linux-edison.git
12 years ago[SCSI] qla4xxx: Fix iocb_cnt calculation in qla4xxx_send_mbox_iocb()
Vikas Chaudhary [Thu, 2 May 2013 07:42:10 +0000 (03:42 -0400)]
[SCSI] qla4xxx: Fix iocb_cnt calculation in qla4xxx_send_mbox_iocb()

Increment 'ha->iocb_cnt' before adding mbox_iocb to iocb queue.

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ufs: Correct the expected data transfersize
Vinayak Holikatti [Tue, 26 Feb 2013 12:34:46 +0000 (18:04 +0530)]
[SCSI] ufs: Correct the expected data transfersize

This patch corrects the expected data transfer size of the
command UPIU. The current implementation of cmd->transfersize
is wrong as it probably equal to sector size. With this
implementation the transfer size is updated correctly

Reported-by: KOBAYASHI Yoshitake <yoshitake.kobayashi@toshiba.co.jp>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Namjae Jeon <linkinjeon@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Tested-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Santosh Yaraganavi <santoshsy@gmail.com>
Signed-off-by: Vinayak Holikatti <vinholikatti@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ufs: Add Platform glue driver for ufshcd
Vinayak Holikatti [Tue, 26 Feb 2013 12:34:45 +0000 (18:04 +0530)]
[SCSI] ufs: Add Platform glue driver for ufshcd

This patch adds Platform glue driver for ufshcd.

Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Namjae Jeon <linkinjeon@gmail.com>
Reviewed-by: Subhash Jadavani <subhashj@codeaurora.org>
Reviewed-by: Sujit Reddy Thumma <sthumma@codeaurora.org>
Tested-by: Maya Erez <merez@codeaurora.org>
Signed-off-by: Vinayak Holikatti <vinholikatti@gmail.com>
Signed-off-by: Santosh Yaraganavi <santoshsy@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] sd: fix array cache flushing bug causing performance problems
James Bottomley [Wed, 24 Apr 2013 21:02:53 +0000 (14:02 -0700)]
[SCSI] sd: fix array cache flushing bug causing performance problems

Some arrays synchronize their full non volatile cache when the sd driver sends
a SYNCHRONIZE CACHE command.  Unfortunately, they can have Terrabytes of this
and we send a SYNCHRONIZE CACHE for every barrier if an array reports it has a
writeback cache.  This leads to massive slowdowns on journalled filesystems.

The fix is to allow userspace to turn off the writeback cache setting as a
temporary measure (i.e. without doing the MODE SELECT to write it back to the
device), so even though the device reported it has a writeback cache, the
user, knowing that the cache is non volatile and all they care about is
filesystem correctness, can turn that bit off in the kernel and avoid the
performance ruinous (and safety irrelevant) SYNCHRONIZE CACHE commands.

The way you do this is add a 'temporary' prefix when performing the usual
cache setting operations, so

echo temporary write through > /sys/class/scsi_disk/<disk>/cache_type

Reported-by: Ric Wheeler <rwheeler@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla2xxx: qla2x00_sp_compl can be static.
Fengguang Wu [Thu, 25 Apr 2013 05:29:19 +0000 (01:29 -0400)]
[SCSI] qla2xxx: qla2x00_sp_compl can be static.

Signed-off-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla2xxx: fix sparse warning "large integer implicitly truncated to unsigned...
Armen Baloyan [Thu, 25 Apr 2013 05:29:18 +0000 (01:29 -0400)]
[SCSI] qla2xxx: fix sparse warning "large integer implicitly truncated to unsigned type"

Found by 0 day test project

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Signed-off-by: Armen Baloyan <armen.baloyan@qlogic.com>
Signed-off-by: Saurav Kashyap <saurav.kashyap@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] scsi_debug: fix logical block provisioning support
Akinobu Mita [Tue, 16 Apr 2013 13:11:58 +0000 (22:11 +0900)]
[SCSI] scsi_debug: fix logical block provisioning support

provisioning map (map_storep) is a bitmap accessed by bitops.

So the allocation size should be a multiple of sizeof(unsigned long) and
also the bitmap should be cleared by using bitmap_clear() instead of
memset().

Otherwise it will cause problem on big-endian architecture if the number of
bits is not a multiple of BITS_PER_LONG.

I tried testing the logical block provisioning support in scsi_debug,
but it didn't work as I expected.

For example, load scsi_debug module with UNMAP command supported
and fill the storage with random data.

        # modprobe scsi_debug lbpu=1
        # dd if=/dev/urandom of=/dev/sdb

Then, try to unmap LBA 0, but Get LBA status reports:

        # sg_unmap --lba=0 --num=1 /dev/sdb
        # sg_get_lba_status --lba=0 /dev/sdb
        descriptor LBA: 0x0000000000000000  blocks: 16384  mapped

This is unexpected result.  Because UNMAP command to LBA 0 finished
without any errors, but Get LBA status shows that LBA 0 is still mapped.

This problem is due to the wrong translation between LBA and index of
provisioning map.  Fix it by using correct translation functions.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] scsi_debug: clear correct memory region when LBPRZ is enabled
Akinobu Mita [Tue, 16 Apr 2013 13:11:57 +0000 (22:11 +0900)]
[SCSI] scsi_debug: clear correct memory region when LBPRZ is enabled

The function unmap_region() clears memory region specified as the logical
block address and the number of logical blocks in ramdisk storage
(fake_storep) if lbpu and lbprz module parameters are enabled.

In the while loop of unmap_region(), it advances optimal unmap granularity
in logical blocks.  But it only clears one logical block at LBA 'block' per
loop iteration.  And furthermore, the 'block' is not pointing to a logical
block address which should be cleared, it is a index of probisioning map
(map_storep).

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] scsi_debug: prohibit scsi_debug_unmap_granularity == scsi_debug_unmap_alignment
Akinobu Mita [Tue, 16 Apr 2013 13:11:56 +0000 (22:11 +0900)]
[SCSI] scsi_debug: prohibit scsi_debug_unmap_granularity == scsi_debug_unmap_alignment

scsi_debug prohibits setting scsi_debug_unmap_alignment to be greater
than scsi_debug_unmap_granularity.  But setting them to be the same value
is not prohibited.  In this case, the only difference with
scsi_debug_unmap_alignment == 0 is the logical blocks from 0 to
scsi_debug_unmap_alignment - 1 cannot be unmapped.  But the difference is
not properly handled in the current code.

So this prohibits such unusual setting.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] scsi_debug: call map_region() and unmap_region() only when needed
Akinobu Mita [Tue, 16 Apr 2013 13:11:55 +0000 (22:11 +0900)]
[SCSI] scsi_debug: call map_region() and unmap_region() only when needed

If the logical block provisioning is not enabled, map_region() and
unmap_region() have no effect and they don't need to be called.

So this makes map_region() and unmap_region() to be called only
when scsi_debug_lbp() returns true, i.e. logical block provisioning is
enabled.

While I'm at it, this also removes meaningless non-zero check for
scsi_debug_unmap_granularity.

Because scsi_debug_unmap_granularity cannot be zero with usual setting:
scsi_debug_unmap_granularity is 1 by default, and it can be changed to
zero with explicit module parameter setting only when the logical block
provisioning is disabled.  But it is only meaningful module parameter
when the logical block provisioning is enabled.

Signed-off-by: Akinobu Mita <akinobu.mita@gmail.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] sd_dif: problem with verify of type 1 protection information (PI)
Jeremy Higdon [Thu, 18 Apr 2013 09:55:23 +0000 (02:55 -0700)]
[SCSI] sd_dif: problem with verify of type 1 protection information (PI)

It appears to me that there is a problem with handling of type 1 protection
information.

It is considering a logical block reference tag of 0xffffffff to be an error,
but it is actually valid any time ((lba & 0xffffffff) == 0xffffffff) [for
example, 2TiB-1, 4TiB-1, 6TiB-1, etc.].

I'm going by what's written in 4.18.3 of SBC3, where there doesn't appear
to be any invalid value for the reference tag.

Signed-off-by: Jeremy Higdon <jeremy@sgi.com>
Acked-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ipr: SATA DVD probing failed with 64bit adapter
wenxiong@linux.vnet.ibm.com [Fri, 19 Apr 2013 02:32:48 +0000 (21:32 -0500)]
[SCSI] ipr: SATA DVD probing failed with 64bit adapter

Driver passed the wrong IOADL address to IOA adapter. The patch
fixes the issue.

Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] scsi_transport_iscsi: fix error return code in iscsi_transport_init()
Wei Yongjun [Fri, 19 Apr 2013 01:23:28 +0000 (09:23 +0800)]
[SCSI] scsi_transport_iscsi: fix error return code in iscsi_transport_init()

Fix to return -ENOMEM in the create workqueue error case
instead of 0, as done elsewhere in this function.

Signed-off-by: Wei Yongjun <yongjun_wei@trendmicro.com.cn>
Reviewed-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 2 May 2013 21:56:31 +0000 (14:56 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Ingo Molnar:
 "This fixes the cputime scaling overflow problems for good without
  having bad 32-bit overhead, and gets rid of the div64_u64_rem() helper
  as well."

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  Revert "math64: New div64_u64_rem helper"
  sched: Avoid prev->stime underflow
  sched: Do not account bogus utime
  sched: Avoid cputime scaling overflow

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Thu, 2 May 2013 21:53:12 +0000 (14:53 -0700)]
Merge git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto update from Herbert Xu:

 - XTS mode optimisation for twofish/cast6/camellia/aes on x86

 - AVX2/x86_64 implementation for blowfish/twofish/serpent/camellia

 - SSSE3/AVX/AVX2 optimisations for sha256/sha512

 - Added driver for SAHARA2 crypto accelerator

 - Fix for GMAC when used in non-IPsec secnarios

 - Added generic CMAC implementation (including IPsec glue)

 - IP update for crypto/atmel

 - Support for more than one device in hwrng/timeriomem

 - Added Broadcom BCM2835 RNG driver

 - Misc fixes

* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (59 commits)
  crypto: caam - fix job ring cleanup code
  crypto: camellia - add AVX2/AES-NI/x86_64 assembler implementation of camellia cipher
  crypto: serpent - add AVX2/x86_64 assembler implementation of serpent cipher
  crypto: twofish - add AVX2/x86_64 assembler implementation of twofish cipher
  crypto: blowfish - add AVX2/x86_64 implementation of blowfish cipher
  crypto: tcrypt - add async cipher speed tests for blowfish
  crypto: testmgr - extend camellia test-vectors for camellia-aesni/avx2
  crypto: aesni_intel - fix Kconfig problem with CRYPTO_GLUE_HELPER_X86
  crypto: aesni_intel - add more optimized XTS mode for x86-64
  crypto: x86/camellia-aesni-avx - add more optimized XTS code
  crypto: cast6-avx: use new optimized XTS code
  crypto: x86/twofish-avx - use optimized XTS code
  crypto: x86 - add more optimized XTS-mode for serpent-avx
  xfrm: add rfc4494 AES-CMAC-96 support
  crypto: add CMAC support to CryptoAPI
  crypto: testmgr - add empty test vectors for null ciphers
  crypto: testmgr - add AES GMAC test vectors
  crypto: gcm - fix rfc4543 to handle async crypto correctly
  crypto: gcm - make GMAC work when dst and src are different
  hwrng: timeriomem - added devicetree hooks
  ...

12 years agoMerge tag 'for-linus-v3.10-rc1' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Thu, 2 May 2013 21:49:33 +0000 (14:49 -0700)]
Merge tag 'for-linus-v3.10-rc1' of git://oss.sgi.com/xfs/xfs

Pull xfs update from Ben Myers:
 "For 3.10-rc1 we have a number of bug fixes and cleanups and a
  currently experimental feature from David Chinner, CRCs protection for
  metadata.  CRCs are enabled by using mkfs.xfs to create a filesystem
  with the feature bits set.

   - numerous fixes for speculative preallocation
   - don't verify buffers on IO errors
   - rename of random32 to prandom32
   - refactoring/rearrangement in xfs_bmap.c
   - removal of unused m_inode_shrink in struct xfs_mount
   - fix error handling of xfs_bufs and readahead
   - quota driven preallocation throttling
   - fix WARN_ON in xfs_vm_releasepage
   - add ratelimited printk for different alert levels
   - fix spurious forced shutdowns due to freed Extent Free Intents
   - remove some obsolete XLOG_CIL_HARD_SPACE_LIMIT() macros
   - remove some obsoleted comments
   - (experimental) CRC support for metadata"

* tag 'for-linus-v3.10-rc1' of git://oss.sgi.com/xfs/xfs: (46 commits)
  xfs: fix da node magic number mismatches
  xfs: Remote attr validation fixes and optimisations
  xfs: Teach dquot recovery about CONFIG_XFS_QUOTA
  xfs: add metadata CRC documentation
  xfs: implement extended feature masks
  xfs: add CRC checks to the superblock
  xfs: buffer type overruns blf_flags field
  xfs: add buffer types to directory and attribute buffers
  xfs: add CRC protection to remote attributes
  xfs: split remote attribute code out
  xfs: add CRCs to attr leaf blocks
  xfs: add CRCs to dir2/da node blocks
  xfs: shortform directory offsets change for dir3 format
  xfs: add CRC checking to dir2 leaf blocks
  xfs: add CRC checking to dir2 data blocks
  xfs: add CRC checking to dir2 free blocks
  xfs: add CRC checks to block format directory blocks
  xfs: add CRC checks to remote symlinks
  xfs: split out symlink code into it's own file.
  xfs: add version 3 inode format with CRCs
  ...

12 years agoMerge tag 'for-linus-docs-2012-05-02' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 2 May 2013 21:45:11 +0000 (14:45 -0700)]
Merge tag 'for-linus-docs-2012-05-02' of git://git./linux/kernel/git/sarah/xhci

Pull ReportingBugs rewrite from Sarah Sharp:
 "Here are the updates to ReportingBugs that were discussed and acked a
  couple weeks ago.  I've updated the fifth patch with your ack, as
  requested"

* tag 'for-linus-docs-2012-05-02' of git://git.kernel.org/pub/scm/linux/kernel/git/sarah/xhci:
  Docs: Move ref to Frohwalt Egerer to end of REPORTING-BUGS
  Docs: Add a tips section to REPORTING-BUGS.
  Docs: Expectations for bug reporters and maintainers
  Docs: Add info on supported kernels to REPORTING-BUGS.
  Docs: Add "Gather info" section to REPORTING-BUGS.
  Docs: Step-by-step directions for reporting bugs.
  Trivial: docs: Remove six-space indentation in REPORTING-BUGS.

12 years agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 2 May 2013 21:43:58 +0000 (14:43 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Peter Anvin:
 "Two regression fixes:

  1. On 64 bits, we would set NX on non-NX-capable hardware (very rare
     in 64-bit land, but a nonzero subset.)

  2. Fix suspend/resume across kernel versions"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86-64, init: Do not set NX bits on non-NX capable hardware
  x86, gdt, hibernate: Store/load GDT for hibernate path.

12 years agoMerge tag 'pinctrl-fixes-v3.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 2 May 2013 21:42:46 +0000 (14:42 -0700)]
Merge tag 'pinctrl-fixes-v3.10-1' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pinctrl fixes from Linus Walleij:
 "Two fixes to the pinctrl subsystem for v3.10:
   - A quite apparent mutex fix in an untested codepath
   - A compile warning fix in the plgpio driver"

* tag 'pinctrl-fixes-v3.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: fix mutex deadlock in get_pinctrl_dev_from_of_node()
  pinctrl: plgpio: add CONFIG_PM_SLEEP to suspend/resume functions

12 years agoMerge branch 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa...
Linus Torvalds [Thu, 2 May 2013 21:38:53 +0000 (14:38 -0700)]
Merge branch 'i2c/for-next' of git://git./linux/kernel/git/wsa/linux

Pull i2c changes from Wolfram Sang:

 - an arbitration driver.  While the driver is quite simple, it caused
   discussion if we need additional arbitration on top of the one
   specified in the I2C standard.  Conclusion is that I accept a few
   generic mechanisms, but not very specific ones.

 - the core lost the detach_adapter() call.  It has no users anymore and
   was in the way for other cleanups.  attach_adapter() is sadly still
   there since there are users waiting to be converted.

 - the core gained a bus recovery infrastructure.  I2C defines a way to
   recover if the data line is stalled.  This mechanism is now in the
   core and drivers can now pass some data to make use of it.

 - bigger driver cleanups for designware, s3c2410

 - removing superfluous refcounting from drivers

 - removing Ben Dooks as second maintainer due to inactivity.  Thanks
   for all your work so far, Ben!

 - bugfixes, feature additions, devicetree fixups, simplifications...

* 'i2c/for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux: (38 commits)
  i2c: xiic: must always write 16-bit words to TX_FIFO
  i2c: octeon: use HZ in timeout value
  i2c: octeon: Fix i2c fail problem when a process is terminated by a signal
  i2c: designware-pci: drop superfluous {get|put}_device
  i2c: designware-plat: drop superfluous {get|put}_device
  i2c: davinci: drop superfluous {get|put}_device
  MAINTAINERS: Ben Dooks is inactive regarding I2C
  i2c: mux: Add i2c-arb-gpio-challenge 'mux' driver
  i2c: at91: convert to dma_request_slave_channel_compat()
  i2c: mxs: do error checking and handling in PIO mode
  i2c: mxs: remove races in PIO code
  i2c-designware: switch to use runtime PM autosuspend
  i2c-designware: use usleep_range() in the busy-loop
  i2c-designware: enable/disable the controller properly
  i2c-designware: use dynamic adapter numbering on Lynxpoint
  i2c-designware-pci: use managed functions pcim_* and devm_*
  i2c-designware-pci: use dev_err() instead of printk()
  i2c-designware: move to managed functions (devm_*)
  i2c: remove CONFIG_HOTPLUG ifdefs
  i2c: s3c2410: Add SMBus emulation for block read
  ...

12 years agoMerge tag 'virtio-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 2 May 2013 21:14:04 +0000 (14:14 -0700)]
Merge tag 'virtio-next-for-linus' of git://git./linux/kernel/git/rusty/linux

Pull virtio & lguest updates from Rusty Russell:
 "Lots of virtio work which wasn't quite ready for last merge window.

  Plus I dived into lguest again, reworking the pagetable code so we can
  move the switcher page: our fixmaps sometimes take more than 2MB now..."

Ugh.  Annoying conflicts with the tcm_vhost -> vhost_scsi rename.
Hopefully correctly resolved.

* tag 'virtio-next-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rusty/linux: (57 commits)
  caif_virtio: Remove bouncing email addresses
  lguest: improve code readability in lg_cpu_start.
  virtio-net: fill only rx queues which are being used
  lguest: map Switcher below fixmap.
  lguest: cache last cpu we ran on.
  lguest: map Switcher text whenever we allocate a new pagetable.
  lguest: don't share Switcher PTE pages between guests.
  lguest: expost switcher_pages array (as lg_switcher_pages).
  lguest: extract shadow PTE walking / allocating.
  lguest: make check_gpte et. al return bool.
  lguest: assume Switcher text is a single page.
  lguest: rename switcher_page to switcher_pages.
  lguest: remove RESERVE_MEM constant.
  lguest: check vaddr not pgd for Switcher protection.
  lguest: prepare to make SWITCHER_ADDR a variable.
  virtio: console: replace EMFILE with EBUSY for already-open port
  virtio-scsi: reset virtqueue affinity when doing cpu hotplug
  virtio-scsi: introduce multiqueue support
  virtio-scsi: push vq lock/unlock into virtscsi_vq_done
  virtio-scsi: pass struct virtio_scsi to virtqueue completion function
  ...

12 years agoMerge tag 'vfio-for-v3.10' of git://github.com/awilliam/linux-vfio
Linus Torvalds [Thu, 2 May 2013 21:02:32 +0000 (14:02 -0700)]
Merge tag 'vfio-for-v3.10' of git://github.com/awilliam/linux-vfio

Pull vfio updates from Alex Williamson:
 "Changes include extension to support PCI AER notification to
  userspace, byte granularity of PCI config space and access to
  unarchitected PCI config space, better protection around IOMMU driver
  accesses, default file mode fix, and a few misc cleanups."

* tag 'vfio-for-v3.10' of git://github.com/awilliam/linux-vfio:
  vfio: Set container device mode
  vfio: Use down_reads to protect iommu disconnects
  vfio: Convert container->group_lock to rwsem
  PCI/VFIO: use pcie_flags_reg instead of access PCI-E Capabilities Register
  vfio-pci: Enable raw access to unassigned config space
  vfio-pci: Use byte granularity in config map
  vfio: make local function vfio_pci_intx_unmask_handler() static
  VFIO-AER: Vfio-pci driver changes for supporting AER
  VFIO: Wrapper for getting reference to vfio_device

12 years agodrivers: net: cpsw: irq not disabled in cpsw isr in particular sequence
Mugunthan V N [Thu, 2 May 2013 01:52:11 +0000 (01:52 +0000)]
drivers: net: cpsw: irq not disabled in cpsw isr in particular sequence

In CPSW NAPI, after processing all interrupts IRQ is enabled and then book
keeping irq_enabled is updated. In random cases when a packet is transmitted
or received between processing packets and IRQ enabled, then just after
enabled IRQ and before irq_enabled is updated, ISR is called so IRQs are
not disabled as irq_enabled is still false and CPU gets locked in CPSW ISR.

By changing the sequence as update the irq_enabled and then enable IRQ
fixes the issue. This issue is not captured always as it is a timing issue
whether Tx or Rx IRQ is invoked between packet processing and enable IRQ.

Cc: Sebastian Siewior <bigeasy@linutronix.de>
Signed-off-by: Mugunthan V N <mugunthanvnm@ti.com>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoxen-netback: better names for thresholds
Wei Liu [Thu, 2 May 2013 00:43:59 +0000 (00:43 +0000)]
xen-netback: better names for thresholds

This patch only changes some names to avoid confusion.

In this patch we have:

  MAX_SKB_SLOTS_DEFAULT -> FATAL_SKB_SLOTS_DEFAULT
  max_skb_slots -> fatal_skb_slots
  #define XEN_NETBK_LEGACY_SLOTS_MAX XEN_NETIF_NR_SLOTS_MIN

The fatal_skb_slots is the threshold to determine whether a packet is
malicious.

XEN_NETBK_LEGACY_SLOTS_MAX is the maximum slots a valid packet can have at
this point. It is defined to be XEN_NETIF_NR_SLOTS_MIN because that's
guaranteed to be supported by all backends.

Suggested-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoxen-netback: avoid allocating variable size array on stack
Wei Liu [Thu, 2 May 2013 00:43:58 +0000 (00:43 +0000)]
xen-netback: avoid allocating variable size array on stack

Tune xen_netbk_count_requests to not touch working array beyond limit, so that
we can make working array size constant.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoxen-netback: remove redundent parameter in netbk_count_requests
Wei Liu [Thu, 2 May 2013 00:43:57 +0000 (00:43 +0000)]
xen-netback: remove redundent parameter in netbk_count_requests

Tracking down from the caller, first_idx is always equal to vif->tx.req_cons.
Remove it to avoid confusion.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Wei Liu <wei.liu2@citrix.com>
Acked-by: Ian Campbell <ian.campbell@citrix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoARM: SPEAr: conditionalize SMP code
Arnd Bergmann [Tue, 30 Apr 2013 12:53:19 +0000 (14:53 +0200)]
ARM: SPEAr: conditionalize SMP code

Some constant definitions are only defined for spear13xx, so
we must not attempt to build SPEAr SMP support when that
SoC is not enabled.

arch/arm/mach-spear/platsmp.c:25:35:
 error: 'VA_SCU_BASE' undeclared here (not in a function)
 arch/arm/mach-spear/platsmp.c: In function 'spear13xx_smp_prepare_cpus':
 arch/arm/mach-spear/platsmp.c:111:58: error: 'SYS_LOCATION' undeclared (first use in this function)

Cc: Viresh Kumar <viresh.linux@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
12 years agodrm: don't check modeset locks in panic handler
Daniel Vetter [Thu, 2 May 2013 07:43:05 +0000 (09:43 +0200)]
drm: don't check modeset locks in panic handler

Since we know that locking is broken in that case and it's more
important to not flood the dmesg with random gunk.

Cc: Dave Airlie <airlied@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
References: http://lkml.kernel.org/r/20130502000206.GH15623@pd.tnic
Cc: stable@vger.kernel.org
Reported-and-tested-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
12 years agobe2net: Fix to fail probe if MSI-X enable fails for a VF
Somnath Kotur [Thu, 2 May 2013 03:37:08 +0000 (03:37 +0000)]
be2net: Fix to fail probe if MSI-X enable fails for a VF

As per SPEC, INTx mode is not supported on VFs. So if enable_msix fails,
then just fail probe. Also bail out of be_open if irq_register fails.

Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: avoid napi_disable() when it has not been enabled
Somnath Kotur [Thu, 2 May 2013 03:36:55 +0000 (03:36 +0000)]
be2net: avoid napi_disable() when it has not been enabled

When RQ creation fails in be_open(), driver jumps to be_close() where
napi_disable() is done without a prior napi_enable(); leading to a hang.
This change is needed as there is no way to see if napi is enable/disabled.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: Fix firmware download for Lancer
Somnath Kotur [Thu, 2 May 2013 03:36:34 +0000 (03:36 +0000)]
be2net: Fix firmware download for Lancer

Increasing the timeout value of write_object command to 60 seconds as
30 second timeout was found to be not enough for the command to complete.

Signed-off-by: Kalesh AP <kalesh.purayil@emulex.com>
Signed-off-by: Somnath Kotur <somnath.kotur@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: Fix to receive Multicast Packets when Promiscuous mode is enabled on certain...
Ajit Khaparde [Wed, 1 May 2013 09:38:24 +0000 (09:38 +0000)]
be2net: Fix to receive Multicast Packets when Promiscuous mode is enabled on certain devices

When a user requests Promiscuous mode on SkyHawk-R devices,
the FW enables Desparate Promiscuous mode.
Due to this, we need to explicitly enable Multicast Promiscuous mode
so that Multicast packets are received as well.

This patch fixes that.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: Fix to show tx priority pause counter in ethtool -S
Ajit Khaparde [Wed, 1 May 2013 09:38:00 +0000 (09:38 +0000)]
be2net: Fix to show tx priority pause counter in ethtool -S

The tx priority pause count is available with the driver.
But we were not showing it.
This change should fix it and display it in ethtool stats.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: Fix to use 32-bit stats to report rx_drops_no_fragment
Ajit Khaparde [Wed, 1 May 2013 09:37:27 +0000 (09:37 +0000)]
be2net: Fix to use 32-bit stats to report rx_drops_no_fragment

Only BE devices provide 16-bit counter for rx_drops_no_fragment.
All other devices provide a 32-bit counter for this stat.
Use the 32-bit value where available.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agobe2net: Fix to use version 2 of cq_create for SkyHawk-R devices
Ajit Khaparde [Wed, 1 May 2013 09:37:17 +0000 (09:37 +0000)]
be2net: Fix to use version 2 of cq_create for SkyHawk-R devices

SkyHawk-R devices should use v2 of cq_create command.
BE3/BE2 devices can use version 0 of the command to create a CQ.

Signed-off-by: Ajit Khaparde <ajit.khaparde@emulex.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agodrm/nouveau: fix build with nv50->nvc0
Dave Airlie [Thu, 2 May 2013 20:40:37 +0000 (06:40 +1000)]
drm/nouveau: fix build with nv50->nvc0

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoMerge branch 'cleanup' into for-linus
Russell King [Thu, 2 May 2013 20:31:29 +0000 (21:31 +0100)]
Merge branch 'cleanup' into for-linus

Conflicts:
arch/arm/plat-omap/dmtimer.c

12 years agoMerge branches 'devel-stable', 'entry', 'fixes', 'mach-types', 'misc' and 'smp-hotplu...
Russell King [Thu, 2 May 2013 20:30:36 +0000 (21:30 +0100)]
Merge branches 'devel-stable', 'entry', 'fixes', 'mach-types', 'misc' and 'smp-hotplug' into for-linus

12 years agoMerge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Linus Torvalds [Thu, 2 May 2013 20:29:14 +0000 (13:29 -0700)]
Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull vhost cleanups and fixes from Michael Tsirkin:
 "Here are vhost cleanups and fixes by Asias He and myself.  They affect
  both vhost-net and vhost-scsi devices.  They also *depend* on both
  net-next and target-pending, where the net and target commits these
  changes depend on are already merged.  So merging through the common
  vhost tree."

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  vhost_scsi: module rename
  tcm_vhost: header split up
  vhost: src file renames
  vhost: fix error handling in RESET_OWNER ioctl
  tcm_vhost: remove virtio-net.h dependency
  vhost: move per-vq net specific fields out to net
  tcm_vhost: document inflight ref-counting use
  vhost: move vhost-net zerocopy fields to net.c
  tcm_vhost: Wait for pending requests in vhost_scsi_flush()
  vhost: Allow device specific fields per vq

12 years agoMerge commit '73053d973' into smp-hotplug
Russell King [Thu, 2 May 2013 20:20:46 +0000 (21:20 +0100)]
Merge commit '73053d973' into smp-hotplug

This is to fix a merge problem with mach-highbank/hotplug.c, which git
silently resolves, but wrongly.  This commit contains the correct
resolution.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
12 years agotile: support new Tilera hypervisor
Chris Metcalf [Thu, 2 May 2013 19:29:04 +0000 (15:29 -0400)]
tile: support new Tilera hypervisor

The Tilera hypervisor shipped in releases up through MDE 4.1 launches
the client operating system (i.e. Linux) at privilege level 1 (PL1).
Starting with MDE 4.2, as part of the work to enable KVM, the
Tilera hypervisor launches Linux at PL2 instead.

This commit makes the KERNEL_PL option default to 2 for tilegx, while
still saying at 1 for tilepro, which doesn't have an updated hypervisor.
It also explains how and when you might want to choose another value.
In addition, we change a small buglet in the on-chip Ethernet driver,
where we were failing to use the KERNEL_PL constant in an API call.

To make the transition cleaner, this change also provides the updated
hv_init() API for the new hypervisor that supports announcing Linux's
compiled-in PL, so the hypervisor can generate a suitable error in the
case of a mismatched hypervisor and Linux binary.

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Cc: stable@vger.linux.org
12 years ago[SCSI] lpfc 8.3.39: Update lpfc version for 8.3.39 driver release
James Smart [Thu, 18 Apr 2013 00:20:10 +0000 (20:20 -0400)]
[SCSI] lpfc 8.3.39: Update lpfc version for 8.3.39 driver release

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed driver handling of CLEAR_LA with NPIV enabled causing SID...
James Smart [Thu, 18 Apr 2013 00:20:00 +0000 (20:20 -0400)]
[SCSI] lpfc 8.3.39: Fixed driver handling of CLEAR_LA with NPIV enabled causing SID=0 frames out

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Reduced tmo value set to FLOGI WQE for quick recovery from FLOGI...
James Smart [Thu, 18 Apr 2013 00:19:44 +0000 (20:19 -0400)]
[SCSI] lpfc 8.3.39: Reduced tmo value set to FLOGI WQE for quick recovery from FLOGI sequence timeout

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Add log message when completes with clean address bit set to...
James Smart [Thu, 18 Apr 2013 00:19:25 +0000 (20:19 -0400)]
[SCSI] lpfc 8.3.39: Add log message when completes with clean address bit set to zero

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed driver vector mapping to CPU affinity
James Smart [Thu, 18 Apr 2013 00:19:16 +0000 (20:19 -0400)]
[SCSI] lpfc 8.3.39: Fixed driver vector mapping to CPU affinity

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed iocb flags not being reset for scsi commands
James Smart [Thu, 18 Apr 2013 00:19:00 +0000 (20:19 -0400)]
[SCSI] lpfc 8.3.39: Fixed iocb flags not being reset for scsi commands

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed system panic during EEH recovery due to midlayer acting...
James Smart [Thu, 18 Apr 2013 00:18:47 +0000 (20:18 -0400)]
[SCSI] lpfc 8.3.39: Fixed system panic during EEH recovery due to midlayer acting on outstanding I/O

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed not returning FAILED status when SCSI invoking host reset...
James Smart [Thu, 18 Apr 2013 00:18:39 +0000 (20:18 -0400)]
[SCSI] lpfc 8.3.39: Fixed not returning FAILED status when SCSI invoking host reset handler failed

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed bad book keeping in posting els sgls to port
James Smart [Thu, 18 Apr 2013 00:18:29 +0000 (20:18 -0400)]
[SCSI] lpfc 8.3.39: Fixed bad book keeping in posting els sgls to port

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed deadlock between hbalock and nlp_lock use
James Smart [Thu, 18 Apr 2013 00:18:19 +0000 (20:18 -0400)]
[SCSI] lpfc 8.3.39: Fixed deadlock between hbalock and nlp_lock use

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed BlockGuard to take advantage of rdprotect/wrprotect info...
James Smart [Thu, 18 Apr 2013 00:18:07 +0000 (20:18 -0400)]
[SCSI] lpfc 8.3.39: Fixed BlockGuard to take advantage of rdprotect/wrprotect info when available

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Reduced spinlock contention on SCSI buffer list
James Smart [Thu, 18 Apr 2013 00:17:40 +0000 (20:17 -0400)]
[SCSI] lpfc 8.3.39: Reduced spinlock contention on SCSI buffer list

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed crash when processing bsg's sg list with high memory pages
James Smart [Thu, 18 Apr 2013 00:17:26 +0000 (20:17 -0400)]
[SCSI] lpfc 8.3.39: Fixed crash when processing bsg's sg list with high memory pages

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Remove lpfc_fcp_look_ahead module parameter
James Smart [Thu, 18 Apr 2013 00:17:13 +0000 (20:17 -0400)]
[SCSI] lpfc 8.3.39: Remove lpfc_fcp_look_ahead module parameter

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fix driver issues with SCSI Host reset
James Smart [Thu, 18 Apr 2013 00:16:51 +0000 (20:16 -0400)]
[SCSI] lpfc 8.3.39: Fix driver issues with SCSI Host reset

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Doorbell formation information logged in dual-chute mode WQ and...
James Smart [Thu, 18 Apr 2013 00:16:37 +0000 (20:16 -0400)]
[SCSI] lpfc 8.3.39: Doorbell formation information logged in dual-chute mode WQ and RQ setup

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years agoARM: imx: reset_controller may be disabled
Arnd Bergmann [Tue, 30 Apr 2013 12:58:31 +0000 (14:58 +0200)]
ARM: imx: reset_controller may be disabled

The new reset controller API is optional, so if that is disabled,
we must not call it from platform code.

arch/arm/mach-imx/built-in.o: In function
 `imx_src_init': /git/arm-soc/arch/arm/mach-imx/src.c:144:
undefined reference to `reset_controller_register'

Cc: Sascha Hauer <kernel@pengutronix.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
12 years ago[SCSI] lpfc 8.3.39: Fix driver issues with large s/g lists for BlockGuard
James Smart [Thu, 18 Apr 2013 00:16:15 +0000 (20:16 -0400)]
[SCSI] lpfc 8.3.39: Fix driver issues with large s/g lists for BlockGuard

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fix driver issues with large lpfc_sg_seg_cnt values
James Smart [Thu, 18 Apr 2013 00:16:05 +0000 (20:16 -0400)]
[SCSI] lpfc 8.3.39: Fix driver issues with large lpfc_sg_seg_cnt values

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed pt2pt and loop discovery problems on topology changes.
James Smart [Thu, 18 Apr 2013 00:15:19 +0000 (20:15 -0400)]
[SCSI] lpfc 8.3.39: Fixed pt2pt and loop discovery problems on topology changes.

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Remove driver dependency on HZ
James Smart [Thu, 18 Apr 2013 00:14:58 +0000 (20:14 -0400)]
[SCSI] lpfc 8.3.39: Remove driver dependency on HZ

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed BlockGuard error reporting
James Smart [Thu, 18 Apr 2013 00:14:49 +0000 (20:14 -0400)]
[SCSI] lpfc 8.3.39: Fixed BlockGuard error reporting

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] lpfc 8.3.39: Fixed VPI allocation issues after firmware dump is performed
James Smart [Thu, 18 Apr 2013 00:14:38 +0000 (20:14 -0400)]
[SCSI] lpfc 8.3.39: Fixed VPI allocation issues after firmware dump is performed

Signed-off-by: James Smart <james.smart@emulex.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years agoNVMe: Schedule timeout for sync commands
Keith Busch [Fri, 19 Apr 2013 20:11:06 +0000 (14:11 -0600)]
NVMe: Schedule timeout for sync commands

Schedule a timeout on sync commands in case the command times out and
the device is not being polled for timeouts. This prevents device removal
from hanging forever if the device has stopped responding.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Meta-data support in NVME_IOCTL_SUBMIT_IO
Keith Busch [Tue, 23 Apr 2013 23:23:59 +0000 (17:23 -0600)]
NVMe: Meta-data support in NVME_IOCTL_SUBMIT_IO

This adds support for namespaces with separate meta-data formats in the
submit io ioctl. The meta-data buffer has to be a contiguous, so such
a buffer is allocated and the mapped user pages are copied to/from this
buffer for write/read commands.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years ago[SCSI] ipr: Need to reset adapter after the 6th EEH error
wenxiong@linux.vnet.ibm.com [Wed, 17 Apr 2013 14:34:06 +0000 (09:34 -0500)]
[SCSI] ipr: Need to reset adapter after the 6th EEH error

Add reset adapter after the 6th EEH errors in ipr driver. This triggers
the adapter reset via the PCI config space even when the slot is frozen.

Signed-off-by: Wen Xiong <wenxiong@linux.vnet.ibm.com>
Acked-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla4xxx: Update driver version to 5.03.00-k9
Vikas Chaudhary [Wed, 17 Apr 2013 09:15:30 +0000 (05:15 -0400)]
[SCSI] qla4xxx: Update driver version to 5.03.00-k9

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla4xxx: Assign values using correct datatype
Adheer Chandravanshi [Wed, 17 Apr 2013 09:15:29 +0000 (05:15 -0400)]
[SCSI] qla4xxx: Assign values using correct datatype

Assign values using correct datatype in function qla4xxx_copy_to_fwddb_param()

Signed-off-by: Adheer Chandravanshi <adheer.chandravanshi@qlogic.com>
Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla4xxx: Fix smatch warnings
Adheer Chandravanshi [Wed, 17 Apr 2013 09:15:28 +0000 (05:15 -0400)]
[SCSI] qla4xxx: Fix smatch warnings

Fix following smatch warnings:-
drivers/scsi/qla4xxx/ql4_os.c:6573
qla4xxx_sysfs_ddb_set_param() warn: possible memory leak of 'fw_ddb_entry'
drivers/scsi/qla4xxx/ql4_os.c:6596
qla4xxx_sysfs_ddb_delete() warn: variable dereferenced before check 'fnode_sess'
(see line 6584)
drivers/scsi/qla4xxx/ql4_os.c:6632
qla4xxx_sysfs_ddb_delete() error: potential NULL dereference 'fw_ddb_entry'.

Signed-off-by: Adheer Chandravanshi <adheer.chandravanshi@qlogic.com>
Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla4xxx: Fix sparse warning for qla4xxx_sysfs_ddb_tgt_create
Vikas Chaudhary [Wed, 17 Apr 2013 09:15:27 +0000 (05:15 -0400)]
[SCSI] qla4xxx: Fix sparse warning for qla4xxx_sysfs_ddb_tgt_create

Fix following warning:
drivers/scsi/qla4xxx/ql4_os.c:5507:5:
warning: symbol 'qla4xxx_sysfs_ddb_tgt_create' was not declared. Should it be
static?

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] qla4xxx: Silence gcc warning
Vikas Chaudhary [Wed, 17 Apr 2013 09:15:26 +0000 (05:15 -0400)]
[SCSI] qla4xxx: Silence gcc warning

Fix followig gcc warning:-
drivers/scsi/qla4xxx/ql4_os.c: In function ‘qla4xxx_sysfs_ddb_get_param’:
drivers/scsi/qla4xxx/ql4_os.c:6279:
warning: comparison is always true due to limited range of data type
drivers/scsi/qla4xxx/ql4_os.c:6290:
warning: comparison is always true due to limited range of data type
drivers/scsi/qla4xxx/ql4_os.c: In function ‘qla4xxx_sysfs_ddb_delete’:
drivers/scsi/qla4xxx/ql4_os.c:6593:
warning: ‘ddb_size’ may be used uninitialized in this function

Signed-off-by: Vikas Chaudhary <vikas.chaudhary@qlogic.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] libosd: remover duplicate __bitwise annotation
Dan Carpenter [Wed, 17 Apr 2013 07:19:05 +0000 (10:19 +0300)]
[SCSI] libosd: remover duplicate __bitwise annotation

__be32 is already a __bitwise type so we don't need the second __bitwise
here.  It causes a Sparse error:
include/scsi/osd_protocol.h:110:26: error: invalid modifier

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] megaraid_sas: release lock on error path
Dan Carpenter [Tue, 16 Apr 2013 07:44:19 +0000 (10:44 +0300)]
[SCSI] megaraid_sas: release lock on error path

We should unlock here before returning.

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Adam Radford <aradford@gmail.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ibmvfc: Driver version 1.0.11
Brian King [Fri, 12 Apr 2013 13:25:19 +0000 (08:25 -0500)]
[SCSI] ibmvfc: Driver version 1.0.11

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ibmvfc: Suppress ABTS if target gone
Brian King [Fri, 12 Apr 2013 13:25:18 +0000 (08:25 -0500)]
[SCSI] ibmvfc: Suppress ABTS if target gone

Adds support for a new VIOS feature that allows ibmvfc to
optimize terminate_rport_io by telling the VIOS the target
is no longer accessible on the fabric and that it should
not send an ABTS out on the fabric to the device.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ibmvfc: Send cancel when link is down
Brian King [Fri, 12 Apr 2013 13:25:17 +0000 (08:25 -0500)]
[SCSI] ibmvfc: Send cancel when link is down

If attempting to abort requests due to a fail fail timeout
or error handling while the link is down, we cannot send
an abort out on the fabric. We can, however, send a cancel
to the VIOS. This fixes ibmvfc to send a cancel in this
case to prevent error handling from failing and/or
escalating.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years ago[SCSI] ibmvfc: Support FAST_IO_FAIL in EH handlers
Brian King [Fri, 12 Apr 2013 13:25:16 +0000 (08:25 -0500)]
[SCSI] ibmvfc: Support FAST_IO_FAIL in EH handlers

Adds support for receiving FAST_IO_FAIL from fc_block_scsi_eh
when in error recovery. This fixes cases of devices being
taken offline when they are no longer accessible on the fabric,
preventing them from coming back online when the fabric recovers.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years agoARM: EXYNOS: remove unnecessary use of IS_ERR_VALUE()
Russell King [Wed, 13 Mar 2013 20:50:02 +0000 (20:50 +0000)]
ARM: EXYNOS: remove unnecessary use of IS_ERR_VALUE()

s5p_register_gpio_interrupt() returns 0 or positive for success, and
-ve for errors, so just use the standard >= 0 test.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
12 years agoARM: IMX: remove unnecessary use of IS_ERR_VALUE()
Russell King [Wed, 13 Mar 2013 20:49:10 +0000 (20:49 +0000)]
ARM: IMX: remove unnecessary use of IS_ERR_VALUE()

device_register() returns -ve values for errors, and zero for success.
There's no need to obfuscate the code with IS_ERR_VALUE().

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
12 years agoARM: OMAP: use consistent error checking
Russell King [Wed, 13 Mar 2013 20:44:21 +0000 (20:44 +0000)]
ARM: OMAP: use consistent error checking

Consistently check errors using the usual method used in the kernel
for much of its history.  For instance:

int gpmc_cs_set_timings(int cs, const struct gpmc_timings *t)
{
int div;
div = gpmc_calc_divider(t->sync_clk);
if (div < 0)
return div;
static int gpmc_set_async_mode(int cs, struct gpmc_timings *t)
{
...
return gpmc_cs_set_timings(cs, t);

.....
ret = gpmc_set_async_mode(gpmc_onenand_data->cs, &t);
if (IS_ERR_VALUE(ret))
return ret;

So, gpmc_cs_set_timings() thinks any negative return value is an error,
but where we check that in higher levels, only a limited range are
errors...

There is only _one_ use of IS_ERR_VALUE() in arch/arm which is really
appropriate, and that is in arch/arm/include/asm/syscall.h:

static inline long syscall_get_error(struct task_struct *task,
     struct pt_regs *regs)
{
unsigned long error = regs->ARM_r0;
return IS_ERR_VALUE(error) ? error : 0;
}

because this function really does have to differentiate between error
return values and addresses which look like negative numbers (eg, from
mmap()).

So, here's a patch to remove them from OMAP, except for the above.

Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
12 years agoARM: cleanup: OMAP hwmod error checking
Russell King [Sun, 24 Feb 2013 10:56:59 +0000 (10:56 +0000)]
ARM: cleanup: OMAP hwmod error checking

omap_hwmod_lookup() only returns NULL on error, never an error pointer.
Checking the returned pointer using IS_ERR_OR_NULL() is needless
overhead.  Use a simple !ptr check instead.

OMAP devices (oh->od) always have a valid platform device attached (see
omap_device_alloc()) so there's no point validating the platform device
pointer (we will have already oopsed long before if this is not the
case here.)

Lastly, oh->od is only ever NULL or a valid omap device pointer - 'oh'
comes from the statically declared hwmod tables, and the pointer is
only filled in by omap_device_alloc() at a point where the omap device
pointer must be valid.

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
12 years ago[SCSI] ibmvfc: Properly set cancel flags when cancelling abort
Brian King [Fri, 12 Apr 2013 13:25:15 +0000 (08:25 -0500)]
[SCSI] ibmvfc: Properly set cancel flags when cancelling abort

The flags on a cancel operation are intended to indicate what,
if any, TMF will follow the cancel request. This fixes a case
where we were incorrectly setting the abort task set flag on
the cancel flag when we were cancelling an abort task set.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Acked-by: Robert Jennings <rcj@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <JBottomley@Parallels.com>
12 years agoNVMe: Device specific stripe size handling
Keith Busch [Tue, 9 Apr 2013 23:13:20 +0000 (17:13 -0600)]
NVMe: Device specific stripe size handling

We have an nvme device that has a concept of a stripe size. IO requests
that do not transfer data crossing a stripe boundary has greater
performance compared to IO that does cross it. This patch sets the
stripe size for the device if the device and vendor ids match one with
this feature and splits IO requests that cross the stripe boundary.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Split non-mergeable bio requests
Keith Busch [Tue, 9 Apr 2013 17:59:32 +0000 (11:59 -0600)]
NVMe: Split non-mergeable bio requests

It is possible a bio request can not be submitted as a single NVMe IO
command if the bio_vec is not mergeable with the NVMe PRP alignement
constraints. This condition was handled by submitting an IO for the
mergeable portion then submitting a follow on IO for the remaining data
after the previous IO completes. The remainder to be sent was tracked
by manipulating the bio->bi_idx and bio->bi_sector. This patch splits
the request as many times as necessary and submits the bios together.

Since submitting the bio may cause it to be requeued on split,
nvme_resubmit_bios had to be modified to remove the wait queue when
the bio list is empty prior to submitting the bio since a split would
have added the wait queue a second time, corrupting the wait queue head
task list.

There are a few other benefits from doing this: it fixes a potential
issue with the previous handling of a non-mergeable bio as the requeuing
method could would use an unlocked nvme_queue if the callback isn't
invoked on the queue's associated cpu; it will be possible to retry a
failed bio if desired at some later time since it does not manipulate
the original bio; the bio integrity extensions require the bio to be in
its original condition for the checks to work correctly if we implement
the end-to-end data protection in the future.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Remove dead code in nvme_dev_add
Keith Busch [Wed, 1 May 2013 19:07:49 +0000 (13:07 -0600)]
NVMe: Remove dead code in nvme_dev_add

There is no situation that could occur where we could error out of this
function and require cleaning up allocated namespaces.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Check for NULL memory in nvme_dev_add
Keith Busch [Wed, 1 May 2013 19:07:48 +0000 (13:07 -0600)]
NVMe: Check for NULL memory in nvme_dev_add

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Fix error clean-up on nvme_alloc_queue
Keith Busch [Wed, 1 May 2013 19:07:47 +0000 (13:07 -0600)]
NVMe: Fix error clean-up on nvme_alloc_queue

The nvme_queue's depth is not set if we fail to allocate submission queue
entries, which was being used to determine how much coherent memory to
free on error. Use the depth variable instead.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Free admin queue on request_irq error
Keith Busch [Wed, 1 May 2013 19:07:51 +0000 (13:07 -0600)]
NVMe: Free admin queue on request_irq error

Fixes a potential memory leak if requesting the admin queue irq fails.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: Add scsi unmap to SG_IO
Keith Busch [Wed, 24 Apr 2013 21:44:24 +0000 (15:44 -0600)]
NVMe: Add scsi unmap to SG_IO

Translates a scsi unmap request from SG_IO ioctl to NVMe
data-set-management deallocate.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Acked-by: Vishal Verma <vishal.l.verma@linux.intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agoNVMe: queue usage fixes in nvme-scsi
Keith Busch [Thu, 25 Apr 2013 20:39:27 +0000 (14:39 -0600)]
NVMe: queue usage fixes in nvme-scsi

Fixes nvme queue usages in scsi-to-nvme translation code to not get
a queue more often than it is being put, and not use the queue in an
unsafe way without it being locked.

Signed-off-by: Keith Busch <keith.busch@intel.com>
Acked-by: Vishal Verma <vishal.l.verma@linux.intel.com>
Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
12 years agox86-64, init: Do not set NX bits on non-NX capable hardware
H. Peter Anvin [Thu, 2 May 2013 17:33:46 +0000 (10:33 -0700)]
x86-64, init: Do not set NX bits on non-NX capable hardware

During early init, we would incorrectly set the NX bit even if the NX
feature was not supported.  Instead, only set this bit if NX is
actually available and enabled.  We already do very early detection of
the NX bit to enable it in EFER, this simply extends this detection to
the early page table mask.

Reported-by: Fernando Luis Vázquez Cao <fernando@oss.ntt.co.jp>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Link: http://lkml.kernel.org/r/1367476850.5660.2.camel@nexus
Cc: <stable@vger.kernel.org> v3.9
12 years agox86, gdt, hibernate: Store/load GDT for hibernate path.
Konrad Rzeszutek Wilk [Thu, 2 May 2013 01:53:30 +0000 (21:53 -0400)]
x86, gdt, hibernate: Store/load GDT for hibernate path.

The git commite7a5cd063c7b4c58417f674821d63f5eb6747e37
("x86-64, gdt: Store/load GDT for ACPI S3 or hibernate/resume path
is not needed.") assumes that for the hibernate path the booting
kernel and the resuming kernel MUST be the same. That is certainly
the case for a 32-bit kernel (see check_image_kernel and
CONFIG_ARCH_HIBERNATION_HEADER config option).

However for 64-bit kernels it is OK to have a different kernel
version (and size of the image) of the booting and resuming kernels.
Hence the above mentioned git commit introduces an regression.

This patch fixes it by introducing a 'struct desc_ptr gdt_desc'
back in the 'struct saved_context'. However instead of having in the
'save_processor_state' and 'restore_processor_state' the
store/load_gdt calls, we are only saving the GDT in the
save_processor_state.

For the restore path the lgdt operation is done in
hibernate_asm_[32|64].S in the 'restore_registers' path.

The apt reader of this description will recognize that only 64-bit
kernels need this treatment, not 32-bit. This patch adds the logic
in the 32-bit path to be more similar to 64-bit so that in the future
the unification process can take advantage of this.

[ hpa: this also reverts an inadvertent on-disk format change ]

Suggested-by: "H. Peter Anvin" <hpa@zytor.com>
Acked-by: "Rafael J. Wysocki" <rjw@sisk.pl>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Link: http://lkml.kernel.org/r/1367459610-9656-2-git-send-email-konrad.wilk@oracle.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
12 years agonet: vlan,ethtool: netdev_features_t is more than 32 bit
Bjørn Mork [Wed, 1 May 2013 23:06:42 +0000 (23:06 +0000)]
net: vlan,ethtool: netdev_features_t is more than 32 bit

Signed-off-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agonet: use netdev_features_t in skb_needs_linearize()
Patrick McHardy [Wed, 1 May 2013 22:36:49 +0000 (22:36 +0000)]
net: use netdev_features_t in skb_needs_linearize()

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoMerge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Thu, 2 May 2013 17:16:16 +0000 (10:16 -0700)]
Merge branch 'next' of git://git./linux/kernel/git/benh/powerpc

Pull powerpc update from Benjamin Herrenschmidt:
 "The main highlights this time around are:

   - A pile of addition POWER8 bits and nits, such as updated
     performance counter support (Michael Ellerman), new branch history
     buffer support (Anshuman Khandual), base support for the new PCI
     host bridge when not using the hypervisor (Gavin Shan) and other
     random related bits and fixes from various contributors.

   - Some rework of our page table format by Aneesh Kumar which fixes a
     thing or two and paves the way for THP support.  THP itself will
     not make it this time around however.

   - More Freescale updates, including Altivec support on the new e6500
     cores, new PCI controller support, and a pile of new boards support
     and updates.

   - The usual batch of trivial cleanups & fixes"

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (156 commits)
  powerpc: Fix build error for book3e
  powerpc: Context switch the new EBB SPRs
  powerpc: Turn on the EBB H/FSCR bits
  powerpc: Replace CPU_FTR_BCTAR with CPU_FTR_ARCH_207S
  powerpc: Setup BHRB instructions facility in HFSCR for POWER8
  powerpc: Fix interrupt range check on debug exception
  powerpc: Update tlbie/tlbiel as per ISA doc
  powerpc: Print page size info during boot
  powerpc: print both base and actual page size on hash failure
  powerpc: Fix hpte_decode to use the correct decoding for page sizes
  powerpc: Decode the pte-lp-encoding bits correctly.
  powerpc: Use encode avpn where we need only avpn values
  powerpc: Reduce PTE table memory wastage
  powerpc: Move the pte free routines from common header
  powerpc: Reduce the PTE_INDEX_SIZE
  powerpc: Switch 16GB and 16MB explicit hugepages to a different page table format
  powerpc: New hugepage directory format
  powerpc: Don't truncate pgd_index wrongly
  powerpc: Don't hard code the size of pte page
  powerpc: Save DAR and DSISR in pt_regs on MCE
  ...

12 years agorbd: fix image request leak on parent read
Alex Elder [Thu, 2 May 2013 02:37:07 +0000 (21:37 -0500)]
rbd: fix image request leak on parent read

When a read for a layered image object finds the target object
doesn't exist, a read image request for the parent image is created
and submitted.  When that completes, the callback routine was
not releasing that parent image request.  Fix that.

The slab allocation stuff just added has greatly simplified the
search for the source of this memory leak.

This resolves:
    http://tracker.ceph.com/issues/4803

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
12 years agolibceph: use slab cache for osd client requests
Alex Elder [Wed, 1 May 2013 17:43:04 +0000 (12:43 -0500)]
libceph: use slab cache for osd client requests

Create a slab cache to manage allocation of ceph_osdc_request
structures.

This resolves:
    http://tracker.ceph.com/issues/3926

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
12 years agolibceph: allocate ceph message data with a slab allocator
Alex Elder [Wed, 1 May 2013 17:43:04 +0000 (12:43 -0500)]
libceph: allocate ceph message data with a slab allocator

Create a slab cache to manage ceph_msg_data structure allocation.

This is part of:
    http://tracker.ceph.com/issues/3926

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>
12 years agolibceph: allocate ceph messages with a slab allocator
Alex Elder [Wed, 1 May 2013 17:43:04 +0000 (12:43 -0500)]
libceph: allocate ceph messages with a slab allocator

Create a slab cache to manage ceph_msg structure allocation.

This is part of:
    http://tracker.ceph.com/issues/3926

Signed-off-by: Alex Elder <elder@inktank.com>
Reviewed-by: Josh Durgin <josh.durgin@inktank.com>