The problem is that when loading next IDE PCI module memory used by
previously registered ide_pci_host_proc_list entry is already unmapped
(because of __{dev}initdata).  This doesn't happen in built-in case
because this memory is freed after all drivers are initialized.
Fix it by removing __{dev}initdata from all ide_pci_host_proc_t.

Thanks to Andre Tomt <andre@tomt.net> for help in debugging this.

The recent change to ide-cd using sector div is passing
the wrong type to do_div().

into sgi.com:/source2/xfs-linux-2.6

SGI Modid: xfs-linux:xfs-kern:164051a6

SGI Modid: xfs-linux:xfs-kern:165155a

SGI Modid: xfs-linux:xfs-kern:166324a

SGI Modid: xfs-linux:xfs-kern:166323a

SGI Modid: xfs-linux:xfs-kern:165783a

SGI Modid: xfs-linux:xfs-kern:166301a

SGI Modid: xfs-linux:xfs-kern:166207a

SGI Modid: xfs-linux:xfs-kern:166109a

SGI Modid: xfs-linux:xfs-kern:166087a

SGI Modid: xfs-linux:xfs-kern:166046a

All metadata buffers are lockable these days.

SGI Modid: xfs-linux:xfs-kern:166042a

SGI Modid: xfs-linux:xfs-kern:166032a

users of this that were very confused indeed.

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

into neo.rr.com:/home/ambx1/linux/bk/linus-2.5

This patch cleans up the kconfig options for the pnp subsystem.  It
updates the comments and makes pnpbios proc support an optional
feature.

This patch adds some aditional information to sysfs for pnp cards.  It
should be useful for userland tools.

This patch updates the matching code to ensure that all requested
devices are present on the card, even if they are in use.  It is
necessary for some ALSA drivers to work properly because early vendors
would have different sets of devices on the same card ids.  It is from
Takashi Iwai <tiwai@suse.de>.

This patch updates file2alias.c to support pnp ids.  It is from
Takashi Iwai <tiwai@suse.de>.

This patch moves the PnP ID declarations to mod_devicetable.h like
most of the other buses.  It is from Takashi Iwai <tiwai@suse.de>.

Recently many PnPBIOS bugs have been triggered by static resource
information requests.  This patch makes an effort to further avoid
making them.

Some PnPBIOSes do not follow the specifications with regard to
disabling devices.  This patch preserves the tag bits, while zeroing
the resource settings.  Previously we would zero the entire buffer.
It has been tested and appears to correct the issue while remaining
compatible with unbroken PnPBIOSes.

This patch reorganizes resource flags to ensure that manual resource
settings are properly recognized.  This fix is necessary for many ALSA
drivers.  It also prevents comparisons between unset resource
structures.  The bug was discovered by Rene Herman
<rene.herman@keyaccess.nl>, who also wrote an initial version of this
patch.  I made further improvements to ensure that the pnp subsystem 
was compatible with this initial change.

The serial driver currently fails to unregister its pnp driver upon
module unload.  This patch corrects the problem by calling
pnp_unregister_driver and implementing a proper remove function.

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

into kernel.bkbits.net:/home/davem/net-2.6

into kernel.bkbits.net:/home/davem/sparc-2.6

The rpciod shutdown code gives ugly sleep_on without BKL warnings in
-mm.  And it looks indeed somewhat racy.

The easy fix would be to simply use a completion as in the patch below,
but that removes all the signal fuzzing semantics the current code has.
I don't really understand why we want to cancel the operation by
signals, but I think it'd be better to leave that to people familar with
the code anyway..