OSDL is new building as of May-2003

consistently.

The driver obviously cannot rely on the interrupt handler
when it is probing for interrupts, so the identify code is
written to not use interrupts and the probing code will
disable the interrupt after having figured out which one it
is.

The non-probe code should do the same, otherwise confusion
happens. 

into home.osdl.org:/home/torvalds/v2.5/linux

confusion in the future over which constants apply to which initializers.

SGI Modid: 2.5.x-xfs:slinx:152354a

into home.osdl.org:/home/torvalds/v2.5/linux

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

- CPUFREQ_ALL_CPUS is not a valid argument any more, don't mention it
  in include/linux/cpufreq.h
- without a driver->init() function a cpufreq driver isn't even
  loaded, so remove that check in the per-CPU initialization
- whitespace and clarification changes to linux/cpufreq.

All from Dominik again..

More from Dominik

From Dominik.

Wait with the destruction of cpufreq objects until all references
are gone (code partly based on rmk's comparable patch for the PCMCIA
subsystem.

From Dominik.

into home.osdl.org:/home/torvalds/v2.5/linux

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

into home.osdl.org:/home/torvalds/v2.5/linux

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

    
This patch fixes BUG when pushing IPv6 destination options over an
IPv6 raw socket.  Patch is based on one from Thomas Graf <tgr@reeler.org>.

And spell it correctly :-)

into home.osdl.org:/home/torvalds/v2.5/linux

into home.osdl.org:/home/torvalds/v2.5/linux

 - export the __div64_32 symbol for modules;

 - add likely() to the fast path (divisor>>32 == 0);

 - add __attribute__((pure)) to __div64_32() prototype so
   the compiler knows global memory isn't clobbered;

 - avoid building __div64_32() on 64bit architectures.

* whitespace cleanup
* removal of unused members of netdev priv struct
* extendable arrangement of h/w reset logic

* Perf cleanup: s/int/unsigned int/ for descriptor ring indexes
  [suggestion by Jeff Garzik].
* Perf cleanup: cache references to ring elements using local pointer

* h/w workaround: several 10's of thousands of 82547 controllers where
  mis-fused during manufacturing, resulting in PHY Tx amplitude to be
  too high and out of spec.  This workaround detects those parts, and
  compensates the Tx amplitude by subtracting ~80mV.

* Cleanup: ethtool diags: only reset if not if_running.

* Bug fix: alloc_etherdev failure didn't cleanup regions in probe.

* Bug fix: missing Tx cleanup opportunities during interrupt handling.

* Bug fix: fix VLAN support on PPC64 [Mark Rakes (mrakes@vivato.net)]

* Bug fix: request_irq() failure resulted in freeing resources twice!
  [Don Fry (brazilnut@us.ibm.com)]

This patch addresses a minor regression reported by Rhine-I users (leading
to occasional Tx timeouts).

I also merged some cosmetic changes.

This patch finally adds policy expiration.

Note that it resends soft policy expire messages every 30 seconds.  This
is needed as when "soft use expire" is used for dead peer detection,
a lost message could lead to a dead peer that isn't discovered until the
SAs expire.

I've only implemented notification for XFRM as I didn't want to just add
another PFKEY extension in case it collides with something else.  Of
course it could be easily done for PFKEY with an extension too.