pandora-kernel.git
12 years agoNFS: Use the nfs_pageio_descriptor->pg_bsize in the read/write request
Trond Myklebust [Tue, 12 Jul 2011 17:42:02 +0000 (13:42 -0400)]
NFS: Use the nfs_pageio_descriptor->pg_bsize in the read/write request

Instead of looking up the rsize and wsize, the routines that generate the
RPC requests should really be using the pg_bsize, since that is what we
use when deciding whether or not to coalesce write requests...

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Cache rpc_ops in struct nfs_pageio_descriptor
Trond Myklebust [Tue, 12 Jul 2011 17:42:02 +0000 (13:42 -0400)]
NFS: Cache rpc_ops in struct nfs_pageio_descriptor

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Clean up: split out the RPC transmission from nfs_pagein_multi/one
Trond Myklebust [Tue, 12 Jul 2011 17:42:02 +0000 (13:42 -0400)]
NFS: Clean up: split out the RPC transmission from nfs_pagein_multi/one

...and do the same for nfs_flush_multi/one.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: fix return value of nfs_pagein_one/nfs_flush_one
Peng Tao [Fri, 15 Jul 2011 07:33:42 +0000 (03:33 -0400)]
NFS: fix return value of nfs_pagein_one/nfs_flush_one

Signed-off-by: Peng Tao <peng_tao@emc.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Clean up nfs_read_rpcsetup and nfs_write_rpcsetup
Trond Myklebust [Tue, 12 Jul 2011 17:42:02 +0000 (13:42 -0400)]
NFS: Clean up nfs_read_rpcsetup and nfs_write_rpcsetup

Split them up into two parts: one which sets up the struct nfs_read/write_data,
the other which sets up the actual RPC call or pNFS call.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Don't use DATA_SYNC writes
Trond Myklebust [Tue, 12 Jul 2011 17:42:01 +0000 (13:42 -0400)]
NFS: Don't use DATA_SYNC writes

If we're writing back data, and the FLUSH_STABLE flag is set, then we
always want to use NFS_FILE_SYNC, since we're always in a situation where
we're doing page reclaim, and so we want to free up the page as quickly
as possible.

If we're in the FLUSH_COND_STABLE case, then we either want to use another
unstable write (if we have to do a commit anyway) or again, we want to
use NFS_FILE_SYNC because we know that we have no more pages to write
out.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: do not use deviceids after MDS clientid invalidation
Andy Adamson [Wed, 15 Jun 2011 21:52:40 +0000 (17:52 -0400)]
NFSv4.1: do not use deviceids after MDS clientid invalidation

Mark all deviceids established under an expired MDS clientid as invalid.
Stop all new i/o through DS and send through the MDS.
Don't use any new LAYOUTGETs that use the invalid deviceid. Purge all layouts
established under the expired MDS clientid.
Remove the MDS clientid deviceid and data servers reference

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: Clean up layoutreturn
Trond Myklebust [Wed, 15 Jun 2011 15:59:10 +0000 (11:59 -0400)]
NFSv4.1: Clean up layoutreturn

Since we take a reference to it, we really ought to pass the a pointer to
the layout header in the arguments instead of assuming that
NFS_I(inode)->layout will forever point to the correct object.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agopnfs-obj: pnfs_osd_xdr: Remove dead code and cleanup
Boaz Harrosh [Mon, 13 Jun 2011 21:52:55 +0000 (17:52 -0400)]
pnfs-obj: pnfs_osd_xdr: Remove dead code and cleanup

* Some leftovers from ancient times.
* This file will only define common types and client API.
  Remove server from comments

Signed-off-by: Boaz Harrosh <Boaz Harrosh bharrosh@panasas.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: File layout only supports whole file layouts
Andy Adamson [Mon, 13 Jun 2011 22:22:38 +0000 (18:22 -0400)]
NFSv4.1: File layout only supports whole file layouts

Ask for whole file layouts. Until support for layout segments is fully
supported in the file layout code, discard non-whole file layouts.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Fred Isaman <iisaman@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: Clean ups for the device id cache
Trond Myklebust [Tue, 14 Jun 2011 16:18:11 +0000 (12:18 -0400)]
NFSv4.1: Clean ups for the device id cache

The fact that the global device id cache holds a reference to the
nfs4_deviceid_node until it is invisible to rcu lookups implies that
we can always assume that the reference count is non-zero in
_find_get_deviceid.

Also clean up nfs4_put_deviceid_node and the removal of the device id
from the cache.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: Fall back to ordinary i/o through the mds if we have no layout segment
Trond Myklebust [Fri, 10 Jun 2011 17:30:23 +0000 (13:30 -0400)]
NFSv4.1: Fall back to ordinary i/o through the mds if we have no layout segment

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: Add an initialisation callback for pNFS
Trond Myklebust [Fri, 10 Jun 2011 17:30:23 +0000 (13:30 -0400)]
NFSv4.1: Add an initialisation callback for pNFS

Ensure that we always get a layout before setting up the i/o request.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Cleanup of the nfs_pageio code in preparation for a pnfs bugfix
Trond Myklebust [Fri, 10 Jun 2011 17:30:23 +0000 (13:30 -0400)]
NFS: Cleanup of the nfs_pageio code in preparation for a pnfs bugfix

We need to ensure that the layouts are set up before we can decide to
coalesce requests. To do so, we want to further split up the struct
nfs_pageio_descriptor operations into an initialisation callback, a
coalescing test callback, and a 'do i/o' callback.

This patch cleans up the existing callback methods before adding the
'initialisation' callback.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: test and free stateids during recovery
Bryan Schumaker [Thu, 2 Jun 2011 18:59:10 +0000 (14:59 -0400)]
NFS: test and free stateids during recovery

When recovering open files and locks, the stateid should be tested
against the server and freed if it is invalid.  This patch adds new
recovery functions for NFS v4.1.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: added FREE_STATEID call
Bryan Schumaker [Thu, 2 Jun 2011 18:59:09 +0000 (14:59 -0400)]
NFS: added FREE_STATEID call

FREE_STATEID is used to tell the server that we want to free a stateid
that no longer has any locks associated with it.  This allows the client
to reclaim locks without encountering edge conditions documented in
section 8.4.3 of RFC 5661.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Added TEST_STATEID call
Bryan Schumaker [Thu, 2 Jun 2011 18:59:08 +0000 (14:59 -0400)]
NFS: Added TEST_STATEID call

This patch adds in the xdr for doing a TEST_STATEID call with a single
stateid. RFC 5661 allows multiple stateids to be tested in a single
call, but only testing one keeps things simpler for now.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Add SECINFO_NO_NAME procedure
Bryan Schumaker [Thu, 2 Jun 2011 18:59:07 +0000 (14:59 -0400)]
NFS: Add SECINFO_NO_NAME procedure

If the client is using NFS v4.1, then we can use SECINFO_NO_NAME to find
the secflavor for the initial mount.  If the server doesn't support
SECINFO_NO_NAME then I fall back on the "guess and check" method used
for v4.0 mounts.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: move pnfs layouts to nfs_server structure
Weston Andros Adamson [Wed, 1 Jun 2011 20:44:44 +0000 (16:44 -0400)]
NFS: move pnfs layouts to nfs_server structure

Layouts should be tracked per nfs_server (aka superblock)
instead of per struct nfs_client, which may have multiple FSIDs associated
with it.

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: fix comment
Weston Andros Adamson [Wed, 1 Jun 2011 20:32:21 +0000 (16:32 -0400)]
NFS: fix comment

We support IPv4 and IPv6 now.

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: use scope from exchange_id to skip reclaim
Weston Andros Adamson [Tue, 31 May 2011 23:05:47 +0000 (19:05 -0400)]
NFS: use scope from exchange_id to skip reclaim

can be skipped if the "eir_server_scope" from the exchange_id proc differs from
previous calls.

Also, in the future server_scope will be useful for determining whether client
trunking is available

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: pnfs: loop over multipath addrs on connect
Weston Andros Adamson [Tue, 31 May 2011 22:48:58 +0000 (18:48 -0400)]
NFS: pnfs: loop over multipath addrs on connect

Don't just use the first addr in the multipath list - instead, loop
over addresses when calling nfs4_set_ds_client() (which calls connect)
until it is successful.

Although this is not real multipath support, it's a quick fix to handle when
an MDS sends a list of addresses for a DS and some of the addr families are
unsupported or misconfigured (like no routable ipv6 addr assigned).
This will attempt all paths to the DS before giving up, instead of immediately
falling back to the MDS.

As before, an error encountered after a successful connect() will cause all
i/o to fall back to the MDS.

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Parse and store all multipath DS addresses
Weston Andros Adamson [Tue, 31 May 2011 22:48:57 +0000 (18:48 -0400)]
NFS: Parse and store all multipath DS addresses

This parses and stores all addresses associated with each data server,
laying the groundwork for supporting multipath to data servers.

 - Skips over addresses that cannot be parsed (ie IPv6 addrs if v6 is not
   enabled).  Only fails if none of the addresses are recognizable
 - Currently only uses the first address that parsed cleanly
 - Tested against pynfs server (modified to support multipath)

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: pnfs IPv6 support
Weston Andros Adamson [Tue, 31 May 2011 22:48:56 +0000 (18:48 -0400)]
NFS: pnfs IPv6 support

Handle ipv6 remote addresses from GETDEVICEINFO

 - supports netid "tcp" for ipv4 and "tcp6" for ipv6 as rfc 5665 specifies
 - added ds_remotestr to avoid having to handle different AFs in every dprintk
 - tested against pynfs 4.1 server, submitting ipv6 support patch to pynfs
 - tested with IPv6 disabled, it compiles cleanly and relies on rpc_pton to
   refuse to accept IPv6 addresses

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agolockd: properly convert be32 values in debug messages
Vasily Averin [Wed, 1 Jun 2011 12:54:32 +0000 (16:54 +0400)]
lockd: properly convert be32 values in debug messages

lockd: server returns status 50331648
it's quite hard to understand that number in this message is 3 in big endian

Signed-off-by: Vasily Averin <vvs@sw.ru>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agosunrpc: use dprint_status() macro in call_decode()
Vasily Averin [Wed, 1 Jun 2011 12:23:59 +0000 (16:23 +0400)]
sunrpc: use dprint_status() macro in call_decode()

common dprint_status() macro is used in all callbacks but not in call_decode()

Signed-off-by: Vasily Averin <vvs@sw.ru>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoSUNRPC: Fix use of static variable in rpcb_getport_async
Ben Greear [Tue, 12 Jul 2011 17:27:55 +0000 (10:27 -0700)]
SUNRPC: Fix use of static variable in rpcb_getport_async

Because struct rpcbind_args *map was declared static, if two
threads entered this method at the same time, the values
assigned to map could be sent two two differen tasks.
This could cause all sorts of problems, include use-after-free
and double-free of memory.

Fix this by removing the static declaration so that the map
pointer is on the stack.

Signed-off-by: Ben Greear <greearb@candelatech.com>
Cc: stable@kernel.org
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: update nfs4_fattr_bitmap_maxsz
Andy Adamson [Mon, 11 Jul 2011 21:17:42 +0000 (17:17 -0400)]
NFSv4.1: update nfs4_fattr_bitmap_maxsz

Attribute IDs assigned in RFC 5661 now require three bitmaps.
Fixes hitting a BUG_ON in xdr_shrink_bufhead when getting ACLs.

Signed-off-by: Andy Adamson <andros@netapp.com>
Cc:stable@kernel.org [2.6.39]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoSUNRPC: Fix a race between work-queue and rpc_killall_tasks
Trond Myklebust [Wed, 6 Jul 2011 23:58:23 +0000 (19:58 -0400)]
SUNRPC: Fix a race between work-queue and rpc_killall_tasks

Since rpc_killall_tasks may modify the rpc_task's tk_action field
without any locking, we need to be careful when dereferencing it.

Reported-by: Ben Greear <greearb@candelatech.com>
Tested-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: stable@kernel.org
12 years agopnfs: write: Set mds_offset in the generic layer - it is needed by all LDs
Boaz Harrosh [Thu, 16 Jun 2011 15:35:46 +0000 (11:35 -0400)]
pnfs: write: Set mds_offset in the generic layer - it is needed by all LDs

In current pnfs tree, all the layouts set mds_offset in their
.write_pagelist member.
mds_offset is only used by generic layer and should be handled by it.

This patch is for upstream. It is needed in this -rc series to fix a
bug in objects layout_commit.

I'll send patches for objects and blocks to be
squashed into current pnfs tree.

TODO: It looks like the read path needs the same patch.

Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoLinux 3.0-rc5 v3.0-rc5
Linus Torvalds [Tue, 28 Jun 2011 02:12:22 +0000 (19:12 -0700)]
Linux 3.0-rc5

12 years agodrm/i915: more struct_mutex locking
Hugh Dickins [Mon, 27 Jun 2011 23:18:20 +0000 (16:18 -0700)]
drm/i915: more struct_mutex locking

When auditing the locking in i915_gem.c (for a prospective change which
I then abandoned), I noticed two places where struct_mutex is not held
across GEM object manipulations that would usually require it.

Since one is in initial setup and the other in driver unload, I'm
guessing the mutex is not required for either; but post a patch in case
it is.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrm/i915: use shmem_truncate_range
Hugh Dickins [Mon, 27 Jun 2011 23:18:19 +0000 (16:18 -0700)]
drm/i915: use shmem_truncate_range

The interface to ->truncate_range is changing very slightly: once "tmpfs:
take control of its truncate_range" has been applied, this can be applied.
 For now there is only a slight inefficiency while this remains unapplied,
but it will soon become essential for managing shmem's use of swap.

Change i915_gem_object_truncate() to use shmem_truncate_range() directly:
which should also spare i915 later change if we switch from
inode_operations->truncate_range to file_operations->fallocate.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrm/i915: use shmem_read_mapping_page
Hugh Dickins [Mon, 27 Jun 2011 23:18:18 +0000 (16:18 -0700)]
drm/i915: use shmem_read_mapping_page

Soon tmpfs will stop supporting ->readpage and read_cache_page_gfp(): once
"tmpfs: add shmem_read_mapping_page_gfp" has been applied, this patch can
be applied to ease the transition.

Make i915_gem_object_get_pages_gtt() use shmem_read_mapping_page_gfp() in
the one place it's needed; elsewhere use shmem_read_mapping_page(), with
the mapping's gfp_mask properly initialized.

Forget about __GFP_COLD: since tmpfs initializes its pages with memset,
asking for a cold page is counter-productive.

Include linux/shmem_fs.h also in drm_gem.c: with shmem_file_setup() now
declared there too, we shall remove the prototype from linux/mm.h later.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Keith Packard <keithp@keithp.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrm/ttm: use shmem_read_mapping_page
Hugh Dickins [Mon, 27 Jun 2011 23:18:17 +0000 (16:18 -0700)]
drm/ttm: use shmem_read_mapping_page

Soon tmpfs will stop supporting ->readpage and read_mapping_page(): once
"tmpfs: add shmem_read_mapping_page_gfp" has been applied, this patch can
be applied to ease the transition.

ttm_tt_swapin() and ttm_tt_swapout() use shmem_read_mapping_page() in
place of read_mapping_page(), since their swap_space has been created with
shmem_file_setup().

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: Thomas Hellstrom <thellstrom@vmware.com>
Cc: Dave Airlie <airlied@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/tty/serial/8250_pci.c: fix warning
Andrew Morton [Mon, 27 Jun 2011 23:18:16 +0000 (16:18 -0700)]
drivers/tty/serial/8250_pci.c: fix warning

Fis the warning

  drivers/tty/serial/8250_pci.c:1457: warning: initialization from incompatible pointer type

Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/misc/ioc4.c: fix section mismatch / race condition
Ralf Baechle [Mon, 27 Jun 2011 23:18:15 +0000 (16:18 -0700)]
drivers/misc/ioc4.c: fix section mismatch / race condition

Fix this section mismatch:

  WARNING: drivers/misc/ioc4.o(.data+0x144): Section mismatch in reference from the variable ioc4_load_modules_work to the function .devinit.text:ioc4_load_modules()
  The variable ioc4_load_modules_work references
  the function __devinit ioc4_load_modules()
  If the reference is valid then annotate the
  variable with __init* or __refdata (see linux/init.h) or name the variable:
  *driver, *_template, *_timer, *_sht, *_ops, *_probe, *_probe_one, *_console

This one is potentially fatal; by the time ioc4_load_modules is invoked
it may already have been freed.  For that reason ioc4_load_modules_work
can't be turned to __devinitdata but also because it's referenced in
ioc4_exit.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Acked-by: Brent Casavant <bcasavan@sgi.com>
Cc: Tejun Heo <tj@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/leds/leds-lp5523.c: fix section mismatches
Ralf Baechle [Mon, 27 Jun 2011 23:18:14 +0000 (16:18 -0700)]
drivers/leds/leds-lp5523.c: fix section mismatches

Fix this section mismatch:

  WARNING: drivers/leds/leds-lp5523.o(.text+0x12f4): Section mismatch in reference from the function lp5523_probe() to the function .init.text:lp5523_init_led()
  The function lp5523_probe() references
  the function __init lp5523_init_led().
  This is often because lp5523_probe lacks a __init
  annotation or the annotation of lp5523_init_led is wrong.

Fixing this one triggers one more mismatch, fix that one as well.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Cc: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/leds/leds-lp5521.c: fix section mismatches
Ralf Baechle [Mon, 27 Jun 2011 23:18:13 +0000 (16:18 -0700)]
drivers/leds/leds-lp5521.c: fix section mismatches

Fix this section mismatch:

  WARNING: drivers/leds/leds-lp5521.o(.text+0xf2c): Section mismatch in reference from the function lp5521_probe() to the function .init.text:lp5521_init_led()
  The function lp5521_probe() references
  the function __init lp5521_init_led().
  This is often because lp5521_probe lacks a __init
  annotation or the annotation of lp5521_init_led is wrong.

Fixing this mismatch triggers one more mismatch, fix that one as well.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Cc: Richard Purdie <rpurdie@rpsys.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomemcg: fix direct softlimit reclaim to be called in limit path
KAMEZAWA Hiroyuki [Mon, 27 Jun 2011 23:18:12 +0000 (16:18 -0700)]
memcg: fix direct softlimit reclaim to be called in limit path

Commit d149e3b25d7c ("memcg: add the soft_limit reclaim in global direct
reclaim") adds a softlimit hook to shrink_zones().  By this, soft limit
is called as

   try_to_free_pages()
       do_try_to_free_pages()
           shrink_zones()
               mem_cgroup_soft_limit_reclaim()

Then, direct reclaim is memcg softlimit hint aware, now.

But, the memory cgroup's "limit" path can call softlimit shrinker.

   try_to_free_mem_cgroup_pages()
       do_try_to_free_pages()
           shrink_zones()
               mem_cgroup_soft_limit_reclaim()

This will cause a global reclaim when a memcg hits limit.

This is bug. soft_limit_reclaim() should be called when
scanning_global_lru(sc) == true.

And the commit adds a variable "total_scanned" for counting softlimit
scanned pages....it's not "total".  This patch removes the variable and
update sc->nr_scanned instead of it.  This will affect shrink_slab()'s
scan condition but, global LRU is scanned by softlimit and I think this
change makes sense.

TODO: avoid too much scanning of a zone when softlimit did enough work.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Cc: Ying Han <yinghan@google.com>
Cc: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agotaskstats: don't allow duplicate entries in listener mode
Vasiliy Kulikov [Mon, 27 Jun 2011 23:18:11 +0000 (16:18 -0700)]
taskstats: don't allow duplicate entries in listener mode

Currently a single process may register exit handlers unlimited times.
It may lead to a bloated listeners chain and very slow process
terminations.

Eg after 10KK sent TASKSTATS_CMD_ATTR_REGISTER_CPUMASKs ~300 Mb of
kernel memory is stolen for the handlers chain and "time id" shows 2-7
seconds instead of normal 0.003.  It makes it possible to exhaust all
kernel memory and to eat much of CPU time by triggerring numerous exits
on a single CPU.

The patch limits the number of times a single process may register
itself on a single CPU to one.

One little issue is kept unfixed - as taskstats_exit() is called before
exit_files() in do_exit(), the orphaned listener entry (if it was not
explicitly deregistered) is kept until the next someone's exit() and
implicit deregistration in send_cpu_listeners().  So, if a process
registered itself as a listener exits and the next spawned process gets
the same pid, it would inherit taskstats attributes.

Signed-off-by: Vasiliy Kulikov <segooon@gmail.com>
Cc: Balbir Singh <bsingharora@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: fix assertion mapping->nrpages == 0 in end_writeback()
Jan Kara [Mon, 27 Jun 2011 23:18:10 +0000 (16:18 -0700)]
mm: fix assertion mapping->nrpages == 0 in end_writeback()

Under heavy memory and filesystem load, users observe the assertion
mapping->nrpages == 0 in end_writeback() trigger.  This can be caused by
page reclaim reclaiming the last page from a mapping in the following
race:

CPU0 CPU1
  ...
  shrink_page_list()
    __remove_mapping()
      __delete_from_page_cache()
        radix_tree_delete()
evict_inode()
  truncate_inode_pages()
    truncate_inode_pages_range()
      pagevec_lookup() - finds nothing
  end_writeback()
    mapping->nrpages != 0 -> BUG
        page->mapping = NULL
        mapping->nrpages--

Fix the problem by doing a reliable check of mapping->nrpages under
mapping->tree_lock in end_writeback().

Analyzed by Jay <jinshan.xiong@whamcloud.com>, lost in LKML, and dug out
by Miklos Szeredi <mszeredi@suse.de>.

Cc: Jay <jinshan.xiong@whamcloud.com>
Cc: Miklos Szeredi <mszeredi@suse.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm/memory-failure.c: fix spinlock vs mutex order
Peter Zijlstra [Mon, 27 Jun 2011 23:18:09 +0000 (16:18 -0700)]
mm/memory-failure.c: fix spinlock vs mutex order

We cannot take a mutex while holding a spinlock, so flip the order and
fix the locking documentation.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/misc/lkdtm.c: fix race when crashpoint is hit multiple times before checking...
Josh Hunt [Mon, 27 Jun 2011 23:18:08 +0000 (16:18 -0700)]
drivers/misc/lkdtm.c: fix race when crashpoint is hit multiple times before checking count

We observed the crash point count going negative in cases where the
crash point is hit multiple times before the check of "count == 0" is
done.  Because of this we never call lkdtm_do_action().  This patch just
adds a spinlock to protect count.

Reported-by: Tapan Dhimant <tdhimant@akamai.com>
Signed-off-by: Josh Hunt <johunt@akamai.com>
Acked-by: Ankita Garg <ankita@in.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoinclude/linux/compat.h: declare compat_sys_sendmmsg()
Chris Metcalf [Mon, 27 Jun 2011 23:18:07 +0000 (16:18 -0700)]
include/linux/compat.h: declare compat_sys_sendmmsg()

This is required for tilegx to be able to use the compat unistd.h header
where compat_sys_sendmmsg() is now mentioned.

Signed-off-by: Chris Metcalf <cmetcalf@tilera.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoromfs: fix romfs_get_unmapped_area() argument check
Bob Liu [Mon, 27 Jun 2011 23:18:06 +0000 (16:18 -0700)]
romfs: fix romfs_get_unmapped_area() argument check

romfs_get_unmapped_area() checks argument `len' without considering
PAGE_ALIGN which will cause do_mmap_pgoff() return -EINVAL error after
commit f67d9b1576c ("nommu: add page_align to mmap").

Fix the check by changing it in same way ramfs_nommu_get_unmapped_area()
was changed in ramfs/file-nommu.c.

Signed-off-by: Bob Liu <lliubbo@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Acked-by: Greg Ungerer <gerg@snapgear.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoum: add asm/percpu.h
Richard Weinberger [Mon, 27 Jun 2011 23:18:05 +0000 (16:18 -0700)]
um: add asm/percpu.h

To make SLUB work on UML we need this_cpu_cmpxchg from
asm-generic/percpu.h.

Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodrivers/rtc/rtc-ds1307.c: add support for RTC device pt7c4338
Priyanka Jain [Mon, 27 Jun 2011 23:18:04 +0000 (16:18 -0700)]
drivers/rtc/rtc-ds1307.c: add support for RTC device pt7c4338

PT7C4338 chip is being manufactured by Pericom Technology Inc.  It is a
serial real-time clock which provides:

1) Low-power clock/calendar.
2) Programmable square-wave output.

It has 56 bytes of nonvolatile RAM.  Its register set is same as that of
rtc device: DS1307.

Signed-off-by: Priyanka Jain <Priyanka.Jain@freescale.com>
Acked-by: Timur Tabi <timur@freescale.com>
Reviewed-by: Wolfram Sang <w.sang@pengutronix.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agotmpfs: add shmem_read_mapping_page_gfp
Hugh Dickins [Mon, 27 Jun 2011 23:18:04 +0000 (16:18 -0700)]
tmpfs: add shmem_read_mapping_page_gfp

Although it is used (by i915) on nothing but tmpfs, read_cache_page_gfp()
is unsuited to tmpfs, because it inserts a page into pagecache before
calling the filesystem's ->readpage: tmpfs may have pages in swapcache
which only it knows how to locate and switch to filecache.

At present tmpfs provides a ->readpage method, and copes with this by
copying pages; but soon we can simplify it by removing its ->readpage.
Provide shmem_read_mapping_page_gfp() now, ready for that transition,

Export shmem_read_mapping_page_gfp() and add it to list in shmem_fs.h,
with shmem_read_mapping_page() inline for the common mapping_gfp case.

(shmem_read_mapping_page_gfp or shmem_read_cache_page_gfp? Generally the
read_mapping_page functions use the mapping's ->readpage, and the
read_cache_page functions use the supplied filler, so I think
read_cache_page_gfp was slightly misnamed.)

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agotmpfs: take control of its truncate_range
Hugh Dickins [Mon, 27 Jun 2011 23:18:03 +0000 (16:18 -0700)]
tmpfs: take control of its truncate_range

2.6.35's new truncate convention gave tmpfs the opportunity to control
its file truncation, no longer enforced from outside by vmtruncate().
We shall want to build upon that, to handle pagecache and swap together.

Slightly redefine the ->truncate_range interface: let it now be called
between the unmap_mapping_range()s, with the filesystem responsible for
doing the truncate_inode_pages_range() from it - just as the filesystem
is nowadays responsible for doing that from its ->setattr.

Let's rename shmem_notify_change() to shmem_setattr().  Instead of
calling the generic truncate_setsize(), bring that code in so we can
call shmem_truncate_range() - which will later be updated to perform its
own variant of truncate_inode_pages_range().

Remove the punch_hole unmap_mapping_range() from shmem_truncate_range():
now that the COW's unmap_mapping_range() comes after ->truncate_range,
there is no need to call it a third time.

Export shmem_truncate_range() and add it to the list in shmem_fs.h, so
that i915_gem_object_truncate() can call it explicitly in future; get
this patch in first, then update drm/i915 once this is available (until
then, i915 will just be doing the truncate_inode_pages() twice).

Though introduced five years ago, no other filesystem is implementing
->truncate_range, and its only other user is madvise(,,MADV_REMOVE): we
expect to convert it to fallocate(,FALLOC_FL_PUNCH_HOLE,,) shortly,
whereupon ->truncate_range can be removed from inode_operations -
shmem_truncate_range() will help i915 across that transition too.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: move shmem prototypes to shmem_fs.h
Hugh Dickins [Mon, 27 Jun 2011 23:18:02 +0000 (16:18 -0700)]
mm: move shmem prototypes to shmem_fs.h

Before adding any more global entry points into shmem.c, gather such
prototypes into shmem_fs.h.  Remove mm's own declarations from swap.h,
but for now leave the ones in mm.h: because shmem_file_setup() and
shmem_zero_setup() are called from various places, and we should not
force other subsystems to update immediately.

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agomm: move vmtruncate_range to truncate.c
Hugh Dickins [Mon, 27 Jun 2011 23:18:01 +0000 (16:18 -0700)]
mm: move vmtruncate_range to truncate.c

You would expect to find vmtruncate_range() next to vmtruncate() in
mm/truncate.c: move it there.

Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Christoph Hellwig <hch@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoFix some kernel-doc warnings
Vitaliy Ivanov [Mon, 27 Jun 2011 16:07:08 +0000 (19:07 +0300)]
Fix some kernel-doc warnings

Fix 'make htmldocs' warnings:

  Warning(/include/linux/hrtimer.h:153): No description found for parameter 'clockid'
  Warning(/include/linux/device.h:604): Excess struct/union/enum/typedef member 'of_match' description in 'device'
  Warning(/include/net/sock.h:349): Excess struct/union/enum/typedef member 'sk_rmem_alloc' description in 'sock'

Signed-off-by: Vitaliy Ivanov <vitalivanov@gmail.com>
Acked-by: Grant Likely <grant.likely@secretlab.ca>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc
Linus Torvalds [Mon, 27 Jun 2011 21:55:43 +0000 (14:55 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/cjb/mmc

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/cjb/mmc:
  mmc: queue: bring discard_granularity/alignment into line with SCSI
  mmc: queue: append partition subname to queue thread name
  mmc: core: make erase timeout calculation allow for gated clock
  mmc: block: switch card to User Data Area when removing the block driver
  mmc: sdio: reset card during power_restore
  mmc: cb710: fix #ifdef HAVE_EFFICIENT_UNALIGNED_ACCESS
  mmc: sdhi: DMA slave ID 0 is invalid
  mmc: tmio: fix regression in TMIO_MMC_WRPROTECT_DISABLE handling
  mmc: omap_hsmmc: use original sg_len for dma_unmap_sg
  mmc: omap_hsmmc: fix ocr mask usage
  mmc: sdio: fix runtime PM path during driver removal
  mmc: Add PCI fixup quirks for Ricoh 1180:e823 reader
  mmc: sdhi: fix module unloading
  mmc: of_mmc_spi: add NO_IRQ define to of_mmc_spi.c
  mmc: vub300: fix null dereferences in error handling

12 years agoFix node_start/end_pfn() definition for mm/page_cgroup.c
KAMEZAWA Hiroyuki [Thu, 16 Jun 2011 08:28:07 +0000 (17:28 +0900)]
Fix node_start/end_pfn() definition for mm/page_cgroup.c

commit 21a3c96 uses node_start/end_pfn(nid) for detection start/end
of nodes. But, it's not defined in linux/mmzone.h but defined in
/arch/???/include/mmzone.h which is included only under
CONFIG_NEED_MULTIPLE_NODES=y.

Then, we see
  mm/page_cgroup.c: In function 'page_cgroup_init':
  mm/page_cgroup.c:308: error: implicit declaration of function 'node_start_pfn'
  mm/page_cgroup.c:309: error: implicit declaration of function 'node_end_pfn'

So, fixiing page_cgroup.c is an idea...

But node_start_pfn()/node_end_pfn() is a very generic macro and
should be implemented in the same manner for all archs.
(m32r has different implementation...)

This patch removes definitions of node_start/end_pfn() in each archs
and defines a unified one in linux/mmzone.h. It's not under
CONFIG_NEED_MULTIPLE_NODES, now.

A result of macro expansion is here (mm/page_cgroup.c)

for !NUMA
 start_pfn = ((&contig_page_data)->node_start_pfn);
  end_pfn = ({ pg_data_t *__pgdat = (&contig_page_data); __pgdat->node_start_pfn + __pgdat->node_spanned_pages;});

for NUMA (x86-64)
  start_pfn = ((node_data[nid])->node_start_pfn);
  end_pfn = ({ pg_data_t *__pgdat = (node_data[nid]); __pgdat->node_start_pfn + __pgdat->node_spanned_pages;});

Changelog:
 - fixed to avoid using "nid" twice in node_end_pfn() macro.

Reported-and-acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Reported-and-tested-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs...
Linus Torvalds [Mon, 27 Jun 2011 20:32:14 +0000 (13:32 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mason/btrfs-unstable

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable:
  btrfs: fix inconsonant inode information
  Btrfs: make sure to update total_bitmaps when freeing cache V3
  Btrfs: fix type mismatch in find_free_extent()
  Btrfs: make sure to record the transid in new inodes

12 years agoMerge branch 'for-linus' of git://oss.sgi.com/xfs/xfs
Linus Torvalds [Mon, 27 Jun 2011 16:01:29 +0000 (09:01 -0700)]
Merge branch 'for-linus' of git://oss.sgi.com/xfs/xfs

* 'for-linus' of git://oss.sgi.com/xfs/xfs:
  xfs: prevent bogus assert when trying to remove non-existent attribute
  xfs: clear XFS_IDIRTY_RELEASE on truncate down
  xfs: reset inode per-lifetime state when recycling it

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid
Linus Torvalds [Mon, 27 Jun 2011 16:00:50 +0000 (09:00 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/jikos/hid

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/hid:
  HID: hid-multitouch: add support for a new Lumio dual-touch panel
  HID: hid-multitouch: correct VID for Stantum panels
  HID: hid-multitouch: ensure slots are initialized

12 years agoMerge branch 'fixes' of master.kernel.org:/home/rmk/linux-2.6-arm
Linus Torvalds [Mon, 27 Jun 2011 15:58:23 +0000 (08:58 -0700)]
Merge branch 'fixes' of /home/rmk/linux-2.6-arm

* 'fixes' of master.kernel.org:/home/rmk/linux-2.6-arm:
  ARM: pm: ensure ARMv7 CPUs save and restore the TLS register
  ARM: pm: proc-v7: fix missing struct processor pointers for suspend code
  ARM: 6969/1: plat-iop: fix build error
  ARM: 6961/1: zImage: Add build-time check for correctly-sized proc_type entries
  ARM: SMP: wait for CPU to be marked active
  ARM: 6963/1: Thumb-2: Relax relocation requirements for non-function symbols
  ARM: 6962/1: mach-h720x: fix build error
  ARM: 6959/1: SMP build fix for entry-macro-multi.S

12 years agoMerge branch 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6
Linus Torvalds [Mon, 27 Jun 2011 15:57:46 +0000 (08:57 -0700)]
Merge branch 'for-linus' of git://git390.marist.edu/linux-2.6

* 'for-linus' of git://git390.marist.edu/pub/scm/linux-2.6:
  [S390] allow setting of upper 32 bit in smp_ctl_set_bit
  [S390] hwsampler: Set a sane default sampling rate
  [S390] s390: enforce HW limits for the initial sampling rate
  [S390] kvm-s390: fix kconfig dependencies

12 years agobtrfs: fix inconsonant inode information
Miao Xie [Thu, 23 Jun 2011 07:27:13 +0000 (07:27 +0000)]
btrfs: fix inconsonant inode information

When iputting the inode, We may leave the delayed nodes if they have some
delayed items that have not been dealt with. So when the inode is read again,
we must look up the relative delayed node, and use the information in it to
initialize the inode. Or we will get inconsonant inode information, it may
cause that the same directory index number is allocated again, and hit the
following oops:

[ 5447.554187] err add delayed dir index item(name: pglog_0.965_0) into the
insertion tree of the delayed node(root id: 262, inode id: 258, errno: -17)
[ 5447.569766] ------------[ cut here ]------------
[ 5447.575361] kernel BUG at fs/btrfs/delayed-inode.c:1301!
[SNIP]
[ 5447.790721] Call Trace:
[ 5447.793191]  [<ffffffffa0641c4e>] btrfs_insert_dir_item+0x189/0x1bb [btrfs]
[ 5447.800156]  [<ffffffffa0651a45>] btrfs_add_link+0x12b/0x191 [btrfs]
[ 5447.806517]  [<ffffffffa0651adc>] btrfs_add_nondir+0x31/0x58 [btrfs]
[ 5447.812876]  [<ffffffffa0651d6a>] btrfs_create+0xf9/0x197 [btrfs]
[ 5447.818961]  [<ffffffff8111f840>] vfs_create+0x72/0x92
[ 5447.824090]  [<ffffffff8111fa8c>] do_last+0x22c/0x40b
[ 5447.829133]  [<ffffffff8112076a>] path_openat+0xc0/0x2ef
[ 5447.834438]  [<ffffffff810c58e2>] ? __perf_event_task_sched_out+0x24/0x44
[ 5447.841216]  [<ffffffff8103ecdd>] ? perf_event_task_sched_out+0x59/0x67
[ 5447.847846]  [<ffffffff81121a79>] do_filp_open+0x3d/0x87
[ 5447.853156]  [<ffffffff811e126c>] ? strncpy_from_user+0x43/0x4d
[ 5447.859072]  [<ffffffff8111f1f5>] ? getname_flags+0x2e/0x80
[ 5447.864636]  [<ffffffff8111f179>] ? do_getname+0x14b/0x173
[ 5447.870112]  [<ffffffff8111f1b7>] ? audit_getname+0x16/0x26
[ 5447.875682]  [<ffffffff8112b1ab>] ? spin_lock+0xe/0x10
[ 5447.880882]  [<ffffffff81112d39>] do_sys_open+0x69/0xae
[ 5447.886153]  [<ffffffff81112db1>] sys_open+0x20/0x22
[ 5447.891114]  [<ffffffff813b9aab>] system_call_fastpath+0x16/0x1b

Fix it by reusing the old delayed node.

Reported-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Miao Xie <miaox@cn.fujitsu.com>
Tested-by: Jim Schutt <jaschut@sandia.gov>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6
Linus Torvalds [Mon, 27 Jun 2011 02:40:31 +0000 (19:40 -0700)]
Merge git://git./linux/kernel/git/sfrench/cifs-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/sfrench/cifs-2.6:
  cifs: mark CONFIG_CIFS_NFSD_EXPORT as BROKEN
  cifs: free blkcipher in smbhash

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6
Linus Torvalds [Mon, 27 Jun 2011 02:39:22 +0000 (19:39 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/viro/vfs-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6:
  cifs: propagate errors from cifs_get_root() to mount(2)
  cifs: tidy cifs_do_mount() up a bit
  cifs: more breakage on mount failures
  cifs: close sget() races
  cifs: pull freeing mountdata/dropping nls/freeing cifs_sb into cifs_umount()
  cifs: move cifs_umount() call into ->kill_sb()
  cifs: pull cifs_mount() call up
  sanitize cifs_umount() prototype
  cifs: initialize ->tlink_tree in cifs_setup_cifs_sb()
  cifs: allocate mountdata earlier
  cifs: leak on mount if we share superblock
  cifs: don't pass superblock to cifs_mount()
  cifs: don't leak nls on mount failure
  cifs: double free on mount failure
  take bdi setup/destruction into cifs_mount/cifs_umount

Acked-by: Steve French <smfrench@gmail.com>
12 years agommc: queue: bring discard_granularity/alignment into line with SCSI
Adrian Hunter [Thu, 23 Jun 2011 10:40:29 +0000 (13:40 +0300)]
mmc: queue: bring discard_granularity/alignment into line with SCSI

SCSI defines discard alignment as the offset to the first
optimal discard.  In the case of SD/MMC, that is always zero
which is the default.

SCSI defines discard granularity as a hint of a optimal
discard size.  That is much better expressed by the MMC
"preferred erase size" (pref_erase) field.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: queue: append partition subname to queue thread name
Adrian Hunter [Thu, 23 Jun 2011 10:40:28 +0000 (13:40 +0300)]
mmc: queue: append partition subname to queue thread name

For example, an eMMC with 2 boot partitions will have 3 threads.
The names change from:

   40 ?        00:00:00 mmcqd/0
   41 ?        00:00:00 mmcqd/0
   42 ?        00:00:00 mmcqd/0

to:

   40 ?        00:00:00 mmcqd/0
   41 ?        00:00:00 mmcqd/0boot0
   42 ?        00:00:00 mmcqd/0boot1

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andrei Warkentin <andreiw@motorola.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: core: make erase timeout calculation allow for gated clock
Adrian Hunter [Thu, 23 Jun 2011 10:40:27 +0000 (13:40 +0300)]
mmc: core: make erase timeout calculation allow for gated clock

The erase timeout calculation may depend on clock rate
which is zero if the clock is gated, so use
mmc_host_clk_rate() which allows for that case.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: block: switch card to User Data Area when removing the block driver
Adrian Hunter [Thu, 23 Jun 2011 10:40:26 +0000 (13:40 +0300)]
mmc: block: switch card to User Data Area when removing the block driver

The MMC block driver and other drivers (e.g. mmc-test) will expect
the card to be switched to the User Data Area eMMC partition when
they start.  Hence the MMC block driver should ensure it is that
way when it is removed.

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Andrei Warkentin <andreiw@motorola.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdio: reset card during power_restore
Daniel Drake [Sat, 25 Jun 2011 18:20:11 +0000 (19:20 +0100)]
mmc: sdio: reset card during power_restore

mmc_sdio_power_restore() skips some steps that are performed in other
power-related codepaths which are necessary to fully reset the card.
Without this, runtime PM fails for SD8686 SDIO wifi on OLPC XO-1.5.

Signed-off-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: cb710: fix #ifdef HAVE_EFFICIENT_UNALIGNED_ACCESS
James Hogan [Tue, 21 Jun 2011 09:55:34 +0000 (10:55 +0100)]
mmc: cb710: fix #ifdef HAVE_EFFICIENT_UNALIGNED_ACCESS

HAVE_EFFICIENT_UNALIGNED_ACCESS is a config option, therefore it needs
the CONFIG_ before it when used by the preprocessor.

Signed-off-by: James Hogan <james@albanarts.com>
Acked-by: Michał Mirosław <mirq-linux@rere.qmqm.pl>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdhi: DMA slave ID 0 is invalid
Guennadi Liakhovetski [Tue, 24 May 2011 10:24:07 +0000 (12:24 +0200)]
mmc: sdhi: DMA slave ID 0 is invalid

Don't try to allocate DMA resources if the platform didn't specify
positive DMA slave IDs.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: tmio: fix regression in TMIO_MMC_WRPROTECT_DISABLE handling
Guennadi Liakhovetski [Mon, 20 Jun 2011 14:51:10 +0000 (16:51 +0200)]
mmc: tmio: fix regression in TMIO_MMC_WRPROTECT_DISABLE handling

Commit b6147490e6aac82 ("mmc: tmio: split core functionality, DMA and
MFD glue") broke handling of the TMIO_MMC_WRPROTECT_DISABLE flag by
the tmio-mmc driver. This patch restores the original behaviour.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Cc: <stable@kernel.org>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: omap_hsmmc: use original sg_len for dma_unmap_sg
Per Forlin [Fri, 17 Jun 2011 18:14:21 +0000 (20:14 +0200)]
mmc: omap_hsmmc: use original sg_len for dma_unmap_sg

Don't use the returned sg_len from dma_map_sg() as inparameter
to dma_unmap_sg(). Use the original sg_len for both dma_map_sg
and dma_unmap_sg according to the documentation in DMA-API.txt.

Signed-off-by: Per Forlin <per.forlin@linaro.org>
Reviewed-by: Venkatraman S <svenkatr@ti.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: omap_hsmmc: fix ocr mask usage
Anand Gadiyar [Tue, 14 Jun 2011 10:29:59 +0000 (15:59 +0530)]
mmc: omap_hsmmc: fix ocr mask usage

The OMAP HSMMC driver uses an ocr_mask to determine the list of voltages
supported by the card. It populates this mask based on the list of
voltages supported by the regulator that supplies the voltage.

Commit 64be97822b (omap4 hsmmc: Update ocr mask for MMC2 for regulator
to use) passed a fixed ocr_mask from the OMAP4 SDP board file to limit
the voltage to 2.9-3.0 Volts, and updated the driver to use this mask
if provided, instead of using the regulator's supported voltages.

However the commit is buggy - the ocr_mask is overridden by the
regulator's capabilities anyway. Fix this.

(The bug shows up when a system-wide suspend is attempted on the OMAP4
SDP/Blaze platforms. The eMMC card comes up at 3V, but drops to 1.65V
after the system resumes).

Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Acked-by: Balaji T K <balajitk@ti.com>
Acked-by: Venkatraman S <svenkatr@ti.com>
Tested-by: Kishore Kadiyala <kishore.kadiyala@ti.com>
Signed-off-by: Sourav Poddar <sourav.poddar@ti.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agommc: sdio: fix runtime PM path during driver removal
Ohad Ben-Cohen [Thu, 9 Jun 2011 23:40:27 +0000 (23:40 +0000)]
mmc: sdio: fix runtime PM path during driver removal

After commit e1866b3 "PM / Runtime: Rework runtime PM handling
during driver removal" was introduced, the driver core stopped
incrementing the runtime PM usage counter of the device during
the invocation of the ->remove() callback.

This indirectly broke SDIO's runtime PM path during driver removal,
because no one calls _put_sync() anymore after ->remove() completes.

This means that the power of runtime-PM-managed SDIO cards is kept
high after their driver is removed (even if it was powered down
beforehand).

Fix that by directly calling _put_sync() when the last usage
counter is downref'ed by the SDIO bus.

Reported-and-tested-by: Daniel Drake <dsd@laptop.org>
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Signed-off-by: Chris Ball <cjb@laptop.org>
12 years agoMerge branch 'master' of /pub/scm/linux/kernel/git/torvalds/linux-2.6
Steve French [Sat, 25 Jun 2011 19:22:28 +0000 (19:22 +0000)]
Merge branch 'master' of /linux/kernel/git/torvalds/linux-2.6

12 years agoMerge branch 'timer-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 25 Jun 2011 14:23:59 +0000 (07:23 -0700)]
Merge branch 'timer-fixes-for-linus' of git://git./linux/kernel/git/tip/linux-2.6-tip

* 'timer-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
  rtc: vt8500: Fix build error & cleanup rtc_class_ops->update_irq_enable()
  alarmtimers: Return -ENOTSUPP if no RTC device is present
  alarmtimers: Handle late rtc module loading

12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6
Linus Torvalds [Sat, 25 Jun 2011 14:23:27 +0000 (07:23 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/tiwai/sound-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6:
  ALSA: Remove unneeded version.h includes from sound/
  ASoC: pxa-ssp: Correct check for stream presence
  ASoC: imx: add missing module informations
  ASoC: imx: Remove unused Kconfig SND_MXC_SOC_SSI entry
  ALSA: HDA: Pinfix quirk for HP Z200 Workstation
  ALSA: VIA HDA: Create a master amplifier control for VT1718S.
  ALSA: VIA HDA: Mute/unmute mixer conncted to Headphone for VT1718S.
  ALSA: VIA HDA: Modify initial verbs list for VT1718S.
  ALSA: hda - Remove ALC268 model override for CPR2000
  ALSA: HDA: Remove quirk for an HP device
  ASoC: Remove unused and about to be broken SND_SOC_CUSTOM I/O bus

12 years agoBtrfs: make sure to update total_bitmaps when freeing cache V3
Josef Bacik [Fri, 24 Jun 2011 16:02:51 +0000 (16:02 +0000)]
Btrfs: make sure to update total_bitmaps when freeing cache V3

A user reported this bug again where we have more bitmaps than we are supposed
to.  This is because we failed to load the free space cache, but don't update
the ctl->total_bitmaps counter when we remove entries from the tree.  This patch
fixes this problem and we should be good to go again.  Thanks,

Signed-off-by: Josef Bacik <josef@redhat.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
12 years agoBtrfs: fix type mismatch in find_free_extent()
Ilya Dryomov [Sat, 18 Jun 2011 20:26:38 +0000 (20:26 +0000)]
Btrfs: fix type mismatch in find_free_extent()

data parameter should be u64 because a full-sized chunk flags field is
passed instead of 0/1 for distinguishing data from metadata.  All
underlying functions expect u64.

Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
12 years agoMerge branch 'fortglx/3.0/tip/timers/rtc' of git://git.linaro.org/people/jstultz...
Thomas Gleixner [Sat, 25 Jun 2011 08:30:00 +0000 (10:30 +0200)]
Merge branch 'fortglx/3.0/tip/timers/rtc' of git://git.linaro.org/people/jstultz/linux into timers/urgent

  * rtc: vt8500: Fix build error & cleanup rtc_class_ops->update_irq_enable()

12 years agoMerge branch 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keith...
Linus Torvalds [Sat, 25 Jun 2011 05:01:40 +0000 (22:01 -0700)]
Merge branch 'drm-intel-fixes' of git://git./linux/kernel/git/keithp/linux-2.6

* 'drm-intel-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/keithp/linux-2.6:
  drm/i915: save/resume forcewake lock fixes
  Revert "drm/i915: Kill GTT mappings when moving from GTT domain"
  drm/i915: Apply HWSTAM workaround for BSD ring on SandyBridge
  drm/i915: Call intel_enable_plane from i9xx_crtc_mode_set (again)

12 years agocifs: propagate errors from cifs_get_root() to mount(2)
Al Viro [Fri, 17 Jun 2011 14:02:59 +0000 (10:02 -0400)]
cifs: propagate errors from cifs_get_root() to mount(2)

... instead of just failing with -EINVAL

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: tidy cifs_do_mount() up a bit
Al Viro [Fri, 17 Jun 2011 13:56:55 +0000 (09:56 -0400)]
cifs: tidy cifs_do_mount() up a bit

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: more breakage on mount failures
Al Viro [Fri, 17 Jun 2011 13:50:44 +0000 (09:50 -0400)]
cifs: more breakage on mount failures

if cifs_get_root() fails, we end up with ->mount() returning NULL,
which is not what callers expect.  Moreover, in case of superblock
reuse we end up leaking a superblock reference...

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: close sget() races
Al Viro [Fri, 17 Jun 2011 13:47:23 +0000 (09:47 -0400)]
cifs: close sget() races

have ->s_fs_info set by the set() callback passed to sget()

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: pull freeing mountdata/dropping nls/freeing cifs_sb into cifs_umount()
Al Viro [Fri, 17 Jun 2011 13:42:43 +0000 (09:42 -0400)]
cifs: pull freeing mountdata/dropping nls/freeing cifs_sb into cifs_umount()

all callers of cifs_umount() proceed to do the same thing; pull it into
cifs_umount() itself.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: move cifs_umount() call into ->kill_sb()
Al Viro [Fri, 17 Jun 2011 13:32:10 +0000 (09:32 -0400)]
cifs: move cifs_umount() call into ->kill_sb()

instead of calling it manually in case if cifs_read_super() fails
to set ->s_root, just call it from ->kill_sb().  cifs_put_super()
is gone now *and* we have cifs_sb shutdown and destruction done
after the superblock is gone from ->s_instances.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: pull cifs_mount() call up
Al Viro [Fri, 17 Jun 2011 13:29:57 +0000 (09:29 -0400)]
cifs: pull cifs_mount() call up

... to the point prior to sget().  Now we have cifs_sb set up early
enough.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agosanitize cifs_umount() prototype
Al Viro [Fri, 17 Jun 2011 13:27:16 +0000 (09:27 -0400)]
sanitize cifs_umount() prototype

a) superblock argument is unused
b) it always returns 0

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: initialize ->tlink_tree in cifs_setup_cifs_sb()
Al Viro [Fri, 17 Jun 2011 13:20:04 +0000 (09:20 -0400)]
cifs: initialize ->tlink_tree in cifs_setup_cifs_sb()

no need to wait until cifs_read_super() and we need it done
by the time cifs_mount() will be called.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: allocate mountdata earlier
Al Viro [Fri, 17 Jun 2011 13:17:28 +0000 (09:17 -0400)]
cifs: allocate mountdata earlier

pull mountdata allocation up, so that it won't stand in the way when
we lift cifs_mount() to location before sget().

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: leak on mount if we share superblock
Al Viro [Fri, 17 Jun 2011 13:14:27 +0000 (09:14 -0400)]
cifs: leak on mount if we share superblock

cifs_sb and nls end up leaked...

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: don't pass superblock to cifs_mount()
Al Viro [Fri, 17 Jun 2011 13:05:48 +0000 (09:05 -0400)]
cifs: don't pass superblock to cifs_mount()

To close sget() races we'll need to be able to set cifs_sb up before
we get the superblock, so we'll want to be able to do cifs_mount()
earlier.  Fortunately, it's easy to do - setting ->s_maxbytes can
be done in cifs_read_super(), ditto for ->s_time_gran and as for
putting MS_POSIXACL into ->s_flags, we can mirror it in ->mnt_cifs_flags
until cifs_read_super() is called.  Kill unused 'devname' argument,
while we are at it...

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: don't leak nls on mount failure
Al Viro [Fri, 17 Jun 2011 12:49:12 +0000 (08:49 -0400)]
cifs: don't leak nls on mount failure

if cifs_sb allocation fails, we still need to drop nls we'd stashed
into volume_info - the one we would've copied to cifs_sb if we could
allocate the latter.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: double free on mount failure
Al Viro [Fri, 17 Jun 2011 12:34:57 +0000 (08:34 -0400)]
cifs: double free on mount failure

if we get to out_super with ->s_root already set (e.g. with
cifs_get_root() failure), we'll end up with cifs_put_super()
called and ->mountdata freed twice.  We'll also get cifs_sb
freed twice and cifs_sb->local_nls dropped twice.  The problem
is, we can get to out_super both with and without ->s_root,
which makes ->put_super() a bad place for such work.

Switch to ->kill_sb(), have all that work done there after
kill_anon_super().  Unlike ->put_super(), ->kill_sb() is
called by deactivate_locked_super() whether we have ->s_root
or not.

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agotake bdi setup/destruction into cifs_mount/cifs_umount
Al Viro [Fri, 17 Jun 2011 12:24:42 +0000 (08:24 -0400)]
take bdi setup/destruction into cifs_mount/cifs_umount

Acked-by: Pavel Shilovsky <piastryyy@gmail.com>
Reviewed-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
12 years agocifs: mark CONFIG_CIFS_NFSD_EXPORT as BROKEN
Jeff Layton [Tue, 21 Jun 2011 11:18:26 +0000 (07:18 -0400)]
cifs: mark CONFIG_CIFS_NFSD_EXPORT as BROKEN

This does not work properly with CIFS as current servers do not
enable support for the FILE_OPEN_BY_FILE_ID on SMB NTCreateX
and not all NFS clients handle ESTALE.

For now, it just plain doesn't work. Mark it BROKEN to discourage
distros from enabling it.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
12 years agoBtrfs: make sure to record the transid in new inodes
Chris Mason [Fri, 24 Jun 2011 17:13:29 +0000 (13:13 -0400)]
Btrfs: make sure to record the transid in new inodes

When we create a new inode, we aren't filling in the
field that records the transaction that last changed this
inode.

If we then go to fsync that inode, it will be skipped because the field
isn't filled in.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
12 years agocifs: free blkcipher in smbhash
Jeff Layton [Mon, 20 Jun 2011 18:33:16 +0000 (14:33 -0400)]
cifs: free blkcipher in smbhash

This is currently leaked in the rc == 0 case.

Reported-by: J. Bruce Fields <bfields@fieldses.org>
Signed-off-by: Jeff Layton <jlayton@redhat.com>
Reviewed-by: Shirish Pargaonkar <shirishpargaonkar@gmail.com>
Signed-off-by: Steve French <sfrench@us.ibm.com>
12 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 24 Jun 2011 15:42:35 +0000 (08:42 -0700)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-block

* 'for-linus' of git://git.kernel.dk/linux-block:
  block: add REQ_SECURE to REQ_COMMON_MASK
  block: use the passed in @bdev when claiming if partno is zero
  block: Add __attribute__((format(printf...) and fix fallout
  block: make disk_block_events() properly wait for work cancellation
  block: remove non-syncing __disk_block_events() and fold it into disk_block_events()
  block: don't use non-syncing event blocking in disk_check_events()
  cfq-iosched: fix locking around ioc->ioc_data assignment