Originally from Matt Dobson. I've been running with this for a while
in -dj, with no noticable side-effects.

Matt:

node_zonelists looks like it should really be declared of size
MAX_NR_ZONES, not GFP_ZONEMASK.  GFP_ZONEMASK is currently 15, making
node_zonelists an array of 16 elements.  The extra zonelists are all
just duplicates of the *real* zonelists, namely the first 3 entries.
Again, if anyone can explain to me why I'm wrong in my thinking, I'd
love to know.  There's certainly no way you could bitwise-and something
with any combination of the GFP_DMA and GFP_HIGHMEM flags to refer to
the 12th zonelist or some such!  Or am I crazy?

Syncs up with 2.4

From 2.4 long long ago.

From 2.4 from way back 13 months ago..

Single spin_unlock path cuts this down a little..

Patch from Steven Cole to fix up ver_linux output on a system
with no module-init-tools, just modutils.

As noted in bugzilla #267 and at
http://marc.theaimsgroup.com/?l=linux-kernel&m=104492524815220&w=2

Erk, that's a really funny looking switch.
Every case fell through..

From Dominik Brodowski, comments says it all..

This went into 2.4 nearly a year back with the wonderfully
descriptive  "Fix from maintainer" comment.

This went into 2.4 back last August with the comment in $subject.

There are still a few places where we play with the RTC
directly, with no locking. This catches some of them.

This went into 2.4 with the following comments..

ide-scsi driver starts DMA as soon as it writes the ATAPI PACKET command
in command register and before sending the ATAPI command. This will
cause problems on many drives. Right way to do it is to start DMA after
sending the ATAPI command. I am attaching a patch that fixes this. This
patch will allow many more CD-RW drives to work reliably in DMA mode
than do today

Alan's comment to this diff previously..
"Thats the least of the 2.5 ide-scsi problems, but yes its probably one to add"

wrmsr is ok, but needs cleans up, second part (rdmsr)
is currently broken.

This patch from Alan went into 2.4 last august with the comment
"get the types right on the lib/inflate.c constants"

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

into kroah.com:/home/linux/linux/BK/gregkh-2.5

This driver is doing a writel to some random u32, rather than to a device
register.

This fixes a few problems with ide tcq, I don't know of any other known
ones (and it is solid here, survives ours of beating on it). Basically:

- Don't enable tcq if the drive isn't alone on the channel. This raises
  all sorts of fun that really requires hardware support (auto poll), or
  it's going to _suck_. I never wanted to do that, and no hardware I
  know of supports auto-poll.

- Introduce a drive blacklist so we don't enable tcq on known broken
  drives. Or enable with restrictions on some models.

- Add a check for pdc4030, apparently tcq doesn't work there (hell knows
  who would be crazy enough to pull such a stunt).

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

With the dynamic request allocation, we get the direction bits set for
us. This breaks the scsi_ioctl stuff, since we always pass in WRITE
there. So actually pass in the right direction instead.

Quiet down the TEST_UNIT_READY commands, we know these may fail (that's
the whole purpose of the command :-).

This is the small patch that we all agreed on. With this patch, we do
nice big writes/reads on ide disks that support 48-bit lba.

Add bio traversal functionality. This is a prereq for doing ide
multiwrites safely and sanely. Patch was originally done by Suparna,
Bartlomiej picked it up and changed the design somewhat. From Bart:

Main idea is now reversed - instead of introducing rq->hard_bio as
pointer for bio to be completed and using rq->bio as pointer for bio
to be submitted, rq->cbio is introduced for submissions and rq->bio
is used for completions

This minimizes changes to block layer and assures that all existing
block users are not affected by this patch.

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

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

This avoids a problem with the original ptrace code using a system
call (SWI) to implement single stepping; programs such as ltrace
do not expect to receive system call trace traps when breakpoints
are hit.

Request the memory region used for the keyboard and mouse ports,
and ioremap.

Also remove uninitialised variable warning and update mach-types.

Patch from Nicolas Pitre

From: Nicolas Pitre.

(manual entry since bk openlogging crapped out again)

Patch from Nicolas Pitre

Of course, the PLD macro is always defined even if it's empty.
Without this fix anything below ARMv5 is broken.