This reinstates the count parameter to store() methods.

This updates new cpufreq sysfs code to new rules.

The following patch provides minimal client support for the
(mandatory) Kerberos V5 authentication mechanism under RPCSEC_GSS.
See RFC2623 and RFC3010 for protocol details.

Only authentication is supported for the moment. Data integrity and/or
data privacy (encryption) will be implemented at a later stage.

This patches the RPCSEC_GSS client to make use of the upcall mechanism
that was provided by patch [3/6].

If an RPC task presents a non-uptodate credential to call_refresh(),
a user daemon is contacted by means of a dedicated rpc_pipefs pipe.
The daemon is then fed the uid for which it must establish a new RPCSEC
security context.

While the daemon goes about its business, the RPC task is put to sleep
on a wait queue in order to allow the 'rpciod' process to service other
requests. If another task wants to use the same credential, it too will
be put to sleep once it reaches call_refresh(). A timeout mechanism
ensures that requests are retried (or that 'soft' mounts fail) if the
daemon crashes / is killed.

Once the daemon has established the RPCSEC context, it writes the result
back to the pipe, causing the credential to be updated. Those RPC tasks
that were sleeping on the context are automatically woken up, and
their execution can proceed.

This patch provides the basic framework for RPCSEC_GSS authentication
in the RPC client. The protocol is fully described in RFC-2203.
Sun has supported it in their commercial NFSv3 and v2 implementations
for quite some time, and it has been specified in RFC3010 as being
mandatory for NFSv4.

  - Update the mount_data struct for NFSv2 and v3 in order to allow them
    to pass an RPCSEC_GSS security flavour. Compatibility with existing
    versions of the 'mount' program is ensured by requiring that RPCSEC
    support be enabled using the new flag NFS_MOUNT_SECFLAVOUR.
  - Provide secure authentication, and later data encryption on
    a per-user basis. A later patch will an provide an implementation
    of the Kerberos 5 security mechanism. SPKM and LIPKEY are still
    being planned.
  - Security context negotiation and initialization are all assumed
    to be done in userland. A later patch will provide the actual upcall
    mechanisms to allow for this.

This patch provides the upcall mechanism that will be used for communicating
with the RPCSEC client user daemons.

It sets up a 'ramfs' style filesystem (rpc_pipefs) that is populated with
named pipes. Each time the kernel initializes a new NFS, lockd, statd or
portmapper client, a directory automatically gets set up in this fs.
The directory is initially only populated with a single file "info"
that provides information such as the server IP address, the port number
and the RPC service for the benefit of the user daemon.

When an RPCSEC_GSS mechanism needs to communicate with the daemon, it
is provided with a toolkit for setting up a named pipe in the same
directory. It can then perform upcalls/downcalls in order to talk to the
daemon in much the same way as is done by CODA.

The NFSv4 client will also need to use this same filesystem to communicate
with its user daemon in order to do name-to-uid/name-from-uid and
name-to-gid/name-from-gid translation.

The RPCSEC_GSS user context defines a 'sequence number' in the AUTH header
fields in order to provide protection against replay attacks. This
number needs to lie within a given 'window', and is required to be updated
even when retransmitting dropped requests.

In order to allow this update to occur, move the XDR 'encode' phase
so that it is done immediately before writing the data to the socket.

Clean up RPC client credcache lookups.

    - Remove the limitation whereby the RPC client may only look up
      credentials for the current task.

The ability to lookup arbitrary credentials is needed in order to allow
a user daemon to set the RPCSEC_GSS private information once it
has finished negotiating the RPCSEC user context with the server.

This patch has a bunch of io scheduler goodies that are, by now, well
tested in -mm and by self and Nick Piggin. In order of interest:

- Use rbtree data structure for sorting of requests. Even with the
  default queue lengths that are fairly short, this cuts a lot of run
  time for io scheduler intensive work loads. If we go to longer queue
  lengths, it very quickly becomes a necessity.

- Add sysfs interface for the tunables. At the same time, finally kill
  the BLKELVGET/BLKELVSET completely. I made these return -ENOTTY in
  2.5.1, but there are left-overs around the kernel. This old interface
  was never any good, it was centered around just one io scheduler.

The io scheduler core itself has received count less hours of tuning by
myself and Nick, should be in pretty good shape. Please apply.

Andrew, I made some sysfs changes to the version from 2.5.56-mm1. It
didn't even compile without warnings (or work, for that matter), as the
sysfs store/show procedures needed updating. Hmm?

Updates for 2.5.56:
 - Integrate Stephen Rothwell's compat code
 - OProfile support for Randolph Chung
 - Makefile updates from Sam Ravnborg
 - Regenerated defconfig as requested by Rusty
 - Generic device model updates (James Bottomley)
 - And lots of general updating bugfixing, etc.

This patch by Hiroshi Miura adds a cpufreq driver for Cyrix MediaGX and
National Semiconductor Geode processors using "Suspend Modulation". It's
partly based on Zwane Mwaikambo's work.

This patch adds a sysfs interface to the cpufreq core, and marks the
previous /proc/cpufreq interface as deprecated.

As in drivers/base/cpu.c a "CPU driver" is registered, cpufreq acts as
"interface" to this, offering the following files for each CPU
(in /system/devices/sys/cpu.../) where CPUfreq support is present

cpuinfo_min_freq (ro)	- minimum frequency (in kHz) the CPU supports
cpuinfo_max_freq (ro)	- maximum frequency (in kHz) the CPU supports
scaling_min_freq (rw)	- minimum frequency (in kHz) cpufreq may scale
			     the CPU core to
scaling_max_freq (rw)	- maximum frequency (in kHz) cpufreq may scale
			     the CPU core to
scaling_governor (rw)	- governor == "A feedback device on a machine
			      or engine that is used to provide
			      automatic control, as of speed,
			      pressure, or temperature" [1, as noted
			      by David Kimdon]. Decides what frequency
			      is used. Currently, only "performance"
			      and "powersave" are supported, more may
			      be added later.

(In future, a file scaling_driver (ro) which shows what CPUfreq driver
is used (arm-sa1100,  gx-suspmod, speedstep, longrun, powernow-k6,
...) might be added, and this driver will be allowed to add files
scaling_driver_* for driver-specific settings like "prefer fast FSB".
And scaling_governor_* files might offer settings for the governor.)

To implement this sysfs interface, the driver model "interface" code
is used. Unfortunately, it has a non-trivial locking bug in
drivers/base/intf.c: there's a down_write call for
cls->subsys.rwsem in add_intf(), which then calls add(), which may call
intf->add_device(), which may call interface_add_data(), which calls
kobject_register(), which calls kobject_add(), which then tries to
down_write cls->subsys.rwsem. Remember, that was already locked writable
in add_intf().

Because of that, interface_add_data() is commented out; this means
that no link in /system/class/cpu/cpufreq is added, and that the
dev-removal code isn't called. This shouldn't be a problem yet,
though; as no cpufreq driver I know of is capable of CPU hotplugging.

    Dominik

[1] http://dictionary.reference.com/search?q=governor

When mounting a Linux partition with NFSv2, READDIR fills in the last
cookie without properly byte-swapping it.  The following READDIR will
fail, so the NFS client sees a truncated directory.

Fix it thus.

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

into osdl.org:/home/mochel/src/kernel/devel/linux-2.5-core

- Remove count and off parameters from per-node meminfo show() method.

- Remove count and off parameters from show() methods.
- Remove manual handling of reading from an offset, since the sysfs core
  handles that now. 
- Remove temp. buffer. 

- Remove count and off parameters from show() methods.
- Remove off parameter from store() methods.

Note 1: These have not been tested, but should be obviously correct.

Note 2: snprintf() was replaced with sprintf() where the filled buffer 
would obviously be < PAGE_SIZE. (like when printing single integer values).
In places where strings were printed, PAGE_SIZE is used as the max string
size.

- Reinstate count parameter for store() methods.
- Remove off parameter from st.c and osst.c sysfs methods.
- Remove count parameter from st.c and osst.c show() methods.

- Fixup struct device_attribute.
- Fix the default device attributes.

- Fixup bus, driver, and class methods.

- Fixup subsys_sysfs_ops along the way.

...instead of returning the error value (which is number of bytes remaining).

into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

  - ISA PnP - removed isapnp_card_protocol reference
  - NE1000/2000 - fixed exit sequence and bugs in PnP code
  - aironet4500 - fixed exit sequence
  - ISDN - hisax_fcpcipnp - fixed compilation
  - OSS sound drivers - ad1848, cs4232 - updated to latest PnP code

As you mentioned, we do not correctly identify the P4-based Celeron in
the kernel configuration help.  Unfortunately, Intel has called all
Celeron products simply "Celeron", so we call these "P4-based Celerons".

Check for PnP BIOS in all fault paths, not just in do_trap().

into home.transmeta.com:/home/torvalds/v2.5/linux

into samba.org:/scratch/anton/sfr

into samba.org:/scratch/anton/sfr

into samba.org:/scratch/anton/for-alan

into samba.org:/scratch/anton/for-alan