Andrej Lajovic observes that in kd.h the definitions of K_NUMLOCK and
K_CAPSLOCK are interchanged.  He is right (and this is the case since
1.1.54; nobody noticed because setleds.c does not use these defines).
So, here is the fix.  No kernel behaviour is changed.

xprt.c:
  Merge *_write_space() in order to fix compile warning and save space.

Cheers,
  Trond

xprt.c:
  Fix the RPC over TCP socket reconnect code.

Cheers,
  Trond

xprt.c:
  Speed up synchronous RPC over TCP calls by having the
  replies delivered by the IPV4 "bottom half", instead of
  switching to the rpciod process in order to call recvmsg().
   - Remove sock_recvmsg() interface.
   - Remove rpc_xprt_pending list and rpciod_tcp_dispatcher() interface.
   - Use the new tcp_read_sock() interface to deliver data directly
     from within tcp_data_ready().
sched.c:
   - Remove references to rpciod_tcp_dispatcher.
xprt.h:
   - New set of flags to reflect the TCP record read state.

Cheers,
  Trond

Fix silly truncate status update

i810-tco: Merge changes of the 2.4 kernel into the i810-tco module. Changes included are:
Support for 82801CA(M) chipset, timer margin needs to be > 3, support for WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT.

Merge updates from the 2.4 kernel into watchdog.h .

Update of PCI-IDS for the 82801BA(M), 82801CA(M) and 82801E I/O Controller Hub's.

i810_rng: add support for other i8xx chipsets to the Random Number Generator module.
This is being done by adding the detection of the 82801E I/O Controller Hub.

This updates 2.5.x Bluetooth subsystem and removes the EXPERIMENTAL
status of Bluetooth support.

         BlueZ Core:
                 New generic HCI connection manager.
                 Complete role switch and link policy support.
                 Security mode 1 and 3 support.
                 L2CAP service level security support.
                 HCI filter support.
                 HCI frame time-stamps.
                 SCO (voice links) support.
                 Improved HCI device unregistration (device destructors).
                 Support for L2CAP signalling frame fragmentation.
                 Improved L2CAP timeout handling.
                 New HCI ioctls for changing ACL and SCO MTU.
                 Killed HCI_MAX_DEV limit.
                 Security fixes.

         HCI USB driver:
                 Performance improvements.
                 Firmware loading support.
                 Stability fixes. URB and disconnect handling rewrite.

         HCI UART driver:
                 Support for multiple UART protocols.

         HCI PCMCIA driver:
                 Support for Nokia Bluetooth PC Cards.
                 Support for Anycom Bluetooth PC/CF Cards.

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

Here is the patch to replace mk_pte_phys() with pfn_pte() and so creates
the counterpart to pte_pfn().

This patch removes VALID_PAGE(), as the test was always too late for
discontinous memory configuration. It is replaced with pfn_valid()/
virt_addr_valid(), which are used to test the original input value.
Other helper functions:
pte_pfn() - extract the page number from a pte
pfn_to_page()/page_to_pfn() - convert a page number to/from a page struct

into heisenberg.transvirtual.com:/tmp/fbdev-2.5

PRD_SEGMENTS was a leftover from struct ata_request an its per-command
sg tables.

 - Synchronize with Jens.  Applying tons of janitorian stuff to his TCQ
   code.  Making functions static where appropriate and so on...
   Marking the config entry for it experimental and so on.  His
   changelog:

	2.5.13 now has the generic tag support that I wrote included,
	here's an IDE TCQ that uses that.  Changes since the version
	posted for 2.5.12:

	Fix the ide_tcq_invalidate_queue() WIN_NOP usage needed to clear
	the internal queue on errors.  It was disabled in the last
	version due to the ata_request changes, it should work now.

	Remove Promise tcq disable check, it works just fine on Promise
	as long as we handle the two-drives-with-tcq case like we
	currently do.

 - Fix wrong usage of time_after in ide.c. This should cure the drive seek
   timeout problems some people where expierencing. This was clarified to me by
   Bartek, who apparently checked whatever the actual code is consistent with
   the comments in front of it. Thank you Bartlomiej Zolnierkiewicz.

   I think now that we should have time_past(xxx) in <linux/timer.h>.

 - Fix hpt34x.c compilation.

 - Minor improvements in ide-pci and some cleanups in ide-probe.c

 - Nuke some vastly outdated comments.

D'oh.  Missing commas...

Needed both in 2.4 and 2.5

Nothing serious. Whoever it was that did that global replace missed a
spot is all...

The following patch fixes this, and also fixes the similar problem in
scsi_debug.c:

into heisenberg.transvirtual.com:/tmp/fbdev-2.5

into hostme.bitkeeper.com:/ua/repos/f/fbdev/fbdev-2.5

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

Upgrade is imperative!

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

Minor cleanups to a scsi fibre channel HBA driver that was broken by the
introduction of SCSI_REPORT_LUNS to scsi.h in a patch that went into
2.5.11 .

[ Another transfer from Dave's tree ]

Documentation enhancement from Kai.Makisara@metla.fi.

[ This is a transfer from Dave's tree ]

Fix oopses on ISA scsi adapters.

[ This is a transfer from Dave's tree ]

- The balance_dirty_pages() logic is simply wrong.  It goes:

	if (value > threshold)
		go_and_write(value - threshold);

  which is just fine for a single process writing data.  But
  for many processes, they *all* go and bring things back into
  balance, and too much data gets written out.

- The

	go_and_write(this much)

  logic is inoperative, because I turned off the ->writeback_mapping()
  function in ext2.  So a call to writeback_unlocked_inodes(this_much)
  doesn't actually decrement and test *this_much.  It will walk every
  inode, all the time.  Silly.

So quickly fixing the above things, the amount of dirty+writeback
memory in the machine nicely stabilises at 500 megabytes across
the run.

The new consistency checks in buffer.c are complaining about
block_symlink's ->readpage against an already uptodate
page.