Pull more xfs updates from Ben Myers:
 "Here are several fixes for filesystems with CRC support turned on:
  fixes for quota, remote attributes, and recovery.  There is also some
  feature work related to CRCs: the implementation of CRCs for the inode
  unlinked lists, disabling noattr2/attr2 options when appropriate, and
  bumping the maximum number of ACLs.

  I would have preferred to defer this last category of items to 3.11.
  This would require setting a feature bit for the on-disk changes, so
  there is some pressure to get these in 3.10.  I believe this
  represents the end of the CRC related queue.

   - Rework of dquot CRCs
   - Fix for remote attribute invalidation of a leaf
   - Fix ordering of transaction replay in recovery
   - Implement CRCs for inode unlinked list
   - Disable noattr2/attr2 mount options when CRCs are enabled
   - Bump the limitation of ACL entries for v5 superblocks"

* tag 'for-linus-v3.10-rc5' of git://oss.sgi.com/xfs/xfs:
  xfs: increase number of ACL entries for V5 superblocks
  xfs: disable noattr2/attr2 mount options for CRC enabled filesystems
  xfs: inode unlinked list needs to recalculate the inode CRC
  xfs: fix log recovery transaction item reordering
  xfs: fix remote attribute invalidation for a leaf
  xfs: rework dquot CRCs

The limit of 25 ACL entries is arbitrary, but baked into the on-disk
format.  For version 5 superblocks, increase it to the maximum nuber
of ACLs that can fit into a single xattr.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Mark Tinguely <tinuguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit 5c87d4bc)

attr2 format is always enabled for v5 superblock filesystems, so the
mount options to enable or disable it need to be cause mount errors.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit d3eaace8)

The inode unlinked list manipulations operate directly on the inode
buffer, and so bypass the inode CRC calculation mechanisms. Hence an
inode on the unlinked list has an invalid CRC. Fix this by
recalculating the CRC whenever we modify an unlinked list pointer in
an inode, ncluding during log recovery. This is trivial to do and
results in  unlinked list operations always leaving a consistent
inode in the buffer.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit 0a32c26e)

There are several constraints that inode allocation and unlink
logging impose on log recovery. These all stem from the fact that
inode alloc/unlink are logged in buffers, but all other inode
changes are logged in inode items. Hence there are ordering
constraints that recovery must follow to ensure the correct result
occurs.

As it turns out, this ordering has been working mostly by chance
than good management. The existing code moves all buffers except
cancelled buffers to the head of the list, and everything else to
the tail of the list. The problem with this is that is interleaves
inode items with the buffer cancellation items, and hence whether
the inode item in an cancelled buffer gets replayed is essentially
left to chance.

Further, this ordering causes problems for log recovery when inode
CRCs are enabled. It typically replays the inode unlink buffer long before
it replays the inode core changes, and so the CRC recorded in an
unlink buffer is going to be invalid and hence any attempt to
validate the inode in the buffer is going to fail. Hence we really
need to enforce the ordering that the inode alloc/unlink code has
expected log recovery to have since inode chunk de-allocation was
introduced back in 2003...
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Mark Tinguely <tinguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit a775ad77)

When invalidating an attribute leaf block block, there might be
remote attributes that it points to. With the recent rework of the
remote attribute format, we have to make sure we calculate the
length of the attribute correctly. We aren't doing that in
xfs_attr3_leaf_inactive(), so fix it.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Mark Tinguely <tinuguely@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit 59913f14)

Calculating dquot CRCs when the backing buffer is written back just
doesn't work reliably. There are several places which manipulate
dquots directly in the buffers, and they don't calculate CRCs
appropriately, nor do they always set the buffer up to calculate
CRCs appropriately.

Firstly, if we log a dquot buffer (e.g. during allocation) it gets
logged without valid CRC, and so on recovery we end up with a dquot
that is not valid.

Secondly, if we recover/repair a dquot, we don't have a verifier
attached to the buffer and hence CRCs are not calculated on the way
down to disk.

Thirdly, calculating the CRC after we've changed the contents means
that if we re-read the dquot from the buffer, we cannot verify the
contents of the dquot are valid, as the CRC is invalid.

So, to avoid all the dquot CRC errors that are being detected by the
read verifier, change to using the same model as for inodes. That
is, dquot CRCs are calculated and written to the backing buffer at
the time the dquot is flushed to the backing buffer. If we modify
the dquot directly in the backing buffer, calculate the CRC
immediately after the modification is complete. Hence the dquot in
the on-disk buffer should always have a valid CRC.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Ben Myers <bpm@sgi.com>
Signed-off-by: Ben Myers <bpm@sgi.com>

(cherry picked from commit 6fcdc59d)

Since the introduction of preemptible mmu_gather TLB fast mode has been
broken. TLB fast mode relies on there being absolutely no concurrency;
it frees pages first and invalidates TLBs later.

However now we can get concurrency and stuff goes *bang*.

This patch removes all tlb_fast_mode() code; it was found the better
option vs trying to patch the hole by entangling tlb invalidation with
the scheduler.

Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Tony Luck <tony.luck@intel.com>
Reported-by: Max Filippov <jcmvbkbc@gmail.com>
Signed-off-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Pull kbuild fixes from Michal Marek:
 "There is one fix for a kbuild regression, plus three kconfig fixes for
  bugs that have alway been there, but are simple enough to be fixed in
  an -rc"

* 'rc-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild:
  kconfig/menu.c: fix multiple references to expressions in menu_add_prop()
  mconf: handle keys in empty dialogs
  kbuild: Don't assume dts files live in arch/*/boot/dts
  scripts/config: fix assignment of parameters for short version of --*-after options

Pull networking fixes from David Miller:

 1) Fix timeouts with direct mode authentication in mac80211, from
    Stanislaw Gruszka.

 2) Aggregation sessions can deadlock in ath9k, from Felix Fietkau.

 3) Netfilter's xt_addrtype doesn't work with ipv6 due to route lookups
    creating undesirable cache entries, from Florian Westphal.

 4) Fix netfilter's ipt_ULOG from generating non-NULL terminated
    strings.

 5) Fix netdev transmit queue crashes in mac80211, from Johannes Berg.

 6) Fix copy and paste error in 802.11 stack that broke reporting of
    64-bit station tx statistics, from Felix Fietkau.

 7) When qlge_probe fails, it leaks the netdev.  Fix from Wei Yongjun.

 8) SKB control block (where we store the IP options information,
    amongst other things) must be cleared properly otherwise ICMP
    sending can crash for IP tunnels.  Fix from Eric Dumazet.

 9) Verification of Energy Efficient Ether support was coded wrongly,
    the test was inversed.  Fix from Giuseppe CAVALLARO.

10) TCP handles redirects improperly because the wrong flow key is used
    for the route lookup.  From Michal Kubecek.

11) Don't interpret MSG_CMSG_COMPAT from userspace, fix from Andy
    Lutomirski.

12) The new AF_VSOCK was missing from the lockdep string table, fix from
    Federico Vaga.

13) be2net doesn't handle checksumming of IP fragments properly, from
    Somnath Kotur.

14) Fix several bugs in the device address list code that lead to
    crashes and other misbehaviors.  From Jay Vosburgh.

15) Fix ipv6 segmentation handling of fragmented GRE tunnel traffic,
    from Pravin B Shalr.

16) Fix usage of stale policies in IPSEC layer, from Paul Moore.

17) Fix team driver dump of ports when there are a large number of them,
    from Jiri Pirko.

18) Fix softlockups in UDP ipv4 socket lookup causes by and error in the
    hlist_nulls_for_each_entry_rcu() macro.  From Eric Dumazet.

19) Fix several regressions added by the high rate accuracy changes to
    the htb packet scheduler.  From Eric Dumazet.

20) Fix DMA'ing onto the stack in esd_usb2 and peak_usb CAN drivers,
    from Olivier Sobrie and Marc Kleine-Budde.

21) Fix unremovable network devices due to missing route pointer
    installation in the per-device ipv6 address list entries.  From Gao
    feng.

22) Apply the tg3 5719 DMA workaround on 5720 chips as well, otherwise
    we get stalls.  From Nithin Sujir.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (68 commits)
  net_sched: htb: do not mix 1ns and 64ns time units
  net: fix sk_buff head without data area
  tg3: Add read dma workaround for 5720
  net: ethernet: xilinx_emaclite: set protocol selector bits when writing ANAR
  bnx2x: Fix bridged GSO for 57710/57711 chips
  net: fec: add fallback to random MAC address
  bnx2x: fix TCP offload for tunneling ipv4 over ipv6
  ipv6: assign rt6_info to inet6_ifaddr in init_loopback
  net/mlx4_core: Keep VF assigned MAC in the PF admin table
  net/mlx4_en: Handle unassigned VF MAC address correctly
  net/mlx4_core: Return -EPROBE_DEFER when a VF is probed before PF is sufficiently initialized
  net/mlx4_en: Fix adaptive moderation cq update
  net: can: peak_usb: Do not do dma on the stack
  net: can: esd_usb2: Do not do dma on the stack
  net: can: kvaser_usb: fix reception on "USBcan Pro" and "USBcan R" type hardware.
  net_sched: restore "overhead xxx" handling
  net: force a reload of first item in hlist_nulls_for_each_entry_rcu
  hyperv: Fix vlan_proto setting in netvsc_recv_callback()
  team: fix port list dump for big number of ports
  list: introduce list_first_entry_or_null
  ...

commit 56b765b7 ("htb: improved accuracy at high rates") added another
regression for low rates, because it mixes 1ns and 64ns time units.

So the maximum delay (mbuffer) was not 60 second, but 937 ms.

Lets convert all time fields to 1ns as 64bit arches are becoming the
norm.
Reported-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Tested-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Eric Dumazet spotted that we have to check skb->head instead
of skb->data as skb->head points to the beginning of the
data area of the skbuff. Similarly, we have to initialize the
skb->head pointer, not skb->data in __alloc_skb_head.

After this fix, netlink crashes in the release path of the
sk_buff, so let's fix that as well.

This bug was introduced in (0ebd0ac5 net: add function to
allocate sk_buff head without data area).
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 091f0ea3 "tg3: Add New 5719 Read
DMA workaround" added a workaround for TX DMA stall on the 5719. This
workaround needs to be applied to the 5720 as well.

Cc: stable@vger.kernel.org
Reported-by: Roland Dreier <roland@purestorage.com>
Tested-by: Roland Dreier <roland@purestorage.com>
Signed-off-by: Nithin Nayak Sujir <nsujir@broadcom.com>
Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

This patch sets the protocol selector bits (4:0) of the PHY's MII_ADVERTISE
register (ANAR) when writing ADVERTISE_ALL. The protocol selector bits are
indicating IEEE 803.3u support and are fixed / read-only on some PHYs. Not
setting them correctly on others (like TI DP83630) makes the PHY fall back
to 10M HDX mode which should be avoided.

Tested for TI DP83630 PHY on Microblaze platform.
Signed-off-by: Jens Renner <renner@efe-gmbh.de>
Tested-by: Michal Simek <monstr@monstr.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>

It was recently found out that GSO on 57710/57711 was broken, due to packets
being sent without a valid IP checksum.

Commit 057cf65e "bnx2x: Fix GSO for 57710/57711 chips" partially fixed this
issue, but failed to set the correct IP checksum when receiving GSO packets
via bridges, as such packets enter bnx2x_tx_split() and the FW flags needed
to calculate IP checksum were erroneously set in the incorrect
buffer descriptor.

This patch re-enables GSO in said scenario for 57710/57711 chips.
Signed-off-by: Yuval Mintz <yuvalmin@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Pull MMC fixes from Chris Ball:
 - sdhci-acpi: Fix initial runtime PM status, add more ACPI IDs
 - atmel-mci, omap_hsmmc: DT handling fixes
 - esdhc-imx: Fix SDIO IRQs, fix multiblock reads (both h/w errata)

* tag 'mmc-fixes-for-3.10-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: omap_hsmmc: Skip platform_get_resource_byname() for dt case
  mmc: omap_hsmmc: convert to dma_request_slave_channel_compat
  mmc: omap_hsmmc: Fix the DT pbias workaround for MMC controllers 2 to 5
  mmc: sdhci-pci: add more device ids
  mmc: sdhci-acpi: add more device ids
  mmc: sdhci-acpi: fix initial runtime pm status
  mmc: atmel-mci: convert to dma_request_slave_channel_compat()
  mmc: sdhci-esdhc-imx: fix multiblock reads on i.MX53
  mmc: sdhci-esdhc-imx: Fix SDIO interrupts

Pull input fixes from Dmitry Torokhov:
 "Just a 2 small driver fixups here"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: wacom - fix a typo for Cintiq 22HDT
  Input: synaptics - fix sync lost after resume on some laptops

Pull kvm bugfixes from Gleb Natapov:
 "The bulk of the fixes is in MIPS KVM kernel<->userspace ABI.  MIPS KVM
  is new for 3.10 and some problems were found with current ABI.  It is
  better to fix them now and do not have a kernel with broken one"

* 'fixes' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: Fix race in apic->pending_events processing
  KVM: fix sil/dil/bpl/spl in the mod/rm fields
  KVM: Emulate multibyte NOP
  ARM: KVM: be more thorough when invalidating TLBs
  ARM: KVM: prevent NULL pointer dereferences with KVM VCPU ioctl
  mips/kvm: Use ENOIOCTLCMD to indicate unimplemented ioctls.
  mips/kvm: Fix ABI by moving manipulation of CP0 registers to KVM_{G,S}ET_ONE_REG
  mips/kvm: Use ARRAY_SIZE() instead of hardcoded constants in kvm_arch_vcpu_ioctl_{s,g}et_regs
  mips/kvm: Fix name of gpr field in struct kvm_regs.
  mips/kvm: Fix ABI for use of 64-bit registers.
  mips/kvm: Fix ABI for use of FPU.

Pull gfs2 fixes from Steven Whitehouse:
 "There are four patches this time.

  The first fixes a problem where the wrong descriptor type was being
  written into the log for journaled data blocks.

  The second fixes a race relating to the deallocation of allocator
  data.

  The third provides a fallback if kmalloc is unable to satisfy a
  request to allocate a directory hash table.

  The fourth fixes the iopen glock caching so that inodes are deleted in
  a more timely manner after rmdir/unlink"

* git://git.kernel.org/pub/scm/linux/kernel/git/steve/gfs2-3.0-fixes:
  GFS2: Don't cache iopen glocks
  GFS2: Fall back to vmalloc if kmalloc fails for dir hash tables
  GFS2: Increase i_writecount during gfs2_setattr_size
  GFS2: Set log descriptor type for jdata blocks

Pull fuse fixes from Miklos Szeredi:
 "One patch fixes an Oops introduced in 3.9 with the readdirplus
  feature.  The rest are fixes for async-dio in 3.10"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
  fuse: fix alignment in short read optimization for async_dio
  fuse: return -EIOCBQUEUED from fuse_direct_IO() for all async requests
  fuse: fix readdirplus Oops in fuse_dentry_revalidate
  fuse: update inode size and invalidate attributes on fallocate
  fuse: truncate pagecache range on hole punch
  fuse: allocate for_background dio requests based on io->async state

Pull microblaze fixes from Michal Simek:
 "One is fixing warning reported by sparse and the second warning was
  reported by Geert in his build regressions/improvements status update
  for -rc4."

* 'next' of git://git.monstr.eu/linux-2.6-microblaze:
  microblaze: Use static inline functions in cacheflush.h
  microblaze: Fix sparse warnings

If no valid MAC address could be obtained from the hardware,
fall back to a randomly generated one.
Signed-off-by: Pavel Machek <pavel@denx.de>
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>

FW was initialized with data from wrong header, this caused TSO packets
have wrong IP csum.
Signed-off-by: Dmitry Kravkov <dmitry@broadcom.com>
Signed-off-by: Ariel Elior <ariele@broadcom.com>
Signed-off-by: Eilon Greenstein <eilong@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Commit 25fb6ca4
"net IPv6 : Fix broken IPv6 routing table after loopback down-up"
forgot to assign rt6_info to the inet6_ifaddr.
When disable the net device, the rt6_info which allocated
in init_loopback will not be destroied in __ipv6_ifa_notify.

This will trigger the waring message below
[23527.916091] unregister_netdevice: waiting for tap0 to become free. Usage count = 1
Reported-by: Arkadiusz Miskiewicz <a.miskiewicz@gmail.com>
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Marc Kleine-Budde says:

====================
here are there fixes for the v3.10 release cycle:

The first patch by Jonas Peterson and Olivier Sobrie fixes the reception of CAN
frames on Kvaser's "USBcan Pro" and "USBcan R" type hardware.

The last two patches by Olivier Sobrie (for esd_usb2) and me (for peak_usb)
change the memory handling for the USB messages from stack to kmalloc(), as
memory used for DMA should not be allocated on stack.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

MAC addresses assigned by the PF to VFs were not kept in the PF driver
admin table. As a result, displaying the VF MACs from the PF interface
to user space showed zero address where in fact the VF got non-zero
address from the PF, fix that.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When a VF sense they didn't get MAC address, use random one. This will
address the case of administrator not assigning MAC to the VF through
the PF OS APIs and keep udev happy.
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

In the PF initialization, SRIOV is enabled before the PF is fully initialized.
This allows the kernel to probe the newly-exposed VFs before the PF is ready
to handle them (nested probes).

Have the probe method return the -EPROBE_DEFER value in this situation (instead
of the VF probe method retrying its initialization in a loop, and returning -EIO
on failure). When -EPROBE_DEFER is returned by the VF probe method, the kernel
itself will retry the probe after a suitable delay.

Based upon a suggestion by Ben Hutchings <bhutchings@solarflare.com>
Signed-off-by: Jack Morgenstein <jackm@dev.mellanox.co.il>
Signed-off-by: Or Gerlitz <ogerlitz@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

When turning on adaptive_rx under adaptive moderation, the CQ's moderation
count wasn't updated according to rx_frames which resulted in too many
interrupts and bandwidth drop.
Signed-off-by: Sagi Grimberg <sagig@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

And make the lines easier to read.
Signed-off-by: Ping Cheng <pingc@wacom.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

In summary, the symptom is intermittent key events lost after resume
on some machines with synaptics touchpad (seems this is synaptics _only_),
and key events loss is due to serio port reconnect after psmouse sync lost.
Removing psmouse and inserting it back during the suspend/resume process
is able to work around the issue, so the difference between psmouse_connect()
and psmouse_reconnect() is the key to the root cause of this problem.

After comparing the two different paths, synaptics driver has its own
implementation of synaptics_reconnect(), and the missing psmouse_probe()
seems significant, the patch below added psmouse_probe() to the reconnect
process, and has been verified many times that the issue could not be reliably
reproduced.

There are two PS/2 commands in psmouse_probe():

  1. PSMOUSE_CMD_GETID
  2. PSMOUSE_CMD_RESET_DIS

Only the PSMOUSE_CMD_GETID seems to be significant. The
PSMOUSE_CMD_RESET_DIS is irrelevant to this issue after trying
several times.  So we have only implemented this patch to issue
the PSMOUSE_CMD_GETID so far.
Tested-by: Daniel Manrique <daniel.manrique@canonical.com>
Signed-off-by: James M Leddy <james.leddy@canonical.com>
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

Pull regulator fixes from Mark Brown:
 "A few small fixes for v3.10, documentation things in the core and a
  few driver bugs."

* tag 'regulator-v3.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: palmas: Fix "enable_reg" to point to the correct reg for SMPS10
  regulator: palmas: Fix incorrect condition
  regulator: core: Correct spelling mistake in comment
  regulator: dbx500: Make local symbol static
  regulator: Fix kernel-doc generation warnings.

Pull jfs bugfixes from David Kleikamp:
 "A couple jfs bug fixes for 3.10-rc5"

* tag 'jfs-3.10-rc5' of git://github.com/kleikamp/linux-shaggy:
  fs/jfs: Add check if journaling to disk has been disabled in lbmRead()
  jfs: Several bugs in jfs_freeze() and jfs_unfreeze()

This patch makes GFS2 immediately reclaim/delete all iopen glocks
as soon as they're dequeued. This allows deleters to get an
EXclusive lock on iopen so files are deleted properly instead of
being set as unlinked.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

This version has one more correction: the vmalloc calls are replaced
by __vmalloc calls to preserve the GFP_NOFS flag.

When GFS2's directory management code allocates buffers for a
directory hash table, if it can't get the memory it needs, it
currently gives a bad return code. Rather than giving an error,
this patch allows it to use virtual memory rather than kernel
memory for the hash table. This should make it possible for
directories to function properly, even when kernel memory becomes
very fragmented.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

This patch calls get_write_access in a few functions. This
merely increases inode->i_writecount for the duration of the function.
That will ensure that any file closes won't delete the inode's
multi-block reservation while the function is running.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

This patch sets the log descriptor type according to whether the
journal commit is for (journaled) data or metadata. This was
recently broken when the functions to process data and metadata
log ops were combined.
Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Steven Whitehouse <swhiteho@redhat.com>

The bug was introduced with async_dio feature: trying to optimize short reads,
we cut number-of-bytes-to-read to i_size boundary. Hence the following example:

	truncate --size=300 /mnt/file
	dd if=/mnt/file of=/dev/null iflag=direct

led to FUSE_READ request of 300 bytes size. This turned out to be problem
for userspace fuse implementations who rely on assumption that kernel fuse
does not change alignment of request from client FS.

The patch turns off the optimization if async_dio is disabled. And, if it's
enabled, the patch fixes adjustment of number-of-bytes-to-read to preserve
alignment.

Note, that we cannot throw out short read optimization entirely because
otherwise a direct read of a huge size issued on a tiny file would generate
a huge amount of fuse requests and most of them would be ACKed by userspace
with zero bytes read.
Signed-off-by: Maxim Patlasov <MPatlasov@parallels.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>

If request submission fails for an async request (i.e.,
get_user_pages() returns -ERESTARTSYS), we currently skip the
-EIOCBQUEUED return and drop into wait_for_sync_kiocb() forever.

Avoid this by always returning -EIOCBQUEUED for async requests. If
an error occurs, the error is passed into fuse_aio_complete(),
returned via aio_complete() and thus propagated to userspace via
io_getevents().
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Maxim Patlasov <MPatlasov@parallels.com>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>

Fix bug introduced by commit 4582a4ab "FUSE: Adapt readdirplus to application
usage patterns".

We need to check for a positive dentry; negative dentries are not added by
readdirplus.  Secondly we need to advise the use of readdirplus on the *parent*,
otherwise the whole thing is useless.  Thirdly all this is only relevant if
"readdirplus_auto" mode is selected by the filesystem.

We advise the use of readdirplus only if the dentry was still valid.  If we had
to redo the lookup then there was no use in doing the -plus version.
Reported-by: Bernd Schubert <bernd.schubert@itwm.fraunhofer.de>
Signed-off-by: Miklos Szeredi <mszeredi@suse.cz>
CC: Feng Shuo <steve.shuo.feng@gmail.com>
CC: stable@vger.kernel.org