linux-2.6
17 years agoNFS: Fix over-conservative attribute invalidation in nfs_update_inode()
Trond Myklebust [Wed, 15 Aug 2007 16:49:17 +0000 (12:49 -0400)] 
NFS: Fix over-conservative attribute invalidation in nfs_update_inode()

We should always be declaring the attribute cache as valid after having
updated it.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFSv4: Make NFSv4 ACCESS calls return attributes too...
Trond Myklebust [Fri, 10 Aug 2007 21:45:11 +0000 (17:45 -0400)] 
NFSv4: Make NFSv4 ACCESS calls return attributes too...

It doesn't really make sense to cache an access call without also
revalidating the attributes.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFSv4: Simplify _nfs4_do_access()
Trond Myklebust [Fri, 10 Aug 2007 21:45:10 +0000 (17:45 -0400)] 
NFSv4: Simplify _nfs4_do_access()

Currently, _nfs4_do_access() is just a copy of nfs_do_access() with added
conversion of the open flags into an access mask. This patch merges the
duplicate functionality.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Replace file->private_data with calls to nfs_file_open_context()
Trond Myklebust [Fri, 10 Aug 2007 21:44:32 +0000 (17:44 -0400)] 
NFS: Replace file->private_data with calls to nfs_file_open_context()

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Add a helper to extract the nfs_open_context from a struct file
Trond Myklebust [Fri, 10 Aug 2007 21:44:28 +0000 (17:44 -0400)] 
NFS: Add a helper to extract the nfs_open_context from a struct file

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Eliminate nfs_refresh_verifier()
Chuck Lever [Mon, 24 Sep 2007 19:40:16 +0000 (15:40 -0400)] 
NFS: Eliminate nfs_refresh_verifier()

nfs_set_verifier() and nfs_refresh_verifier() do exactly the same thing, so
replace one with the other.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Eliminate nfs_renew_times()
Chuck Lever [Mon, 24 Sep 2007 19:40:11 +0000 (15:40 -0400)] 
NFS: Eliminate nfs_renew_times()

The nfs_renew_times() function plants the current time in jiffies in
dentry->d_time.  But a call to nfs_renew_times() is always followed by
another call that overwrites dentry->d_time.  Get rid of the
nfs_renew_times() calls.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Don't call nfs_renew_times() in nfs_dentry_iput()
Chuck Lever [Mon, 24 Sep 2007 19:40:06 +0000 (15:40 -0400)] 
NFS: Don't call nfs_renew_times() in nfs_dentry_iput()

Negative dentries need to be reverified after an asynchronous unlink.

Quoth Trond:

"Unfortunately I don't think that we can avoid revalidating the
resulting negative dentry since the UNLINK call is asynchronous,
and so the new verifier on the directory will only be known a
posteriori."

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Fix bytes-per-op accounting for RPC over UDP
Chuck Lever [Mon, 24 Sep 2007 19:40:00 +0000 (15:40 -0400)] 
SUNRPC: Fix bytes-per-op accounting for RPC over UDP

NFS performance metrics reported zero bytes sent per op when mounting with
UDP.  The UDP socket transport wasn't properly counting the number of bytes
sent.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Show "nointr" mount option
Chuck Lever [Mon, 24 Sep 2007 19:39:55 +0000 (15:39 -0400)] 
NFS: Show "nointr" mount option

The default "intr" setting is different for NFS and NFSv4.  To avoid
confusion on this issue, don't hide the "nointr" option in /proc/mounts.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Verify server address before invoking in-kernel mount client
Chuck Lever [Mon, 24 Sep 2007 19:39:50 +0000 (15:39 -0400)] 
NFS: Verify server address before invoking in-kernel mount client

Re-order mount option sanity checking slightly to ensure we have a valid
server address *before* trying to do the mountd RPC call.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Add RDMA dependency to SUNRPC_XPRT_RDMA
\"Talpey, Thomas\ [Thu, 20 Sep 2007 21:37:58 +0000 (17:37 -0400)] 
SUNRPC: Add RDMA dependency to SUNRPC_XPRT_RDMA

Add a dependency on RDMA before enabling SUNRPC_XPRT_RDMA
Yes, "INFINIBAND" also turns on iWARP and other RDMA support.

Signed-off-by: Tom Talpey <talpey@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoRPCRDMA: rpc rdma verbs interface implementation
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:51:18 +0000 (13:51 -0400)] 
RPCRDMA: rpc rdma verbs interface implementation

This implements the interface from rpcrdma to the RDMA verbs interface
supported by Infniband and iWARP.

Signed-off-by: Tom Talpey <talpey@netapp.com>
Signed-off-by: James Lentini <jlentini@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoRPCRDMA: rpc rdma protocol implementation
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:50:42 +0000 (13:50 -0400)] 
RPCRDMA: rpc rdma protocol implementation

This implements the marshaling and unmarshaling of the rpcrdma transport
headers. Connection management is also addressed.

Signed-off-by: Tom Talpey <talpey@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoRPCRDMA: rpc rdma transport switch
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:50:12 +0000 (13:50 -0400)] 
RPCRDMA: rpc rdma transport switch

This implements the configuration and building of the core transport
switch implementation of the rpcrdma transport. Stubs are provided for
the rpcrdma protocol handling, and the infiniband/iwarp verbs interface.
These are provided in following patches.

Signed-off-by: Tom Talpey <talpey@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: support RDMA mounts
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:49:41 +0000 (13:49 -0400)] 
NFS: support RDMA mounts

Adds hooks to the string-based NFS mount to support an "rdma" protocol option.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoRPCRDMA: Kconfig and header file with rpcrdma protocol definitions
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:49:15 +0000 (13:49 -0400)] 
RPCRDMA: Kconfig and header file with rpcrdma protocol definitions

This file implements the configuration target, protocol template and
constants for the rpcrdma transport framing, for use by the xprtrdma
rpc transport implementation.

Signed-off-by: Tom Talpey <talpey@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS - print accurate transport protocol
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:48:47 +0000 (13:48 -0400)] 
NFS - print accurate transport protocol

Use the per-transport strings to display the transport protocol accurately.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS/SUNRPC: use transport protocol naming
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:48:23 +0000 (13:48 -0400)] 
NFS/SUNRPC: use transport protocol naming

Instead of an { address family, raw IP protocol number }-tuple, use the
newly-defined RPC identifier when creating clients in the upper layers.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS/SUNRPC: support transport protocol naming
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:47:57 +0000 (13:47 -0400)] 
NFS/SUNRPC: support transport protocol naming

To prepare for including non-sockets-based RPC transports, select
RPC transports by an identifier (to be used in following patches).

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: rearrange RPC sockets definitions
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:47:31 +0000 (13:47 -0400)] 
SUNRPC: rearrange RPC sockets definitions

To prepare for including non-sockets-based RPC transports, move the
sockets-dependent definitions into their own file.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: rename the rpc_xprtsock_create structure
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:47:07 +0000 (13:47 -0400)] 
SUNRPC: rename the rpc_xprtsock_create structure

To prepare for including non-sockets-based RPC transports, change the
overly suggestive name of the transport creation arguments struct.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Finish API to load RPC transport implementations dynamically
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:46:39 +0000 (13:46 -0400)] 
SUNRPC: Finish API to load RPC transport implementations dynamically

Allow RPC client transport implementations to be loaded as needed, or
as they become available from distributors or third-party vendors.

Note that we leave the IP sockets implementation in sunrpc.o
permanently, as IP functionality is always available in any
kernel that runs NFS.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Provide a new API for registering transport implementations
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:46:00 +0000 (13:46 -0400)] 
SUNRPC: Provide a new API for registering transport implementations

To allow transport capabilities to be loaded dynamically, provide an API
for registering and unregistering the transports with the RPC client.
Eventually xprt_create_transport() will be changed to search the list of
registered transports when initializing a fresh transport.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: add EXPORT_SYMBOL_GPL for generic transport functions
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:45:36 +0000 (13:45 -0400)] 
SUNRPC: add EXPORT_SYMBOL_GPL for generic transport functions

SUNRPC: add EXPORT_SYMBOL_GPL for generic transport functions

As a preface to allowing arbitrary transport modules to be loaded
dynamically, add EXPORT_SYMBOL_GPL for all generic transport functions
that a transport implementation might want to use.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Acked-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: mark bulk read/write data in xdrbuf
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:44:58 +0000 (13:44 -0400)] 
SUNRPC: mark bulk read/write data in xdrbuf

Adds a flag word to the xdrbuf struct which indicates any bulk
disposition of the data. This enables RPC transport providers to
marshal it efficiently/appropriately, and may enable other
optimizations.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFSv4: Fix a bug in nfs4_validate_mount_data()
Trond Myklebust [Fri, 21 Sep 2007 00:23:51 +0000 (20:23 -0400)] 
NFSv4: Fix a bug in nfs4_validate_mount_data()

The previous patch introduced a bug when copying the server address.

Also clarify a copy into the auth_flavours array: currently the two
size calculations are equivalent, but we may decide to change the size
of auth_flavors[] at some point.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: use in-kernel mount argument structure for nfsv4 mounts
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:44:33 +0000 (13:44 -0400)] 
NFS: use in-kernel mount argument structure for nfsv4 mounts

The user-visible nfs4_mount_data does not contain sufficient data to
describe new mount options, and also is now a legacy structure. Replace
it with the internal nfs_parsed_mount_data for nfsv4 in-kernel use.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Acked-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: use in-kernel mount argument structure for nfsv[23] mounts
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:43:56 +0000 (13:43 -0400)] 
NFS: use in-kernel mount argument structure for nfsv[23] mounts

The user-visible nfs_mount_data does not contain sufficient data to
describe new mount options, and also is now a legacy structure. Replace
it with the internal nfs_parsed_mount_data for nfsv[23] in-kernel use.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Acked-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: move nfs_parsed_mount_data structure definition
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:43:29 +0000 (13:43 -0400)] 
NFS: move nfs_parsed_mount_data structure definition

In preparation for rearranging the nfs mount argument passing, make the
nfs_parsed_mount_data struct visible across nfs kernel files.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Acked-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: export per-transport rpcbind netid's
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:43:05 +0000 (13:43 -0400)] 
SUNRPC: export per-transport rpcbind netid's

The rpcbind (v3+) netid is provided by each RPC client transport. This fixes
an omission in IPv6 rpcbind client support, and enables future extension.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: move per-transport rpcbind netid's
\"Talpey, Thomas\ [Mon, 10 Sep 2007 17:42:38 +0000 (13:42 -0400)] 
SUNRPC: move per-transport rpcbind netid's

Move the TCP/UDP rpcbind netid's from the rpcbind client to a global header.

Signed-off-by: Tom Talpey <tmt@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFSD: Convert printk's to dprintk's in NFSD's nfs4xdr
Chuck Lever [Tue, 11 Sep 2007 22:01:20 +0000 (18:01 -0400)] 
NFSD: Convert printk's to dprintk's in NFSD's nfs4xdr

Due to recent edict to remove or replace printk's that can flood the system
log.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoLOCKD: Convert printk's to dprintk's in lockd XDR routines
Chuck Lever [Tue, 11 Sep 2007 22:01:15 +0000 (18:01 -0400)] 
LOCKD: Convert printk's to dprintk's in lockd XDR routines

Due to recent edict to remove or replace printk's that might flood the
system log.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Convert printk's to dprintk's in fs/nfs/nfs?xdr.c
Chuck Lever [Tue, 11 Sep 2007 22:01:10 +0000 (18:01 -0400)] 
NFS: Convert printk's to dprintk's in fs/nfs/nfs?xdr.c

Due to recent edict to replace or remove printk's that can be triggered en
masse by remote misbehavior.  Left a few that only occur just before a BUG.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Add new 'mountaddr=' mount option
Chuck Lever [Tue, 11 Sep 2007 22:01:04 +0000 (18:01 -0400)] 
NFS: Add new 'mountaddr=' mount option

I got the 'mounthost=' option wrong - it shouldn't look for an address
value, but rather a hostname value.  However, the in-kernel mount client
and NFS client cannot resolve a hostname by themselves; they rely on
user-land to pass in the resolved address.

Create a new mount option that does take an address so that the mount
program's address can be passed in.  The mount hostname is now ignored
by the kernel.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years ago[NFS] [PATCH] NFS: initialize default port in kernel mount client
James Lentini [Mon, 24 Sep 2007 21:32:49 +0000 (17:32 -0400)] 
[NFS] [PATCH] NFS: initialize default port in kernel mount client

If no mount server port number is specified, the previous change to the
kernel mount client inadvertently allows the NFS server's port number to be
the used as the mount server's port number. If the user specifies an NFS
server port (-o port=x), the mount will fail.

The fix below sets the mount server's port to 0 if no mount server
port is specified by the user.

Signed-off-by: James Lentini <jlentini@netapp.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Kernel mount client should use async bind
Chuck Lever [Tue, 11 Sep 2007 22:00:58 +0000 (18:00 -0400)] 
NFS: Kernel mount client should use async bind

Simplify the in-kernel mount client by using autobind instead of an
explicit call to rpc_getport_sync.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: RPC bind failures should be permanent for NULL requests
Chuck Lever [Tue, 11 Sep 2007 22:00:52 +0000 (18:00 -0400)] 
SUNRPC: RPC bind failures should be permanent for NULL requests

The purpose of an RPC ping (a NULL request) is to determine whether the
remote end is operating and supports the RPC program and version of the
request.

If we do an RPC bind and the remote's rpcbind service says "this
program or service isn't supported" then we have our answer already,
and we should give up immediately.

This is good for the kernel mount client, as it will cause the request
to fail, and then allow an immediate retry with different options.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Split another new rpcbind retry error code from EACCES
Chuck Lever [Tue, 11 Sep 2007 22:00:47 +0000 (18:00 -0400)] 
SUNRPC: Split another new rpcbind retry error code from EACCES

Add more new error code processing to the kernel's rpcbind client
and to call_bind_status() to distinguish two cases:

Case 1: the remote has replied that the program/version tuple is not
registered (returns EACCES)

Case 2: retry with a lesser rpcbind version (rpcb now returns EPFNOSUPPORT)

This change allows more specific error processing for each of these two
cases.  We now fail case 2 instead of retrying... it's a server
configuration error not to support even rpcbind version 2.  And don't
expose this new error code to user land -- convert it to EIO before
failing the RPC.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Add a new error code for retry waiting for another binder
Chuck Lever [Tue, 11 Sep 2007 22:00:41 +0000 (18:00 -0400)] 
SUNRPC: Add a new error code for retry waiting for another binder

Add new error code processing to the kernel's rpcbind client and to
call_bind_status() to distinguish two cases:

Case 1: the remote has replied that the program/version tuple is not
registered (returns -EACCES)

Case 2: another process is already in the middle of binding on this
transport (now returns -EAGAIN)

This change allows more specific retry processing for each of these two
cases.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Retry bad rpcbind replies
Chuck Lever [Tue, 11 Sep 2007 22:00:36 +0000 (18:00 -0400)] 
SUNRPC: Retry bad rpcbind replies

When a server returns a bad rpcbind reply, make rpcbind client recovery logic
retry with an older protocol version.  Older versions are more likely to work
correctly.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Make rpcb_decode_getaddr more picky about universal addresses
Chuck Lever [Tue, 11 Sep 2007 22:00:31 +0000 (18:00 -0400)] 
SUNRPC: Make rpcb_decode_getaddr more picky about universal addresses

Add better sanity checking of server replies to the GETVERSADDR reply
decoder.  Change the error return code: EIO is what other XDR decoding
routines return if there is a failure while decoding.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Clean up in rpc_show_tasks
Chuck Lever [Tue, 11 Sep 2007 22:00:25 +0000 (18:00 -0400)] 
SUNRPC: Clean up in rpc_show_tasks

/home/cel/linux/net/sunrpc/clnt.c: In function ‘rpc_show_tasks’:
/home/cel/linux/net/sunrpc/clnt.c:1538: warning:
signed and unsigned type in conditional expression

This points out another case where a conditional expression returns a
signed value in one arm and an unsigned value in the other.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Make sure server name is reasonable before trying to print it
Chuck Lever [Tue, 11 Sep 2007 22:00:20 +0000 (18:00 -0400)] 
SUNRPC: Make sure server name is reasonable before trying to print it

Check the length of the passed-in server name before trying to print it in
the log.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Use correct argument type in memcpy()
Chuck Lever [Tue, 11 Sep 2007 22:00:15 +0000 (18:00 -0400)] 
SUNRPC: Use correct argument type in memcpy()

Noticed by Tom Talpey <tmt@netapp.com>:

OBTW, there's a nit on that memcpy, too. The r_addr is an array, so

memcpy(&map->r_addr

is passing the address of the array as a char **. It's the same as
map->r_addr, but technically the wrong type.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: fix a signed v. unsigned comparison nit in rpc_bind_new_program
Chuck Lever [Tue, 11 Sep 2007 22:00:09 +0000 (18:00 -0400)] 
SUNRPC: fix a signed v. unsigned comparison nit in rpc_bind_new_program

/home/cel/linux/net/sunrpc/clnt.c: In function ‘rpc_bind_new_program’:
/home/cel/linux/net/sunrpc/clnt.c:445: warning:
comparison between signed and unsigned

RPC version numbers are u32, not int.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Only one dprintk is needed during client creation
Chuck Lever [Tue, 11 Sep 2007 22:00:03 +0000 (18:00 -0400)] 
SUNRPC: Only one dprintk is needed during client creation

Remove one of two identical dprintk's that occur when an RPC client is
created.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Fix generation of universal addresses for
Chuck Lever [Thu, 16 Aug 2007 20:03:31 +0000 (16:03 -0400)] 
SUNRPC: Fix generation of universal addresses for

Fix some problems with rpcbind v3 and v4 queries from the in-kernel rpcbind
client:

1.  The r_addr argument must be a full universal address, not just an IP
address, and

2.  The universal address in r_addr is the address of the remote rpcbind
server, not the RPC service being requested

This addresses bugzilla.kernel.org report 8891 for 2.6.23-rc and greater.

In addition, if the rpcbind client is unable to start the rpcbind request,
make sure not to leak the xprt.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
17 years agoSUNRPC: Add support for formatted universal addresses
Chuck Lever [Thu, 16 Aug 2007 20:03:26 +0000 (16:03 -0400)] 
SUNRPC: Add support for formatted universal addresses

"Universal addresses" are a string representation of an IP address and
port.  They are described fully in RFC 3530, section 2.2.  Add support
for generating them in the RPC client's socket transport module.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
17 years agoSUNRPC: Split xs_reclassify_socket into an IPv4 and IPv6 version
Chuck Lever [Mon, 6 Aug 2007 15:58:04 +0000 (11:58 -0400)] 
SUNRPC: Split xs_reclassify_socket into an IPv4 and IPv6 version

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Add a helper for extracting the address using the correct type
Chuck Lever [Mon, 6 Aug 2007 15:57:58 +0000 (11:57 -0400)] 
SUNRPC: Add a helper for extracting the address using the correct type

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Add IPv6 address support to net/sunrpc/xprtsock.c
Chuck Lever [Mon, 6 Aug 2007 15:57:53 +0000 (11:57 -0400)] 
SUNRPC: Add IPv6 address support to net/sunrpc/xprtsock.c

Finalize support for setting up RPC client transports to remote RPC
services addressed via IPv6.

Based on work done by Gilles Quillard at Bull Open Source.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: create connect workers for IPv6
Chuck Lever [Mon, 6 Aug 2007 15:57:48 +0000 (11:57 -0400)] 
SUNRPC: create connect workers for IPv6

Clone separate connect worker functions for connecting AF_INET6 sockets.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Rename IPv4 connect workers
Chuck Lever [Mon, 6 Aug 2007 15:57:43 +0000 (11:57 -0400)] 
SUNRPC: Rename IPv4 connect workers

Prepare for introduction of IPv6 versions of same.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Refactor a part of socket connect logic into a helper function
Chuck Lever [Mon, 6 Aug 2007 15:57:38 +0000 (11:57 -0400)] 
SUNRPC: Refactor a part of socket connect logic into a helper function

Finishing a socket connect is the same for IPv4 and IPv6, so split it out
into a helper.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: create an IPv6-savvy mechanism for binding to a reserved port
Chuck Lever [Mon, 6 Aug 2007 15:57:33 +0000 (11:57 -0400)] 
SUNRPC: create an IPv6-savvy mechanism for binding to a reserved port

Clone xs_bindresvport into two functions, one that can handle IPv4
addresses, and one that can handle IPv6 addresses.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Rename xs_bind() to prepare for IPv6-specific bind method
Chuck Lever [Mon, 6 Aug 2007 15:57:28 +0000 (11:57 -0400)] 
SUNRPC: Rename xs_bind() to prepare for IPv6-specific bind method

Prepare for introduction of IPv6-specific socket bind function.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Introduce support for setting the port number in IPv6 addresses
Chuck Lever [Mon, 6 Aug 2007 15:57:23 +0000 (11:57 -0400)] 
SUNRPC: Introduce support for setting the port number in IPv6 addresses

We could clone xs_set_port, but this is easier overall.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: add support for IPv6 to the kernel's rpcbind client
Chuck Lever [Mon, 6 Aug 2007 15:57:18 +0000 (11:57 -0400)] 
SUNRPC: add support for IPv6 to the kernel's rpcbind client

Prepare for adding IPv6 support to the RPC client by adding IPv6
capabilities to rpcbind.  Note that this is support on the query side
only; registering IPv6 addresses with the local portmapper will come
later.

Note we have to take care not to fall back to using version 2 of the
rpcbind protocol if we're dealing with IPv6 address.  Version 2 doesn't
support IPv6 at all.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: add a function to format IPv6 addresses
Chuck Lever [Mon, 6 Aug 2007 15:57:12 +0000 (11:57 -0400)] 
SUNRPC: add a function to format IPv6 addresses

Clone xs_format_ipv4_peer_addresses into an IPv6 version.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Rename xs_format_peer_addresses
Chuck Lever [Mon, 6 Aug 2007 15:57:07 +0000 (11:57 -0400)] 
SUNRPC: Rename xs_format_peer_addresses

Prepare to add an IPv6 version of xs_format_peer_addresses by renaming it
to xs_format_ipv4_peer_addresses.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Add hex-formatted address support to rpc_peeraddr2str()
Chuck Lever [Mon, 6 Aug 2007 15:57:02 +0000 (11:57 -0400)] 
SUNRPC: Add hex-formatted address support to rpc_peeraddr2str()

Add support for the NFS client's need to export volume information
with IP addresses formatted in hex instead of decimal.

This isn't used yet, but subsequent patches (not in this series) will
change the NFS client to use this functionality.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Free address buffers in a loop
Chuck Lever [Mon, 6 Aug 2007 15:56:57 +0000 (11:56 -0400)] 
SUNRPC: Free address buffers in a loop

Use more generic logic to free buffers holding formatted addresses.  This
makes it less likely a bug will be introduced when adding additional buffer
types in xs_format_peer_address().

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Use standard macros for printing IP addresses
Chuck Lever [Mon, 6 Aug 2007 15:56:52 +0000 (11:56 -0400)] 
SUNRPC: Use standard macros for printing IP addresses

include/linux/kernel.h gives us some nice macros for formatting IP
addresses.  Use them.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Fix a signed v. unsigned comparison in net/sunrpc/xprtsock.c
Chuck Lever [Mon, 6 Aug 2007 15:56:42 +0000 (11:56 -0400)] 
SUNRPC: Fix a signed v. unsigned comparison in net/sunrpc/xprtsock.c

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Fix a signed v. unsigned comparison in rpcbind's XDR routines
Chuck Lever [Mon, 6 Aug 2007 15:56:31 +0000 (11:56 -0400)] 
SUNRPC: Fix a signed v. unsigned comparison in rpcbind's XDR routines

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years ago[NFS] [PATCH] NFS: show addr=ipaddr in /proc/mounts rather than
Jeff Layton [Mon, 30 Jul 2007 12:47:38 +0000 (08:47 -0400)] 
[NFS] [PATCH] NFS: show addr=ipaddr in /proc/mounts rather than

A minor thing, but useful when working with a server with multiple
addrs. This looks like it might also be necessary if Miklos' effort
to eliminate /etc/mtab ever comes to fruition.

When displaying mount options in /proc/mounts, the kernel prints
"addr=hostname". This info is redundant since we already have the
hostname displayed as part of the "device" section of the mount. This
patch changes it to display the IP address to which the socket is
connected.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years ago[NFS] [PATCH] nfs: tiny makefile cleanup
Christoph Hellwig [Fri, 3 Aug 2007 14:20:32 +0000 (16:20 +0200)] 
[NFS] [PATCH] nfs: tiny makefile cleanup

no need to set up foo-objs these days.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoRe: [NFS] [PATCH] Attribute timeout handling and wrapping u32 jiffies
Fabio Olive Leite [Fri, 27 Jul 2007 01:59:00 +0000 (22:59 -0300)] 
Re: [NFS] [PATCH] Attribute timeout handling and wrapping u32 jiffies

I would like to discuss the idea that the current checks for attribute
timeout using time_after are inadequate for 32bit architectures, since
time_after works correctly only when the two timestamps being compared
are within 2^31 jiffies of each other. The signed overflow caused by
comparing values more than 2^31 jiffies apart will flip the result,
causing incorrect assumptions of validity.

2^31 jiffies is a fairly large period of time (~25 days) when compared
to the lifetime of most kernel data structures, but for long lived NFS
mounts that can sit idle for months (think that for some reason autofs
cannot be used), it is easy to compare inode attribute timestamps with
very disparate or even bogus values (as in when jiffies have wrapped
many times, where the comparison doesn't even make sense).

Currently the code tests for attribute timeout by simply adding the
desired amount of jiffies to the stored timestamp and comparing that
with the current timestamp of obtained attribute data with time_after.
This is incorrect, as it returns true for the desired timeout period
and another full 2^31 range of jiffies.

In testing with artificial jumps (several small jumps, not one big
crank) of the jiffies I was able to reproduce a problem found in a
server with very long lived NFS mounts, where attributes would not be
refreshed even after touching files and directories in the server:

Initial uptime:
03:42:01 up 6 min, 0 users, load average: 0.01, 0.12, 0.07

NFS volume is mounted and time is advanced:
03:38:09 up 25 days, 2 min, 0 users, load average: 1.22, 1.05, 1.08

# ls -l /local/A/foo/bar /nfs/A/foo/bar
-rw-r--r--  1 root root 0 Dec 17 03:38 /local/A/foo/bar
-rw-r--r--  1 root root 0 Nov 22 00:36 /nfs/A/foo/bar

# touch /local/A/foo/bar

# ls -l /local/A/foo/bar /nfs/A/foo/bar
-rw-r--r--  1 root root 0 Dec 17 03:47 /local/A/foo/bar
-rw-r--r--  1 root root 0 Nov 22 00:36 /nfs/A/foo/bar

We can see the local mtime is updated, but the NFS mount still shows
the old value. The patch below makes it work:

Initial setup...
07:11:02 up 25 days, 1 min,  0 users,  load average: 0.15, 0.03, 0.04

# ls -l /local/A/foo/bar /nfs/A/foo/bar
-rw-r--r--  1 root root 0 Jan 11 07:11 /local/A/foo/bar
-rw-r--r--  1 root root 0 Jan 11 07:11 /nfs/A/foo/bar

# touch /local/A/foo/bar

# ls -l /local/A/foo/bar /nfs/A/foo/bar
-rw-r--r--  1 root root 0 Jan 11 07:14 /local/A/foo/bar
-rw-r--r--  1 root root 0 Jan 11 07:14 /nfs/A/foo/bar

Signed-off-by: Fabio Olive Leite <fleite@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years ago64 bit ino support for NFS client
Peter Staubach [Fri, 3 Aug 2007 19:07:10 +0000 (15:07 -0400)] 
64 bit ino support for NFS client

Hi.

Attached is a patch to modify the NFS client code to support
64 bit ino's, as appropriate for the system and the NFS
protocol version.

The code basically just expand the NFS interfaces for routines
which handle ino's from using ino_t to u64 and then uses the
fileid in the nfs_inode instead of i_ino in the inode.  The
code paths that were updated are in the getattr method and
the readdir methods.

This should be no real change on 64 bit platforms.  Since
the ino_t is an unsigned long, it would already be 64 bits
wide.

    Thanx...

           ps

Signed-off-by: Peter Staubach <staubach@redhat.com>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoSUNRPC: Convert rpc_pipefs to use the generic filesystem notification hooks
Trond Myklebust [Fri, 8 Jun 2007 02:44:34 +0000 (22:44 -0400)] 
SUNRPC: Convert rpc_pipefs to use the generic filesystem notification hooks

This will allow rpc.gssd to use inotify instead of dnotify in order to
locate new rpc upcall pipes.

This also requires the exporting of __audit_inode_child(), which is used by
fsnotify_create() and fsnotify_mkdir(). Ccing David Woodhouse.

Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Fall back to synchronous writes when a background write errors...
Trond Myklebust [Wed, 25 Jul 2007 18:09:54 +0000 (14:09 -0400)] 
NFS: Fall back to synchronous writes when a background write errors...

This helps prevent huge queues of background writes from building up
whenever the server runs out of disk or quota space, or if someone changes
the file access modes behind our backs.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Writeback optimisation
Trond Myklebust [Wed, 25 Jul 2007 18:09:54 +0000 (14:09 -0400)] 
NFS: Writeback optimisation

Schedule writes using WB_SYNC_NONE first, then come back for a second pass
using WB_SYNC_ALL.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Clean up NFS writeback flush code
Trond Myklebust [Fri, 20 Jul 2007 17:13:28 +0000 (13:13 -0400)] 
NFS: Clean up NFS writeback flush code

The only user of nfs_sync_mapping_range() is nfs_getattr(), which uses it
to flush out the entire inode without sending a commit. We therefore
replace nfs_sync_mapping_range with a more appropriate helper.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoVFS: Remove writeback_control->fs_private
Trond Myklebust [Sun, 22 Jul 2007 23:27:46 +0000 (19:27 -0400)] 
VFS: Remove writeback_control->fs_private

The only user of this field was NFS.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Clean up nfs_writepages()
Trond Myklebust [Sun, 22 Jul 2007 23:27:32 +0000 (19:27 -0400)] 
NFS: Clean up nfs_writepages()

Just call write_cache_pages directly instead of hacking the writeback
control structure in order to find out if we were called from writepages()
or directly from the VM.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Clean up write code...
Trond Myklebust [Sun, 22 Jul 2007 21:09:05 +0000 (17:09 -0400)] 
NFS: Clean up write code...

The addition of nfs_page_mkwrite means that We should no longer need to
create requests inside nfs_writepage()

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoNFS: Add the helper nfs_vm_page_mkwrite
Trond Myklebust [Sun, 22 Jul 2007 21:09:05 +0000 (17:09 -0400)] 
NFS: Add the helper nfs_vm_page_mkwrite

This is needed in order to set up a proper nfs_page request for mmapped
files.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
17 years agoLinux 2.6.23 v2.6.23
Linus Torvalds [Tue, 9 Oct 2007 20:31:38 +0000 (13:31 -0700)] 
Linux 2.6.23

17 years agoMerge branch 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus
Linus Torvalds [Tue, 9 Oct 2007 19:38:44 +0000 (12:38 -0700)] 
Merge branch 'upstream' of git://ftp.linux-mips.org/upstream-linus

* 'upstream' of git://ftp.linux-mips.org/pub/scm/upstream-linus:
  [MIPS] Au1000: set the PCI controller IO base
  [MIPS] Alchemy: Fix USB initialization.
  [MIPS] IP32: Fix fatal typo in address computation.

17 years agoNLM: Fix a memory leak in nlmsvc_testlock
Trond Myklebust [Tue, 9 Oct 2007 15:04:57 +0000 (11:04 -0400)] 
NLM: Fix a memory leak in nlmsvc_testlock

The recent fix for a circular lock dependency unfortunately introduced a
potential memory leak in the event where the call to nlmsvc_lookup_host
fails for some reason.

Thanks to Roel Kluin for spotting this.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agosata_mv: correct S/G table limits
Jeff Garzik [Tue, 9 Oct 2007 17:51:57 +0000 (13:51 -0400)] 
sata_mv: correct S/G table limits

The recent mv_fill_sg() rewrite, to fix a data corruption problem
related to IOMMU virtual merging, forgot to account for the
potentially-increased size of the scatter/gather table after its run.

Additionally, the DMA boundary is reduced from 0xffffffff to 0xffff
to more closely match the needs of mv_fill_sg().

Signed-off-by: Jeff Garzik <jgarzik@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[MIPS] Au1000: set the PCI controller IO base
Florian Fainelli [Tue, 25 Sep 2007 15:07:30 +0000 (17:07 +0200)] 
[MIPS] Au1000: set the PCI controller IO base

The PCI controller IO base was not set in the au1000 pci code.

Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Florian Fainelli <florian.fainelli@telecomint.eu>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] Alchemy: Fix USB initialization.
Florian Fainelli [Tue, 25 Sep 2007 15:07:24 +0000 (17:07 +0200)] 
[MIPS] Alchemy: Fix USB initialization.

This patch fixes a wrong ifdef in the board setup code, leading to the GPIO
pin not being pulled high, and thus the USB switch not being powered at all.

This finishes the rename of CONFIG_USB_OHCI to CONFIG_USB_OHCI_HCD, which
started in 2005 (before 2.6.12-rc2), then probably because things were
working anyway for most people got forgotten.

[Ralf: Paolo's original patch didn't fix the module case, Florian's patch
only fixed MTX1 etc. so this is a combined patch plus some cleanups.]

Cc: Giuseppe Patanè <giuseppe.patane@tvblob.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John Crispin <blogic@openwrt.org>
Signed-off-by: Florian Fainelli <florian.fainelli@telecomint.eu>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years ago[MIPS] IP32: Fix fatal typo in address computation.
Giuseppe Sacco [Sat, 6 Oct 2007 17:55:03 +0000 (19:55 +0200)] 
[MIPS] IP32: Fix fatal typo in address computation.

Signed-off-by: Giuseppe Sacco <eppesuig@debian.org>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
17 years agoCorrect Makefile rule for generating custom keymap
Maarten Bressers [Mon, 8 Oct 2007 22:59:13 +0000 (15:59 -0700)] 
Correct Makefile rule for generating custom keymap

When building a custom keymap, after setting GENERATE_KEYMAP := 1 in
drivers/char/Makefile, the kernel build fails like this:

    CC      drivers/char/vt.o
  make[2]: *** No rule to make target `drivers/char/%.map', needed by `drivers/char/defkeymap.c'.  Stop.
  make[1]: *** [drivers/char] Error 2
  make: *** [drivers] Error 2

This was caused by commit af8b128719f5248e542036ea994610a29d0642a6, which
deleted a necessary colon from the Makefile rule that generates the keymap,
since that rule contains both a target and a target-pattern.  The following
patch puts the colon back:

Signed-off-by: Maarten Bressers <mbres@gentoo.org>
Cc: Yoichi Yuasa <yoichi_yuasa@tripeaks.co.jp>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoISDN: Fix data access out of array bounds
Karsten Keil [Mon, 8 Oct 2007 10:52:09 +0000 (12:52 +0200)] 
ISDN: Fix data access out of array bounds

Fix against access random data bytes outside the dev->chanmap array.
Thanks to Oliver Neukum for pointing me to this issue.

Signed-off-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Mon, 8 Oct 2007 19:59:10 +0000 (12:59 -0700)] 
Merge branch 'master' of /linux/kernel/git/davem/net-2.6

* 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [IPv6]: Fix ICMPv6 redirect handling with target multicast address
  [PKT_SCHED] cls_u32: error code isn't been propogated properly
  [ROSE]: Fix rose.ko oops on unload
  [TCP]: Fix fastpath_cnt_hint when GSO skb is partially ACKed

17 years agoAIO: fix cleanup in io_submit_one(...)
Yan Zheng [Mon, 8 Oct 2007 19:16:20 +0000 (12:16 -0700)] 
AIO: fix cleanup in io_submit_one(...)

When IOCB_FLAG_RESFD flag is set and iocb->aio_resfd is incorrect,
statement 'goto out_put_req' is executed. At label 'out_put_req',
aio_put_req(..) is called, which requires 'req->ki_filp' set.

Signed-off-by: Yan Zheng<yanzheng@21cn.com>
Cc: Zach Brown <zach.brown@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agofix page release issue in filemap_fault
Yan Zheng [Mon, 8 Oct 2007 17:08:37 +0000 (10:08 -0700)] 
fix page release issue in filemap_fault

find_lock_page increases page's usage count, we should decrease it
before return VM_FAULT_SIGBUS

Signed-off-by: Yan Zheng<yanzheng@21cn.com>
Cc: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agofix VM_CAN_NONLINEAR check in sys_remap_file_pages
Yan Zheng [Mon, 8 Oct 2007 17:05:48 +0000 (10:05 -0700)] 
fix VM_CAN_NONLINEAR check in sys_remap_file_pages

The test for VM_CAN_NONLINEAR always fails

Signed-off-by: Yan Zheng<yanzheng@21cn.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agomm: set_page_dirty_balance() vs ->page_mkwrite()
Peter Zijlstra [Mon, 8 Oct 2007 16:54:37 +0000 (18:54 +0200)] 
mm: set_page_dirty_balance() vs ->page_mkwrite()

All the current page_mkwrite() implementations also set the page dirty. Which
results in the set_page_dirty_balance() call to _not_ call balance, because the
page is already found dirty.

This allows us to dirty a _lot_ of pages without ever hitting
balance_dirty_pages().  Not good (tm).

Force a balance call if ->page_mkwrite() was successful.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years ago[IPv6]: Fix ICMPv6 redirect handling with target multicast address
Brian Haley [Mon, 8 Oct 2007 07:12:05 +0000 (00:12 -0700)] 
[IPv6]: Fix ICMPv6 redirect handling with target multicast address

When the ICMPv6 Target address is multicast, Linux processes the
redirect instead of dropping it.  The problem is in this code in
ndisc_redirect_rcv():

         if (ipv6_addr_equal(dest, target)) {
                 on_link = 1;
         } else if (!(ipv6_addr_type(target) & IPV6_ADDR_LINKLOCAL)) {
                 ND_PRINTK2(KERN_WARNING
                            "ICMPv6 Redirect: target address is not
link-local.\n");
                 return;
         }

This second check will succeed if the Target address is, for example,
FF02::1 because it has link-local scope.  Instead, it should be checking
if it's a unicast link-local address, as stated in RFC 2461/4861 Section
8.1:

       - The ICMP Target Address is either a link-local address (when
         redirected to a router) or the same as the ICMP Destination
         Address (when redirected to the on-link destination).

I know this doesn't explicitly say unicast link-local address, but it's
implied.

This bug is preventing Linux kernels from achieving IPv6 Logo Phase II
certification because of a recent error that was found in the TAHI test
suite - Neighbor Disovery suite test 206 (v6LC.2.3.6_G) had the
multicast address in the Destination field instead of Target field, so
we were passing the test.  This won't be the case anymore.

The patch below fixes this problem, and also fixes ndisc_send_redirect()
to not send an invalid redirect with a multicast address in the Target
field.  I re-ran the TAHI Neighbor Discovery section to make sure Linux
passes all 245 tests now.

Signed-off-by: Brian Haley <brian.haley@hp.com>
Acked-by: David L Stevens <dlstevens@us.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[PKT_SCHED] cls_u32: error code isn't been propogated properly
Stephen Hemminger [Mon, 8 Oct 2007 06:57:45 +0000 (23:57 -0700)] 
[PKT_SCHED] cls_u32: error code isn't been propogated properly

Signed-off-by: Stephen Hemminger <shemminger@linux-foundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[ROSE]: Fix rose.ko oops on unload
Alexey Dobriyan [Mon, 8 Oct 2007 06:44:17 +0000 (23:44 -0700)] 
[ROSE]: Fix rose.ko oops on unload

Commit a3d384029aa304f8f3f5355d35f0ae274454f7cd aka
"[AX.25]: Fix unchecked rose_add_loopback_neigh uses"
transformed rose_loopback_neigh var into statically allocated one.
However, on unload it will be kfree's which can't work.

Steps to reproduce:

modprobe rose
rmmod rose

BUG: unable to handle kernel NULL pointer dereference at virtual address 00000008
 printing eip:
c014c664
*pde = 00000000
Oops: 0000 [#1]
PREEMPT DEBUG_PAGEALLOC
Modules linked in: rose ax25 fan ufs loop usbhid rtc snd_intel8x0 snd_ac97_codec ehci_hcd ac97_bus uhci_hcd thermal usbcore button processor evdev sr_mod cdrom
CPU:    0
EIP:    0060:[<c014c664>]    Not tainted VLI
EFLAGS: 00210086   (2.6.23-rc9 #3)
EIP is at kfree+0x48/0xa1
eax: 00000556   ebx: c1734aa0   ecx: f6a5e000   edx: f7082000
esi: 00000000   edi: f9a55d20   ebp: 00200287   esp: f6a5ef28
ds: 007b   es: 007b   fs: 0000  gs: 0033  ss: 0068
Process rmmod (pid: 1823, ti=f6a5e000 task=f7082000 task.ti=f6a5e000)
Stack: f9a55d20 f9a5200c 00000000 00000000 00000000 f6a5e000 f9a5200c f9a55a00
       00000000 bf818cf0 f9a51f3f f9a55a00 00000000 c0132c60 65736f72 00000000
       f69f9630 f69f9528 c014244a f6a4e900 00200246 f7082000 c01025e6 00000000
Call Trace:
 [<f9a5200c>] rose_rt_free+0x1d/0x49 [rose]
 [<f9a5200c>] rose_rt_free+0x1d/0x49 [rose]
 [<f9a51f3f>] rose_exit+0x4c/0xd5 [rose]
 [<c0132c60>] sys_delete_module+0x15e/0x186
 [<c014244a>] remove_vma+0x40/0x45
 [<c01025e6>] sysenter_past_esp+0x8f/0x99
 [<c012bacf>] trace_hardirqs_on+0x118/0x13b
 [<c01025b6>] sysenter_past_esp+0x5f/0x99
 =======================
Code: 05 03 1d 80 db 5b c0 8b 03 25 00 40 02 00 3d 00 40 02 00 75 03 8b 5b 0c 8b 73 10 8b 44 24 18 89 44 24 04 9c 5d fa e8 77 df fd ff <8b> 56 08 89 f8 e8 84 f4 fd ff e8 bd 32 06 00 3b 5c 86 60 75 0f
EIP: [<c014c664>] kfree+0x48/0xa1 SS:ESP 0068:f6a5ef28

Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years ago[TCP]: Fix fastpath_cnt_hint when GSO skb is partially ACKed
Ilpo Järvinen [Mon, 8 Oct 2007 06:43:10 +0000 (23:43 -0700)] 
[TCP]: Fix fastpath_cnt_hint when GSO skb is partially ACKed

When only GSO skb was partially ACKed, no hints are reset,
therefore fastpath_cnt_hint must be tweaked too or else it can
corrupt fackets_out. The corruption to occur, one must have
non-trivial ACK/SACK sequence, so this bug is not very often
that harmful. There's a fackets_out state reset in TCP because
fackets_out is known to be inaccurate and that fixes the issue
eventually anyway.

In case there was also at least one skb that got fully ACKed,
the fastpath_skb_hint is set to NULL which causes a recount for
fastpath_cnt_hint (the old value won't be accessed anymore),
thus it can safely be decremented without additional checking.

Reported by Cedric Le Goater <clg@fr.ibm.com>

Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Signed-off-by: David S. Miller <davem@davemloft.net>
17 years agoDriver core: fix SYSF_DEPRECATED breakage for nested classdevs
Dmitry Torokhov [Sun, 7 Oct 2007 16:22:21 +0000 (12:22 -0400)] 
Driver core: fix SYSF_DEPRECATED breakage for nested classdevs

We should only reparent to a class former class devices that
form the base of class hierarchy. Nested devices should still
grow from their real parents.

Signed-off-by: Dmitry Torokhov <dtor@mail.ru>
Tested-by: Andrey Borzenkov <arvidjaar@mail.ru>
Tested-by: Anssi Hannula <anssi.hannula@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394...
Linus Torvalds [Sun, 7 Oct 2007 23:41:09 +0000 (16:41 -0700)] 
Merge branch 'for-linus' of git://git./linux/kernel/git/ieee1394/linux1394-2.6

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ieee1394/linux1394-2.6:
  firewire: point to migration document

17 years agoAdd manufacturer and card id of teltonica pcmcia modems
Attila Kinali [Sun, 7 Oct 2007 07:24:38 +0000 (00:24 -0700)] 
Add manufacturer and card id of teltonica pcmcia modems

Add the manufacturer and card id of teltonica pcmcia modems to serial_cs.c

Signed-off-by: Attila Kinali <attila@kinali.ch>
Acked-by: Alan Cox <alan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>