GitHub/moto-9609/android_kernel_motorola_exynos9610.git
12 years agonfs4.1: add BIND_CONN_TO_SESSION operation
Weston Andros Adamson [Thu, 24 May 2012 17:22:50 +0000 (13:22 -0400)]
nfs4.1: add BIND_CONN_TO_SESSION operation

This patch adds the BIND_CONN_TO_SESSION operation which is needed for
upcoming SP4_MACH_CRED work and useful for recovering from broken connections
without destroying the session.

Signed-off-by: Weston Andros Adamson <dros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 test the mdsthreshold hint parameters
Andy Adamson [Wed, 23 May 2012 09:02:37 +0000 (05:02 -0400)]
NFSv4.1 test the mdsthreshold hint parameters

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 add nfs_inode book keeping for mdsthreshold
Andy Adamson [Thu, 24 May 2012 17:13:24 +0000 (13:13 -0400)]
NFSv4.1 add nfs_inode book keeping for mdsthreshold

Keep track of the number of bytes read or written via buffered, direct, and
mem-mapped i/o for use by mdsthreshold size_io hints.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 cache mdsthreshold values on OPEN
Andy Adamson [Wed, 23 May 2012 09:02:35 +0000 (05:02 -0400)]
NFSv4.1 cache mdsthreshold values on OPEN

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 mdsthreshold attribute xdr
Andy Adamson [Wed, 23 May 2012 09:02:34 +0000 (05:02 -0400)]
NFSv4.1 mdsthreshold attribute xdr

We only support one layout type per file system, so one threshold_item4 per
mdsthreshold4.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Add memory barriers to the nfs_client->cl_cons_state initialisation
Trond Myklebust [Wed, 23 May 2012 17:26:10 +0000 (13:26 -0400)]
NFS: Add memory barriers to the nfs_client->cl_cons_state initialisation

Ensure that a process that uses the nfs_client->cl_cons_state test
for whether the initialisation process is finished does not read
stale data.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4: Fix a race in the net namespace mount notification
Trond Myklebust [Wed, 23 May 2012 17:24:36 +0000 (13:24 -0400)]
NFSv4: Fix a race in the net namespace mount notification

Since the struct nfs_client gets added to the global nfs_client_list
before it is initialised, it is possible that rpc_pipefs_event can
end up trying to create idmapper entries on such a thing.

The solution is to have the mount notification wait for the
initialisation of each nfs_client to complete, and then to
skip any entries for which the it failed.

Reported-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Acked-by: Stanislav Kinsbursky <skinsbursky@parallels.com>
12 years agoNFSv4.1: Fix session initialisation races
Trond Myklebust [Wed, 23 May 2012 17:23:31 +0000 (13:23 -0400)]
NFSv4.1: Fix session initialisation races

Session initialisation is not complete until the lease manager
has run. We need to ensure that both nfs4_init_session and
nfs4_init_ds_session do so, and that they check for any resulting
errors in clp->cl_cons_state.

Only after this is done, can nfs4_ds_connect check the contents
of clp->cl_exchange_flags.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Cc: Andy Adamson <andros@netapp.com>
12 years agoNFS: EXCHANGE_ID should save the server major and minor ID
Chuck Lever [Tue, 22 May 2012 02:46:16 +0000 (22:46 -0400)]
NFS: EXCHANGE_ID should save the server major and minor ID

Save the server major and minor ID results from EXCHANGE_ID, as they
are needed for detecting server trunking.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Add nfs_client behavior flags
Chuck Lever [Tue, 22 May 2012 02:46:07 +0000 (22:46 -0400)]
NFS: Add nfs_client behavior flags

"noresvport" and "discrtry" can be passed to nfs_create_rpc_client()
by setting flags in the passed-in nfs_client.  This change makes it
easy to add new flags.

Note that these settings are now "sticky" over the lifetime of a
struct nfs_client, and may even be copied when an nfs_client is
cloned.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Refactor nfs_get_client(): initialize nfs_client
Chuck Lever [Tue, 22 May 2012 02:45:59 +0000 (22:45 -0400)]
NFS: Refactor nfs_get_client(): initialize nfs_client

Clean up: Continue to rationalize the locking in nfs_get_client() by
moving the logic that handles the case where a matching server IP
address is not found.

When we support server trunking detection, client initialization may
return a different nfs_client struct than was passed to it.  Change
the synopsis of the init_client methods to return an nfs_client.

The client initialization logic in nfs_get_client() is not much more
than a wrapper around ->init_client.  It's simpler to keep the little
bits of error handling in the version-specific init_client methods.

No behavior change is expected.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Refactor nfs_get_client(): add nfs_found_client()
Chuck Lever [Tue, 22 May 2012 02:45:50 +0000 (22:45 -0400)]
NFS: Refactor nfs_get_client(): add nfs_found_client()

Clean up: Code that takes and releases nfs_client_lock remains in
nfs_get_client().  Logic that handles a pre-existing nfs_client is
moved to a separate function.

No behavior change is expected.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Always use the same SETCLIENTID boot verifier
Chuck Lever [Tue, 22 May 2012 02:45:41 +0000 (22:45 -0400)]
NFS: Always use the same SETCLIENTID boot verifier

Currently our NFS client assigns a unique SETCLIENTID boot verifier
for each server IP address it knows about.  It's set to CURRENT_TIME
when the struct nfs_client for that server IP is created.

During the SETCLIENTID operation, our client also presents an
nfs_client_id4 string to servers, as an identifier on which the server
can hang all of this client's NFSv4 state.  Our client's
nfs_client_id4 string is unique for each server IP address.

An NFSv4 server is obligated to wipe all NFSv4 state associated with
an nfs_client_id4 string when the client presents the same
nfs_client_id4 string along with a changed SETCLIENTID boot verifier.

When our client unmounts the last of a server's shares, it destroys
that server's struct nfs_client.  The next time the client mounts that
NFS server, it creates a fresh struct nfs_client with a fresh boot
verifier.  On seeing the fresh verifer, the server wipes any previous
NFSv4 state associated with that nfs_client_id4.

However, NFSv4.1 clients are supposed to present the same
nfs_client_id4 string to all servers.  And, to support Transparent
State Migration, the same nfs_client_id4 string should be presented
to all NFSv4.0 servers so they recognize that migrated state for this
client belongs with state a server may already have for this client.
(This is known as the Uniform Client String model).

If the nfs_client_id4 string is the same but the boot verifier changes
for each server IP address, SETCLIENTID and EXCHANGE_ID operations
from such a client could unintentionally result in a server wiping a
client's previously obtained lease.

Thus, if our NFS client is going to use a fixed nfs_client_id4 string,
either for NFSv4.0 or NFSv4.1 mounts, our NFS client should use a
boot verifier that does not change depending on server IP address.
Replace our current per-nfs_client boot verifier with a per-nfs_net
boot verifier.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Force server to drop NFSv4 state
Chuck Lever [Tue, 22 May 2012 02:45:33 +0000 (22:45 -0400)]
NFS: Force server to drop NFSv4 state

nfs4_reset_all_state() refreshes the boot verifier a server sees to
trigger that server to wipe this client's state.  This function is
invoked when an NFSv4.1 server reports that it has revoked some or
all of a client's NFSv4 state.

To facilitate server trunking discovery, we will eventually want to
move the cl_boot_time field to a more global structure.  The Uniform
Client String model (and specifically, server trunking detection)
requires that all servers see the same boot verifier until the client
actually does reboot, and not a fresh verifier every time the client
unmounts and remounts the server.

Without the cl_boot_time field, however, nfs4_reset_all_state() will
have to find some other way to force the server to purge the client's
NFSv4 state.

Because these verifiers are opaque (ie, the server doesn't know or
care that they happen to be timestamps), we can force the server
to wipe NFSv4 state by updating the boot verifier as we do now, then
immediately afterwards establish a fresh client ID using the old boot
verifier again.

Hopefully there are no extra paranoid server implementations that keep
track of the client's boot verifiers and prevent clients from reusing
a previous one.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Remove nfs_unique_id
Chuck Lever [Tue, 22 May 2012 02:45:08 +0000 (22:45 -0400)]
NFS: Remove nfs_unique_id

Clean up:  this structure is unused.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Clean up return code checking in nfs4_proc_exchange_id()
Chuck Lever [Tue, 22 May 2012 02:44:58 +0000 (22:44 -0400)]
NFS: Clean up return code checking in nfs4_proc_exchange_id()

Clean up: update to use matching types in "if" expressions.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Use proper naming conventions for the nfs_client.net field
Chuck Lever [Tue, 22 May 2012 02:44:50 +0000 (22:44 -0400)]
NFS: Use proper naming conventions for the nfs_client.net field

Clean up:  When naming fields and data types, follow established
conventions to facilitate accurate grep/cscope searches.

Introduced by commit e50a7a1a "NFS: make NFS client allocated per
network namespace context," Tue Jan 10, 2012.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Use proper naming conventions for nfs_client.impl_id field
Chuck Lever [Tue, 22 May 2012 02:44:41 +0000 (22:44 -0400)]
NFS: Use proper naming conventions for nfs_client.impl_id field

Clean up:  When naming fields and data types, follow established
conventions to facilitate accurate grep/cscope searches.

Additionally, for consistency, move the impl_id field into the NFSv4-
specific part of the nfs_client, and free that memory in the logic
that shuts down NFSv4 nfs_clients.

Introduced by commit 7d2ed9ac "NFSv4: parse and display server
implementation ids," Fri Feb 17, 2012.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Use proper naming conventions for NFSv4.1 server scope fields
Chuck Lever [Tue, 22 May 2012 02:44:31 +0000 (22:44 -0400)]
NFS: Use proper naming conventions for NFSv4.1 server scope fields

Clean up:  When naming fields and data types, follow established
conventions to facilitate accurate grep/cscope searches.

Additionally, for consistency, move the scope field into the NFSv4-
specific part of the nfs_client, and free that memory in the logic
that shuts down NFSv4 nfs_clients.

Introduced by commit 99fe60d0 "nfs41: exchange_id operation", April
1 2009.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Fix comment misspelling in struct nfs_client definition
Chuck Lever [Tue, 22 May 2012 02:44:22 +0000 (22:44 -0400)]
NFS: Fix comment misspelling in struct nfs_client definition

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Add NFSDBG_STATE
Chuck Lever [Tue, 22 May 2012 02:45:24 +0000 (22:45 -0400)]
NFS: Add NFSDBG_STATE

fs/nfs/nfs4state.c does not yet have any dprintk() call sites, and I'm
about to introduce some.  We will need a new flag for enabling them.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Don't swap bytes in nfs4_construct_boot_verifier()
Chuck Lever [Tue, 22 May 2012 02:45:16 +0000 (22:45 -0400)]
NFS: Don't swap bytes in nfs4_construct_boot_verifier()

The SETCLIENTID boot verifier is opaque to NFSv4 servers, thus there
is no requirement for byte swapping before the client puts the
verifier on the wire.

This treatment is similar to other timestamp-based verifiers.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Fix compiler warnings
Bryan Schumaker [Tue, 22 May 2012 14:10:03 +0000 (10:10 -0400)]
NFS: Fix compiler warnings

The "struct inode *inode" was only used in a dprintk, so compiling with
CONFIG_SUNRPC_DEBUG off triggers a warning.  To get around this, I
remove the "struct inode *inode" variable and instead change the
dprintk()s to use hdr->inode instead.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 skip rpc_call_done only on disconnected DS slot_table_waitq tasks
Andy Adamson [Tue, 22 May 2012 12:09:27 +0000 (08:09 -0400)]
NFSv4.1 skip rpc_call_done only on disconnected DS slot_table_waitq tasks

We reset all I/O on a disconnected data server through the pgio layer indicated
by the NFS_IOHDR_REDO flag.

Differentiate between on-the-wire tasks returning with an error which must
call rpc_call_done and tasks woken from the data server slot_table_waitq
waiting for a session slot with a status of zero which call rpc_exit in
rpc_prepare and need to skip rpc_call_done.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 Just use nfs_put_client in filelayout release
Andy Adamson [Tue, 22 May 2012 12:09:26 +0000 (08:09 -0400)]
NFSv4.1 Just use nfs_put_client in filelayout release

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 fix null state reference in filelayout_async_handle_error
Andy Adamson [Tue, 22 May 2012 12:09:28 +0000 (08:09 -0400)]
NFSv4.1 fix null state reference in filelayout_async_handle_error

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: Fix a bad reference count issue in the pNFS commit code
Trond Myklebust [Tue, 22 May 2012 20:36:27 +0000 (16:36 -0400)]
NFSv4.1: Fix a bad reference count issue in the pNFS commit code

filelayout_scan_commit_lists needs to bump the reference count on
the struct nfs_page just like nfs_scan_commit_list().

Reported-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoMerge branch 'bugfixes' into nfs-for-next
Trond Myklebust [Mon, 21 May 2012 14:12:39 +0000 (10:12 -0400)]
Merge branch 'bugfixes' into nfs-for-next

12 years agoNFSv4.1 resend LAYOUTGET on data server invalid layout errors
Andy Adamson [Fri, 27 Apr 2012 21:53:53 +0000 (17:53 -0400)]
NFSv4.1 resend LAYOUTGET on data server invalid layout errors

The "invalid layout" class of errors is handled by destroying the layout and
getting a new layout from the server.  Currently, the layout must be
destroyed before a new layout can be obtained.

This means that all references (e.g.lsegs) to the "to be destroyed" layout
header must be dropped before it can be destroyed. This in turn means waiting
for all in flight RPC's using the old layout as well as draining the data
server session slot table wait queue.

Set the NFS_LAYOUT_INVALID flag to redirect I/O to the MDS while waiting for
the old layout to be destroyed.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 dereference a disconnected data server client record
Andy Adamson [Fri, 27 Apr 2012 21:53:52 +0000 (17:53 -0400)]
NFSv4.1 dereference a disconnected data server client record

When the last DS io is processed, the data server client record will be
freed.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 ref count nfs_client across filelayout data server io
Andy Adamson [Fri, 27 Apr 2012 21:53:51 +0000 (17:53 -0400)]
NFSv4.1 ref count nfs_client across filelayout data server io

Prepare to put a dis-connected DS client record.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 send layoutreturn to fence disconnected data server
Andy Adamson [Fri, 27 Apr 2012 21:53:50 +0000 (17:53 -0400)]
NFSv4.1 send layoutreturn to fence disconnected data server

Let the MDS know that you are redirecting I/O from pNFS to MDS.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 wake up all tasks on un-connected DS slot table waitq
Andy Adamson [Fri, 27 Apr 2012 21:53:49 +0000 (17:53 -0400)]
NFSv4.1 wake up all tasks on un-connected DS slot table waitq

The DS has a connection error (invalid deviceid). Drain the fore channel
slot table waitq.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 Check invalid deviceid upon slot table waitq wakeup
Andy Adamson [Fri, 27 Apr 2012 21:53:48 +0000 (17:53 -0400)]
NFSv4.1 Check invalid deviceid upon slot table waitq wakeup

Tasks sleeping on the slot table waitq wake to the rpc_prepare_task state.
Reset the task for io through the MDS if the deviceid is invalid.

The reset functions put the io pages through the pageio layer which has the
advantage of re-coalescing which allows for the MDS and DS having different
r/wsizes. Exit the awakened task without executing the rpc_call_done routine.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 remove nfs4_reset_write and nfs4_reset_read
Andy Adamson [Fri, 27 Apr 2012 21:53:47 +0000 (17:53 -0400)]
NFSv4.1 remove nfs4_reset_write and nfs4_reset_read

Replaced by filelayout_reset_write and filelayout_reset_read

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: mark deviceid invalid on filelayout DS connection errors
Andy Adamson [Fri, 27 Apr 2012 21:53:46 +0000 (17:53 -0400)]
NFSv4.1: mark deviceid invalid on filelayout DS connection errors

This prevents the use of any layout for i/o that references the deviceid.
I/O is redirected through the MDS.

Redirect the unhandled failed I/O to the MDS without marking either the
layout or the deviceid invalid.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 data server timeo and retrans module parameters
Andy Adamson [Fri, 27 Apr 2012 21:53:45 +0000 (17:53 -0400)]
NFSv4.1 data server timeo and retrans module parameters

Set the recovery parameters for data servers.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 set RPC_TASK_SOFTCONN for filelayout DS RPC calls
Andy Adamson [Fri, 27 Apr 2012 21:53:44 +0000 (17:53 -0400)]
NFSv4.1 set RPC_TASK_SOFTCONN for filelayout DS RPC calls

RPC_TASK_SOFTCONN returns connection errors to the caller which allows the pNFS
file layout to quickly try the MDS or perhaps another DS.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1 cleanup filelayout invalid layout handling
Andy Adamson [Fri, 27 Apr 2012 21:53:43 +0000 (17:53 -0400)]
NFSv4.1 cleanup filelayout invalid layout handling

The invalid layout bits are should only be used to block LAYOUTGETs.

Do not invalidate a layout on deviceid invalidation.
Do not invalidate a layout on un-handled READ, WRITE, COMMIT errors.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFSv4.1: cleanup filelayout invalid deviceid handling
Andy Adamson [Fri, 27 Apr 2012 21:53:42 +0000 (17:53 -0400)]
NFSv4.1: cleanup filelayout invalid deviceid handling

Move the invalid deviceid test into nfs4_fl_prepare_ds, called by the
filelayout read, write, and commit routines. NFS4_DEVICE_ID_NEG_ENTRY
is no longer needed.
Remove redundant printk's - filelayout_mark_devid_invalid prints a KERN_WARNING.

An invalid device prevents pNFS io.

Signed-off-by: Andy Adamson <andros@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoFixed goto readability in nfs_update_inode.
Matthew Treinish [Mon, 30 Apr 2012 15:32:57 +0000 (11:32 -0400)]
Fixed goto readability in nfs_update_inode.

Simplified error gotos to make it slightly easier to read,
it doesn't affect the functionality of the routine.

Signed-off-by: Matthew Treinish <treinish@linux.vnet.ibm.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agosunrpc: fix loss of task->tk_status after rpc_delay call in xprt_alloc_slot
Trond Myklebust [Sat, 19 May 2012 16:12:53 +0000 (12:12 -0400)]
sunrpc: fix loss of task->tk_status after rpc_delay call in xprt_alloc_slot

xprt_alloc_slot will call rpc_delay() to make the task wait a bit before
retrying when it gets back an -ENOMEM error from xprt_dynamic_alloc_slot.
The problem is that rpc_delay will clear the task->tk_status, causing
call_reserveresult to abort the task.

The solution is simply to let call_reserveresult handle the ENOMEM error
directly.

Reported-by: Jeff Layton <jlayton@redhat.com>
Cc: stable@vger.kernel.org [>= 3.1]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agosunrpc: suppress page allocation warnings in xprt_alloc_slot()
Jeff Layton [Wed, 16 May 2012 17:30:35 +0000 (13:30 -0400)]
sunrpc: suppress page allocation warnings in xprt_alloc_slot()

It's easily possible for these allocations to fail since we're using
GFP_NOWAIT here. We don't want to spam the logs with warnings about
that though.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: fix unsigned comparison in nfs4_create_sec_client
Benny Halevy [Wed, 16 May 2012 08:35:36 +0000 (11:35 +0300)]
NFS: fix unsigned comparison in nfs4_create_sec_client

fs/nfs/nfs4namespace.c: In function ‘nfs4_create_sec_client’:
fs/nfs/nfs4namespace.c:171:2: error: comparison of unsigned expression < 0 is always false [-Werror=type-limits]

Introduced by commit 72de53ec4bca39c26709122a8f78bfefe7b6bca4
"NFS: Do secinfo as part of lookup"

Signed-off-by: Benny Halevy <bhalevy@tonian.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Fix a compile issue when CONFIG_NFS_FSCACHE was undefined
Trond Myklebust [Wed, 16 May 2012 17:21:30 +0000 (10:21 -0700)]
NFS: Fix a compile issue when CONFIG_NFS_FSCACHE was undefined

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agorpc_pipefs: clear write bit from top level rpc_pipefs directory
Jeff Layton [Wed, 16 May 2012 17:04:26 +0000 (13:04 -0400)]
rpc_pipefs: clear write bit from top level rpc_pipefs directory

We can't create new files or directories here from userspace, so let's
not pretend that this directory is writable.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agosunrpc: fix kernel-doc warnings
Randy Dunlap [Sun, 13 May 2012 17:35:40 +0000 (10:35 -0700)]
sunrpc: fix kernel-doc warnings

Fix kernel-doc warnings in sunrpc/rpc_pipe.c and
sunrpc/rpcb_clnt.c:

Warning(net/sunrpc/rpcb_clnt.c:428): No description found for parameter 'net'
Warning(net/sunrpc/rpcb_clnt.c:567): No description found for parameter 'net'

Warning(net/sunrpc/rpc_pipe.c:133): No description found for parameter 'pipe'
Warning(net/sunrpc/rpc_pipe.c:133): Excess function parameter 'inode' description in 'rpc_queue_upcall'
Warning(net/sunrpc/rpc_pipe.c:839): No description found for parameter 'pipe'
Warning(net/sunrpc/rpc_pipe.c:839): Excess function parameter 'ops' description in 'rpc_mkpipe_dentry'
Warning(net/sunrpc/rpc_pipe.c:839): Excess function parameter 'flags' description in 'rpc_mkpipe_dentry'
Warning(net/sunrpc/rpc_pipe.c:949): No description found for parameter 'dentry'
Warning(net/sunrpc/rpc_pipe.c:949): Excess function parameter 'clnt' description in 'rpc_remove_client_dir'

Signed-off-by: Randy Dunlap <rdunlap@xenotime.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: kmalloc() doesn't return an ERR_PTR()
Dan Carpenter [Mon, 14 May 2012 19:45:28 +0000 (22:45 +0300)]
NFS: kmalloc() doesn't return an ERR_PTR()

Obviously we should check for NULL here instead of IS_ERR().

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Cc: stable@vger.kernel.org [3.4]
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Turn v3 on by default
Bryan Schumaker [Thu, 10 May 2012 20:47:20 +0000 (16:47 -0400)]
NFS: Turn v3 on by default

Most users will use NFS v3 or possibly v4 so this makes it easier for
them.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Make v2 configurable
Bryan Schumaker [Thu, 10 May 2012 20:47:19 +0000 (16:47 -0400)]
NFS: Make v2 configurable

With this patch NFS v2 can be disabled during Kconfig.  I default the
option to "y" to match the current behavior.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create an NFS v3 stat_to_errno()
Bryan Schumaker [Thu, 10 May 2012 20:47:18 +0000 (16:47 -0400)]
NFS: Create an NFS v3 stat_to_errno()

In theory, NFS v3 can have different error versions than NFS v2. v4 is
already using its own nfs4_stat_to_errno() to map error codes, so
rather than create something in the generic client for v2 and v3 to
share I instead give v3 its own function.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Pass mntfh as part of the nfs_mount_info structure
Bryan Schumaker [Thu, 10 May 2012 19:07:43 +0000 (15:07 -0400)]
NFS: Pass mntfh as part of the nfs_mount_info structure

This allows me to use the filehandle allocated in nfs_fs_mount() for nfs
v4 mounts instead of allocating a new one.  Rather than change
nfs4_mount() to look almost exactly like nfs_fs_mount(), I instead
remove the function.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Allocate parsed mount data directly to the nfs_mount_info structure
Bryan Schumaker [Thu, 10 May 2012 19:07:42 +0000 (15:07 -0400)]
NFS: Allocate parsed mount data directly to the nfs_mount_info structure

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a single nfs_validate_mount_data() function
Bryan Schumaker [Thu, 10 May 2012 19:07:41 +0000 (15:07 -0400)]
NFS: Create a single nfs_validate_mount_data() function

This new function chooses between the v2/3 parser and the v4 parser by
filesystem type.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a single function for text mount data
Bryan Schumaker [Thu, 10 May 2012 19:07:40 +0000 (15:07 -0400)]
NFS: Create a single function for text mount data

The v2/3 and v4 cases were very similar, with just a few parameters
changed.  This makes it easy to share code.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a new nfs_try_mount()
Bryan Schumaker [Thu, 10 May 2012 19:07:39 +0000 (15:07 -0400)]
NFS: Create a new nfs_try_mount()

This function returns the same same return type as nfs4_try_mount() so
they two can be more easily substituted.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Let mount data parsing set the NFS version
Bryan Schumaker [Thu, 10 May 2012 19:07:38 +0000 (15:07 -0400)]
NFS: Let mount data parsing set the NFS version

This field is unconditionally set while parsing mount data, so there is
no need to fill it in here.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Use nfs_fs_mount_common() for remote referral mounts
Bryan Schumaker [Thu, 10 May 2012 19:07:37 +0000 (15:07 -0400)]
NFS: Use nfs_fs_mount_common() for remote referral mounts

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Use nfs_fs_mount_common() for xdev mounts
Bryan Schumaker [Thu, 10 May 2012 19:07:36 +0000 (15:07 -0400)]
NFS: Use nfs_fs_mount_common() for xdev mounts

At this point, there are only a few small differences between these two
functions.  I can set a few function pointers in the nfs_mount_info
struct to get around these differences.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a common xdev_mount() function
Bryan Schumaker [Thu, 10 May 2012 19:07:35 +0000 (15:07 -0400)]
NFS: Create a common xdev_mount() function

The only difference between nfs_xdev_mount() and nfs4_xdev_mount() is the
clone_super() function called to clone the super block.  I can combine
these two functions by using the fill_super field in the mount_info
structure.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a common fs_mount() function
Bryan Schumaker [Thu, 10 May 2012 19:07:34 +0000 (15:07 -0400)]
NFS: Create a common fs_mount() function

The nfs4_remote_mount() function was only slightly different from the
nfs_fs_mount() function used by the generic client.  I created a new
nfs_mount_info structure to set different parameters to help combine
these functions.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Remove NFS4_MOUNT_UNSHARED
Bryan Schumaker [Thu, 10 May 2012 19:07:33 +0000 (15:07 -0400)]
NFS: Remove NFS4_MOUNT_UNSHARED

This flag is numerically equivalent to NFS_MOUNT_UNSHARED, so I can
remove it to make collapsing functions more straightforward.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Don't pass mount data to nfs_fscache_get_super_cookie()
Bryan Schumaker [Thu, 10 May 2012 19:07:32 +0000 (15:07 -0400)]
NFS: Don't pass mount data to nfs_fscache_get_super_cookie()

I intend on creating a single nfs_fs_mount() function used by all our
mount paths.  To avoid checking between new mounts and clone mounts, I
instead pass both structures to a new function in super.c that finds the
cache key and then looks up the super cookie.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Create a single nfs_get_root()
Bryan Schumaker [Thu, 10 May 2012 19:07:31 +0000 (15:07 -0400)]
NFS: Create a single nfs_get_root()

This patch splits out the NFS v4 specific functionality of
nfs4_get_root() into its own rpc_op called by the generic client, and
leaves nfs4_proc_get_rootfh() as its own stand alone function.  This
also allows me to change nfs4_remote_mount(), nfs4_xdev_mount() and
nfs4_remote_referral_mount() to use the generic client's nfs_get_root()
function.  Later patches in this series will collapse these functions
into one common function, so using the same get_root() function
everywhere simplifies future changes.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoNFS: Rename nfs4_proc_get_root()
Bryan Schumaker [Thu, 10 May 2012 19:07:30 +0000 (15:07 -0400)]
NFS: Rename nfs4_proc_get_root()

This function is really getting the root filehandle and not the root
dentry of the filesystem.  I also removed the rpc_ops lookup from
nfs4_get_rootfh() under the assumption that if we reach this function
then we already know we are using NFS v4.

Signed-off-by: Bryan Schumaker <bjschuma@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
12 years agoLinux 3.4-rc7
Linus Torvalds [Sun, 13 May 2012 01:37:47 +0000 (18:37 -0700)]
Linux 3.4-rc7

.. and this should hopefully be the last -rc before final 3.4 release.

12 years agoMerge tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm...
Linus Torvalds [Sun, 13 May 2012 00:27:41 +0000 (17:27 -0700)]
Merge tag 'fixes-for-linus' of git://git./linux/kernel/git/arm/arm-soc

Pull ARM: SoC fixes from Olof Johansson:
 "I was hoping to be done with fixes for 3.4 but we got two branches
  from subarch maintainers the last couple of days.  So here is one
  last(?) pull request for arm-soc containing 7 patches:

   - Five of them are for shmobile dealing with SMP setup and compile
     failures
   - The remaining two are for regressions on the Samsung platforms"

* tag 'fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/arm/arm-soc:
  ARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1
  ARM: EXYNOS: use s5p-timer for UniversalC210 board
  ARM / mach-shmobile: Invalidate caches when booting secondary cores
  ARM / mach-shmobile: sh73a0 SMP TWD boot regression fix
  ARM / mach-shmobile: r8a7779 SMP TWD boot regression fix
  ARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper
  ARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper

12 years agoMerge tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Sun, 13 May 2012 00:24:29 +0000 (17:24 -0700)]
Merge tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6

Pull a few more GPIO bug fixes from Grant Likely:
 "Oops, missed a couple.  Here's an updated pull req for GPIO"

A set of PCH bug fixes, and one patch to fix up compile warnings

* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  gpio/exynos: Fix compiler warnings when non-exynos machines are selected
  gpio: pch9: Use proper flow type handlers

12 years agoMerge branch 'v3.4-samsung-fixes-5' of git://git.kernel.org/pub/scm/linux/kernel...
Olof Johansson [Sat, 12 May 2012 22:41:22 +0000 (15:41 -0700)]
Merge branch 'v3.4-samsung-fixes-5' of git://git./linux/kernel/git/kgene/linux-samsung into fixes

* 'v3.4-samsung-fixes-5' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1
  ARM: EXYNOS: use s5p-timer for UniversalC210 board

12 years agoARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1
Kukjin Kim [Sat, 12 May 2012 07:45:47 +0000 (16:45 +0900)]
ARM: EXYNOS: fix ctrlbit for exynos5_clk_pdma1

It should be (1 << 2) for ctrlbit of exynos5_clk_pdma1.

Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
12 years agoARM: EXYNOS: use s5p-timer for UniversalC210 board
Marek Szyprowski [Fri, 11 May 2012 21:17:59 +0000 (06:17 +0900)]
ARM: EXYNOS: use s5p-timer for UniversalC210 board

Commit 069d4e743 ("ARM: EXYNOS4: Remove clock event timers using
ARM private timers") removed support for local timers and forced
to use MCT as event source. However MCT is not operating properly
on early revision of EXYNOS4 SoCs. All UniversalC210 boards are
based on it, so that commit broke support for it. This patch
provides a workaround that enables UniversalC210 boards to boot
again. s5p-timer is used as an event source, it works only for
non-SMP builds.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Kyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
12 years agoMerge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas...
Olof Johansson [Sat, 12 May 2012 22:40:56 +0000 (15:40 -0700)]
Merge branch 'fixes' of git://git./linux/kernel/git/rafael/renesas into fixes

By Guennadi Liakhovetski (2) and others via Rafael J. Wysocki:
"[...] urgent fixes for Renesas ARM-based platforms.  Four of these
commits are fixes of regressions new in 3.4-rc and the last one is
necessary for SMP to work on those systems in general."

* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/renesas:
  ARM / mach-shmobile: Invalidate caches when booting secondary cores
  ARM / mach-shmobile: sh73a0 SMP TWD boot regression fix
  ARM / mach-shmobile: r8a7779 SMP TWD boot regression fix
  ARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper
  ARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper

12 years agoARM / mach-shmobile: Invalidate caches when booting secondary cores
Magnus Damm [Wed, 9 May 2012 07:24:59 +0000 (16:24 +0900)]
ARM / mach-shmobile: Invalidate caches when booting secondary cores

Make sure L1 caches are invalidated when booting secondary
cores. Needed to boot all mach-shmobile SMP systems that
are using Cortex-A9 including sh73a0, r8a7779 and EMEV2.

Thanks to imx and tegra guys for actual code.

Signed-off-by: Magnus Damm <damm@opensource.se>
Tested-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoARM / mach-shmobile: sh73a0 SMP TWD boot regression fix
Kuninori Morimoto [Thu, 10 May 2012 07:26:58 +0000 (00:26 -0700)]
ARM / mach-shmobile: sh73a0 SMP TWD boot regression fix

Fix SMP TWD boot regression on sh73a0 based platforms caused by:

4200b16 ARM: shmobile: convert to twd_local_timer_register() interface

After the merge of the above commit it has been impossible to boot
sh73a0 based SoCs with SMP enabled and CONFIG_HAVE_ARM_TWD=y. The
kernel crashes at smp_init_cpus() timing which is before the console
has been initialized, so to the user this looks like a kernel lock up
without any particular error message.

This patch fixes the regression on sh73a0 by moving the TWD
registration code from smp_init_cpus() to sys_timer->init() time.

This patch removed shmobile_twd_init() which is no longer needed

Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoARM / mach-shmobile: r8a7779 SMP TWD boot regression fix
Magnus Damm [Thu, 10 May 2012 05:57:22 +0000 (14:57 +0900)]
ARM / mach-shmobile: r8a7779 SMP TWD boot regression fix

Fix SMP TWD boot regression on r8a7779 based platforms caused by:

4200b16 ARM: shmobile: convert to twd_local_timer_register() interface

After the merge of the above commit it has been impossible to boot
r8a7779 based SoCs with SMP enabled and CONFIG_HAVE_ARM_TWD=y. The
kernel crashes at smp_init_cpus() timing which is before the console
has been initialized, so to the user this looks like a kernel lock up
without any particular error message.

This patch fixes the regression on r8a7779 by moving the TWD
registration code from smp_init_cpus() to sys_timer->init() time.

Signed-off-by: Magnus Damm <damm@opensource.se>
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper
Guennadi Liakhovetski [Mon, 16 Apr 2012 21:09:19 +0000 (23:09 +0200)]
ARM: mach-shmobile: convert ag5evm to use the generic MMC GPIO hotplug helper

This also fixes the following modular mmc build failure:

arch/arm/mach-shmobile/built-in.o: In function `mackerel_sdhi0_gpio_cd':
pfc-sh7372.c:(.text+0x1138): undefined reference to `mmc_detect_change'

on this platform by eliminating the use of an inline function, which
calls into the mmc core.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Reviewed-by: Simon Horman <horms@verge.net.au>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper
Guennadi Liakhovetski [Mon, 16 Apr 2012 21:09:13 +0000 (23:09 +0200)]
ARM: mach-shmobile: convert mackerel to use the generic MMC GPIO hotplug helper

This also fixes the following modular mmc build failure:

arch/arm/mach-shmobile/built-in.o: In function `ag5evm_sdhi0_gpio_cd':
pfc-sh73a0.c:(.text+0x7c0): undefined reference to `mmc_detect_change'

on this platform by eliminating the use of an inline function, which
calls into the mmc core.

Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Tested-by: Simon Horman <horms@verge.net.au>
Acked-by: Magnus Damm <damm@opensource.se>
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
12 years agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sat, 12 May 2012 20:02:31 +0000 (13:02 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "This is a set of minor qla and virto fixes plus one major regression
  fix (oops in all legacy host drivers)."

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  [SCSI] virtio_scsi: fix TMF use-after-free
  [SCSI] fix oops in all legacy host adapters caused by 6f381fa
  [SCSI] qla2xxx: Update version number to 8.04.00.03-k.
  [SCSI] qla2xxx: Properly check for current state after the fabric-login request.
  [SCSI] qla2xxx: Proper completion to scsi-ml for scsi status task_set_full and busy.
  [SCSI] qla2xxx: Block flash access from application when device is initialized for ISP82xx.
  [SCSI] qla2xxx: Fix reset time out as qla2xxx not ack to reset request.

12 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Sat, 12 May 2012 19:57:01 +0000 (12:57 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David S. Miller:

 1) Since we do RCU lookups on ipv4 FIB entries, we have to test if the
    entry is dead before returning it to our caller.

 2) openvswitch locking and packet validation fixes from Ansis Atteka,
    Jesse Gross, and Pravin B Shelar.

 3) Fix PM resume locking in IGB driver, from Benjamin Poirier.

 4) Fix VLAN header handling in vhost-net and macvtap, from Basil Gor.

 5) Revert a bogus network namespace isolation change that was causing
    regressions on S390 networking devices.

 6) If bonding decides to process and handle a LACPDU frame, we
    shouldn't bump the rx_dropped counter.  From Jiri Bohac.

 7) Fix mis-calculation of available TX space in r8169 driver when doing
    TSO, which can lead to crashes and/or hung device.  From Julien
    Ducourthial.

 8) SCTP does not validate cached routes properly in all cases, from
    Nicolas Dichtel.

 9) Link status interrupt needs to be handled in ks8851 driver, from
    Stephen Boyd.

10) Use capable(), not cap_raised(), in connector/userns netlink code.
    From Eric W. Biederman via Andrew Morton.

11) Fix pktgen OOPS on module unload, from Eric Dumazet.

12) iwlwifi under-estimates SKB truesizes, also from Eric Dumazet.

13) Cure division by zero in SFC driver, from Ben Hutchings.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (26 commits)
  ks8851: Update link status during link change interrupt
  macvtap: restore vlan header on user read
  vhost-net: fix handle_rx buffer size
  bonding: don't increase rx_dropped after processing LACPDUs
  connector/userns: replace netlink uses of cap_raised() with capable()
  sctp: check cached dst before using it
  pktgen: fix crash at module unload
  Revert "net: maintain namespace isolation between vlan and real device"
  ehea: fix losing of NEQ events when one event occurred early
  igb: fix rtnl race in PM resume path
  ipv4: Do not use dead fib_info entries.
  r8169: fix unsigned int wraparound with TSO
  sfc: Fix division by zero when using one RX channel and no SR-IOV
  openvswitch: Validation of IPv6 set port action uses IPv4 header
  net: compare_ether_addr[_64bits]() has no ordering
  cdc_ether: Ignore bogus union descriptor for RNDIS devices
  bnx2x: bug fix when loading after SAN boot
  e1000: Silence sparse warnings by correcting type
  igb, ixgbe: netdev_tx_reset_queue incorrectly called from tx init path
  openvswitch: Release rtnl_lock if ovs_vport_cmd_build_info() failed.
  ...

12 years agoMerge tag 'dm-3.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm
Linus Torvalds [Sat, 12 May 2012 19:56:08 +0000 (12:56 -0700)]
Merge tag 'dm-3.4-fixes' of git://git./linux/kernel/git/agk/linux-dm

Pull device-mapper fixes from Alasdair G Kergon:
 "Fix a couple of serious memory leaks in device-mapper thin
  provisioning and tidy its MODULE_DESCRIPTION.

  Mitigate occasional reported hangs associated with multipath scsi_dh
  module loading."

* tag 'dm-3.4-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/agk/linux-dm:
  dm mpath: check if scsi_dh module already loaded before trying to load
  dm thin: correct module description
  dm thin: fix unprotected use of prepared_discards list
  dm thin: reinstate missing mempool_free in cell_release_singleton

12 years agoMAINTAINERS: Add myself as the cpufreq maintainer
Rafael J. Wysocki [Fri, 11 May 2012 19:35:45 +0000 (21:35 +0200)]
MAINTAINERS: Add myself as the cpufreq maintainer

Since cpufreq has no official maintainer at the moment, I'm willing
to maintain it along some other power management core code I've been
maintaining already.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agodm mpath: check if scsi_dh module already loaded before trying to load
Mike Snitzer [Sat, 12 May 2012 00:43:21 +0000 (01:43 +0100)]
dm mpath: check if scsi_dh module already loaded before trying to load

If the requested scsi_dh module is already loaded then skip
request_module().

Multipath table loads can hang in an unnecessary __request_module.

Reported-by: Ben Marzinski <bmarzins@redhat.com>
Cc: stable@kernel.org
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
12 years agodm thin: correct module description
Alasdair G Kergon [Sat, 12 May 2012 00:43:19 +0000 (01:43 +0100)]
dm thin: correct module description

Remove duplicate copy of string "device-mapper" (DM_NAME) from
MODULE_DESCRIPTION.

Signed-off-by: Alasdair G Kergon <agk@redhat.com>
12 years agodm thin: fix unprotected use of prepared_discards list
Mike Snitzer [Sat, 12 May 2012 00:43:16 +0000 (01:43 +0100)]
dm thin: fix unprotected use of prepared_discards list

Fix two places in commit 104655fd4dce ("dm thin: support discards") that
didn't use pool->lock to protect against concurrent changes to the
prepared_discards list.

Without this fix, thin_endio() can race with process_discard(), leading
to concurrent list_add()s that result in the processes locking up with
an error like the following:

WARNING: at lib/list_debug.c:32 __list_add+0x8f/0xa0()
...
list_add corruption. next->prev should be prev (ffff880323b96140), but was ffff8801d2c48440. (next=ffff8801d2c485c0).
...
Pid: 17205, comm: kworker/u:1 Tainted: G        W  O 3.4.0-rc3.snitm+ #1
Call Trace:
 [<ffffffff8103ca1f>] warn_slowpath_common+0x7f/0xc0
 [<ffffffff8103cb16>] warn_slowpath_fmt+0x46/0x50
 [<ffffffffa04f6ce6>] ? bio_detain+0xc6/0x210 [dm_thin_pool]
 [<ffffffff8124ff3f>] __list_add+0x8f/0xa0
 [<ffffffffa04f70d2>] process_discard+0x2a2/0x2d0 [dm_thin_pool]
 [<ffffffffa04f6a78>] ? remap_and_issue+0x38/0x50 [dm_thin_pool]
 [<ffffffffa04f7c3b>] process_deferred_bios+0x7b/0x230 [dm_thin_pool]
 [<ffffffffa04f7df0>] ? process_deferred_bios+0x230/0x230 [dm_thin_pool]
 [<ffffffffa04f7e42>] do_worker+0x52/0x60 [dm_thin_pool]
 [<ffffffff81056fa9>] process_one_work+0x129/0x450
 [<ffffffff81059b9c>] worker_thread+0x17c/0x3c0
 [<ffffffff81059a20>] ? manage_workers+0x120/0x120
 [<ffffffff8105eabe>] kthread+0x9e/0xb0
 [<ffffffff814ceda4>] kernel_thread_helper+0x4/0x10
 [<ffffffff8105ea20>] ? kthread_freezable_should_stop+0x70/0x70
 [<ffffffff814ceda0>] ? gs_change+0x13/0x13
---[ end trace 7e0a523bc5e52692 ]---

Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
12 years agodm thin: reinstate missing mempool_free in cell_release_singleton
Mike Snitzer [Sat, 12 May 2012 00:43:12 +0000 (01:43 +0100)]
dm thin: reinstate missing mempool_free in cell_release_singleton

Fix a significant memory leak inadvertently introduced during
simplification of cell_release_singleton() in commit
6f94a4c45a6f744383f9f695dde019998db3df55 ("dm thin: fix stacked bi_next
usage").

A cell's hlist_del() must be accompanied by a mempool_free().
Use __cell_release() to do this, like before.

Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Signed-off-by: Alasdair G Kergon <agk@redhat.com>
12 years agogpio/exynos: Fix compiler warnings when non-exynos machines are selected
Sachin Kamat [Mon, 30 Apr 2012 06:52:48 +0000 (12:22 +0530)]
gpio/exynos: Fix compiler warnings when non-exynos machines are selected

Fixes the following compiler warnings:

drivers/gpio/gpio-samsung.c: In function ‘samsung_gpiolib_init’:
drivers/gpio/gpio-samsung.c:2980:1: warning: label ‘err_ioremap1’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2978:1: warning: label ‘err_ioremap2’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2976:1: warning: label ‘err_ioremap3’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2974:1: warning: label ‘err_ioremap4’ defined but not used [-Wunused-label]
drivers/gpio/gpio-samsung.c:2722:55: warning: unused variable ‘gpio_base4’ [-Wunused-variable]

drivers/gpio/gpio-samsung.c:455:32: warning: ‘exynos_gpio_cfg’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2126:33: warning: ‘exynos4_gpios_1’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2228:33: warning: ‘exynos4_gpios_2’ defined but not used [-Wunused-variable]
drivers/gpio/gpio-samsung.c:2373:33: warning: ‘exynos4_gpios_3’ defined but not used [-Wunused-variable]

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
12 years agogpio: pch9: Use proper flow type handlers
Thomas Gleixner [Sat, 28 Apr 2012 08:13:45 +0000 (10:13 +0200)]
gpio: pch9: Use proper flow type handlers

Jean-Francois Dagenais reported:

 Configuring a gpio pin with the gpio-pch driver with
 "IRQF_TRIGGER_LOW | IRQF_ONESHOT" generates an interrupt storm for
 threaded ISR until the ISR thread actually gets to physically clear
 the interrupt on the triggering chip!! The immediate observable
 symptom is the high CPU usage for my ISR thread task and the
 interrupt count in /proc/interrupts incrementing radically.

The driver is wrong in several ways:

1) Using handle_simple_irq() does not provide proper flow control
   handling. In the case of oneshot threaded handlers for the
   demultiplexed interrupts this results in an interrupt storm because
   the simple handler does not deal with masking/unmasking.  Even
   without threaded oneshot handlers an interrupt storm for level type
   interrupts can easily be triggered when the interrupt is disabled
   and the interrupt line is activated from the device.

2) Acknowlegding the demultiplexed interrupt before calling the
   handler is wrong for level type interrupts.

3) The set_type function unconditionally enables the interrupt. It's
   supposed to set the type and nothing else. The unmasking is done by
   the core code.

Move the acknowledge code into a separate function and add it to the
demux irqchip callbacks.

Remove the unconditional enabling from the set_type() callback and set
the proper flow handlers depending on the selected type (level/edge).

Reported-and-tested-by: Jean-Francois Dagenais <jeff.dagenais@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Grant Likely <grant.likely@secretlab.ca>
12 years agoMerge tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6
Linus Torvalds [Fri, 11 May 2012 23:59:07 +0000 (16:59 -0700)]
Merge tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6

Pull GPIO omap bug fix from Grant Likely.

* tag 'gpio-for-linus' of git://git.secretlab.ca/git/linux-2.6:
  gpio/omap: fix incorrect initialization of omap_gpio_mod_init

12 years agoMerge branch 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc
Linus Torvalds [Fri, 11 May 2012 23:58:14 +0000 (16:58 -0700)]
Merge branch 'merge' of git://git./linux/kernel/git/benh/powerpc

Pull another powerpc irq fix from Benjamin Herrenschmidt:
 "It looks like my previous fix for the lazy irq masking problem wasn't
  quite enough.  There was another problem related to performance
  monitor interrupts acting as NMIs leaving the flags in an incorrect
  state.  Here's a fix that finally seems to make perf solid again."

* 'merge' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc:
  powerpc/irq: Fix another case of lazy IRQ state getting out of sync

12 years agoMerge branch '3.4-urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target...
Linus Torvalds [Fri, 11 May 2012 23:49:09 +0000 (16:49 -0700)]
Merge branch '3.4-urgent' of git://git./linux/kernel/git/nab/target-pending

Pull target fix from Nicholas Bellinger:
 "This patch removes some incorrect legacy code to free se_lun_acl
  memory in the NodeACL release path that could potentially trigger an
  OOPS during shutdown once dynamic -> explicit initiator NodeACL
  conversion has occurred.

  That said, we've been able to trigger an OOPS in v4.0 code for this
  special case when the associated MappedLUNs had not also been made
  explicit based on active TPG LUN layout during the conversion, so it
  really makes senses to go ahead and drop this extra cruft to avoid any
  possible issues here.

  This ends up only effecting iscsi-target module code (it's the only
  user) and is CC'ed to stable."

* '3.4-urgent' of git://git.kernel.org/pub/scm/linux/kernel/git/nab/target-pending:
  target: Drop incorrect se_lun_acl release for dynamic -> explict ACL conversion

12 years agopowerpc/irq: Fix another case of lazy IRQ state getting out of sync
Benjamin Herrenschmidt [Thu, 10 May 2012 16:12:38 +0000 (16:12 +0000)]
powerpc/irq: Fix another case of lazy IRQ state getting out of sync

So we have another case of paca->irq_happened getting out of
sync with the HW irq state. This can happen when a perfmon
interrupt occurs while soft disabled, as it will return to a
soft disabled but hard enabled context while leaving a stale
PACA_IRQ_HARD_DIS flag set.

This patch fixes it, and also adds a test for the condition
of those flags being out of sync in arch_local_irq_restore()
when CONFIG_TRACE_IRQFLAGS is enabled.

This helps catching those gremlins faster (and so far I
can't seem see any anymore, so that's good news).

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
12 years agoks8851: Update link status during link change interrupt
Stephen Boyd [Thu, 10 May 2012 12:51:30 +0000 (12:51 +0000)]
ks8851: Update link status during link change interrupt

If a link change interrupt comes in we just clear the interrupt
and continue along without notifying the upper networking layers
that the link has changed. Use the mii_check_link() function to
update the link status whenever a link change interrupt occurs.

Cc: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agomacvtap: restore vlan header on user read
Basil Gor [Thu, 3 May 2012 22:55:24 +0000 (22:55 +0000)]
macvtap: restore vlan header on user read

Ethernet vlan header is not on the packet and kept in the skb->vlan_tci
when it comes from lower dev. This patch inserts vlan header in user
buffer during skb copy on user read.

Signed-off-by: Basil Gor <basil.gor@gmail.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agovhost-net: fix handle_rx buffer size
Basil Gor [Thu, 3 May 2012 22:55:23 +0000 (22:55 +0000)]
vhost-net: fix handle_rx buffer size

Take vlan header length into account, when vlan id is stored as
vlan_tci. Otherwise tagged packets coming from macvtap will be
truncated.

Signed-off-by: Basil Gor <basil.gor@gmail.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agotarget: Drop incorrect se_lun_acl release for dynamic -> explict ACL conversion
Nicholas Bellinger [Fri, 11 May 2012 05:05:49 +0000 (22:05 -0700)]
target: Drop incorrect se_lun_acl release for dynamic -> explict ACL conversion

This patch removes some potentially problematic legacy code within
core_clear_initiator_node_from_tpg() that was originally intended to
release left over se_lun_acl setup during dynamic NodeACL+MappedLUN
generate when running with TPG demo-mode operation.

Since we now only ever expect to allocate and release se_lun_acl from
within target_core_fabric_configfs.c:target_fabric_make_mappedlun() and
target_fabric_drop_mappedlun() context respectively, this code for
demo-mode release is incorrect and needs to be removed.

Cc: Christoph Hellwig <hch@lst.de>
Cc: Andy Grover <agrover@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Nicholas Bellinger <nab@linux-iscsi.org>
12 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu
Linus Torvalds [Fri, 11 May 2012 16:28:35 +0000 (09:28 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/gerg/m68knommu

Pull a m68knommu fix from Greg Ungerer:
 "It contains a single fix for including the ColdFire QSPI interface
  setup code when enabled as a module.  This was broken in the
  consolidation of the ColdFire SoC device tables in the 3.4 merge
  window."

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu:
  m68knommu: enable qspi support when SPI_COLDFIRE_QSPI = m

12 years agomm: raise MemFree by reverting percpu_pagelist_fraction to 0
Hugh Dickins [Fri, 11 May 2012 08:00:07 +0000 (01:00 -0700)]
mm: raise MemFree by reverting percpu_pagelist_fraction to 0

Why is there less MemFree than there used to be?  It perturbed a test,
so I've just been bisecting linux-next, and now find the offender went
upstream yesterday.

Commit 93278814d359 "mm: fix division by 0 in percpu_pagelist_fraction()"
mistakenly initialized percpu_pagelist_fraction to the sysctl's minimum 8,
which leaves 1/8th of memory on percpu lists (on each cpu??); but most of
us expect it to be left unset at 0 (and it's not then used as a divisor).

  MemTotal: 8061476kB  8061476kB  8061476kB  8061476kB  8061476kB  8061476kB
  Repetitive test with percpu_pagelist_fraction 8:
  MemFree:  6948420kB  6237172kB  6949696kB  6840692kB  6949048kB  6862984kB
  Same test with percpu_pagelist_fraction back to 0:
  MemFree:  7945000kB  7944908kB  7948568kB  7949060kB  7948796kB  7948812kB

Signed-off-by: Hugh Dickins <hughd@google.com>
[ We really should fix the crazy sysctl interface too, but that's a
  separate thing - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
12 years agobonding: don't increase rx_dropped after processing LACPDUs
Jiri Bohac [Wed, 9 May 2012 01:01:40 +0000 (01:01 +0000)]
bonding: don't increase rx_dropped after processing LACPDUs

Since commit 3aba891d, bonding processes LACP frames (802.3ad
mode) with bond_handle_frame(). Currently a copy of the skb is
made and the original is left to be processed by other
rx_handlers and the rest of the network stack by returning
RX_HANDLER_ANOTHER.  As there is no protocol handler for
PKT_TYPE_LACPDU, the frame is dropped and dev->rx_dropped
increased.

Fix this by making bond_handle_frame() return RX_HANDLER_CONSUMED
if bonding has processed the LACP frame.

Signed-off-by: Jiri Bohac <jbohac@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agoconnector/userns: replace netlink uses of cap_raised() with capable()
Eric W. Biederman [Fri, 4 May 2012 11:34:03 +0000 (11:34 +0000)]
connector/userns: replace netlink uses of cap_raised() with capable()

In 2009 Philip Reiser notied that a few users of netlink connector
interface needed a capability check and added the idiom
cap_raised(nsp->eff_cap, CAP_SYS_ADMIN) to a few of them, on the premise
that netlink was asynchronous.

In 2011 Patrick McHardy noticed we were being silly because netlink is
synchronous and removed eff_cap from the netlink_skb_params and changed
the idiom to cap_raised(current_cap(), CAP_SYS_ADMIN).

Looking at those spots with a fresh eye we should be calling
capable(CAP_SYS_ADMIN).  The only reason I can see for not calling capable
is that it once appeared we were not in the same task as the caller which
would have made calling capable() impossible.

In the initial user_namespace the only difference between between
cap_raised(current_cap(), CAP_SYS_ADMIN) and capable(CAP_SYS_ADMIN) are a
few sanity checks and the fact that capable(CAP_SYS_ADMIN) sets
PF_SUPERPRIV if we use the capability.

Since we are going to be using root privilege setting PF_SUPERPRIV seems
the right thing to do.

The motivation for this that patch is that in a child user namespace
cap_raised(current_cap(),...) tests your capabilities with respect to that
child user namespace not capabilities in the initial user namespace and
thus will allow processes that should be unprivielged to use the kernel
services that are only protected with cap_raised(current_cap(),..).

To fix possible user_namespace issues and to just clean up the code
replace cap_raised(current_cap(), CAP_SYS_ADMIN) with
capable(CAP_SYS_ADMIN).

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Patrick McHardy <kaber@trash.net>
Cc: Philipp Reisner <philipp.reisner@linbit.com>
Acked-by: Serge E. Hallyn <serge.hallyn@canonical.com>
Acked-by: Andrew G. Morgan <morgan@kernel.org>
Cc: Vasiliy Kulikov <segoon@openwall.com>
Cc: David Howells <dhowells@redhat.com>
Reviewed-by: James Morris <james.l.morris@oracle.com>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
12 years agosctp: check cached dst before using it
Nicolas Dichtel [Fri, 4 May 2012 05:24:54 +0000 (05:24 +0000)]
sctp: check cached dst before using it

dst_check() will take care of SA (and obsolete field), hence
IPsec rekeying scenario is taken into account.

Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Acked-by: Vlad Yaseivch <vyasevich@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>