9 years ago[SCSI] zfcp: Add allow_lun_scan module parameter
Christof Schmitt [Wed, 8 Dec 2010 16:30:46 +0000]
[SCSI] zfcp: Add allow_lun_scan module parameter

The zfcpdump tool requires a method to attach exactly one LUN. The
easiest way to achieve this is to add a new zfcp module parameter.
When allow_lun_scan is set to "false", zfcp only accepts LUNs that
have been configured through the unit_add sysfs interface.

Reviewed-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Add __init declaration to zfcp_cache_hw_align
Christof Schmitt [Thu, 2 Dec 2010 14:16:18 +0000]
[SCSI] zfcp: Add __init declaration to zfcp_cache_hw_align

The function zfcp_cache_hw_align is only called from zfcp_module_init,
so it should be declared with __init as well.

Reviewed-by: Steffen Maier <maier@linux.vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Move qdio setup from erp to zfcp_qdio.c
Christof Schmitt [Thu, 2 Dec 2010 14:16:17 +0000]
[SCSI] zfcp: Move qdio setup from erp to zfcp_qdio.c

Initialization of the qdio waitqueue should happen when the qdio data
is initialized and the QDIOUP flag should be handled in the qdio code
as well. Adjust the code accordingly and remove the superfluos
function zfcp_erp_adapter_strategy_open_qdio.

Reviewed-by: Steffen Maier <maier@linux.vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Redesign of the debug tracing final cleanup.
Swen Schillig [Thu, 2 Dec 2010 14:16:16 +0000]
[SCSI] zfcp: Redesign of the debug tracing final cleanup.

This patch is the final cleanup of the redesign from the zfcp tracing.
Structures and elements which were used by multiple areas of the
former debug tracing are now changed to the new scheme.

Signed-off-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Redesign of the debug tracing for SCSI records.
Swen Schillig [Thu, 2 Dec 2010 14:16:15 +0000]
[SCSI] zfcp: Redesign of the debug tracing for SCSI records.

This patch is the continuation to redesign the zfcp tracing to a more
straight-forward and easy to extend scheme.

This patch deals with all trace records of the zfcp SCSI area.

Signed-off-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Redesign of the debug tracing for HBA records.
Swen Schillig [Thu, 2 Dec 2010 14:16:14 +0000]
[SCSI] zfcp: Redesign of the debug tracing for HBA records.

This patch is the continuation to redesign the zfcp tracing to a more
straight-forward and easy to extend scheme.

This patch deals with all trace records of the zfcp HBA area.

Signed-off-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Redesign of the debug tracing for SAN records.
Swen Schillig [Thu, 2 Dec 2010 14:16:13 +0000]
[SCSI] zfcp: Redesign of the debug tracing for SAN records.

This patch is the continuation to redesign the zfcp tracing to a more
straight-forward and easy to extend scheme.

This patch deals with all trace records of the zfcp SAN area.

Signed-off-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] zfcp: Redesign of the debug tracing for recovery actions.
Swen Schillig [Thu, 2 Dec 2010 14:16:12 +0000]
[SCSI] zfcp: Redesign of the debug tracing for recovery actions.

The tracing environment of the zfcp LLD has become very bulky and hard
to maintain. Small changes involve a large modification process which
is error-prone and not effective.  This patch is the first of a set to
redesign the zfcp tracing to a more straight-forward and easy to
extend scheme.  It removes all interpretation and visualization parts
and focuses on bare logging of the information.

This patch deals with all trace records of the zfcp error recovery.

Signed-off-by: Swen schillig <swen@vnet.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: Update driver version to 5.02.00-k5
Vikas Chaudhary [Fri, 3 Dec 2010 06:13:00 +0000]
[SCSI] qla4xxx: Update driver version to 5.02.00-k5

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: Updated the Copyright header
Vikas Chaudhary [Fri, 3 Dec 2010 06:12:51 +0000]
[SCSI] qla4xxx: Updated the Copyright header

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: do not reset hba if ql4xdontresethba is set
Vikas Chaudhary [Fri, 3 Dec 2010 06:12:47 +0000]
[SCSI] qla4xxx: do not reset hba if ql4xdontresethba is set

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: do not check for fw hung if reset retry is in progress
Lalit Chandivade [Fri, 3 Dec 2010 06:12:45 +0000]
[SCSI] qla4xxx: do not check for fw hung if reset retry is in progress

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Lalit Chandivade <lalit.chandivade@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: cache new IP address acquired via DHCP
Prasanna Mumbai [Fri, 3 Dec 2010 06:12:43 +0000]
[SCSI] qla4xxx: cache new IP address acquired via DHCP

Prior to firmware state change from ACQUIRING to READY, an
0x8029 AEN is received. Added code to check previous state
being ACQUIRING in order to update the ip address in the driver.

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Prasanna Mumbai <prasanna.mumbai@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: Fail initialization if qla4_8xxx_pci_mem_write_2M fails
Lalit Chandivade [Fri, 3 Dec 2010 06:12:40 +0000]
[SCSI] qla4xxx: Fail initialization if qla4_8xxx_pci_mem_write_2M fails

Since if fw load is failing, running on incomplete fw load would
be fatal.

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Lalit Chandivade <lalit.chandivade@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: put device in FAILED state for 82XX initialization failure
Lalit Chandivade [Fri, 3 Dec 2010 06:12:36 +0000]
[SCSI] qla4xxx: put device in FAILED state for 82XX initialization failure

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Lalit Chandivade <lalit.chandivade@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: do not process interrupt unconditionally
Lalit Chandivade [Fri, 3 Dec 2010 06:12:27 +0000]
[SCSI] qla4xxx: do not process interrupt unconditionally

in mailbox command do not process interrupt unconditionally,
process interrupt only in polling mode

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Lalit Chandivade <lalit.chandivade@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: use correct fw_ddb_index in abort task
Karen Higgins [Fri, 3 Dec 2010 06:12:22 +0000]
[SCSI] qla4xxx: use correct fw_ddb_index in abort task

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Karen Higgins <karen.higgins@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: memory wedge with peg_halt test in loop
Swapnil Nagle [Fri, 3 Dec 2010 06:12:15 +0000]
[SCSI] qla4xxx: memory wedge with peg_halt test in loop

Signed-off-by: Swapnil Nagle <swapnil.nagle@qlogic.com>
Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: initialize MSI in correct way
Shyam Sundar [Fri, 3 Dec 2010 06:12:08 +0000]
[SCSI] qla4xxx: initialize MSI in correct way

IRQF_SHARED flag should not be set when calling request_irq for MSI since
this interrupt mechanism cannot be shared like standard INTx

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Shyam Sundar <shyam.sundar@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] qla4xxx: Drop use of IRQF_DISABLE
Vikas Chaudhary [Fri, 3 Dec 2010 06:12:03 +0000]
[SCSI] qla4xxx: Drop use of IRQF_DISABLE

IRQF_DISABLE flag is deprecated and this flag is a NOOP in kernel.

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: Ravi Anand <ravi.anand@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix statistics for FCP input/output megabytes
Joe Eykholt [Wed, 1 Dec 2010 00:20:18 +0000]
[SCSI] libfc: fix statistics for FCP input/output megabytes

The statistics for InputMegabytes and OutputMegabytes are
misnamed.  They're accumulating bytes, not megabytes.

The statistic returned via /sys must be in megabytes, however,
which is what the HBA-API wants.  The FCP code needs to accumulate
it in bytes and then divide by 1,000,000 (not 2^20) before it
presented via sysfs.

This affects fcoe.ko only, not fnic.  The fnic driver
correctly by accumulating bytes and then converts to megabytes.

I checked that libhbalinux is using the /sys file directly without
conversion.

BTW, qla2xxx does divide by 2^20, which I'm not fixing here.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: change fip_select to return new FCF
Joe Eykholt [Wed, 1 Dec 2010 00:20:12 +0000]
[SCSI] libfcoe: change fip_select to return new FCF

Neaten several calls to fip_select() by having it return the
pointer to the new FCF.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: reorder FCF list to put latest advertiser first
Joe Eykholt [Wed, 1 Dec 2010 00:20:07 +0000]
[SCSI] libfcoe: reorder FCF list to put latest advertiser first

When there are several FCFs to choose from, the one most likely
to accept a FLOGI on certian switches is the one that last
answered a multicast solicit.

So, when receiving an advertisement, move the FCF to the front
of the list so that it gets chosen first among those with the
same priority.

Without this, more FLOGIs need to be sent in a test with
multiple FCFs and a switch in NPV mode, but it still
eventually finds one that accepts the FLOGI.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: add debug message for FCF destination MAC
Joe Eykholt [Wed, 1 Dec 2010 00:20:02 +0000]
[SCSI] libfcoe: add debug message for FCF destination MAC

When multiple FCFs to the same fabric exist, the debug messages
all look alike.   Change the message to include the MAC address.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: retry rejected FLOGI to another FCF if possible
Joe Eykholt [Wed, 1 Dec 2010 00:19:56 +0000]
[SCSI] libfcoe: retry rejected FLOGI to another FCF if possible

Switches using multiple-FCFs may reject FLOGI in order to
balance the load between multiple FCFs.  Even though the FCF
was available, it may have more load at the point we actually
send the FLOGI.

If the FLOGI fails, select a different FCF
if possible, among those with the same priority.  If no other
FCF is available, just deliver the reject to libfc for retry.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: fix checking of conflicting fabrics in fcoe_ctlr_select()
Joe Eykholt [Wed, 1 Dec 2010 00:19:51 +0000]
[SCSI] libfcoe: fix checking of conflicting fabrics in fcoe_ctlr_select()

The check for conflicting fabrics in fcoe_ctlr_select()
ignores any FCFs that aren't usable.  This is a minor
problem now but becomes more pronounced after later patches.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: move some timer code to make it reusable.
Joe Eykholt [Wed, 1 Dec 2010 00:19:46 +0000]
[SCSI] libfcoe: move some timer code to make it reusable.

Move some of the code in fcoe_ctlr_timer_work() to
fcoe_ctlr_select() so that it can be shared
with another function in a forthcoming patch.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfcoe: update FIP FCF announcements
Joe Eykholt [Wed, 1 Dec 2010 00:19:40 +0000]
[SCSI] libfcoe: update FIP FCF announcements

Move the announcement code to a separate function for reuse in
a forthcoming patch.

For messages regarding FCF timeout and selection, use the
previously-announced FCF MAC address (dest_addr) in the fcoe_ctlr struct.
Only print (announce) the FCF if it is new.   Print MAC for
timed-out or deselected FCFs.

Signed-off-by: Joe Eykholt <jeykholt@cisco.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix fc_tm_done not freeing the allocated fsp pkt
Yi Zou [Wed, 1 Dec 2010 00:19:35 +0000]
[SCSI] libfc: fix fc_tm_done not freeing the allocated fsp pkt

Frame should be freed in fc_tm_done, this is an updated patch on the one
initially submitted by Hillf Danton.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: the timeout for the REC itself is 2 * R_A_TOV_els
Yi Zou [Wed, 1 Dec 2010 00:19:30 +0000]
[SCSI] libfc: the timeout for the REC itself is 2 * R_A_TOV_els

The timeout for the exchange carrying REC itself is 2 * R_A_TOV_els.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix exchange being deleted when the abort itself is timed out
Yi Zou [Wed, 1 Dec 2010 00:19:25 +0000]
[SCSI] libfc: fix exchange being deleted when the abort itself is timed out

Should not continue when the abort itself is being timeout since in that case
the exchange will be deleted and relesased. We still want to call the
associated response handler to let the layer, e.g., fcp, know the exchange
itself is being timed out.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: do not fc_io_compl on fsp w/o any scsi_cmnd associated
Yi Zou [Wed, 1 Dec 2010 00:19:20 +0000]
[SCSI] libfc: do not fc_io_compl on fsp w/o any scsi_cmnd associated

Do not call fc_io_compl() on fsp w/o any scsi_cmnd, e.g., lun reset is built
inside fc_fcp, not from a scsi command from queuecommnd from scsi-ml, so in
in case target is buggy that is invalid flags in the FCP_RSP, as we have seen
in some SAN Blaze target where all bits in flags are 0, we do not want to call
io_compl on this fsp.

[ Comment block added by Robert Love ]

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: add print of exchange id for debugging fc_fcp
Yi Zou [Wed, 1 Dec 2010 00:19:15 +0000]
[SCSI] libfc: add print of exchange id for debugging fc_fcp

This is very helpful to match up the corresponding exchange to the actual I/O
described by the fsp, particularly when you do a side-by-side comparison of
the syslog with your trace.

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] drivers/scsi/fcoe: Update WARN uses
Joe Perches [Wed, 1 Dec 2010 00:19:09 +0000]
[SCSI] drivers/scsi/fcoe: Update WARN uses

Add missing newlines.

Signed-off-by: Joe Perches <joe@perches.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix memory leakage in remote port
Hillf Danton [Wed, 1 Dec 2010 00:19:04 +0000]
[SCSI] libfc: fix memory leakage in remote port

There seems rdata should get put before return.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix memory leakage in local port
Hillf Danton [Wed, 1 Dec 2010 00:18:59 +0000]
[SCSI] libfc: fix memory leakage in local port

There seems info should get freed when error encountered.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix memory leakage in local port
Hillf Danton [Wed, 1 Dec 2010 00:18:54 +0000]
[SCSI] libfc: fix memory leakage in local port

There seems info should get freed when error encountered.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: remove tgt_flags from fc_fcp_pkt struct
john fastabend [Wed, 1 Dec 2010 00:18:49 +0000]
[SCSI] libfc: remove tgt_flags from fc_fcp_pkt struct

We can easily remove the tgt_flags from fc_fcp_pkt struct
and use rpriv->tgt_flags directly where needed.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: use rport timeout values for fcp recovery
john fastabend [Wed, 1 Dec 2010 00:18:44 +0000]
[SCSI] libfc: use rport timeout values for fcp recovery

Use the rport value for rec_tov for timeout values when
sending fcp commands. Currently, defaults are being used
which may or may not match the advertised values.

The default may cause i/o to timeout on networks that
set this value larger then the default value. To make
the timeout more configurable in the non-REC mode we
remove the FC_SCSI_ER_TIMEOUT completely allowing the
scsi-ml to do the timeout. This removes an unneeded
timer and allows the i/o timeout to be configured
using the scsi-ml knobs.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: incorrect scsi host byte codes returned to scsi-ml
john fastabend [Wed, 1 Dec 2010 00:18:39 +0000]
[SCSI] libfc: incorrect scsi host byte codes returned to scsi-ml

The fcp packet recovery handler fc_fcp_recover() is called
when errors occurr in a fcp session. Currently it is
generically setting the status code to FC_CMD_RECOVERY for
all error types. This results in DID_BUS_BUSY errors
being returned to the scsi-ml.

DID_BUS_BUSY errors indicate "BUS stayed busy through time
out period" according to scsi.h. Many of the error reported
by fc_rcp_recovery() are pkt errors. Here we update
fc_fcp_recovery to use better host byte codes.

With certain FAST FAIL flags set DID_BUS_BUSY and DID_ERROR
will have different behaviors this was causing dm multipath
to fail quickly in some cases where a retry would be a
better action.

Signed-off-by: John Fastabend <john.r.fastabend@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix stats computation in fc_queuecommand()
Hillf Danton [Wed, 1 Dec 2010 00:18:33 +0000]
[SCSI] libfc: fix stats computation in fc_queuecommand()

There seems accumulation needed.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix mem leak in fc_seq_assign()
Hillf Danton [Wed, 1 Dec 2010 00:18:28 +0000]
[SCSI] libfc: fix mem leak in fc_seq_assign()

There is a typo cleaned, which triggers memory leakage.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: Fix incorrect locking and unlocking in FCP
Robert Love [Wed, 1 Dec 2010 00:18:23 +0000]
[SCSI] libfc: Fix incorrect locking and unlocking in FCP

The error handler grabs the si->scsi_queue_lock, but
in the case where the fsp pointer is NULL it releases
the scsi_host lock. This can lead to a variety of
system hangs depending on which is used first- the
scsi_host lock or the scsi_queue_lock.

This patch simply unlocks the correct lock when fcp
is NULL.

Signed-off-by: Robert Love <robert.w.love@intel.com>
Tested-by: Ross Brattain <ross.b.brattain@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: tune fc_exch_em_alloc() to be O(2)
Hillf Danton [Wed, 1 Dec 2010 00:18:17 +0000]
[SCSI] libfc: tune fc_exch_em_alloc() to be O(2)

For allocating new exch from pool,  scanning for free slot in exch
array fluctuates when exch pool is close to exhaustion.

The fluctuation is smoothed, and the scan looks to be O(2).

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix mem leak in fc_exch_recv_seq_resp()
Hillf Danton [Wed, 1 Dec 2010 00:18:12 +0000]
[SCSI] libfc: fix mem leak in fc_exch_recv_seq_resp()

There seems that ep should get released, or it will no longer get freed.

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: fix NULL pointer dereference bug in fc_fcp_pkt_release
Yi Zou [Wed, 1 Dec 2010 00:18:07 +0000]
[SCSI] libfc: fix NULL pointer dereference bug in fc_fcp_pkt_release

This happens when then tearing down the fcoe interface with active I/O.
The back trace shows dead000000200200 in RAX, i.e., LIST_POISON2, indicating
that the fsp is already being dequeued, which is probably why no complaining
was seen in fc_fcp_destroy() about outstanding fsp not freed, since we dequeue
it in the end of fc_io_compl() before releasing it. The bug is due to the
fact that we have already destroyed lport's scsi_pkt_pool while on-going i/o
is still accessing it through fc_fcp_pkt_release(), like this trace or the
similar code path from scsi-ml to fc_eh_abort, etc. This is fixed by moving
the fc_fcp_destroy() after lport is detached from scsi-ml since fc_fcp_destroy
is supposed to called only once where no lport lock is taken, otherwise the
fc_fcp_pkt_release() would have to grab the lport lock.

 BUG: unable to handle kernel NULL pointer dereference at (null)
 .......
 RIP: 0010:[<0000000000000000>]
 [<(null)>] (null)
 RSP: 0018:ffff8803270f7b88  EFLAGS: 00010282
 RAX: dead000000200200 RBX: ffff880197d2fbc0 RCX: 0000000000005908
 RDX: ffff880195ea6d08 RSI: 0000000000000282 RDI: ffff880180f4fec0
 RBP: ffff8803270f7bc0 R08: ffff880197d2fbe0 R09: 0000000000000000
 R10: ffff88032867f090 R11: 0000000000000000 R12: ffff880195ea6d08
 R13: 0000000000000282 R14: ffff880180f4fec0 R15: 0000000000000000
 FS:  0000000000000000(0000) GS:ffff8801b5820000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
 CR2: 0000000000000000 CR3: 00000001a6eae000 CR4: 00000000000006e0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
 Process fc_rport_eq (pid: 5278, threadinfo ffff8803270f6000, task ffff880326254ab0)
 Stack:
 ffffffffa02c39ca ffff8803270f7ba0 ffff88019331cbc0 ffff880197d2fbc0
 0000000000000000 ffff8801a8c895e0 ffff8801a8c895e0 ffff8803270f7c10
 ffffffffa02c4962 ffff8803270f7be0 ffffffff814c94ab ffff8803270f7c10
 Call Trace:
 [<ffffffffa02c39ca>] ? fc_io_compl+0x10a/0x530 [libfc]
 [<ffffffffa02c4962>] fc_fcp_complete_locked+0x72/0x150 [libfc]
 [<ffffffff814c94ab>] ? _spin_unlock_bh+0x1b/0x20
 [<ffffffffa02b98ff>] ? fc_exch_done+0x3f/0x60 [libfc]
 [<ffffffffa02c4a8f>] fc_fcp_retry_cmd+0x4f/0x60 [libfc]
 [<ffffffffa02c6150>] fc_fcp_recv+0x9b0/0xc30 [libfc]
 [<ffffffff8106ba7a>] ? _call_console_drivers+0x4a/0x80
 [<ffffffff8107d5ec>] ? lock_timer_base+0x3c/0x70
 [<ffffffff8107e06b>] ? try_to_del_timer_sync+0x7b/0xe0
 [<ffffffffa02b9dcf>] fc_exch_mgr_reset+0x1df/0x250 [libfc]
 [<ffffffffa02c57a0>] ? fc_fcp_recv+0x0/0xc30 [libfc]
 [<ffffffffa02c1042>] fc_rport_work+0xf2/0x4e0 [libfc]
 [<ffffffff8109203e>] ? prepare_to_wait+0x4e/0x80
 [<ffffffffa02c0f50>] ? fc_rport_work+0x0/0x4e0 [libfc]
 [<ffffffff8108c6c0>] worker_thread+0x170/0x2a0
 [<ffffffff81091d50>] ? autoremove_wake_function+0x0/0x40
 [<ffffffff8108c550>] ? worker_thread+0x0/0x2a0
 [<ffffffff810919e6>] kthread+0x96/0xa0
 [<ffffffff810141ca>] child_rip+0xa/0x20
 [<ffffffff81091950>] ? kthread+0x0/0xa0
 [<ffffffff810141c0>] ? child_rip+0x0/0x20
 Code:
 Bad RIP value.

 RIP
 [<(null)>] (null)
 RSP <ffff8803270f7b88>
 CR2: 0000000000000000

Signed-off-by: Yi Zou <yi.zou@intel.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libfc: remove define of fc_seq_exch in fc_exch.c
Hillf Danton [Wed, 1 Dec 2010 00:18:01 +0000]
[SCSI] libfc: remove define of fc_seq_exch in fc_exch.c

The define for fc_seq_exch is unnecessary, since it also appears in scsi/libfc.h

Signed-off-by: Hillf Danton <dhillf@gmail.com>
Signed-off-by: Robert Love <robert.w.love@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bfa: fix endianess sparse check warnings
Maggie [Tue, 30 Nov 2010 02:26:32 +0000]
[SCSI] bfa: fix endianess sparse check warnings

First round of fix for the endianess check warnings from make C=2 CF="-D__CHECK_ENDIAN__".

Signed-off-by: Maggie <xmzhang@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bfa: fix regular sparse check warnings.
Maggie [Tue, 30 Nov 2010 02:21:32 +0000]
[SCSI] bfa: fix regular sparse check warnings.

Fix all sparse check warnings from make C=2.

Signed-off-by: Maggie <xmzhang@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Updated version to 2.6.2.2
Eddie Wai [Tue, 23 Nov 2010 23:29:32 +0000]
[SCSI] bnx2i: Updated version to 2.6.2.2

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Updated copyright and maintainer info
Eddie Wai [Tue, 23 Nov 2010 23:29:31 +0000]
[SCSI] bnx2i: Updated copyright and maintainer info

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Cleaned up various error conditions in ep_connect/disconnect
Eddie Wai [Tue, 23 Nov 2010 23:29:30 +0000]
[SCSI] bnx2i: Cleaned up various error conditions in ep_connect/disconnect

Various error conditions inside ep_connect and ep_disconnect were
either not being handled or not being handled correctly.  This patch
fixes all those issues.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Anil Veerabhadrappa <anilgv@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Added return code check for chip kwqe submission request
Eddie Wai [Tue, 23 Nov 2010 23:29:29 +0000]
[SCSI] bnx2i: Added return code check for chip kwqe submission request

Added the handling for cases when a chip request is made to the
CNIC module but the hardware is not ready to accept.  This would
lead to many unnecessary wait timeouts.
This code adds check in the connect establishment and destruction
path.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Modified the bnx2i stop path to compensate for in progress ops
Eddie Wai [Tue, 23 Nov 2010 23:29:28 +0000]
[SCSI] bnx2i: Modified the bnx2i stop path to compensate for in progress ops

The stop path has been augmented to wait a max of 10s for all in
progress offload and destroy activities to complete before proceeding
to terminate all active connections (via iscsid or forcefully).

Note that any new offload and destroy requests are now blocked and
return to the caller immediately.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Anil Veerabhadrappa <anilgv@broadcom.com>
Reviewed-by: Benjamin Li <benli@broadcom.com>
Reviewed-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Removed the dynamic registration of CNIC
Eddie Wai [Tue, 23 Nov 2010 23:29:27 +0000]
[SCSI] bnx2i: Removed the dynamic registration of CNIC

The code no longer needs to dynamically register and unregister
the CNIC device.  The CNIC device will be kept registered until
module unload.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Added mutex lock protection to conn_get_param
Eddie Wai [Tue, 23 Nov 2010 23:29:26 +0000]
[SCSI] bnx2i: Added mutex lock protection to conn_get_param

Added net_dev mutex lock protection before accessing the csk
parameters.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Benjamin Li <benli@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Allow to abort the connection if connect request times out
Eddie Wai [Tue, 23 Nov 2010 23:29:25 +0000]
[SCSI] bnx2i: Allow to abort the connection if connect request times out

In the situation where the connect completion response arrives after
the connect request has already timed out, the connection was not being
aborted but only the resource was being freed.  This creates a problem
for 5771X (10g) as the chip flags this with an assertion.

This change will properly aborts the connection before freeing the
resource.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Fixed the remote TCP RST handling for the 570X (1g)
Eddie Wai [Tue, 23 Nov 2010 23:29:24 +0000]
[SCSI] bnx2i: Fixed the remote TCP RST handling for the 570X (1g)

Modified the handling of the remote TCP RST code so the chip can now
flush the tx pipe accordingly upon a remote TCP RST reception.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Fixed a cid leak issue for 5771X (10g)
Eddie Wai [Tue, 23 Nov 2010 23:29:23 +0000]
[SCSI] bnx2i: Fixed a cid leak issue for 5771X (10g)

A cid leak issue was found when the connect destroy request exceeded
the driver's disconnection timeout.  This will lead to a cid resource
leak issue.
The fix is to allow the cid cleanup even when this happens.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Anil Veerabhadrappa <anilgv@broadcom.com>
Reviewed-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Fixed the endian bug in the TMF LUN cmd send
Eddie Wai [Tue, 23 Nov 2010 23:29:22 +0000]
[SCSI] bnx2i: Fixed the endian bug in the TMF LUN cmd send

Added a be32_to_cpu call for the TMF LUN wqe.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Added fix for NOP-Out response panic from unsolicited NOP-In
Eddie Wai [Tue, 23 Nov 2010 23:29:21 +0000]
[SCSI] bnx2i: Added fix for NOP-Out response panic from unsolicited NOP-In

The patch fixes the following situations where NOP-Out pkt is called for:
- local unsolicited NOP-Out requests (requesting no NOP-In response)
- local NOP-Out responses to unsolicited NOP-In requests

kernel panic is observed due to double session spin_lock requests; one in the
bnx2i_process_nopin_local_cmpl routine in bnx2i_hwi.c and the other in the
iscsi_put_task routine in libiscsi.c

The proposed fix is to export the currently static __iscsi_put_task() routine
and have bnx2i call it directly instead of the iscsi_put_task() routine which
holds the session spin lock.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Anil Veerabhadrappa <anilgv@broadcom.com>
Acked-by: Benjamin Li <benli@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bnx2i: Fixed bugs in the handling of unsolicited NOP-Ins
Eddie Wai [Tue, 23 Nov 2010 23:29:20 +0000]
[SCSI] bnx2i: Fixed bugs in the handling of unsolicited NOP-Ins

Unsolicited NOP-Ins are placed in the receive queue of the hardware
which requires to be read out regardless if the receive pipe is suspended
or not.  This patch adds the disposal of this RQ element under this
condition.
Also fixed the bug in the unsolicited NOP-In handling routine which
checks for the RESERVED_ITT.

Signed-off-by: Eddie Wai <eddie.wai@broadcom.com>
Reviewed-by: Michael Chan <mchan@broadcom.com>
Reviewed-by: Benjamin Li <benli@broadcom.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Bump version 07.100.00.00
Kashyap, Desai [Fri, 12 Nov 2010 23:12:28 +0000]
[SCSI] mpt2sas: Bump version 07.100.00.00

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Debug string changes from target to device.
Kashyap, Desai [Fri, 12 Nov 2010 23:11:32 +0000]
[SCSI] mpt2sas: Debug string changes from target to device.

Changing debug print to correct string.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Sanity check for phy count is added using max phy count
Kashyap, Desai [Fri, 12 Nov 2010 23:10:51 +0000]
[SCSI] mpt2sas: Sanity check for phy count is added using max phy count

Fix oops loading driver when there is direct attached
SEP device

The driver set max phys count to the value reported in sas iounit page
zero.  However this page doesn't take into account additional virutal
phys.  When sas topology event arrives, the phy count is larger than
expected, and the driver accesses memory array beyond the end of
allocated space, then oops.  Manufacturing page 8 contains the info
on direct attached phys.

For this fix will making sure that sas topology event is not
processing phys greater than the expected phy count.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: change queue depth with reason argument more appropriately
Kashyap, Desai [Fri, 12 Nov 2010 23:09:11 +0000]
[SCSI] mpt2sas: change queue depth with reason argument more appropriately

change_queue_depth callback API changed
The change_queue_depth callback changed where there is now an additional
parameter called reason, with SCSI_QDEPTH_DEFAULT, SCSI_QDEPTH_QFULL,
and SCSI_QDEPTH_RAMP_UP codes.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Remove code for TASK_SET_FULL from driver.
Kashyap, Desai [Fri, 12 Nov 2010 23:07:36 +0000]
[SCSI] mpt2sas: Remove code for TASK_SET_FULL from driver.

remove support for MPI2_EVENT_TASK_SET_FULL
This event is obsoleted, so this processing of this event
needs to be removed from the driver.  The controller firmware is going
to handle TASK_SET_FULL, the driver doesn't need to do anything.
Even though we are removing the EVENT handling, the behavour has not
changed between driver versions becuase fimrware will still be handling
queue throttling, and retrying of commands when the target device queues
are full.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: MPI 2.0 Header updated
Kashyap, Desai [Fri, 12 Nov 2010 23:06:14 +0000]
[SCSI] mpt2sas: MPI 2.0 Header updated

MPI2 Rev header files.

1) Removed Task Set Full Event. Modified description of Disable SCSI
Initiator Task Set Full Handling bit in the Flags field of IO Unit
Page 1. Modified the descriptions for the three queue depth fields in
SAS IO Unit Page 1.
(2) Added new value for the Current Operation bits of the Flags field
in the RAID Volume Indicator Structure to indicate that the Make Data
Consistent operation is running.
(3) Added a value of 0x6 to various SAS link rate fields to indicate an
attached PHY that is not using any commonly supported settings.
(4) Added Volume Not Consistent bit to the VolumeStatusFlags field of
RAID Volume Page 0.
(5) Added a new value for the IncompatibleReason field of RAID Physical
Disk Page 0 to indicate an incompatible media type.
(6) Added Diagnostic Data Upload tool for the Toolbox Request.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Modify code to support Expander switch
Kashyap, Desai [Fri, 12 Nov 2010 23:05:30 +0000]
[SCSI] mpt2sas: Modify code to support Expander switch

Issue : Switch swap doesn't work when device missing delay is enabled.

(1) add support to individually add and remove phys to and from
existing ports. This replaces the routine
_transport_delete_duplicate_port.
(2) _scsih_sas_host_refresh - was modified to change the link rate
from zero to 1.5 GB rate when the firmware reports there is an
attached device with zero link.
(3) add new function mpt2sas_device_remove, this is wrapper function
deletes some redundant code through out driver by combining into one
subrountine
(4) two subroutines were modified so the sas_device, raid_device, and
port lists are traversed once when objects are deleted from the list.
Previously it was looping back each time an object was deleted from the
list.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Create a pool of chain buffer instead of dedicated per IOs
Kashyap, Desai [Fri, 12 Nov 2010 23:04:06 +0000]
[SCSI] mpt2sas: Create a pool of chain buffer instead of dedicated per IOs

Create a pool of chain buffers, instead of dedicated per IO:
This enahancment is to address memory allocation failure when asking
for more than 2300 IOs per host.   There is just not enough contiquious
DMA physical memory to make one single allocation to hold both message
frames and chain buffers when asking for more than 2300 request. In order
to address this problem we will have to allocate memory for each chain
buffer in a seperate individual memory allocation, placing each chain
element of 128 bytes onto a pool of available chains, which can be
shared amoung all request.
Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Added loadtime para for IOMissingDelay and DMD
Kashyap, Desai [Fri, 12 Nov 2010 23:02:18 +0000]
[SCSI] mpt2sas: Added loadtime para for IOMissingDelay and DMD

Ability to override/set the ReportDeviceMissingDelay and
IODeviceMissingDelay from driver: Add new command line option missing_delay,
this is an array, where the first element is the device missing delay,
and the second element is io missing delay.  The driver will program
sas iounit page 1 with the new setting when the driver loads. This is
programmed to the current and persistent configuration page so this takes
immediately, as will be sticky across host reboots.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Added sanity check for cb_idx and smid access.
Kashyap, Desai [Fri, 12 Nov 2010 23:01:14 +0000]
[SCSI] mpt2sas: Added sanity check for cb_idx and smid access.

Sometime it is seen that controller
firmware returns an invalid system message id (smid).

the oops is occurring becuase mpt_callbacks pointer is referenced to
either null or invalid virtual address.  this is due to cb_idx set
incorrectly from routine _base_get_cb_idx.  the cb_idx was set incorrectly
becuase there is no check to make sure smid is less than maxiumum
anticapted smid.   to fix this issue, we add a check in
_base_get_cb_idx to make sure smid is not greater than
ioc->hba_queue_depth.   in addition, a similar check was added to make
sure the reply address was less than the largest anticapated address.

Newer firmware has sovled this issue, however it good to have this sanity
check.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] mpt2sas: Removed compiler warnnings when logging is disabled
Kashyap, Desai [Fri, 12 Nov 2010 22:59:20 +0000]
[SCSI] mpt2sas: Removed compiler warnnings when logging is disabled

The compiler throws warning messages while compiling without
CONFIG_SCSI_MPT2SAS_LOGGING.
Set proper ifdef for CONFIG_SCSI_MPT2SAS_LOGGING to avoid warnnings.

Signed-off-by: Kashyap Desai <kashyap.desai@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Update lpfc driver version to 8.3.19
James Smart [Sun, 21 Nov 2010 04:12:11 +0000]
[SCSI] lpfc 8.3.19: Update lpfc driver version to 8.3.19

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Change heartbeat to default to disabled
James Smart [Sun, 21 Nov 2010 04:12:02 +0000]
[SCSI] lpfc 8.3.19: Change heartbeat to default to disabled

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Added support for ELS RRQ command
James Smart [Sun, 21 Nov 2010 04:11:55 +0000]
[SCSI] lpfc 8.3.19: Added support for ELS RRQ command

Added support for ELS RRQ command

- Add new routine lpfc_set_rrq_active() to track XRI qualifier state.
- Add new module parameter lpfc_enable_rrq to control RRQ operation.
- Add logic to ELS RRQ completion handler and xri qualifier timeout
  to clear XRI qualifier state.
- Use OX_ID from XRI_ABORTED_CQE for RRQ payload.
- Tie abort and XRI_ABORTED_CQE andler to RRQ generation.

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Add SLI4 FC Discovery support
James Smart [Sun, 21 Nov 2010 04:11:48 +0000]
[SCSI] lpfc 8.3.19: Add SLI4 FC Discovery support

Add SLI4 FC Discovery support

- Replace READ_LA and READ_LA64 with READ_TOPOLOGY mailbox command.
- Converted the old READ_LA structure to use bf_set/get instead of bit fields.
- Rename HBA_FCOE_SUPPORT flag to HBA_FCOE_MODE. Flag now indicates function
  is running as SLI-4 FC or FCoE port. Make sure flag reset each time
  READ_REV completed as it can dynamically change.
- Removed BDE union in the READ_TOPOLOGY mailbox command and added a define to
  define the ALPA MAP SIZE. Added FC Code for async events.
- Added code to support new 16G link speed.
- Define new set of values to keep track of valid user settable link speeds.
- Used new link speed definitions to define link speed max and bitmap.
- Redefined FDMI Port sppeds to be hax values and added the 16G value.
- Added new CQE trailer code for FC Events.
- Add lpfc_issue_init_vfi and lpfc_init_vfi_cmpl routines.
- Replace many calls to the initial_flogi routine with lpfc_issue_init_vfi.
- Add vp and vpi fields to the INIT_VFI mailbox command.
- Addapt lpfc_hba_init_link routine for SLI4 use.
- Use lpfc_hba_init_link call from lpfc_sli4_hba_setup.
- Add a check for FC mode to register the FCFI before init link.
- Convert lpfc_sli4_init_vpi to be called without a vpi (get it from vport).

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Add latest SLI4 Hardware initialization support
James Smart [Sun, 21 Nov 2010 04:11:37 +0000]
[SCSI] lpfc 8.3.19: Add latest SLI4 Hardware initialization support

- Add the Lancer FC and FCoE PCI IDs
- Add new SLI4 INTF register definitions
- Implement new SLI4 doorbell register

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] lpfc 8.3.19: Fix critical errors and crashes
James Smart [Sun, 21 Nov 2010 04:14:19 +0000]
[SCSI] lpfc 8.3.19: Fix critical errors and crashes

Fix critical errors and crashes

- Replace LOF_SECURITY with LOG_SECURITY
- When calculating diag test memory size, use full size with header.
- Return LS_RJT with status=UNSUPPORTED on unrecognized ELS's
- Correct NULL pointer dereference when lpfc_create_vport_work_array()
  returns NULL.
- Added code to handle CVL when port is in LPFC_VPORT_FAILED state.
- In lpfc_do_scr_ns_plogi, check the nodelist for FDMI_DID and reuse
  the resource.
- Check for generic request 64 and calculate the sgl offset for the request
  and reply sgls, also calculate the xmit length using only the request bde.

Signed-off-by: Alex Iannicelli <alex.iannicelli@emulex.com>
Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] fix id computation in scsi_eh_target_reset()
James Bottomley [Mon, 25 Oct 2010 20:53:41 +0000]
[SCSI] fix id computation in scsi_eh_target_reset()

The current code in scsi_eh_target_reset() has an off by one error
that actually sends spurious extra resets.  Since there's no real need
to reset the targets in numerical order, simply chunk up the command
recovery list doing target resets and pulling matching targets out of
the list (that also makes the loop O(N) instead of O(N^2).

[mike christie found and fixed a list_splice -> list_splice_init problem]

Reported-by: Hillf Danton<dhillf@gmail.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] ipr: fix mailbox register definition and add a delay before reading
Wayne Boyer [Thu, 4 Nov 2010 16:36:16 +0000]
[SCSI] ipr: fix mailbox register definition and add a delay before reading

The definition for the mailbox register for new adapters was incorrect.  The
value has been updated to the correct offset.

After an adapter reset, the mailbox register on the new adapters takes a
number of seconds to stabilize.  A delay has been added before reading the
register.

Signed-off-by: Wayne Boyer <wayneb@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] ipr: fix lun assignment and comparison
Wayne Boyer [Thu, 4 Nov 2010 16:35:58 +0000]
[SCSI] ipr: fix lun assignment and comparison

The lun value was not getting set up correctly for all devices attached to the
new 64 bit adapters.  The fix is to move the logic to earlier in the
ipr_init_res_entry routine such that the value does get set correctly for all
devices.

Then the ipr_is_same_device comparison function was using the wrong lun value
in the logic for the new adapters.  Change this to use the correct lun value.

Signed-off-by: Wayne Boyer <wayneb@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] Retrieve the Caching mode page
Luben Tuikov [Sat, 23 Oct 2010 04:24:36 +0000]
[SCSI] Retrieve the Caching mode page

Some kernel transport drivers unconditionally disable
retrieval of the Caching mode page. One such for example is
the BBB/CBI transport over USB.  Such a restraint is too
harsh as some devices do support the Caching mode
page. Unconditionally enabling the retrieval of this mode
page over those transports at their transport code level may
result in some devices failing and becoming unusable.

This patch implements a method of retrieving the Caching
mode page without unconditionally enabling it in the
transports which unconditionally disable it. The idea is to
ask for all supported pages, page code 0x3F, and then search
for the Caching mode page in the mode parameter data
returned. The sd driver already asks for all the mode pages
supported by the attached device by setting the page code to
0x3F in order to find out if the media is write protected by
reading the WP bit in the Device Specific Parameter
field. It then attempts to retrieve only the Caching mode
page by setting the page code to 8 and actually attempting
to retrieve it if and only if the transport allows it.

The method implemented here is that if the transport doesn't
allow retrieval of the Caching mode page and the device is
not RBC, then we ask for all pages supported by setting the
page code to 0x3F (similarly to how the WP bit is retrieved
above), and then we search for the Caching mode page in the
mode parameter data returned.

With this patch, devices over SATA, report this (no change):

Oct 22 18:45:58 localhost kernel: sd 0:0:0:0: [sda] 976773168 512-byte logical blocks: (500 GB/465 GiB)
Oct 22 18:45:58 localhost kernel: sd 0:0:0:0: Attached scsi generic sg0 type 0
Oct 22 18:45:58 localhost kernel: sd 0:0:0:0: [sda] Write Protect is off
Oct 22 18:45:58 localhost kernel: sd 0:0:0:0: [sda] Mode Sense: 00 3a 00 00
Oct 22 18:45:58 localhost kernel: sd 0:0:0:0: [sda] Write cache: enabled, read cache: enabled, doesn't support DPO or FUA

Smart devices report their Caching mode page. This is a
change where we'd previously see the kernel making
assumption about the device's cache being write-through:

Oct 22 18:45:58 localhost kernel: sd 6:0:0:0: Attached scsi generic sg2 type 0
Oct 22 18:45:58 localhost kernel: sd 6:0:0:0: [sdb] 610472646 4096-byte logical blocks: (2.50 TB/2.27 TiB)
Oct 22 18:45:58 localhost kernel: sd 6:0:0:0: [sdb] Write Protect is off
Oct 22 18:45:58 localhost kernel: sd 6:0:0:0: [sdb] Mode Sense: 47 00 10 08
Oct 22 18:45:58 localhost kernel: sd 6:0:0:0: [sdb] Write cache: enabled, read cache: enabled, supports DPO and FUA

And "dumb" devices over BBB, are correctly shown not to
support reporting the Caching mode page:

Oct 22 18:49:06 localhost kernel: sd 7:0:0:0: [sdc] 15663104 512-byte logical blocks: (8.01 GB/7.46 GiB)
Oct 22 18:49:06 localhost kernel: sd 7:0:0:0: [sdc] Write Protect is off
Oct 22 18:49:06 localhost kernel: sd 7:0:0:0: [sdc] Mode Sense: 23 00 00 00
Oct 22 18:49:06 localhost kernel: sd 7:0:0:0: [sdc] No Caching mode page present
Oct 22 18:49:06 localhost kernel: sd 7:0:0:0: [sdc] Assuming drive cache: write through

Signed-off-by: Luben Tuikov <ltuikov@yahoo.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] libsas: fix definition of wideport, include local sas address
Dan Williams [Fri, 1 Oct 2010 20:55:52 +0000]
[SCSI] libsas: fix definition of wideport, include local sas address

To date libsas has only looked at the attached sas address when
determining the formation of wide ports.  The specification and some
hardware expects that phys with different addresses will not form a wide
port unless the local peer phys also match each other.  Introduce a flag
to select stricter behavior at sas_register_ha() time.  The flag can be
dropped once it is known that all libsas users expect the same behavior.

Current drivers just initialize this field to zero and get the
traditional behavior.

Reported-by: Patrick Thomson <patrick.s.thomson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] sd: improve logic and efficiecy of media-change detection
Alan Stern [Tue, 7 Sep 2010 15:44:01 +0000]
[SCSI] sd: improve logic and efficiecy of media-change detection

This patch (as1415) improves the formerly incomprehensible logic in
sd_media_changed() (the current code refers to "changed" as a state,
whereas in fact it is a relation between two states).  It also adds a
big comment so that everyone can understand what is really going on.

The patch also improves efficiency by not reporting a media change
when no medium was ever present.  If no medium was present the last
time we checked and there's still no medium, it's not necessary to
tell the caller that a change occurred.  Doing so merely causes the
caller to attempt to revalidate a non-existent disk, which is a waste
of time.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] fix up documentation for change in ->queuecommand to lockless calling
James Bottomley [Thu, 16 Dec 2010 14:22:24 +0000]
[SCSI] fix up documentation for change in ->queuecommand to lockless calling

The current doc still says we call it with the host lock held, which is
going to cause confusion.

Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years ago[SCSI] bfa: rename log_level to bfa_log_level
Jing Huang [Fri, 10 Dec 2010 01:11:53 +0000]
[SCSI] bfa: rename log_level to bfa_log_level

Rename log_level to bfa_log_level to make the global variable more bfa
specific and avoid clashes with other drivers which was causing a
build failure.

Signed-off-by: Jing Huang <huangj@brocade.com>
Signed-off-by: James Bottomley <James.Bottomley@suse.de>

9 years agoLinux 2.6.37-rc6
Linus Torvalds [Thu, 16 Dec 2010 01:24:48 +0000]
Linux 2.6.37-rc6

9 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Thu, 16 Dec 2010 01:24:05 +0000]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: ghash-intel - ghash-clmulni-intel_glue needs err.h

9 years agoMerge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4
Linus Torvalds [Wed, 15 Dec 2010 20:41:17 +0000]
Merge branch 'for_linus' of git://git./linux/kernel/git/tytso/ext4

* 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
  ext4: fix typo which broke '..' detection in ext4_find_entry()
  ext4: Turn off multiple page-io submission by default

9 years agoxen: Provide a variant of __RING_SIZE() that is an integer constant expression
Jeremy Fitzhardinge [Wed, 8 Dec 2010 20:39:12 +0000]
xen: Provide a variant of __RING_SIZE() that is an integer constant expression

Without this, gcc 4.5 won't compile xen-netfront and xen-blkfront, where
this is being used to specify array sizes.

Signed-off-by: Jan Beulich <jbeulich@novell.com>
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: David Miller <davem@davemloft.net>
Cc: Stable Kernel <stable@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoMAINTAINERS: update MSM git tree
Daniel Walker [Thu, 9 Dec 2010 23:45:27 +0000]
MAINTAINERS: update MSM git tree

The MSM main git tree has changed over to this new address.

Signed-off-by: Daniel Walker <dwalker@codeaurora.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoinstall_special_mapping skips security_file_mmap check.
Tavis Ormandy [Thu, 9 Dec 2010 14:29:42 +0000]
install_special_mapping skips security_file_mmap check.

The install_special_mapping routine (used, for example, to setup the
vdso) skips the security check before insert_vm_struct, allowing a local
attacker to bypass the mmap_min_addr security restriction by limiting
the available pages for special mappings.

bprm_mm_init() also skips the check, and although I don't think this can
be used to bypass any restrictions, I don't see any reason not to have
the security check.

  $ uname -m
  x86_64
  $ cat /proc/sys/vm/mmap_min_addr
  65536
  $ cat install_special_mapping.s
  section .bss
      resb BSS_SIZE
  section .text
      global _start
      _start:
          mov     eax, __NR_pause
          int     0x80
  $ nasm -D__NR_pause=29 -DBSS_SIZE=0xfffed000 -f elf -o install_special_mapping.o install_special_mapping.s
  $ ld -m elf_i386 -Ttext=0x10000 -Tbss=0x11000 -o install_special_mapping install_special_mapping.o
  $ ./install_special_mapping &
  [1] 14303
  $ cat /proc/14303/maps
  0000f000-00010000 r-xp 00000000 00:00 0                                  [vdso]
  00010000-00011000 r-xp 00001000 00:19 2453665                            /home/taviso/install_special_mapping
  00011000-ffffe000 rwxp 00000000 00:00 0                                  [stack]

It's worth noting that Red Hat are shipping with mmap_min_addr set to
4096.

Signed-off-by: Tavis Ormandy <taviso@google.com>
Acked-by: Kees Cook <kees@ubuntu.com>
Acked-by: Robert Swiecki <swiecki@google.com>
[ Changed to not drop the error code - akpm ]
Reviewed-by: James Morris <jmorris@namei.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agocrypto: ghash-intel - ghash-clmulni-intel_glue needs err.h
Randy Dunlap [Wed, 15 Dec 2010 09:58:57 +0000]
crypto: ghash-intel - ghash-clmulni-intel_glue needs err.h

Add missing header file:

arch/x86/crypto/ghash-clmulni-intel_glue.c:256: error: implicit declaration of function 'IS_ERR'
arch/x86/crypto/ghash-clmulni-intel_glue.c:257: error: implicit declaration of function 'PTR_ERR'

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>

9 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq
Linus Torvalds [Wed, 15 Dec 2010 02:50:10 +0000]
Merge branch 'for-linus' of git://git./linux/kernel/git/tj/wq

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: It is likely that WORKER_NOT_RUNNING is true
  MAINTAINERS: Add workqueue entry
  workqueue: check the allocation of system_unbound_wq

9 years agoMerge branch 'for-linus' of git://neil.brown.name/md
Linus Torvalds [Wed, 15 Dec 2010 02:49:40 +0000]
Merge branch 'for-linus' of git://neil.brown.name/md

* 'for-linus' of git://neil.brown.name/md:
  md: protect against NULL reference when waiting to start a raid10.
  md: fix bug with re-adding of partially recovered device.
  md: fix possible deadlock in handling flush requests.
  md: move code in to submit_flushes.
  md: remove handling of flush_pending in md_submit_flush_data

9 years agodw_spi: Fix missing final read in some polling situations
Major Lee [Fri, 10 Dec 2010 10:13:49 +0000]
dw_spi: Fix missing final read in some polling situations

There is a possibility that the last word of a transaction will be lost
if data is not ready.  Re-read in poll_transfer() to solve this issue
when poll_mode is enabled.

Verified on SPI touch screen device.

Signed-off-by: Major Lee <major_lee@wistron.com>
Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoi2c_intel_mid: Fix slash in sysfs name
Alan Cox [Tue, 14 Dec 2010 15:29:08 +0000]
i2c_intel_mid: Fix slash in sysfs name

This gets caught by the new sanity check code. Instead of the slash use a
different symbol. This was originally found by Major Lee who proposed a
rather more complex patch which changed the name according to the chip
type.

On the basis that we are in a late -rc and making Linus grumpy isn't always
a good idea (however fun) this is a simple alternative.

Signed-off-by: Alan Cox <alan@linux.intel.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

9 years agoext4: fix typo which broke '..' detection in ext4_find_entry()
Aaro Koskinen [Wed, 15 Dec 2010 02:45:31 +0000]
ext4: fix typo which broke '..' detection in ext4_find_entry()

There should be a check for the NUL character instead of '0'.

Fortunately the only thing that cares about this is NFS serving, which
is why we didn't notice this in the merge window testing.

Reported-by: Phil Carmody <ext-phil.2.carmody@nokia.com>
Signed-off-by: Aaro Koskinen <aaro.koskinen@nokia.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>

9 years agoMerge branch 'sh-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 15 Dec 2010 01:37:08 +0000]
Merge branch 'sh-fixes-for-linus' of git://git./linux/kernel/git/lethal/sh-2.6

* 'sh-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/lethal/sh-2.6:
  sh: wire up accept4 syscall (non-multiplexed path)
  sh: Enable deprecated IRQ chip APIs for MFD and GPIOLIB drivers.