Before timers were run per-cpu as they are now, one used
to use the SMP_TIMER_* macros to move globally synchronized
timer work into tasklets.

But with per-cpu timers this is useless work and in fact makes
thing run more slowly.

into kernel.bkbits.net:/home/davem/net-2.5

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

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

into tetrachloride.(none):/mnt/raid/src/kernel/2.5/agpgart

The calling task must have a valid reclaim_state when running page
reclaim.  But I had forgotten about shrink_all_memory().

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

Move the PCI core to start using these, enabling proper reference counting
on struct pci_dev.

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

From Jeff Wiedemeier and Richard Henderson

into nuts.ninka.net:/home/davem/src/BK/net-2.5

into tetrachloride.(none):/mnt/raid/src/kernel/2.5/agpgart

Adjusts the IPTables targets to handle non-linear packets.
ipt_ULOG untested.

Adjusts the IPTables matches to handle non-linear packets.
Untested: ipt_ah and ipt_esp.

Adjusts the IPTables core to handle non-linear packets.  Extensions
done separately in next patch.  Also, comments about when to set
"me" field in struct ipt_match, ipt_table and ipt_target are
corrected.

Generic suspend/resume methods are no more.

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

Trivial bugfix: opening a file that doesn't have ->direct_IO() with
O_DIRECT passed in flags fails (as it should) but doesn't call
->release() even though ->open() had been successful.

IOW, we have a leak there - e.g.  open() of a tty with O_DIRECT will
leak a reference to tty and tty_driver.  Besides, quite a few drivers
are _not_ happy with struct file being freed without ->release() after
->open() has returned 0.