW is already started.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
For target mode, when new scsi command arrive, driver first performs
a look up of the SCSI Host. The current look up method is based on
the ALPA portion of the NPort ID. For Cisco switch, the ALPA can
not be used as the index. Instead, the new search method is based
on the full value of the Nport_ID via btree lib.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
After a Qlogic card breaks when initializing (test case), the system can
crash in qla2xxx_eh_abort if processing anything but a scsi command type
srb.
Fixes: 1535aa75a3 ("scsi: qla2xxx: fix invalid DMA access after command aborts in PCI device remove")
Signed-off-by: Bill Kuzeja <william.kuzeja@stratus.com>
Acked-By: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull more SCSI updates from James Bottomley:
"This is the set of stuff that didn't quite make the initial pull and a
set of fixes for stuff which did.
The new stuff is basically lpfc (nvme), qedi and aacraid. The fixes
cover a lot of previously submitted stuff, the most important of which
probably covers some of the failing irq vectors allocation and other
fallout from having the SCSI command allocated as part of the block
allocation functions"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (59 commits)
scsi: qedi: Fix memory leak in tmf response processing.
scsi: aacraid: remove redundant zero check on ret
scsi: lpfc: use proper format string for dma_addr_t
scsi: lpfc: use div_u64 for 64-bit division
scsi: mac_scsi: Fix MAC_SCSI=m option when SCSI=m
scsi: cciss: correct check map error.
scsi: qla2xxx: fix spelling mistake: "seperator" -> "separator"
scsi: aacraid: Fixed expander hotplug for SMART family
scsi: mpt3sas: switch to pci_alloc_irq_vectors
scsi: qedf: fixup compilation warning about atomic_t usage
scsi: remove scsi_execute_req_flags
scsi: merge __scsi_execute into scsi_execute
scsi: simplify scsi_execute_req_flags
scsi: make the sense header argument to scsi_test_unit_ready mandatory
scsi: sd: improve TUR handling in sd_check_events
scsi: always zero sshdr in scsi_normalize_sense
scsi: scsi_dh_emc: return success in clariion_std_inquiry()
scsi: fix memory leak of sdpk on when gd fails to allocate
scsi: sd: make sd_devt_release() static
scsi: qedf: Add QLogic FastLinQ offload FCoE driver framework.
...
Pull SCSI target updates from Nicholas Bellinger:
"The highlights this round include:
- enable dual mode (initiator + target) qla2xxx operation. (Quinn +
Himanshu)
- add a framework for qla2xxx async fabric discovery. (Quinn +
Himanshu)
- enable iscsi PDU DDP completion offload in cxgbit/T6 NICs. (Varun)
- fix target-core handling of aborted failed commands. (Bart)
- fix a long standing target-core issue NULL pointer dereference with
active I/O LUN shutdown. (Rob Millner + Bryant + nab)"
* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending: (44 commits)
target: Add counters for ABORT_TASK success + failure
iscsi-target: Fix early login failure statistics misses
target: Fix NULL dereference during LUN lookup + active I/O shutdown
target: Delete tmr from list before processing
target: Fix handling of aborted failed commands
uapi: fix linux/target_core_user.h userspace compilation errors
target: export protocol identifier
qla2xxx: Fix a warning reported by the "smatch" static checker
target/iscsi: Fix unsolicited data seq_end_offset calculation
target/cxgbit: add T6 iSCSI DDP completion feature
target/cxgbit: Enable DDP for T6 only if data sequence and pdu are in order
target/cxgbit: Use T6 specific macros to get ETH/IP hdr len
target/cxgbit: use cxgb4_tp_smt_idx() to get smt idx
target/iscsi: split iscsit_check_dataout_hdr()
target: Remove command flag CMD_T_DEV_ACTIVE
target: Remove command flag CMD_T_BUSY
target: Move session check from target_put_sess_cmd() into target_release_cmd_kref()
target: Inline transport_cmd_check_stop()
target: Remove an overly chatty debug message
target: Stop execution if CMD_T_STOP has been set
...
This patch cleaned up queue configuration code, such that once
initialized, we should not touch msix_count value. This will prevent
incorrect numbers of MSI-X vectors requested while performing target
mode configuration.
[mkp: fixed Fixes: hash]
Cc: <stable@vger.kernel.org>
Fixes: d74595278f ("scsi: qla2xxx: Add multiple queue pair functionality.")
Signed-off-by: Michael Hernandez <michael.hernandez@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull SCSI updates from James Bottomley:
"This update includes the usual round of major driver updates (ncr5380,
ufs, lpfc, be2iscsi, hisi_sas, storvsc, cxlflash, aacraid,
megaraid_sas, ...).
There's also an assortment of minor fixes and the major update of
switching a bunch of drivers to pci_alloc_irq_vectors from Christoph"
* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi: (188 commits)
scsi: megaraid_sas: handle dma_addr_t right on 32-bit
scsi: megaraid_sas: array overflow in megasas_dump_frame()
scsi: snic: switch to pci_irq_alloc_vectors
scsi: megaraid_sas: driver version upgrade
scsi: megaraid_sas: Change RAID_1_10_RMW_CMDS to RAID_1_PEER_CMDS and set value to 2
scsi: megaraid_sas: Indentation and smatch warning fixes
scsi: megaraid_sas: Cleanup VD_EXT_DEBUG and SPAN_DEBUG related debug prints
scsi: megaraid_sas: Increase internal command pool
scsi: megaraid_sas: Use synchronize_irq to wait for IRQs to complete
scsi: megaraid_sas: Bail out the driver load if ld_list_query fails
scsi: megaraid_sas: Change build_mpt_mfi_pass_thru to return void
scsi: megaraid_sas: During OCR, if get_ctrl_info fails do not continue with OCR
scsi: megaraid_sas: Do not set fp_possible if TM capable for non-RW syspdIO, change fp_possible to bool
scsi: megaraid_sas: Remove unused pd_index from megasas_build_ld_nonrw_fusion
scsi: megaraid_sas: megasas_return_cmd does not memset IO frame to zero
scsi: megaraid_sas: max_fw_cmds are decremented twice, remove duplicate
scsi: megaraid_sas: update can_queue only if the new value is less
scsi: megaraid_sas: Change max_cmd from u32 to u16 in all functions
scsi: megaraid_sas: set pd_after_lb from MR_BuildRaidContext and initialize pDevHandle to MR_DEVHANDLE_INVALID
scsi: megaraid_sas: latest controller OCR capability from FW before sending shutdown DCMD
...
Pull SCSI fixes from James Bottomley:
"Six fairly small fixes. None is a real show stopper, two automation
detected problems: one memory leak, one use after free and four others
each of which fixes something that has been a significant source of
annoyance to someone"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: zfcp: fix use-after-free by not tracing WKA port open/close on failed send
scsi: aacraid: Fix INTx/MSI-x issue with older controllers
scsi: mpt3sas: disable ASPM for MPI2 controllers
scsi: mpt3sas: Force request partial completion alignment
scsi: qla2xxx: Avoid that issuing a LIP triggers a kernel crash
scsi: qla2xxx: Fix a recently introduced memory leak
This patch simplifies SRB structure usage in driver.
- Simplify sp->done() and sp->free() interfaces.
- Remove sp->fcport->vha to use vha pointer from sp.
- Use sp->vha context in qla2x00_rel_sp().
Signed-off-by: Joe Carnuccio <joe.carnuccio@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Current code blindly does State Change Registration when
the link is up. Move SCR behind fabric scan, so that arbitrated
loop scan would not get erroneous error message.
Some of the other improvements are as follows
- Add session deletion for TPRLO and send acknowledgment for TPRLO.
- Enable FW option to move ABTS, RIDA & PUREX from RSPQ to ATIOQ.
- Save NPort ID early in link init.
- Move ABTS & RIDA to ATIOQ helps in keeping command ordering and
link up sequence ordering.
- Save Nport ID and update VP map so that SCSI CMD/ATIO won't be dropped.
- fcport alloc does the initializes memory to zero. Remove memset to
zero since It might corrupt link list.
- Turn off Registration for State Change MB in loop mode.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Currently code performs a full scan of the fabric for
every RSCN. Its an expensive process in a noisy large SAN.
This patch optimizes expensive fabric discovery process by
scanning switch for the affected port when RSCN is received.
Currently Initiator Mode code makes login/logout decision without
knowledge of target mode. This causes driver and firmware to go
out-of-sync. This framework synchronizes both initiator mode
personality and target mode personality in making login/logout
decision.
This patch adds following capabilities in the driver
- Send Notification Acknowledgement asynchronously.
- Update session/fcport state asynchronously.
- Create a session or fcport struct asynchronously.
- Send GNL asynchronously. The command will ask FW to
provide a list of FC Port entries FW knows about.
- Send GPDB asynchronously. The command will ask FW to
provide detail data of an FC Port FW knows about or
perform ADISC to verify the state of the session.
- Send GPNID asynchronously. The command will ask switch
to provide WWPN for provided NPort ID.
- Send GPSC asynchronously. The command will ask switch
to provide registered port speed for provided WWPN.
- Send GIDPN asynchronously. The command will ask the
switch to provide Nport ID for provided WWPN.
- In driver unload path, schedule all session for deletion
and wait for deletion to complete before allowing driver
unload to proceed.
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
[ bvanassche: fixed spelling in patch description ]
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
Instead define the timeout behavior purely based on the host_template
eh_timed_out method and wire up the existing transport implementations
in the host templates. This also clears up the confusion that the
transport template method overrides the host template one, so some
drivers have to re-override the transport template one.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Tyrel Datwyler <tyreld@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Pull SCSI target fixes from Bart Van Assche:
- two small fixes for the ibmvscsis driver
- ten patches with bug fixes for the target mode of the qla2xxx driver
- four patches that avoid that the "sparse" and "smatch" static
analyzer tools report false positives for the qla2xxx code base
* 'scsi-target-for-v4.10' of git://git.kernel.org/pub/scm/linux/kernel/git/bvanassche/linux:
qla2xxx: Disable out-of-order processing by default in firmware
qla2xxx: Fix erroneous invalid handle message
qla2xxx: Reduce exess wait during chip reset
qla2xxx: Terminate exchange if corrupted
qla2xxx: Fix crash due to null pointer access
qla2xxx: Collect additional information to debug fw dump
qla2xxx: Reset reserved field in firmware options to 0
qla2xxx: Set tcm_qla2xxx version to automatically track qla2xxx version
qla2xxx: Include ATIO queue in firmware dump when in target mode
qla2xxx: Fix wrong IOCB type assumption
qla2xxx: Avoid that building with W=1 triggers complaints about set-but-not-used variables
qla2xxx: Move two arrays from header files to .c files
qla2xxx: Declare an array with file scope static
qla2xxx: Fix indentation
ibmvscsis: Fix sleeping in interrupt context
ibmvscsis: Fix max transfer length
During code inspection, while investigating following stack trace
seen on one of the test setup, we found out there was possibility
of memory leak becuase driver was not unwinding the stack properly.
This issue has not been reproduced in a test environment or on a
customer setup.
Here's stack trace that was seen.
[1469877.797315] Call Trace:
[1469877.799940] [<ffffffffa03ab6e9>] qla2x00_mem_alloc+0xb09/0x10c0 [qla2xxx]
[1469877.806980] [<ffffffffa03ac50a>] qla2x00_probe_one+0x86a/0x1b50 [qla2xxx]
[1469877.814013] [<ffffffff813b6d01>] ? __pm_runtime_resume+0x51/0xa0
[1469877.820265] [<ffffffff8157c1f5>] ? _raw_spin_lock_irqsave+0x25/0x90
[1469877.826776] [<ffffffff8157cd2d>] ? _raw_spin_unlock_irqrestore+0x6d/0x80
[1469877.833720] [<ffffffff810741d1>] ? preempt_count_sub+0xb1/0x100
[1469877.839885] [<ffffffff8157cd0c>] ? _raw_spin_unlock_irqrestore+0x4c/0x80
[1469877.846830] [<ffffffff81319b9c>] local_pci_probe+0x4c/0xb0
[1469877.852562] [<ffffffff810741d1>] ? preempt_count_sub+0xb1/0x100
[1469877.858727] [<ffffffff81319c89>] pci_call_probe+0x89/0xb0
Cc: <stable@vger.kernel.org>
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
[ bvanassche: Fixed spelling in patch description ]
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Commit 093df73771 ("scsi: qla2xxx: Fix Target mode handling with
Multiqueue changes.") introduces two bodies of code that look similar
but with s/req/rsp/ in the second instance. But in one case, it looks
like this conversion was missed.
Signed-off-by: Dave Jones <davej@codemonkey.org.uk>
Reviewed-by: Laurence Oberman <loberman@redhat.com>
Acked-by: Quinn Tran <Quinn.Tran@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
- Fix race condition between dpc_thread accessing Multiqueue resources
and qla2x00_remove_one thread trying to free resource.
- Fix out of order free for Multiqueue resources. Also, Multiqueue
interrupts needs a workqueue. Interrupt needed to stop before
the wq can be destroyed.
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Quinn Tran <quinn.tran@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Tell the SCSI layer how many hardware queues we have based on the number
of max queue pairs created. The number of max queue pairs created will
depend on number of MSI-X vector count.
This feature can be turned on via CONFIG_SCSI_MQ_DEFAULT or passing
scsi_mod.use_blk_mq=Y as a parameter to the kernel
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Sawan Chandak <sawan.chandak@cavium.com>
Signed-off-by: Michael Hernandez <michael.hernandez@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Replaced existing multiple queue functionality with framework
that allows for the creation of pairs of request and response queues,
either at start of day or dynamically.
Queue pair creation depend on module parameter "ql2xmqsupport",
which need to be enabled to create queue pair.
Signed-off-by: Sawan Chandak <sawan.chandak@cavium.com>
Signed-off-by: Michael Hernandez <michael.hernandez@cavium.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Acked-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The previous commit 1535aa75a3 ("qla2xxx: fix invalid DMA access after
command aborts in PCI device remove") introduced a regression during an
EEH recovery, since the change to the qla2x00_abort_all_cmds() function
calls qla2xxx_eh_abort(), which verifies the EEH recovery condition but
handles it heavy-handed. (commit a465537ad1 "qla2xxx: Disable the
adapter and skip error recovery in case of register disconnect.")
This problem warrants a more general/optimistic solution right into
qla2xxx_eh_abort() (eg in case a real command abort arrives during EEH
recovery, or if it takes long enough to trigger command aborts); but
it's still worth to add a check to ensure the code added by the previous
commit is correct and contained within its owner function.
This commit just adds a 'if (!ha->flags.eeh_busy)' check around it.
(ahem; a trivial fix for this -rc series; sorry for this oversight.)
With it applied, both PCI device remove and EEH recovery works fine.
Fixes: 1535aa75a3 ("scsi: qla2xxx: fix invalid DMA access after command aborts in PCI device remove")
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
If a command is aborted in the kernel but not in the adapter, it might be
considered complete and its DMA memory released, but it is still alive in
the adapter, which will trigger an invalid DMA access upon its completion
(in the DMA operations to deliver the command response to the driver).
On powerpc platforms with IOMMU/EEH capabilities, the problem is observed
during PCI device removal with ongoing IO requests -- which might trigger
an EEH event very often, pointing to a 'TCE Request Page Access Error'.
In that path, which is qla2x00_remove_one(), the commands are aborted in
qla2x00_abort_all_cmds(), which does not perform an abort in the adapter
as is done in qla2xxx_eh_abort() for example.
So, this patch changes qla2x00_abort_all_cmds() to abort commands in the
adapter too, with a call to qla2xxx_eh_abort(), which already implements
all the logic to submit abort requests and handle responses.
Reported-by: Naresh Bannoth <nbannoth@in.ibm.com>
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
When the driver is unloading, in qla2x00_remove_one(), there is a single
call/point in time to abort ongoing commands, qla2x00_abort_all_cmds(),
which is still several steps away from the call to scsi_remove_host().
If more commands continue to arrive and be processed during that
interval, when the driver is tearing down and releasing its structures,
it might potentially hit an oops due to invalid memory access:
Unable to handle kernel paging request for data at address 0x00000138
<...>
NIP [d000000004700a40] qla2xxx_queuecommand+0x80/0x3f0 [qla2xxx]
LR [d000000004700a10] qla2xxx_queuecommand+0x50/0x3f0 [qla2xxx]
So, fail commands in qla2xxx_queuecommand() if the UNLOADING bit is set.
Signed-off-by: Mauricio Faria de Oliveira <mauricfo@linux.vnet.ibm.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
A system can get hung task timeouts if a qlogic board fails during
initialization (if the board breaks again or fails the init). The hang
involves the scsi scan.
In a nutshell, since commit beb9e315e6 ("qla2xxx: Prevent removal and
board_disable race"):
...it is possible to have freed ha (base_vha->hw) early by a call to
qla2x00_remove_one when pdev->enable_cnt equals zero:
if (!atomic_read(&pdev->enable_cnt)) {
scsi_host_put(base_vha->host);
kfree(ha);
pci_set_drvdata(pdev, NULL);
return;
Almost always, the scsi_host_put above frees the vha structure
(attached to the end of the Scsi_Host we're putting) since it's the last
put, and life is good. However, if we are entering this routine because
the adapter has broken sometime during initialization AND a scsi scan is
already in progress (and has done its own scsi_host_get), vha will not
be freed. What's worse, the scsi scan will access the freed ha structure
through qla2xxx_scan_finished:
if (time > vha->hw->loop_reset_delay * HZ)
return 1;
The scsi scan keeps checking to see if a scan is complete by calling
qla2xxx_scan_finished. There is a timeout value that limits the length
of time a scan can take (hw->loop_reset_delay, usually set to 5
seconds), but this definition is in the data structure (hw) that can get
freed early.
This can yield unpredictable results, the worst of which is that the
scsi scan can hang indefinitely. This happens when the freed structure
gets reused and loop_reset_delay gets overwritten with garbage, which
the scan obliviously uses as its timeout value.
The fix for this is simple: at the top of qla2xxx_scan_finished, check
for the UNLOADING bit in the vha structure (_vha is not freed at this
point). If UNLOADING is set, we exit the scan for this adapter
immediately. After this last reference to the ha structure, we'll exit
the scan for this adapter, and continue on.
This problem is hard to hit, but I have run into it doing negative
testing many times now (with a test specifically designed to bring it
out), so I can verify that this fix works. My testing has been against a
RHEL7 driver variant, but the bug and patch are equally relevant to to
the upstream driver.
Fixes: beb9e315e6 ("qla2xxx: Prevent removal and board_disable race")
Cc: <stable@vger.kernel.org> # v3.18+
Signed-off-by: Bill Kuzeja <william.kuzeja@stratus.com>
Acked-by: Himanshu Madhani <himanshu.madhani@cavium.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
The "if (test_bit(UNLOADING..." line was indented one tab more than it
should have been. There was an extra parenthesis around the
qla2x00_reset_active() function call. I lined up the conditions a bit
so that it shows how they group together.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
During hot swap of PCI device, there can be PCI error on device,
during normal driver unload. The race between normal driver unload and
driver unload due to PCI error, can lead to system crash.Fix is to check
if there is unload going on and allow that function to unload the driver.
Signed-off-by: Sawan Chandak <sawan.chandak@qlogic.com>
Signed-off-by: Himanshu Madhani <himanshu.madhani@qlogic.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
