To recap my argument: the current cdev implementation keeps an uncounted
reference to every cdev in cdev_map.  Creators of cdevs must know to call
cdev_unmap() with the same arguments they passed to cdev_add() before
releasing the device, or that reference will remain and will oops the
kernel should user space attempt to open the (missing) device.  It's an
easy mistake to make, and, IMO, entirely unnecessary; the cdev code should
be able to do its own bookkeeping.

into kroah.com:/home/greg/linux/BK/driver-2.6

From: Marcelo Tosatti <marcelo.tosatti@cyclades.com>

Chris Siebenmann posted a fix to update the old inode ctime on rename().

This fix went into 2.2.13, but did not make into 2.4 or 2.6 kernels (the
2.2.13 diff is at the end of this message).

Here is updated 2.6 version without mark_inode_dirty(), which is not
necessary anymore because its called by "ext2_dec_count()" few lines down.

From: Stelian Pop <stelian@popies.net>

This patch enables sonypi to successfully report P1/P2 programmable keys
events on Sony Vaio Z1 laptops.

Note however that sonypi is unable to distinguish between the two events,
both of them will be reported as SONYPI_EVENT_PKEY_P1, but one event is
better than zero, so there it is.

From: Martin Hicks <mort@wildopensource.com>

Get rid of __exit from mptscsih_exit() to get the kernel to link.

From: jbarnes@sgi.com (Jesse Barnes)

Looks like Jes forgot missed some conversions in his NR_MEMBLKS removal
patch.  Here's are the fixes to get ia64 going again.

From: jbarnes@sgi.com (Jesse Barnes)

I left out some of the necessary machine vector magic in my readX_relaxed
patch, which broke the ia64 generic build.

From: Davide Libenzi <davidel@xmailserver.org>

As suggested by Eric Dumazet the following patch achieve a more compact
struct epitem on 64 bit archs.

From: James Simmons <jsimmons@infradead.org>

Fix this header so that it can again be used by userspace.  It also makes
struct fb_pixmap a internal kernel object.

I exported this because vmware needs it.

Petr has updated the vmware wrapper to open-code the sys_exit() trap so this
export is no longer needed.

The updated vmware wrapper is available at

http://platan.vc.cvut.cz/ftp/pub/vmware/vmware-any-any-update50.tar.gz

From: Andi Kleen <ak@suse.de>

The new linux/elf.h include in linux/mm.h caused all kinds of problems for
the x86-64 32bit emulation code.  This patch avoids the dependency by
moving the depending functions out of line.  It makes x86-64 compile again.

From: Gerd Knorr <kraxel@bytesex.org>

This patch is a update for the saa7134 driver in the linux kernel.

Changes:

* kernel thread cleanups (exit/rmmod sync using completions, wait queue
  fixes).

* add support for more cards.

* improved infrared remote support.

From: Gerd Knorr <kraxel@bytesex.org>

This patch brings a few cleanups/fixes for the v4l-related i2c modules:

* fix "badness in interruptible_sleep_on"

* use completions instead of semaphores to sync rmmod + kernel thread
  exit

* drop some some obsolete code.

* minor tweaks for some tv cards.

From: Stelian Pop <stelian@popies.net>

This driver had an array which contained both dma_addr_t's and kernel virtual
addresses.  Split those up, thus tidying things and avoiding a printk
warning.

From: <gb@phonema.ea.univpm.it>

The "%c" in sscanf actually reads and writes one extra character (i.e.  2
characters insted of just one), and may thus easily overflow caller's
buffer.

Also affects 2.4 tree, even if there "%c" seems not to be used at all.

From: Sergei Golod <rover@tob.ru>

This patch put moxa devices (when DEVFS enabled) under /dev/tts/M* (as
standard serial ports /dev/tts/0,1,...) instead /dev/.  Patch ported from
Stallion serial driver.

From: Martin Hicks <mort@wildopensource.com>

A patch to clean up HWGRAPH_DEBUG.  It also cleans up some compile warning
by changing the prototype of hwgraph_debug().

I ran this patch by Pat & Colin and they asked me to send it to you.

From: Martin Hicks <mort@wildopensource.com>

This is a patch that removes the unused include/asm-ia64/sn/alenlist.h
header file, along with the dead code that still referenced it.

I ran the patch by Pat Gefre and Colin Ngam.

It is performing higher-order atomic allocations then falling back, so
failures are expected.  Suppress the scary warnings.

From: Leann Ogasawara <ogasawara@osdl.org>

Audit fb_alloc_cmap() and incorporate an error path.

From: "Randy.Dunlap" <rddunlap@osdl.org>

From: Timmy Yee <shoujun@masterofpi.org>

This patch adds missing ampersands.

From: "Randy.Dunlap" <rddunlap@osdl.org>

From: Luiz Fernando Capitulino <lcapitulino@prefeitura.sp.gov.br>

this patch fixes this warning:

drivers/ide/pci/triflex.c:49: warning: `triflex_get_info' defined but not used

From: "Randy.Dunlap" <rddunlap@osdl.org>

From: Leann Ogasawara <ogasawara@osdl.org>

Handle kmalloc() failures

From: Torben Mathiasen <torben.mathiasen@hp.com>

After a lot of testing and measuring we decided to increase the read-ahead
for the CISS driver in 2.6 to 1MB.

From: Andreas Gruenbacher <agruen@suse.de>

Here is a fix for extended attributes on ext2 and ext3, reported by
Stephen Tweedie <sct@redhat.com>.

From: Stephen Tweedie <sct@redhat.com>:

When you have an EA block that is shared between multiple inodes; AND you
then change an attribute in that on one inode, AND the new attribute value
is the same as the old, then xattr computes the new EA block, finds it
still in the cache, bumps the reference count on it (and the i_blocks field
on the inode, incidentally), and leaves it incremented because we haven't
changed EA block so there's no need to drop the refcount on the old block.

So *every* time you have more than one inode sharing an EA block and you
perform an identical write to an EA, you get a leak on both i_blocks and
the EA refcount.

This is a big problem for symlinks, which rely on correct i_blocks
accounting to determine the difference between fast and slow symlinks.
With the leak, you end up thinking that a fast symlink (ie.  one small
enough to be stored in the inode direct blocks) is slow, so you dereference
the ascii contents of the symlink as if they were a disk block address.
That typically results in EIO all over the place.

From: "Randy.Dunlap" <rddunlap@osdl.org>

fix dma_addr_t type error with CONFIG_HIGHMEM64G=y

From: Andreas Gruenbacher <agruen@suse.de>

The getxattr (listxattr) syscall returns E2BIG if the buffer passed to them
is bigger than XATTR_SIZE_MAX (XATTR_LIST_MAX), no matter what buffer size is
actually required.  Here is a fix.  It also removes the xattr_alloc and
xattr_free functions which are not of much use anymore.

From: Paul Jackson <pj@sgi.com>

Explain the snprintf() return value.

drivers/atm/atmtcp.c: In function `atmtcp_c_close':
drivers/atm/atmtcp.c:258: error: invalid lvalue in assignment
drivers/atm/atmtcp.c: In function `atmtcp_create':
drivers/atm/atmtcp.c:383: error: invalid lvalue in assignment

From: Benjamin Herrenschmidt <benh@kernel.crashing.org>

This patch adds those for ppc64, except for iSeries which cannot do
readq/writeq easily, at least not as far as I know but I need to ask the
iSeries specialists in Rochester to be sure.  But anyway, iSeries shouldn't
use anything in fb.h anyway ...

It also add the proper tweaks & barriers to make sure reads are actually
done right away and not delayed indefinitely (making the CPU think the read
data is actually used) and add necessary write barriers on IO writes.  For
some reasons, ppc64 in 2.6 lacked some of these, opening potential races
within some drivers.

From: Anton Blanchard <anton@samba.org>

It doesn't link.  Add a chunk which got lost.

From: Benjamin Herrenschmidt <benh@kernel.crashing.org>

This patch was missing from my big merge.  It updates the PowerMac
"dmasound" driver.  Adds input support for some recent machines using the
tas3004 coded/mixer chip.  Code mostly written by Renzo Davoli.

This driver isn't (unfortunately) fully obsoleted by the Alsa one.  There
are lots of reports of the Alsa one not working properly on various
PowerMac machines, and some people are unhappy with Alsa in general, enough
to have ported the messy PowerMac dmasound to 2.6 :)

Adds thermal management drivers for desktop G5, Windtunnel G4s, and
recent laptops (iBook G4, aluminium 15" and 17" powerbooks)

Move the call to mac_hid_mouse_emulate_buttongs to the
right function so it works again in console mode

Add support for Apple latest laptops, add a warning when building
an SMP kernel (the core cpufreq doesn't scale loops_per_jiffies
properly on SMP kernels), cleanup some code.

Forward port Paul's latest fixes from 2.4. This driver still need
significant work to fit into the "macio" model properly, among
others.

Add support for G5s, add an option for controlling the server
mode setting on core99 machines and later.

Most of the /proc stuff still need to be moved to sysfs