isci: remove SCSI host before detaching from SAS transport

commit cff549e4860f ("scsi: proper state checking and module refcount
handling in scsi_device_get") , the reference count of scsi device was
changed, which could lead to when rmmod with at least on drive attached,
SCSI error handle will run into infinite loop, and lockup the system.

Fix it by remove scsi host first, this way scsi core will not send
commands down after detaching SAS transport.

This is a follow up fix for Benjamin's fix for pm80xx.

See also:
http://www.spinics.net/lists/linux-scsi/msg90088.html

Signed-off-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ipr: Driver version 2.6.3.

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ipr: Issue Configure Cache Parameters command.

Some new adapters require a special Configure Cache Parameters command
to enable the adapter write cache, so send this during the adapter
initialization if the adapter requires it.

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ipr: Inquiry IOA page 0xC4 during initialization.

Add an IOA Inquiry command for Page 0xC4 during IOA initialization to
collect cache capabilities, particularly to check if Sync IOA Write
Cache is supported.

Inquiry will happen right after Cap Inquiry on page 0xD0; and will
execute only if the "Supported Pages" field in Inquiry Page 0x0 shows
support for Page 0xC4.  Otherwise, assume Sync IOA Write Cache is
not supported.

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ipr: Don't set NO_ULEN_CHK bit when resource is a vset.

According to the IPR specification, Inhibit Underlength Checking bit
must be disabled when issuing commands to vsets.  Enabling it in this
case might cause SCSI commands to fail with an Illegal Request, so make
sure we keep this bit cleared when resource is a vset.

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

ipr: Add delay to ensure coherent dumps.

Add a holding pattern prior to collecting dump data, to wait for the IOA
indication that the Mailbox register is stable and won't change without
an explicit reset.  This ensures we'll be collecting meaningful dump
data, even when dumping right after an adapter reset.

In the event of a timeout, we still force the dump, since a partial dump
still might be useful.

Signed-off-by: Gabriel Krisman Bertazi <krisman@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs-qcom: add QUniPro hardware support and power optimizations

New revisions of UFS host controller supports the new UniPro
hardware controller (referred as QUniPro). This patch adds
the support to enable this new UniPro controller hardware.

This change also adds power optimization for bus scaling feature,
as well as support for HS-G3 power mode.

Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs-qcom: add debug prints for test bus

Adds support for configuring and reading the test bus and debug
registers. This change also adds another vops in order to print the
debug registers.

Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: make the UFS variant a platform device

This change turns the UFS variant (SCSI_UFS_QCOM) into a UFS
a platform device.
In order to do so a few additional changes are required:
1. The ufshcd-pltfrm is no longer serves as a platform device.
   Now it only serves as a group of platform APIs such as PM APIs
   (runtime suspend/resume, system suspend/resume etc), parsers of
   clocks, regulators and pm_levels from DT.
2. What used to be the old platform "probe" is now "only"
   a pltfrm_init() routine, that does exactly the same, but only
   being called by the new probe function of the UFS variant.

Reviewed-by: Rob Herring <robherring2@gmail.com>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: creates wrapper functions for vops

In order to simplify the code a set of wrapper functions is created
to test and call each of the variant operations.

Reviewed-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs: add ufshcd_get_variant ufshcd_set_variant

This patch adds ufshcd_get_variant() and ufshcd_set_variant()
routines in order to get/set the variant specific data.

Reviewed-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Tested-by: Alim Akhtar <alim.akhtar@samsung.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs-qcom: update configuration option of SCSI_UFS_QCOM component

This change is required in order to be able to build the component
as a module.

Reviewed-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: ufs-qcom: fix compilation warning if compiled as a module

This change fixes a compilation warning that happens if SCSI_UFS_QCOM is
compiled as a module.  Also this patch fixes an error happens when
insmod the module: "ufs_qcom: module license 'unspecified' taints
kernel."

Reviewed-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

phy: qcom-ufs: fix build error when the component is built as a module

Export the following functions in order to avoid build errors
when the component PHY_QCOM_UFS is compiled as a module:

ERROR: "ufs_qcom_phy_disable_ref_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_enable_ref_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_is_pcs_ready"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_disable_iface_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_start_serdes"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_calibrate_phy"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_enable_dev_ref_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_set_tx_lane_enable"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_disable_dev_ref_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_save_controller_version"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
ERROR: "ufs_qcom_phy_enable_iface_clk"
[drivers/scsi/ufs/ufs-qcom.ko] undefined!
make[1]: *** [__modpost] Error 1

Reviewed-by: Akinobu Mita <akinobu.mita@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Gilad Broner <gbroner@codeaurora.org>
Signed-off-by: Yaniv Gardi <ygardi@codeaurora.org>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: move lockup_detected attribute to host attr

This patch fixes a 'general protection fault' issue by
moving the attribute to where it was likely meant.

Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: bump the driver version

Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Gerry Morong <gerry.morong.pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: add in sas transport class

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: fix multiple issues in path_info_show

path_info_show() seems to be broken in multiple ways.

First, there's

  817 return snprintf(buf, output_len+1, "%s%s%s%s%s%s%s%s",
  818       path[0], path[1], path[2], path[3],
  819       path[4], path[5], path[6], path[7]);

so hopefully output_len contains the combined length of the eight
strings. Otherwise, snprintf will stop copying to the output
buffer, but still end up reporting that combined length - which
in turn would result in user-space getting a bunch of useless nul
bytes (thankfully the upper sysfs layer seems to clear the output
buffer before passing it to the various ->show routines). But we have

  767      output_len = snprintf(path[i],
  768                       PATH_STRING_LEN, "[%d:%d:%d:%d] %20.20s ",
  769                       h->scsi_host->host_no,
  770                       hdev->bus, hdev->target, hdev->lun,
  771                       scsi_device_type(hdev->devtype));

so output_len at best contains the length of the last string printed.

Inside the loop, we then otherwise add to output_len. By magic,
we still have PATH_STRING_LEN available every time... This
wouldn't really be a problem if the bean-counting has been done
properly and each line actually does fit in 50 bytes, and maybe
it does, but I don't immediately see why. Suppose we end up
taking this branch:

  802                  output_len += snprintf(path[i] + output_len,
  803                          PATH_STRING_LEN,
  804                          "BOX: %hhu BAY: %hhu %s\n",
  805                          box, bay, active);

An optimistic estimate says this uses strlen("BOX: 1 BAY: 2
Active\n") which is 21. Now add the 20 bytes guaranteed by the
%20.20s and then some for the rest of that format string, and
we're easily over 50 bytes. I don't think we can get over 100
bytes even being pessimistic, so this just means we'll scribble
into the next path[i+1] and maybe get that overwritten later,
leading to some garbled output (in fact, since we'd overwrite the
previous string's 0-terminator, we could end up with one very
long string and then print various suffixes of that, leading to
much more than 400 bytes of output). Except of course when we're
filling path[7], where overrunning it means writing random stuff
to the kernel stack, which is usually a lot of fun.

We can fix all of that and get rid of the 400 byte stack buffer by
simply writing directly to the given output buffer, which the upper
layer guarantees is at least PAGE_SIZE. s[c]nprintf doesn't care where
it is writing to, so this doesn't make the spin lock hold time any
longer. Using scnprintf ensures that output_len always represents the
number of bytes actually written to the buffer, so we'll report the
proper amount to the upper layer.

Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: enhance device messages

Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: disable report lun data caching

When external target arrays are present, disable the firmware's
normal behavior of returning a cached copy of the report lun data,
and force it to collect new data each time we request a report luns.

This is necessary for external arrays, since there may be no
reliable signal from the external array to the smart array when
lun configuration changes, and thus when driver requests
report luns, it may be stale data.

Use diag options to turn off RPL data caching.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: add discovery polling for PT RAID devices.

There are problems with getting configuration change notification
in pass-through RAID environments.  So, activate flag
h->discovery_polling when one of these devices is detected in
update_scsi_devices.

After discovery_polling is set, execute a report luns from
rescan_controller_worker (every 30 seconds).

If the data from report_luns is different than last
time (binary compare), execute a full rescan via update_scsi_devices.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: eliminate fake lun0 enclosures

We don't need to create fake enclosure devices at Lun0
in external target array configurations anymore.
This was done to support Pre-SCSI rev 5 controllers
that didn't suppoprt report luns commands, so the
SCSI layer had to scan targets. If there was no
LUN at LUN 0, then the target scan would stop, and
move to the next target.  Lun0 enclosure device
was added to prevent sparsely-numbered LUNs from
being missed.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: generalize external arrays

External array LUNs must use target and lun numbers assigned by the
external array. So the driver must treat these differently from
local LUNs when assigning lun/target.

LUN's 'model' field has been used to detect Lun types that need
special treatment, but the desire is to eliminate the need to reference
specific array models, and support any external array.

Pass-through RAID (PTRAID) luns are not luns of the local controller,
so they are not reported in LUN count of command 'ID controller'.
However, they ARE reported in "Report logical Luns" command.
Local luns are listed first, then PTRAID LUNs.

The number of luns from "Report LUNs" in excess of those reported by
'ID controller' are therefore the PTRAID LUNS.

We can now remove function is_ext_target, and the 'white list'
array of supported model names.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: move scsi_add_device and scsi_remove_device calls to new function

preparation for adding the sas transport class

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: refactor hpsa_figure_bus_target_lun

setup for sas transport. Need to set the
bus and target accordingly.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: enhance hpsa_get_device_id

use an index into vpd data for SAS/SATA drives

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: add function is_logical_device

simplify checking for logical/physical devices

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: simplify update scsi devices

remove repeated calculation that checks for physical
or logical devices.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: simplify check for device exposure

remove macros and cleanup device exposure checking

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: correct ioaccel2 sg chain len

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: correct check for non-disk devices

The driver is using two MACROs which seemingly are looking in
the wrong location for the device_flags returned from
CISS_REPORT_PHYS. Both MACROs, NON_DISK_PHYS_DEV and
PHYS_IOACCEL, are using the pointer returned from figure_lunaddrbytes
which is the address of the LUN.lunid element in
the extended CISS_REPORT_PHYS.  But the MACROS are using offsets
beyond the range of the element (offset 17 of an 8 byte element).

These MACROs actually are looking at the correct location but
they fail static checker analysis. It also will not work
if any new elements are added to the extended LUN structure.

Change the code to use the structure elements directly
since this MACRO is only used in one location.

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: fix physical target reset

Set reset type in device_reset_handler to do either
logical unit reset for logical devices, or physical
target reset, for physical devices.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: fix hpsa_adjust_hpsa_scsi_table

Fix a NULL pointer issue in the driver when devices are removed
during a reset.

Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: correct transfer length for 6 byte read/write commands

handle block counts of 0. Cleanup block and block count calculations.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: abandon rescans on memory alloaction failures.

Abandon and reschedule rescan process only if device inquiries
fail due to mem alloc failures, which are likely to occur for
all devices.

Otherwise, skip device if inquiry fails for other reasons,
and continue rescanning process for other devices.

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: allow driver requested rescans

Reviewed-by: Scott Teel <scott.teel@pmcs.com>
Reviewed-by: Justin Lindley <justin.lindley@pmcs.com>
Reviewed-by: Kevin Barnett <kevin.barnett@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by; Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: fix null device issues

Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: check for null arguments to dev_printk

Check for NULLs.

Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: change devtype to unsigned

This member is used in calls to scsi_device_type.
It should be unsigned since the kernel checks for upper bounds
and it should never be negative.

Suggested-by: Tomas Henzl <thenzl@redhat.com>
Suggested-by: Hannes Reinecke <hare@suse.de>
Suggested-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: remove unused hpsa_tag_discard_error_bits

This function is no longer used.

Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Manoj Kumar <manoj@linux.vnet.ibm.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: stop zeroing reset_cmds_out and ioaccel_cmds_out during rescan

pulling the rug out from under the reset handler
likewise for ioaccel_cmds_out

Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Don Brace <don.brace@pmcs.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

hpsa: remove unused parameter hostno

This parameter was once used before scan_start was defined
but now it is no longer used.

Signed-off-by: Don Brace <don.brace@pmcs.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi_sysfs: Fix queue_ramp_up_period return code

Writing a number to /sys/bus/scsi/devices/<sdev>/queue_ramp_up_period
returns the value of that number instead of the number of bytes written.
This behavior can confuse programs expecting POSIX write() semantics.
Fix this by returning the number of bytes written instead.

Signed-off-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Cc: stable@vger.kernel.org
Reviewed-by: Matthew R. Ochs <mrochs@linux.vnet.ibm.com>
Reviewed-by: Ewan D. Milne <emilne@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

scsi: Export SCSI Inquiry data to sysfs

Export the RAW SCSI Inquiry to sysfs as binfile. This way the data can be used
by userland without the need to have and ioctl or use the sg_inq tool.

Here is an example of the provided data

linux:~ # hexdump /sys/class/scsi_device/1\:0\:0\:0/device/inquiry
0000000 8005 3205 001f 0000 4551 554d 2020 2020
0000010 4551 554d 4420 4456 522d 4d4f 2020 2020
0000020 2e32 2e33
0000024

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Fix sparse warning

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Make tape drives visible on PERC5 controllers

The DELL PERC5 controller firmware does not list tape drives in response
to MR_DCMD_PD_LIST_QUERY. This causes tape drives not be exposed to the
OS when connected to a PERC5 controller.

This patch permits detection of tape drives connected to a PERC5
controller by exposing non-TYPE_DISK devices unconditionally.

Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

sg: Fix double-free when drives detach during SG_IO

In sg_common_write(), we free the block request and return -ENODEV if
the device is detached in the middle of the SG_IO ioctl().

Unfortunately, sg_finish_rem_req() also tries to free srp->rq, so we
end up freeing rq->cmd in the already free rq object, and then free
the object itself out from under the current user.

This ends up corrupting random memory via the list_head on the rq
object. The most common crash trace I saw is this:

  ------------[ cut here ]------------
  kernel BUG at block/blk-core.c:1420!
  Call Trace:
  [<ffffffff81281eab>] blk_put_request+0x5b/0x80
  [<ffffffffa0069e5b>] sg_finish_rem_req+0x6b/0x120 [sg]
  [<ffffffffa006bcb9>] sg_common_write.isra.14+0x459/0x5a0 [sg]
  [<ffffffff8125b328>] ? selinux_file_alloc_security+0x48/0x70
  [<ffffffffa006bf95>] sg_new_write.isra.17+0x195/0x2d0 [sg]
  [<ffffffffa006cef4>] sg_ioctl+0x644/0xdb0 [sg]
  [<ffffffff81170f80>] do_vfs_ioctl+0x90/0x520
  [<ffffffff81258967>] ? file_has_perm+0x97/0xb0
  [<ffffffff811714a1>] SyS_ioctl+0x91/0xb0
  [<ffffffff81602afb>] tracesys+0xdd/0xe2
    RIP [<ffffffff81281e04>] __blk_put_request+0x154/0x1a0

The solution is straightforward: just set srp->rq to NULL in the
failure branch so that sg_finish_rem_req() doesn't attempt to re-free
it.

Additionally, since sg_rq_end_io() will never be called on the object
when this happens, we need to free memory backing ->cmd if it isn't
embedded in the object itself.

KASAN was extremely helpful in finding the root cause of this bug.

Signed-off-by: Calvin Owens <calvinowens@fb.com>
Acked-by: Douglas Gilbert <dgilbert@interlog.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: avoid a panic if MSI(X) interrupts are disabled

If MSI(X) interrupts are disabled via the kernel command line
(pci=nomsi), the pm8001 driver will kernel panic because it does not
detect that MSI interrupts are disabled and will soldier on and attempt to
configure MSI interrupts anyways.  This leads to a kernel panic, most
likely because a required data structure is not available down the
line.  Using the pci_msi_enabled() function in order to detect if MSI
interrupts are enabled before configuring them resolves this issue and
avoids a kernel panic when the module is loaded.  Additionally, the
irq_vector structure must be initialized when legacy interrupts are
being used otherwise legacy interrupts will simply not function and
result in another panic.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: wait a minimum of 500ms before issuing commands to SPCv

The documentation for the 8070 and 8072 SPCv chip explicitly states that
a minimum of 500ms must elapse before issuing commands, otherwise the
SPCv may not process them and the firmware may get into an unrecoverable
state requiring a reboot.  While the Linux guys will probably think this
is 'racy', it is called out in the chip documentation and inserting this
delay makes power management function properly.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: do not examine registers for iButton feature if ATTO adapter

ATTO adapters do not support this feature.  If the firmware fails to be
ready, it should not check the examined registers in order to examine
the state of the feature in order to prevent undefined behavior.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: set PHY profiles for ATTO 12Gb SAS controllers

PHY profiles are not saved in NVRAM on ATTO 12Gb SAS controllers.
Therefore, in order for the controller to function in a wide range of
configurations, the PHY profiles must be statically set.  This patch
provides the necessary functionality to do so.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: add support for ATTO devices during SAS address initiailization

ATTO SAS controllers retrieve the SAS address from the NVRAM in a location
different from non-ATTO PMC Sierra SAS controllers.  This patch makes the
necessary adjustments in order to retrieve the SAS address on these types
of adapters.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: add ATTO PCI IDs to pm8001_pci_table

These PCI IDs allow the pm8001 driver to load against ATTO 12Gb SAS
controllers that use PMC Sierra 8070 and PMC Sierra 8072 SAS chips.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: add support for PMC Sierra 8070 and PMC Sierra 8072 SAS controllers

These SAS controllers support speeds up to 12Gb.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

pm80xx: configure PHY settings based on subsystem vendor ID

Previuosly, all PMC Sierra 80xx controllers are assumed to be a
motherboard controller, except if the subsystem vendor ID was equal to
PCI_VENDOR_ID_ADAPTEC.  The driver then attempts to load PHY settings
from NVRAM.  While this may be correct behavior for most controllers, it
does not work with Adaptec and ATTO controllers since they do not store
PHY settings in NVRAM and choose to use either custom PHY settings or
chip defaults.  Loading random values from NVRAM may cause the
controllers to malfunction in this edge case.

Signed-off-by: Benjamin Rood <brood@attotech.com>
Reviewed-by: Jack Wang <jinpu.wang@profitbricks.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

SCSI: Increase REPORT_LUNS timeout

This patch fixes an issue seen with an IBM 2145 (SVC) where, following an error
injection test which results in paths going offline, when they came
back online, the path would timeout the REPORT_LUNS issued during the
scan. This timeout situation continued until retries were expired, resulting in
falling back to a sequential LUN scan. Then, since the target responds
with PQ=1, PDT=0 for all possible LUNs, due to the way the sequential
LUN scan code works, we end up adding 512 LUNs for each target, when there
is really only a small handful of LUNs that are actually present.

This patch increases the timeout used on the REPORT_LUNS to 30 seconds.
This patch solves the issue of 512 non existent LUNs showing up after
this event.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas : Remove debug print from function megasas_update_span_set

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas : Driver version upgrade

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas : SMAP restriction--do not access user memory from IOCTL code

This is an issue on SMAP enabled CPUs and 32 bit apps running on 64 bit
OS. Do not access user memory from kernel code. The SMAP bit restricts
accessing user memory from kernel code.

Cc: <stable@vger.kernel.org>
Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Make PI enabled VD 8 byte DMA aligned

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Initialize tasklet before setting up IRQs

It may happen (kdump), that an interrupt is invoked just after the
setup_irqs function was called but before the tasklet was initialised.
At this phase the hw ints should have been disabled, but for unknown
reason this mechanism seems to not work properly.

From: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Indicate online firmware upgrade support for Secure JBOD feature

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Update OCR capability on controller properties change

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Do not use PAGE_SIZE for max_sectors

Do not use PAGE_SIZE marco to calculate max_sectors per I/O
request. Driver code assumes PAGE_SIZE will be always 4096 which can
lead to wrongly calculated value if PAGE_SIZE is not 4096. This issue
was reported in Ubuntu Bugzilla Bug #1475166.

Cc: <stable@vger.kernel.org>
Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Support for Cutlass (12 Gbps) controller

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Support for Intruder (12 Gbps) controller

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Remove PCI id checks

Remove PCI id based checks and use instance->ctrl_context to decide
whether controller is MFI-based or a Fusion adapter. Additionally,
Fusion adapters are divided into two categories: Thunderbolt and
Invader.

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Expose TAPE drives unconditionally

Expose non-disk (TAPE drive, CD-ROM) unconditionally.

Cc: <stable@vger.kernel.org>
Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Tomas Henzl <thenzl@redhat.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Version update

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Code refactor for use of requestorId

Some of these code changes were proposed by David Binderman.

Removed redudant check of requestorId. Redundant condition:
instance.requestorId. Check for plasma firmware 1.11 are now
restructured to support only specific device id.

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Fix validHandles check in I/O path

Syncro firmware supports round robin I/O switching on dual path. Driver
uses validHandles to check for dual path. However, it is supposed to
check for values > 1 (not > 2).

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Print critical firmware event messages

Print firmware events in human-readable form. This will help users track
any critical firmware events without special application support.

Sample syslogd output:

megaraid_sas 0000:02:00.0: 8619 (491648347s/0x0020/WARN) - Controller temperature threshold exceeded. This may indicate inadequate system cooling. Switching to low performance mode.

The format of logged events is:

"<pci_dev_id>: <sequence_number> (<timestamp>/<locale>/<class>) - <description>"

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Chip reset if driver fails to get IOC ready

Fix the issue reported at:

http://marc.info/?l=linux-scsi&m=143694494104544&w=2

Try to do chip reset at driver load time. If firmware fails to reach
ready state, try chip reset using adp_reset() callback. For Fusion
adapters the call back was previously void. Provide a suitable reset
function.

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Support for max_io_size 1MB

Driver will expose max sge = 256 (earlier it was 64) if firmware
supports extended IO size (1M).

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Code cleanup-use local variable drv_ops inside megasas_ioc_init_fusion

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: JBOD sequence number support

Implemented JBOD map which will provide quick access for JBOD path and
also provide sequence number.  This will help hardware to fail command
to the FW in case of any sequence mismatch.

Fast Path I/O for JBOD will refer JBOD map (which has sequence number
per JBOD device) instead of RAID map.  Previously, the driver used RAID
map to get device handle for fast path I/O and this not have sequence
number information. Now, driver will use JBOD map instead.  As part of
error handling, if JBOD map is failed/not supported by firmware, driver
will continue using legacy behavior.

Now there will be three IO paths for JBOD (syspd):

 - JBOD map with sequence number (Fast Path)
 - RAID map without sequence number (Fast Path)
 - FW path via h/w exception queue deliberately setup devhandle
   0xFFFF (FW path).

Relevant data structures:

 - Driver send new DCMD MR_DCMD_SYSTEM_PD_MAP_GET_INFO for this purpose.
 - struct MR_PD_CFG_SEQ- This structure represent map of single physical
   device.
 - struct MR_PD_CFG_SEQ_NUM_SYNC- This structure represent whole JBOD
   map in general(size, count of sysPDs configured, struct MR_PD_CFG_SEQ
   of syspD with 0 index).
 - JBOD sequence map size is: sizeof(struct MR_PD_CFG_SEQ_NUM_SYNC)
   + (sizeof(struct MR_PD_CFG_SEQ) * (MAX_PHYSICAL_DEVICES - 1)) which
   is allocated while setting up JBOD map at driver load time.

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Increase timeout to 60 secs for abort frames during shutdown

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

megaraid_sas: Synchronize driver headers with firmware APIs

Signed-off-by: Sumit Saxena <sumit.saxena@avagotech.com>
Signed-off-by: Kashyap Desai <kashyap.desai@avagotech.com>
Reviewed-by: Martin Petersen <martin.petersen@oracle.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>

SCSI: fix bug in scsi_dev_info_list matching

The "compatible" matching algorithm used for looking up old-style
blacklist entries in a scsi_dev_info_list is buggy.  The core of the
algorithm looks like this:

if (memcmp(devinfo->vendor, vendor,
    min(max, strlen(devinfo->vendor))))
/* not a match */

where max is the length of the device's vendor string after leading
spaces have been removed but trailing spaces have not.  Because of the
min() computation, either entry could be a proper substring of the
other and the code would still think that they match.

In the case originally reported, the device's vendor and product
strings were "Inateck " and "                ".  These matched against
the following entry in the global device list:

{"", "Scanner", "1.80", BLIST_NOLUN}

because "" is a substring of "Inateck " and "" (the result of removing
leading spaces from the device's product string) is a substring of
"Scanner".  The mistaken match prevented the system from scanning and
finding the device's second Logical Unit.

This patch fixes the problem by making two changes.  First, the code
for leading-space removal is hoisted out of the loop.  (This means it
will sometimes run unnecessarily, but since a large percentage of all
lookups involve the "compatible" entries in global device list, this
should be an overall improvement.)  Second and more importantly, the
patch removes trailing spaces and adds a check to verify that the two
resulting strings are exactly the same length.  This prevents matches
where one entry is a proper substring of the other.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Reported-by: Giulio Bernardi <ugilio@gmail.com>
Tested-by: Giulio Bernardi <ugilio@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

SCSI: refactor device-matching code in scsi_devinfo.c

In drivers/scsi/scsi_devinfo.c, the scsi_dev_info_list_del_keyed() and
scsi_get_device_flags_keyed() routines contain a large amount of
duplicate code for finding vendor/product matches in a
scsi_dev_info_list.  This patch factors out the duplicate code and
puts it in a separate function, scsi_dev_info_list_find().

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Suggested-by: Giulio Bernardi <ugilio@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

fnic: check pci_map_single() return value

the kernel prints some warnings when compiled with CONFIG_DMA_API_DEBUG.
This is because the fnic driver doesn't check the return value of
pci_map_single().

[   11.942770] scsi host12: fnic
[   11.950811] ------------[ cut here ]------------
[   11.950818] WARNING: at lib/dma-debug.c:937 check_unmap+0x47b/0x920()
[   11.950821] fnic 0000:0c:00.0: DMA-API: device driver failed to check map error[device address=0x0000002020a30040] [size=44 bytes] [mapped as single]

Signed-off-by: Maurizio Lombardi <mlombard@redhat.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Reviewed By: Tomas Henzl <thenzl@redhat.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

be2iscsi: Revert ownership to Emulex

We would like to get the following updates in:
    Revert ownership to "Emulex" from "Avago Technologies"

Signed-off-by: Ketan Mukadam <ketan.mukadam@avagotech.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

MAINTAINERS: Update be2iscsi driver

Signed-off-by: Ketan Mukadam <ketan.mukadam@avagotech.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

bnx2fc: Do not log error for netevents that need no action

Do not log error for netevents that need no action such as
NETDEV_REGISTER 0x0005, NETDEV_CHANGEADDR, and NETDEV_CHANGENAME.
It results in logging error messages such as these

[   35.315872] bnx2fc: Unknown netevent 5
[   35.315935] bnx2fc: Unknown netevent 8
[   35.353866] bnx2fc: Unknown netevent 10

and generating bug reports.
Remove logging this message as an ERROR instead of turning them into
either DEBUG or INFO level messages.

Signed-off-by: Shirish Pargaonkar <spargaonkar@suse.com>
Acked-by: Eddie Wai <eddie.wai@broadcom.com>
Acked-by: Chad Dupuis <chad.dupuis@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Update version to 11.0.0.0 for upstream patch set

Signed-off-by: Dick Kennedy <dick.kennedy@avagotech.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Fix default RA_TOV and ED_TOV in the FC/FCoE driver for all topologies

Initial link up defaults were not properly being tracked relative to
initial FLOGI or pt2pt PLOGI. Add code to initialize them.

Signed-off-by: Dick Kennedy <dick.kennedy@avagotech.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: The linux driver does not reinitiate discovery after a failed FLOGI

Forgot to clear FCF Discovery in-progress flag upon FLOGI failures.
Thus we didn't restart FLOGI.

Signed-off-by: Dick Kennedy <dick.kennedy@avagotech.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Fix for discovery failure in PT2PT when FLOGI's ELS ACC response gets aborted

Fix for discovery failure in PT2PT when FLOGI's ELS ACC response gets aborted

Change login state machine to:
- Restart FLOGI if prior is ABTS'd
- Reject incoming FLOGIs if we have one pending

The above ensures that we always finish FLOGI processing, regardless
of who initated FLOGI, before processing PLOGI's.

Signed-off-by: Dick Kennedy <dick.kennedy@avagotech.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Add support for Lancer G6 and 32G FC links

Signed-off-by: Dick Kennedy <dick.kennedy@avagotech.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

fix: lpfc_send_rscn_event sends bigger buffer size

lpfc_send_rscn_event() allocates data for sizeof(struct
lpfc_rscn_event_header) + payload_len, but claims that the data has size
of sizeof(struct lpfc_els_event_header) + payload_len. That leads to
buffer overruns.

Signed-off-by: Ales Novak <alnovak@suse.cz>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sebastian Herbszt <herbszt@gmx.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: remove set but not used variables

Remove set but not used variables.

Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc:Make the function lpfc_sli4_mbox_completions_pending static in order to comply with function prototype

This makes the function lpfc_sli4_mbox_completion's definition
static now in order to comply with its prototype being also
declared as static too.

Signed-off-by: Nicholas Krause <xerofoify@gmail.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Sebastian Herbszt <herbszt@gmx.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

Fix kmalloc overflow in LPFC driver at large core count

This patch allows the LPFC to start up without a fatal kernel bug based
on an exceeded KMALLOC_MAX_SIZE and a too large NR_CPU-based maskbits
field. The bug was based on the number of CPU cores in a system.
Using the get_cpu_mask() function declared in kernel/cpu.c allows the
driver to load on the community kernel 4.2 RC1.

Below is the kernel bug reproduced:

8<--------------------------------------------------------------------
2199382.828437 (    0.005216)| lpfc 0003:02:00.0: enabling device (0140 -> 0142)
2199382.999272 (    0.170835)| ------------[ cut here ]------------
2199382.999337 (    0.000065)| WARNING: CPU: 84 PID: 404 at mm/slab_common.c:653 kmalloc_slab+0x2f/0x89()
2199383.004534 (    0.005197)| Modules linked in: lpfc(+) usbcore(+) mptctl scsi_transport_fc sg lpc_ich i2c_i801 usb_common tpm_tis mfd_core tpm acpi_cpufreq button scsi_dh_alua scsi_dh_rdacusbcore: registered new device driver usb
2199383.020568 (    0.016034)|
2199383.020581 (    0.000013)|  scsi_dh_hp_sw scsi_dh_emc scsi_dh gru thermal sata_nv processor piix fan thermal_sysehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
2199383.035288 (    0.014707)|
2199383.035306 (    0.000018)|  hwmon ata_piix
2199383.035336 (    0.000030)| CPU: 84 PID: 404 Comm: kworker/84:0 Not tainted 3.18.0-rc2-gat-00106-ga7ca10f-dirty #178
2199383.047077 (    0.011741)| ehci-pci: EHCI PCI platform driver
2199383.047134 (    0.000057)| Hardware name: SGI UV2000/ROMLEY, BIOS SGI UV 2000/3000 series BIOS 01/15/2013
2199383.056245 (    0.009111)| Workqueue: events work_for_cpu_fn
2199383.066174 (    0.009929)|  000000000000028d ffff88eef827bbe8 ffffffff815a542f 000000000000028d
2199383.069545 (    0.003371)|  ffffffff810ea142 ffff88eef827bc28 ffffffff8104365c ffff88eefe4006c8
2199383.076214 (    0.006669)|  0000000000000000 00000000000080d0 0000000000000000 0000000000000004
2199383.079213 (    0.002999)| Call Trace:
2199383.084084 (    0.004871)|  [<ffffffff815a542f>] dump_stack+0x49/0x62
2199383.087283 (    0.003199)|  [<ffffffff810ea142>] ? kmalloc_slab+0x2f/0x89
2199383.091415 (    0.004132)|  [<ffffffff8104365c>] warn_slowpath_common+0x77/0x92
2199383.095197 (    0.003782)|  [<ffffffff8104368c>] warn_slowpath_null+0x15/0x17
2199383.103336 (    0.008139)|  [<ffffffff810ea142>] kmalloc_slab+0x2f/0x89
2199383.107082 (    0.003746)|  [<ffffffff8110fd9e>] __kmalloc+0x13/0x16a
2199383.112531 (    0.005449)|  [<ffffffffa01a8ed9>] lpfc_pci_probe_one_s4+0x105b/0x1644 [lpfc]
2199383.115316 (    0.002785)|  [<ffffffff81302b92>] ? pci_bus_read_config_dword+0x75/0x87
2199383.123431 (    0.008115)|  [<ffffffffa01a951f>] lpfc_pci_probe_one+0x5d/0xcb5 [lpfc]
2199383.127364 (    0.003933)|  [<ffffffff81497119>] ? dbs_check_cpu+0x168/0x177
2199383.136438 (    0.009074)|  [<ffffffff81496fa5>] ? gov_queue_work+0xb4/0xc0
2199383.140407 (    0.003969)|  [<ffffffff8130b2a1>] local_pci_probe+0x1e/0x52
2199383.143105 (    0.002698)|  [<ffffffff81052c47>] work_for_cpu_fn+0x13/0x1b
2199383.147315 (    0.004210)|  [<ffffffff81054965>] process_one_work+0x222/0x35e
2199383.151379 (    0.004064)|  [<ffffffff81054e76>] worker_thread+0x3d5/0x46e
2199383.159402 (    0.008023)|  [<ffffffff81054aa1>] ? process_one_work+0x35e/0x35e
2199383.163097 (    0.003695)|  [<ffffffff810599c6>] kthread+0xc8/0xd2
2199383.167476 (    0.004379)|  [<ffffffff810598fe>] ? kthread_freezable_should_stop+0x5b/0x5b
2199383.176434 (    0.008958)|  [<ffffffff815a8cac>] ret_from_fork+0x7c/0xb0
2199383.180086 (    0.003652)|  [<ffffffff810598fe>] ? kthread_freezable_should_stop+0x5b/0x5b
2199383.192333 (    0.012247)| ehci-pci 0000:00:1a.0: EHCI Host Controller
-------------------------------------------------------------------->8

The proposed solution was approved by James Smart at Emulex and tested
on a UV2 machine with 6144 cores. With the fix, the LPFC module loads
with no unwanted effects on the system.

Signed-off-by: Ian Mitchell <imitchell@sgi.com>
Signed-off-by: Alex Thorlton <athorlton@sgi.com>
Suggested-by: Robert Elliot <elliott@hp.com>
[james.smart: resolve unused variable warning]
Signed-off-by: James Smart <james.smart@avagotech.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Destroy lpfc_hba_index IDR on module exit

Destroy lpfc_hba_index IDR on module exit, reclaiming the allocated memory.

This was detected by the following semantic patch (written by Luis Rodriguez
<mcgrof@suse.com>)
<SmPL>
@ defines_module_init @
declarer name module_init, module_exit;
declarer name DEFINE_IDR;
identifier init;
@@

module_init(init);

@ defines_module_exit @
identifier exit;
@@

module_exit(exit);

@ declares_idr depends on defines_module_init && defines_module_exit @
identifier idr;
@@

DEFINE_IDR(idr);

@ on_exit_calls_destroy depends on declares_idr && defines_module_exit @
identifier declares_idr.idr, defines_module_exit.exit;
@@

exit(void)
{
 ...
 idr_destroy(&idr);
 ...
}

@ missing_module_idr_destroy depends on declares_idr && defines_module_exit && !on_exit_calls_destroy @
identifier declares_idr.idr, defines_module_exit.exit;
@@

exit(void)
{
 ...
 +idr_destroy(&idr);
}
</SmPL>

Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: in sli3 use configured sg_seg_cnt for sg_tablesize

Currently the module parameter lpfc_sg_seg_count does not have effect
for sli3 devices.

In lpfc_sli_driver_resource_setup(), which is used for sli3, the code
writes the configured sg_seg_cnt into lpfc_template.sg_tablesize.
But lpfc_template is the template used for sli4 only. Thus the value should
correctly be written to lpfc_template_s3->sg_tablesize.

This patch is for kernel 4.1-rc5, but is tested with lpfc 10.2.405.26 only.

Signed-off-by: Bodo Stroesser <bstroesser@ts.fujitsu.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Sebastian Herbszt <herbszt@gmx.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: Remove unnessary cast

kzalloc() returns a void pointer - no need to cast it in
drivers/scsi/lpfc/lpfc_init.c::lpfc_sli_driver_resource_setup()

Signed-off-by: Firo Yang <firogm@gmail.com>
Signed-off-by: James Smart <james.smart@avagotech.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

lpfc: fix model description

Remove trailing space from model description.

Signed-off-by: Sebastian Herbszt <herbszt@gmx.de>
Signed-off-by: James Smart <james.smart@avagotech.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

3w-9xxx: don't unmap bounce buffered commands

3w controller don't dma map small single SGL entry commands but instead
bounce buffer them.  Add a helper to identify these commands and don't
call scsi_dma_unmap for them.

Based on an earlier patch from James Bottomley.

Fixes: 118c85 ("3w-9xxx: fix command completion race")
Reported-by: Tóth Attila <atoth@atoth.sote.hu>
Tested-by: Tóth Attila <atoth@atoth.sote.hu>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Adam Radford <aradford@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

scsi_dh: Use the correct module name when loading device handler

This fixes a bug in recent kernels which results in failure to boot
on systems that have multipath SCSI disks.  I observed this failure
on a POWER8 server where all the disks are multipath SCSI disks.
The symptoms are several messages like this on the console:

[    3.018700] device-mapper: table: 253:0: multipath: error attaching hardware handler
[    3.018828] device-mapper: ioctl: error adding target to table

and the system does not find its disks, and therefore fails to boot.

Bisection revealed that the bug was introduced in commit 566079c849cf,
"dm-mpath, scsi_dh: request scsi_dh modules in scsi_dh, not dm-mpath".
The specific reason for the failure is that where we previously loaded
the "scsi_dh_alua" module, we are now trying to load the "alua" module,
which doesn't exist.

To fix this, we change the request_module call in scsi_dh_lookup()
to prepend "scsi_dh_" to the name, just like the old code in
drivers/md/dm-mpath.c:parse_hw_handler() used to do.

[jejb: also fixes issue spotted by Sasha Levin that formatting
characters could be passed in via sysfs and cause issues with
request_module()]

Fixes: 566079c849cf
Signed-off-by: Paul Mackerras <paulus@ozlabs.org>
Cc: Sasha Levin <sasha.levin@oracle.com>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: James Bottomley <JBottomley@Odin.com>

libiscsi: Fix iscsi_check_transport_timeouts possible infinite loop

Connection last_ping is not being updated when iscsi_send_nopout fails.
Not updating the last_ping will cause firing a timer to a past time
(last_ping + ping_tmo < current_time) which triggers an infinite loop of
iscsi_check_transport_timeouts() and hogs the cpu.

Fix this issue by checking the return value of iscsi_send_nopout.
If it fails set the next_timeout to one second later.

Signed-off-by: Ariel Nahum <arieln@mellanox.com>
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Odin.com>