[PATCH] eCryptfs: xattr flags and mount options

This patch set introduces the ability to store cryptographic metadata into an
lower file extended attribute rather than the lower file header region.

This patch set implements two new mount options:

ecryptfs_xattr_metadata
 - When set, newly created files will have their cryptographic
   metadata stored in the extended attribute region of the file rather
   than the header.

   When storing the data in the file header, there is a minimum of 8KB
   reserved for the header information for each file, making each file at
   least 12KB in size.  This can take up a lot of extra disk space if the user
   creates a lot of small files.  By storing the data in the extended
   attribute, each file will only occupy at least of 4KB of space.

   As the eCryptfs metadata set becomes larger with new features such as
   multi-key associations, most popular filesystems will not be able to store
   all of the information in the xattr region in some cases due to space
   constraints.  However, the majority of users will only ever associate one
   key per file, so most users will be okay with storing their data in the
   xattr region.

   This option should be used with caution.  I want to emphasize that the
   xattr must be maintained under all circumstances, or the file will be
   rendered permanently unrecoverable.  The last thing I want is for a user to
   forget to set an xattr flag in a backup utility, only to later discover
   that their backups are worthless.

ecryptfs_encrypted_view
 - When set, this option causes eCryptfs to present applications a
   view of encrypted files as if the cryptographic metadata were
   stored in the file header, whether the metadata is actually stored
   in the header or in the extended attributes.

   No matter what eCryptfs winds up doing in the lower filesystem, I want
   to preserve a baseline format compatibility for the encrypted files.  As of
   right now, the metadata may be in the file header or in an xattr.  There is
   no reason why the metadata could not be put in a separate file in future
   versions.

   Without the compatibility mode, backup utilities would have to know to
   back up the metadata file along with the files.  The semantics of eCryptfs
   have always been that the lower files are self-contained units of encrypted
   data, and the only additional information required to decrypt any given
   eCryptfs file is the key.  That is what has always been emphasized about
   eCryptfs lower files, and that is what users expect.  Providing the
   encrypted view option will provide a way to userspace applications wherein
   they can always get to the same old familiar eCryptfs encrypted files,
   regardless of what eCryptfs winds up doing with the metadata behind the
   scenes.

This patch:

Add extended attribute support to version bit vector, flags to indicate when
xattr or encrypted view modes are enabled, and support for the new mount
options.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] eCryptfs: Public key; packet management

Public key support code.  This reads and writes packets in the header that
contain public key encrypted file keys.  It calls the messaging code in the
previous patch to send and receive encryption and decryption request
packets from the userspace daemon.

[akpm@osdl.org: cleab fix]
Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] eCryptfs: Public key transport mechanism

This is the transport code for public key functionality in eCryptfs.  It
manages encryption/decryption request queues with a transport mechanism.
Currently, netlink is the only implemented transport.

Each inode has a unique File Encryption Key (FEK).  Under passphrase, a File
Encryption Key Encryption Key (FEKEK) is generated from a salt/passphrase
combo on mount.  This FEKEK encrypts each FEK and writes it into the header of
each file using the packet format specified in RFC 2440.  This is all
symmetric key encryption, so it can all be done via the kernel crypto API.

These new patches introduce public key encryption of the FEK.  There is no
asymmetric key encryption support in the kernel crypto API, so eCryptfs pushes
the FEK encryption and decryption out to a userspace daemon.  After
considering our requirements and determining the complexity of using various
transport mechanisms, we settled on netlink for this communication.

eCryptfs stores authentication tokens into the kernel keyring.  These tokens
correlate with individual keys.  For passphrase mode of operation, the
authentication token contains the symmetric FEKEK.  For public key, the
authentication token contains a PKI type and an opaque data blob managed by
individual PKI modules in userspace.

Each user who opens a file under an eCryptfs partition mounted in public key
mode must be running a daemon.  That daemon has the user's credentials and has
access to all of the keys to which the user should have access.  The daemon,
when started, initializes the pluggable PKI modules available on the system
and registers itself with the eCryptfs kernel module.  Userspace utilities
register public key authentication tokens into the user session keyring.
These authentication tokens correlate key signatures with PKI modules and PKI
blobs.  The PKI blobs contain PKI-specific information necessary for the PKI
module to carry out asymmetric key encryption and decryption.

When the eCryptfs module parses the header of an existing file and finds a Tag
1 (Public Key) packet (see RFC 2440), it reads in the public key identifier
(signature).  The asymmetrically encrypted FEK is in the Tag 1 packet;
eCryptfs puts together a decrypt request packet containing the signature and
the encrypted FEK, then it passes it to the daemon registered for the
current->euid via a netlink unicast to the PID of the daemon, which was
registered at the time the daemon was started by the user.

The daemon actually just makes calls to libecryptfs, which implements request
packet parsing and manages PKI modules.  libecryptfs grabs the public key
authentication token for the given signature from the user session keyring.
This auth tok tells libecryptfs which PKI module should receive the request.
libecryptfs then makes a decrypt() call to the PKI module, and it passes along
the PKI block from the auth tok.  The PKI uses the blob to figure out how it
should decrypt the data passed to it; it performs the decryption and passes
the decrypted data back to libecryptfs.  libecryptfs then puts together a
reply packet with the decrypted FEK and passes that back to the eCryptfs
module.

The eCryptfs module manages these request callouts to userspace code via
message context structs.  The module maintains an array of message context
structs and places the elements of the array on two lists: a free and an
allocated list.  When eCryptfs wants to make a request, it moves a msg ctx
from the free list to the allocated list, sets its state to pending, and fires
off the message to the user's registered daemon.

When eCryptfs receives a netlink message (via the callback), it correlates the
msg ctx struct in the alloc list with the data in the message itself.  The
msg->index contains the offset of the array of msg ctx structs.  It verifies
that the registered daemon PID is the same as the PID of the process that sent
the message.  It also validates a sequence number between the received packet
and the msg ctx.  Then, it copies the contents of the message (the reply
packet) into the msg ctx struct, sets the state in the msg ctx to done, and
wakes up the process that was sleeping while waiting for the reply.

The sleeping process was whatever was performing the sys_open().  This process
originally called ecryptfs_send_message(); it is now in
ecryptfs_wait_for_response().  When it wakes up and sees that the msg ctx
state was set to done, it returns a pointer to the message contents (the reply
packet) and returns.  If all went well, this packet contains the decrypted
FEK, which is then copied into the crypt_stat struct, and life continues as
normal.

The case for creation of a new file is very similar, only instead of a decrypt
request, eCryptfs sends out an encrypt request.

> - We have a great clod of key mangement code in-kernel.  Why is that
>   not suitable (or growable) for public key management?

eCryptfs uses Howells' keyring to store persistent key data and PKI state
information.  It defers public key cryptographic transformations to userspace
code.  The userspace data manipulation request really is orthogonal to key
management in and of itself.  What eCryptfs basically needs is a secure way to
communicate with a particular daemon for a particular task doing a syscall,
based on the UID.  Nothing running under another UID should be able to access
that channel of communication.

> - Is it appropriate that new infrastructure for public key
> management be private to a particular fs?

The messaging.c file contains a lot of code that, perhaps, could be extracted
into a separate kernel service.  In essence, this would be a sort of
request/reply mechanism that would involve a userspace daemon.  I am not aware
of anything that does quite what eCryptfs does, so I was not aware of any
existing tools to do just what we wanted.

>   What happens if one of these daemons exits without sending a quit
>   message?

There is a stale uid<->pid association in the hash table for that user.  When
the user registers a new daemon, eCryptfs cleans up the old association and
generates a new one.  See ecryptfs_process_helo().

> - _why_ does it use netlink?

Netlink provides the transport mechanism that would minimize the complexity of
the implementation, given that we can have multiple daemons (one per user).  I
explored the possibility of using relayfs, but that would involve having to
introduce control channels and a protocol for creating and tearing down
channels for the daemons.  We do not have to worry about any of that with
netlink.

Signed-off-by: Michael Halcrow <mhalcrow@us.ibm.com>
Cc: David Howells <dhowells@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] include/linux/nfsd/const.h: remove NFS_SUPER_MAGIC

NFS_SUPER_MAGIC is already defined in include/linux/magic.h

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: fix up svc_create_socket() to take a sockaddr struct + length

Replace existing svc_create_socket() API to allow callers to pass addresses
larger than a sockaddr_in.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: support IPv6 addresses in RPC server's UDP receive path

Add support for IPv6 addresses in the RPC server's UDP receive path.

[akpm@linux-foundation.org: cleanups]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Support IPv6 addresses in svc_tcp_accept

Modify svc_tcp_accept to support connecting on IPv6 sockets.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: add a "generic" function to see if the peer uses a secure port

The only reason svcsock.c looks at a sockaddr's port is to check whether the
remote peer is connecting from a privileged port.  Refactor this check to hide
processing that is specific to address format.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: teach svc_sendto() to deal with IPv6 addresses

CMSG_DATA comes in different sizes, depending on address family.

[akpm@linux-foundation.org: remove unneeded do/while (0)]
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Make rq_daddr field address-version independent

The rq_daddr field must support larger addresses.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Provide room in svc_rqst for larger addresses

Expand the rq_addr field to allow it to contain larger addresses.

Specifically, we replace a 'sockaddr_in' with a 'sockaddr_storage', then
everywhere the 'sockaddr_in' was referenced, we use instead an accessor
function (svc_addr_in) which safely casts the _storage to _in.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Use sockaddr_storage to store address in svc_deferred_req

Sockaddr_storage will allow us to store arbitrary socket addresses in the
svc_deferred_req struct.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Add a function to format the address in an svc_rqst for printing

There are loads of places where the RPC server assumes that the rq_addr fields
contains an IPv4 address.  Top among these are error and debugging messages
that display the server's IP address.

Let's refactor the address printing into a separate function that's smart
enough to figure out the difference between IPv4 and IPv6 addresses.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Don't set msg_name and msg_namelen when calling sock_recvmsg

Clean-up: msg_name and msg_namelen are not used by sock_recvmsg, so don't
bother to set them in svc_recvfrom.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: Cache remote peer's address in svc_sock

The remote peer's address won't change after the socket has been accepted.  We
don't need to call ->getname on every incoming request.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: aplit svc_sock_enqueue out of svc_setup_socket

Rather than calling svc_sock_enqueue at the end of svc_setup_socket, we now
call it (via svc_sock_recieved) after calling svc_setup_socket at each call
site.

We do this because a subsequent patch will insert some code between the two
calls at one call site.

Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: allow creating an RPC service without registering with portmapper

Sometimes we need to create an RPC service but not register it with the local
portmapper.  NFSv4 delegation callback, for example.

Change the svc_makesock() API to allow optionally creating temporary or
permanent sockets, optionally registering with the local portmapper, and make
it return the ephemeral port of the new socket.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] knfsd: SUNRPC: update internal API: separate pmap register and temp sockets

Currently in the RPC server, registering with the local portmapper and
creating "permanent" sockets are tied together.  Expand the internal APIs to
allow these two socket characteristics to be separately specified.

This will be externalized in the next patch.

Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Cc: Aurelien Charbon <aurelien.charbon@ext.bull.net>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] isdn-eicon: Use ARRAY_SIZE macro when appropriate

Use ARRAY_SIZE macro already defined in kernel.h

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Workaround CAPI subsystem locking issue

I think the following patch should go into the kernel, until the ISDN/CAPI
guys create the real fix for this issue.

The issue is a concurrency issue with some internal CAPI data structure
which can crash the kernel.

On my FritzCard DSL with the AVM driver it crashes about once a day without
this workaround patch.  With this workaround patch it's rock-stable (at
least on UP, but I don't see why this shouldn't work on SMP as well.  But
maybe I'm missing something.)

This workaround is kind of a sledgehammer which inserts a global lock to
wrap around all the critical sections.  Of course, this is a scalability
issue, if you have many ISDN/CAPI cards.  But it prevents a crash.  So I
vote for this fix to get merged, until people come up with a better
solution.  Better have a stable kernel that's less scalable, than a
crashing and useless kernel.

This bug is in the kernel since 2.6.15 (at least).

Signed-off-by: Michael Buesch <mb@bu3sch.de>
Cc: Kai Germaschewski <kai.germaschewski@gmx.de>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/hisax/: Convert to generic boolean-values

Signed-off-by: Richard Knutsson <ricknu-0@student.ltu.se>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/hardware/eicon/: convert to generic boolean-values

Signed-off-by: Richard Knutsson <ricknu-0@student.ltu.se>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ISDN: Rename special macro CONFIG_HISAX_HFC4S8S_PCIMEM

Rename the macro CONFIG_HISAX_HFC4S8S_PCIMEM to simply HISAX_HFC4S8S_PCIMEM so
that it no longer resembles a user-settable kernel config macro.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ISDN: Remove defunct test emulator

Based on advice from K.  Keil, get rid of remaining traces of defunct test
emulator for HISAX.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ISDN: Rename debug option CONFIG_SERIAL_NOPAUSE_IO

Based on advice from K.  Keil, rename the special debug option
CONFIG_SERIAL_NOPAUSE_IO to ELSA_SERIAL_NOPAUSE_IO so it no longer resembles a
user-selectable kernel config option.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ISDN: Rename some debugging macros to not resemble CONFIG options

Rename some of the debugging macros for ISDN AVM so that they don't resemble
kernel config settings, as they're primarily for author debugging instead.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ISDN: Fix typo "CONFIG_HISAX_QUADRO" -> "CONFIG_HISAX_SCT_QUADRO".

Replace misspelled CONFIG_HISAX_QUADRO with CONFIG_HISAX_SCT_QUADRO.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] isdn/capi: use ARRAY_SIZE when appropriate

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Acked-by: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/sc/: proper prototypes

Add proper prototypes in a header file for global code under
drivers/isdn/sc/.

Since the GNU C compiler is now able do tell us that caller and callee
disagreed about the number of arguments of setup_buffers(), this patch
also fixes this bug.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/hisax/: proper prototypes

- add functions prototypes for some global functions to header files

- remove unneeded "extern"s from some function prototypes

You might note that this patch results in a new warning - that's due to the
fact that with a proper prototype gcc is able to discover a broken
work_struct conversion.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/pcbit/: proper prototypes

Add correct prototypes in header files for global functions and variables.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Karsten Keil <kkeil@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] S3C2410 GPIO wrappers

Arch-neutral GPIO calls for S3C24xx.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SA1100 GPIO wrappers

Arch-neutral GPIO calls for SA-1100.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] PXA GPIO wrappers

Arch-neutral GPIO calls for PXA.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] AT91 GPIO wrappers

This is a first cut at making the AT91 code use the generic GPIO calls.

Note that the original AT91 GPIO calls merged the "mux pin as GPIO" and "set
GPIO direction" functionality into one API call, contrary to what's specified
as a cross-platform portable model.  So this involved a few non-inlinable
functions.

[akpm@osdl.org: cleanups]
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] OMAP GPIO wrappers

This teaches OMAP how to implement the cross-platform GPIO interfaces.

[akpm@osdl.org: cleanups]
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] GPIO core

This defines a simple and minimalist programming interface for GPIO APIs:

  - Documentation/gpio.txt ... describes things (read it)

  - include/asm-arm/gpio.h ... defines the ARM hook, which just punts
    to <asm/arch/gpio.h> for any implementation

  - include/asm-generic/gpio.h ... implement "can sleep" variants as calling
    the normal ones, for systems that don't handle i2c expanders.

The immediate need for such a cross-architecture API convention is to support
drivers that work the same on AT91 ARM and AVR32 AP7000 chips, which embed many
of the same controllers but have different CPUs.  However, several other users
have been reported, including a driver for a hardware watchdog chip and some
handhelds.org multi-CPU button drivers.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] EDAC: Add Fully-Buffered DIMM APIs to core

Eric Wollesen ported the Bluesmoke Memory Controller driver for the Intel
5000X/V/P (Blackford/Greencreek) chipset to the in kernel EDAC model.

This patch incorporates those required changes to the edac_mc.c and edac_mc.h
core files by added new Fully Buffered DIMM interface to the EDAC Core module.

Signed-off-by: eric wollesen <ericw@xmtp.net>
Signed-off-by: doug thompson <norsk5@xmission.com>
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>

[PATCH] EDAC: Add memory scrubbing controls API to core

This is an attempt of providing an interface for memory scrubbing control in
EDAC.

This patch modifies the EDAC Core to provide the Interface for memory
controller modules to implment.

The following things are still outstanding:

 - K8 is the first implemenation,

   The patch provide a method of configuring the K8 hardware memory scrubber
   via the 'mcX' sysfs directory.  There should be some fallback to a generic
   scrubber implemented in software if the hardware does not support
   scrubbing.

   Or .. the scrubbing sysfs entry should not be visible at all.

 - Only works with SDRAM, not cache,

   The K8 can scrub cache and l2cache also - but I think this is not so
   useful as the cache is busy all the time (one hopes).

   One would also expect that cache scrubbing requires hardware support.

 - Error Handling,

   I would like that errors are returned to the user in "terms of file
   system".

 - Presentation,

   I chose Bandwidth in Bytes/Second as a representation of the scrubbing
   rate for the following reasons:

   I like that the sysfs entries are sort-of textual, related to something
   that makes sense instead of magical values that must be looked up.

   "My People" wants "% main memory scrubbed per hour" others prefer "%
   memory bandwidth used" as representation, "bandwith used" makes it easy to
   calculate both versions in one-liner scripts.

   If one later wants to scrub cache, the scaling becomes wierd for K8
   changing from "blocks of 64 byte memory" to "blocks of 64 cache lines" to
   "blocks of 64 bit".  Using "bandwidth used" makes sense in all three cases,
   (I.M.O.  anyway ;-).

 - Discovery,

   There is no way to discover the possible settings and what they do
   without reading the code and the documentation.

   *I* do not know how to make that work in a practical way.

 - Bugs(??),

   other tools can set invalid values in the memory scrub control register,
   those will read back as '-1', requiring the user to reset the scrub rate.
   This is how *I* think it should be.

 - Afflicting other areas of code,

   I made changes to edac_mc.c and edac_mc.h which will show up globally -
   this is not nice, it would be better that the memory scrubbing fuctionality
   and interface could be entirely contained within the memory controller it
   applies to.

Frithiof Jensen

edac_mc.c and its .h file is a CORE helper module for EDAC
driver modules. This provides the abstraction for device specific
drivers. It is fine to modify this CORE to provide help for
new features of the the drivers

doug thompson

Signed-off-by: Frithiof Jensen <frithiof.jensen@ericson.com>
Signed-off-by: doug thompson <norsk5@xmission.com>
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>

[PATCH] EDAC: Fix in e752x mc driver

This fix/change returns the offset into the page for the ce/ue error, instead
of just 0.  The e752x dram controller reads 34:6 of the linear address with
the error.

Signed-off-by: Mike Chan <mikechan@google.com>
Signed-off-by: doug thompson <norsk5@xmission.com>
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>

[PATCH] EDAC: e752x byte access fix

The reading of the DRA registers should be a byte at a time (one register at a
time) instead of 4 bytes at a time (four registers).  Reading a dword at a
time retrieves erroneous information from all but the first register.  A
change was made to read in each register in a loop prior to using the data in
those registers.

Signed-off-by: Brian Pomerantz <bapper@mvista.com>
Signed-off-by: Dave Jiang <djiang@mvista.com>
Signed-off-by: Doug Thompson <norsk5@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Andi Kleen <ak@suse.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] EDAC: e752x bit mask fix

The fatal vs.  non-fatal mask for the sysbus FERR status is incorrect
according to the E7520 datasheet.  This patch corrects the mask to correctly
handle fatal and non-fatal errors.

Signed-off-by: Brian Pomerantz <bapper@mvista.com>
Signed-off-by: Dave Jiang <djiang@mvista.com>
Signed-off-by: Doug Thompson <norsk5@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Andi Kleen <ak@suse.de>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: remove the now unused kill_pg kill_pg_info and __kill_pg_info

Now that I have changed all of the in-tree users remove the old version of
these functions.  This should make it clear to any out of tree users that they
should be using kill_pgrp kill_pgrp_info or __kill_pgrp_info instead.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: remove now unused do_each_task_pid and while_each_task_pid

Now that I have changed all of the users remove the old version of these
functions.  This should be a clear hint to any out of tree users that they
should use do_each_pid_task and while_each_pid_task for new code.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: replace do/while_each_task_pid with do/while_each_pid_task

There isn't any real advantage to this change except that it allows the old
functions to be removed.  Which is easier on maintenance and puts the code in
a more uniform style.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] tty: update the tty layer to work with struct pid

Of kernel subsystems that work with pids the tty layer is probably the largest
consumer.  But it has the nice virtue that the assiation with a session only
lasts until the session leader exits.  Which means that no reference counting
is required.  So using struct pid winds up being a simple optimization to
avoid hash table lookups.

In the long term the use of pid_nr also ensures that when we have multiple pid
spaces mixed everything will work correctly.

Signed-off-by: Eric W. Biederman <eric@maxwell.lnxi.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: replace is_orphaned_pgrp with is_current_pgrp_orphaned

Every call to is_orphaned_pgrp passed in process_group(current) which is racy
with respect to another thread changing our process group.  It didn't bite us
because we were dealing with integers and the worse we would get would be a
stale answer.

In switching the checks to use struct pid to be a little more efficient and
prepare the way for pid namespaces this race became apparent.

So I simplified the calls to the more specialized is_current_pgrp_orphaned so
I didn't have to worry about making logic changes to avoid the race.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: use struct pid for talking about process groups in exitc

Modify has_stopped_jobs and will_become_orphan_pgrp to use struct pid based
process groups.  This reduces the number of hash tables looks ups and paves
the way for multiple pid spaces.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] pid: make session_of_pgrp use struct pid instead of pid_t

To properly implement a pid namespace I need to deal exclusively in terms of
struct pid, because pid_t values become ambiguous.

To this end session_of_pgrp is transformed to take and return a struct pid
pointer.  To avoid the need to worry about reference counting I now require my
caller to hold the appropriate locks.  Leaving callers repsonsible for
increasing the reference count if they need access to the result outside of
the locks.

Since session_of_pgrp currently only has one caller and that caller simply
uses only test the result for equality with another process group, the locking
change means I don't actually have to acquire the tasklist_lock at all.

tiocspgrp is also modified to take and release the lock.  The logic there is a
little more complicated but nothing I won't need when I convert pgrp of a tty
to a struct pid pointer.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] signal: rewrite kill_something_info so it uses newer helpers

The goal is to remove users of the old signal helper functions so they can be
removed.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] signal: use kill_pgrp not kill_pg in the sunos compatibility code

I am slowly moving to a model where all process killing is struct pid based
instead of pid_t based.  The sunos compatibility code is one of the last users
of the old pid_t based kill_pg in the kernel.  By being complete I allow for
the future removal of kill_pg from the kernel, which will ensure I don't miss
something.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] tty: fix the locking for signal->session in disassociate_ctty

commit 24ec839c431eb79bb8f6abc00c4e1eb3b8c4d517 while fixing the locking for
signal->tty got the locking wrong for signal->session.  This places our
accesses of signal->session back under the tasklist_lock where they belong.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] tty: clarify disassociate_ctty

The code to look at tty_old_pgrp and send SIGHUP and SIGCONT when it is
present only executes when disassociate_ctty is called from do_exit.  Make
this clear by adding an explict on_exit check, and explicitly setting
tty_old_pgrp to 0.

In addition fix the locking by reading tty_old_pgrp under the siglock.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] tty: make __proc_set_tty static

The aim of this patch set is to start wrapping up the struct pid conversions.
As such this patchset culminates with the removal of kill_pg, kill_pg_info,
__kill_pg_info, do_each_task_pid, and while_each_task_pid.

kill_proc, daemonize, and kernel_thread are still in my sights but there is
still work to get to them.

The first three are basic cleanups around disassociate_ctty, while working on
converting it I found several issues.  tty_old_pgrp can be a tricky concept to
wrap your head around.

 1 tty: Make __proc_set_tty static.
 2 tty: Clarify disassociate_ctty
 3 tty: Fix the locking for signal->session in disassociate_ctty

These just stop using the old helper functions.

 4 signal: Use kill_pgrp not kill_pg in the sunos compatibility code.
 5 signal: Rewrite kill_something_info so it uses newer helpers.

Then the grind to convert the tty layer and all of it's helper functions to
struct pid.

 6 pid: Make session_of_pgrp use struct pid instead of pid_t.
 7 pid: Use struct pid for talking about process groups in exit.c
 8 pid: Replace is_orphaned_pgrp with is_current_pgrp_orphaned
 9 tty: Update the tty layer to work with struct pid.

A final helper function update.

10 pid: Replace do/while_each_task_pid with do/while_each_pid_task

And the removal of the functions that are now unused.
11 pid: Remove now unused do_each_task_pid and while_each_task_pid
12 pid: Remove the now unused kill_pg kill_pg_info and __kill_pg_info

All of these should be fairly simple and to the point.

This patch:

Currently all users of __proc_set_tty are in tty_io.c so make the function
static.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Alan Cox <alan@lxorguk.ukuu.org.uk>
Cc: Oleg Nesterov <oleg@tv-sign.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Minix V3 support

This morning I needed to read a Minix V3 filesystem, but unfortunately my
2.6.19 did not support that, and neither did the downloaded 2.6.20rc4.

Fortunately, google told me that Daniel Aragones had already done the work,
patch found at http://www.terra.es/personal2/danarag/

Unfortunaly, looking at the patch was painful to my eyes, so I polished it
a bit before applying.  The resulting kernel boots, and reads the
filesystem it needed to read.

Signed-off-by: Daniel Aragones <danarag@gmail.com>
Signed-off-by: Andries Brouwer <aeb@cwi.nl>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI eeprom driver

This is adds a simple SPI EEPROM driver, providing access to the EEPROM
through sysfs much like the I2C "eeprom" driver ...  except this driver
supports write access, and multiple EEPROM sizes.

From: "Tuppa, Walter" <walter.tuppa@siemens.com>

Since I have EEPROMs on SPI with different address sizing, I made some
changes to your at25.c to support them.  Works perfectly.  (Also includes a
small bugfix for the "what size address" test.)

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Walter Tuppa <walter.tuppa@siemens.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] RTC gets sysfs wakealarm attribute

This adds a new "wakealarm" sysfs attribute to RTC class devices which support
alarm operations and are wakeup-capable:

 - It reads as either empty, or the scheduled alarm time as seconds
   since the POSIX epoch.  (That time may already have passed, since
   nothing currently enforces one-shot alarm semantics.)

 - It can be written with an alarm time in the future, again seconds
   since the POSIX epoch, which enables the alarm.

 - It can be written with an alarm time not in the future (such as 0,
   the start of the POSIX epoch) to disable the alarm.

Usage examples (some need GNU date) after "cd /sys/class/rtc/rtcN":

    alarm after 10 minutes:
# echo $(( $(cat since_epoch) + 10 * 60 )) > wakealarm
    alarm tuesday evening 10pm:
# date -d '10pm tuesday' "+%s" > wakealarm
    disable alarm:
     # echo 0 > wakealarm

This resembles the /proc/acpi/alarm file in that nothing happens when the
alarm triggers ...  except possibly waking the system from sleep.  It's also
like that in a nasty way: not much can be done to prevent one task from
clobbering another task's alarm settings.

It differs from that file in that there's no in-kernel date parser.

Note that a few RTCs ignore rtc_wkalrm.enabled when setting alarms, or aren't
set up correctly, so they won't yet behave with this attribute.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Acked-by: Pavel Machek <pavel@ucw.cz>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Greg KH <greg@kroah.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI doc clarifications

This clarifies some aspects of the SPI programming interface, based on
feedback from Hans-Peter Nilsson.  The in-memory representation of words is
right-aligned, so for example a twelve bit word is stored using sixteen bits
with four undefined bits in the MSB.  And controller drivers must reject
protocol tweaking modes they do not support.

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI cleanup() method param becomes non-const

I'd like to assign NULL to kfree()d members of a structure.  I can't do
that without ugly casting (see the PXA patch) when the structure pointed to
is const-qualified.  I don't really see a reason why the cleanup method
isn't allowed to alter the object it should clean up.  :-)

No, I didn't test the PXA patch, but I verified that the NULL-assignment
doesn't stop me from doing rmmod/insmodding my own spi_bitbang-based
driver.

Signed-off-by: Hans-Peter Nilsson <hp@axis.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] spi_bitbang(): use overridable setup_transfer() method

A small bug-fix for spi_bitbang: it must always call the setup_transfer
function via the overridable pointer, not assume that its
spi_bitbang_setup_transfer is sufficient.  Otherwise, if all options in the
transfers are default (0), the overrided function will never be called.

Granted, the function replacing it must call spi_bitbang_setup_transfer,
but it might also have other important things to do, even if the second
argument (the spi_transfer) is NULL.  Tested together with the other
patches on the spi_crisv32_sser and spi_crisv32_gpio drivers (not yet in
the kernel, will IIUC be submitted as part of the usual
arch-maintainer-pushes).

Signed-off-by: Hans-Peter Nilsson <hp@axis.com>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] spi: remove return in spi_unregister_driver()

Make the spi_unregister_driver() code fit in with the rest of the header
file, and only do the action if the driver passed is non-NULL.

This also makes the code a line smaller.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] spi: documentation does not need to set driver's bus_type field

The spi_register_driver() sets the bus_type field of the spi_driver being
registered, so there is no need to have it set in the driver itself.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] spi: add spi_set_drvdata() and spi_get_drvdata()

Add wrappers for getting and setting the driver data using spi_device
instead of using dev_{get|set}_drvdata with &spi->dev, to mirror the
platform_{get|set}_drvdata.

Signed-off-by: Ben Dooks <ben-linux@fluff.org>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI: Freescale iMX SPI controller driver (BIS+)

Add the SPI controller driver for Freescale i.MX(S/L/1).
Main features summary:

 > Per chip setup via board specific code and/or protocol driver.
 > Per transfer setup.
 > PIO transfers.
 > DMA transfers.
 > Managing of NULL tx / rx buffer for rd only / wr only transfers.

This patch replace patch-2.6.20-rc4-spi_imx with the following changes:
 > Few cosmetic changes.
 > Function map_dma_buffers now return 0 for success and -1 for failure.
 > Solved a bug inside spi_imx_probe function (wrong error path).
 > Solved a bug inside setup function (bad undo setup for max_speed_hz).
 > For read-only transfers, always write zero bytes.

This is almost the same as the 'BIS' version sent by Andrea, except for
updating the 'DUMMY' byte so that read-only transfers shift out zeroes.
That part of the API changed recently, since some half duplex peripheral
chips require that semantic.

Signed-off-by: Andrea Paterniani <a.paterniani@swapp-eng.it>
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI controller driver for OMAP Microwire

This adds a SPI driver for the Microwire controller on OMAP1 chips.  This
driver has been used in the Linux-OMAP tree for some time now, including
with some of those displays using standardized 9-bit commands followed by
data with 8-bit words.

Microwire only supports half duplex transfers, but that's all that most SPI
protocols need.  When full duplex, or higher speeds, are needed there are
several other controllers that can be used on OMAP.

[akpm@osdl.org: cleanups]
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Imre Deak <imre.deak@solidboot.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] SPI Kconfig fix

Minor Kconfig cleanup ... put the SPI_S3C24XX entry in the
correct location (alphabetical order).

Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] DS1302: local_irq_disable() is redundant after local_irq_save()

drivers/char/ds1302.c::get_rtc_time() contains local_irq_disable() call
after local_irq_save(). This looks redundant.

drivers/char/ds1302.c::rtc_ioctl() contains local_irq_disable() call after
local_irq_save(). This looks redundant.

Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/gigaset: new M101 driver (v2)

This patch adds the line discipline based driver for the Gigaset M101
wireless RS232 adapter. It also improves the documentation a bit.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Hansjoerg Lipp <hjlipp@web.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/telephony/ixj: Convert to generic boolean

Convert:
BOOL  -> bool
FALSE -> false
TRUE  -> true

Change a variable ('mContinue') to boolean from char, since it is used
as boolean.

Signed-off-by: Richard Knutsson <ricknu-0@student.ltu.se>
Acked-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Char: timers cleanup

- Use timer macros to set function and data members and to modify
  expiration time.
- Use DEFINE_TIMER for global timers and do not init them at run-time in
  these cases.
- del_timer_sync is common in most cases -- we want to wait for timer
  function if it's still running.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: Dave Airlie <airlied@linux.ie>
Cc: David Woodhouse <dwmw2@infradead.org>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Cc: Paul Fulghum <paulkf@microgate.com>
Cc: Kylene Jo Hall <kjhall@us.ibm.com>
Cc: Wim Van Sebroeck <wim@iguana.be>
Acked-by: Dmitry Torokhov <dtor@mail.ru> (Input bits)
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Char: specialix, isr have 2 params

specialix, isr have 2 params

pt_regs are no longer the third parameter of isr, call sx_interrupt without
it.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] change __init to __devinit in 2 rtc drivers

Change __init to __devinit in rtc drivers' probe functions.

Resolves MODPOST warnings:

WARNING: drivers/rtc/rtc-ds1553.o - Section mismatch: reference to
.init.text:ds1553_rtc_probe from .data.rel between 'ds1553_rtc_driver' (at
offset 0x0) and 'ds1553_nvram_attr'
WARNING: drivers/rtc/rtc-ds1742.o - Section mismatch: reference to
.init.text:ds1742_rtc_probe from .data.rel between 'ds1742_rtc_driver' (at
offset 0x0) and 'ds1742_nvram_attr'

Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Cc: Alessandro Zummo <a.zummo@towertech.it>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] export ufs_fs.h to userspace

Was ufs_fs.h purposefully not exported to userspace or did it just slip
through the cracks ?  assuming the latter scenario, the attached patch touches
up the relationship between ufs_fs.h and its sub headers (like ufs_fs_sb.h) so
that we can export it ...  the silo bootloader takes advantage of this header
for example.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Cc: Evgeniy Dushistov <dushistov@mail.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] close_files(): add scheduling point

close_files() can sometimes take long enough to trigger the soft lockup
detector.

Cc: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/gigaset: reduce kernel message spam

Reduce the number of kernel messages the Gigaset drivers produce in case of an
excessively long device response, from one per character exceeding the limit
to one per overlong message.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] drivers/isdn/gigaset: reduce mutex scope

Do not lock the cardstate structure mutex earlier than necessary.

Signed-off-by: Tilman Schmidt <tilman@imap.cc>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] FS: speed up rw_verify_area()

oprofile hunting showed a stall in rw_verify_area(), because of triple
indirection and potential cache misses.
(file->f_path.dentry->d_inode->i_flock)

By moving initialization of 'struct inode' pointer before the pos/count
sanity tests, we allow the compiler and processor to perform two loads by
anticipation, reducing stall, without prefetch() hints.  Even x86 arch has
enough registers to not use temporary variables and not increase text size.

I validated this patch running a bench and studied oprofile changes, and
absolute perf of the test program.

Results of my epoll_pipe_bench (source available on request) on a Pentium-M
1.6 GHz machine

Before :
# ./epoll_pipe_bench -l 30 -t 20
Avg: 436089 evts/sec read_count=8843037 write_count=8843040 21.218390 samples
per call
(best value out of 10 runs)

After :
# ./epoll_pipe_bench -l 30 -t 20
Avg: 470980 evts/sec read_count=9549871 write_count=9549894 21.216694 samples
per call
(best value out of 10 runs)

oprofile CPU_CLK_UNHALTED events gave a reduction from 5.3401 % to 2.5851 %
for the rw_verify_area() function.

Signed-off-by: Eric Dumazet <dada1@cosmosbay.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] com20020 build fix

Need to export com20020 symbols for com20020_cs also.

WARNING:  "com20020_found" [drivers/net/pcmcia/com20020_cs.ko] undefined!
WARNING:  "com20020_check" [drivers/net/pcmcia/com20020_cs.ko] undefined!

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Cc: Esben Nielsen <nielsen.esben@googlemail.com>
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Dominik Brodowski <linux@dominikbrodowski.net>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] lockdep: forward declare struct task_struct

3117df0453828bd045c16244e6f50e5714667a8a causes this:

In file included from arch/s390/kernel/early.c:13:
include/linux/lockdep.h:300: warning:
"struct task_struct" declared inside parameter list
include/linux/lockdep.h:300:
warning: its scope is only this definition or
declaration, which is probably not what you want

Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: <stable@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] kexec: fix references to init in documentation for kexec

I've noticed that the boot options are not correct for in the documentation
for kdump.  The "init" keyword is not necessary, and causes a kernel panic
when booting with an initrd on Fedora 5.

[horms@verge.net.au: put original comment with the latest version of the patch]
Signed-off-by: Judith Lebzeelter <judith@osdl.org>
Acked-by: Vivek Goyal <vgoyal@in.ibm.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] swiotlb uninlinings

Optimise swiotlb.c for size.

   text    data     bss     dec     hex filename
   5009      89      64    5162    142a lib/swiotlb.o-before
   4666      89      64    4819    12d3 lib/swiotlb.o-after

For some reason my gcc (4.0.2) doesn't want to tailcall these things.

swiotlb_sync_sg_for_device:
pushq %rbp #
movl $1, %r8d #,
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret
.size swiotlb_sync_sg_for_device, .-swiotlb_sync_sg_for_device
.section .text.swiotlb_sync_sg_for_cpu,"ax",@progbits
.globl swiotlb_sync_sg_for_cpu
.type swiotlb_sync_sg_for_cpu, @function
swiotlb_sync_sg_for_cpu:
pushq %rbp #
xorl %r8d, %r8d #
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret

Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] atmel_serial: Use __raw I/O register access

Access to chip-internal registers should always be native-endian.  This is
especially important for AVR32 since it's a big-endian architecture and the
non-raw readl() and writel() macros are defined to do little-endian
accesses.

Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com>
Acked-by: Andrew Victor <andrew@sanpeople.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] warning fix: unsigned->signed

While compiling my code with -Wconversion using gcc-trunk, I always get a
bunch of warrning from headers, here is fix for them:

__getblk is alawys called with unsigned argument,
but it takes signed, the same story with __bread,__breadahead and so on.

Signed-off-by: Tomasz Kvarsin
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] ipc: save the ipc namespace while reading proc files

The problem we were assuming that current->nsproxy->ipc_ns would never
change while someone has our file in /proc/sysvipc/ file open.  Given that
this can change with both unshare and by passing the file descriptor to
another process that assumption is occasionally wrong.

Therefore this patch causes /proc/sysvipc/* to cache the namespace and
increment it's count when we open the file and to decrement the count when
we close the file, ensuring consistent operation with no surprises.

Signed-off-by: Eric W. Biederman <ebiederm@xmission.com>
Cc: Serge E. Hallyn <serue@us.ibm.com>
Cc: Herbert Poetzl <herbert@13thfloor.at>
Cc: Kirill Korotaev <dev@sw.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] reiserfs: Use ARRAY_SIZE macro when appropriate

Use ARRAY_SIZE macro already defined in kernel.h

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] OSS: Use ARRAY_SIZE macro when appropriate (2)

Use ARRAY_SIZE macro already defined in kernel.h

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] OSS: Use ARRAY_SIZE macro when appropriate

Use ARRAY_SIZE macro already defined in kernel.h

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] w1: Use ARRAY_SIZE macro when appropriate

A patch to use ARRAY_SIZE macro already defined in kernel.h

Signed-off-by: Ahmed S. Darwish <darwish.07@gmail.com>
Acked-by: Evgeniy Polyakov <johnpol@2ka.mipt.ru>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] kernel: shut up the IRQ mismatch messages

The problem is various drivers legally validly and sensibly try to claim
IRQs but the kernel insists on vomiting forth a giant irrelevant debugging
spew when the types clash.

Edit kernel/irq/manage.c go down to mismatch: in setup_irq() and ifdef out
the if clause that checks for mismatches.  It'll then just do the right
thing and work sanely.

For the current -mm kernel this will do the trick (and moves it into shared
irq debugging as in debug mode the info spew is useful).  I've had a
variant of this in my private tree for some time as I got fed up on the
mess on boxes where old legacy IRQs get reused.

Signed-off-by: Alan Cox <alan@redhat.com>
Cc: Arjan van de Ven <arjan@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Debug shared irqs

Drivers registering IRQ handlers with SA_SHIRQ really ought to be able to
handle an interrupt happening before request_irq() returns.  They also
ought to be able to handle an interrupt happening during the start of their
call to free_irq().  Let's test that hypothesis....

[bunk@stusta.de: Kconfig fixes]
Signed-off-by: David Woodhouse <dwmw2@infradead.org>
Cc: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Jesper Juhl <jesper.juhl@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] inotify: read return val fix

Fix for inotify read bug (bugzilla.kernel.org #6999)

Problem Description:
When reading from an inotify device with an insufficient sized buffer, read(2)
will return 0 with no errno set. This is because of an logically incorrect
action from the user program thus should return an more logical value. My
suggestion is return -EINVAL as for bind(2).

This patch is based on the proposal from Ryan <wolf0403@hotmail.com>, and
feedback from John McCutchan <john@johnmccutchan.com>.

Return -EINVAL if we have not passed in enough buffer space to read a single
inotify event, rather than 0 which indicates that there is nothing to read.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Acked-by: "John McCutchan" <john@johnmccutchan.com>
Cc: Ryan <wolf0403@hotmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] remove sb->s_files and file_list_lock usage in dquot.c

Iterate over sb->s_inodes instead of sb->s_files in add_dquot_ref.  This
reduces list search and lock hold time aswell as getting rid of one of the
few uses of file_list_lock which Ingo identified as a scalability problem.

Previously we called dq_op->initialize for every inode handing of a
writeable file that wasn't initialized before.  Now we're calling it for
every inode that has a non-zero i_writecount, aka a writeable file
descriptor refering to it.

Thanks a lot to Jan Kara for running this patch through his quota test
harness.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] move remove_dquot_ref to dqout.c

Remove_dquot_ref can move to dqout.c instead of beeing in inode.c under
#ifdef CONFIG_QUOTA.  Also clean the resulting code up a tiny little bit by
testing sb->dq_op earlier - it's constant over a filesystems lifetime.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Cc: Jan Kara <jack@ucw.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] scnprintf(): fix a comment

The return value of scnprintf() never exceeds @size.

Signed-off-by: Martin Peschke <mp3@de.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] fix the defaults mentioned in Documentation/nfsroot.txt

This patch fixes the documentation of nfsroot to match NFS_DEF_FILE_IO_SIZE.

Or perhaps we need to change NFS_DEF_FILE_IO_SIZE to match the
documentation?

Signed-off-by: Dan Aloni <da-x@monatomic.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] MAINTAINERS: remove two dead e-mail

Cyclades no longer serves the 2 e-mails listed in MAINTAINERS.  Remove them
and mark those entries as Orphaned.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] QUOTA: Have <linux/quota.h> include <linux/rwsem.h> explicitly

Since quota.h declares a R/W semaphore, it should include rwsem.h
explicitly.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] include/linux/kernel.h: Remove labs()

Remove labs() since it is not used/needed.

Signed-off-by: Richard Knutsson <ricknu-0@student.ltu.se>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Char: cyclades, use pci_device_id

Use pci_device_id struct instead of ushort array. Add MODULE_DEVICE_TABLE.

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

[PATCH] Char: use more PCI_DEVICE macro

Use more PCI_DEVICE macro

Signed-off-by: Jiri Slaby <jirislaby@gmail.com>
Acked-by: Wim Van Sebroeck <wim@iguana.be> (alim7101_wdt.c part)
Cc: Michael Buesch <mb@bu3sch.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>