If further finer grained locking is needed, we can add a lock to the sysdev_class to
lock the class drivers list.  But if you do that, remember the global list also is still
there and needs to be protected.  That's why I went with a simple lock for everything.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Heh, "global_drivers" as a static...
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This forces the caller to provide the lock, but as they all already had one, it's not a big change.
It also removes the now-unneeded cdev_subsys.  Thanks to Jon Corbet for reminding me about that.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This is needed if the class code is going to be made easier to use, and it makes the code
smaller and easier to understand.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

This is needed for the upcoming klist code from Pat.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Do this instead of using the rwsem of a subsys.
Smaller, faster, and I'm trying to get rid of the rwsem in the subsys.
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

With this patch the floppy driver creates the usual symlink in sysfs to
the physical device backing the block device:

  $tree /sys/block/
  /sys/block/
  |-- fd0
  |   |-- dev
  |   |-- device -> ../../devices/platform/floppy0
  ...
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

On Tue, Feb 15, 2005 at 09:53:44PM +0100, Kay Sievers wrote:
> Add a "bus" symlink to the class and block devices, just like the "driver"
> and "device" links. This may be a huge speed gain for e.g. udev to determine
> the bus value of a device, as we currently need to do a brute-force scan in
> /sys/bus/* to find this value.

Hmm, while playing around with it, I think we should create the "bus"
link on the physical device on not on the class device.

Also the current "driver" link at the class device should be removed,
cause class devices don't have a driver. Block devices never had this
misleading symlink.

 From the class device we point with the "device" link to the physical
 device, and only the physical device should have the "driver" and the
 "bus" link, as it represents the real relationship.
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Get rid of semaphore abuse by converting device_driver->unload_sem
semaphore to device_driver->unloaded completion.

This should get rid of any confusion as well as save a few bytes in the
process.
Signed-off-by: Mike Waychison <michael.waychison@sun.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Move the creation of the sysfs "dev" file of a class device into the
driver core. The struct class_device contains a dev_t value now.  If set,
the driver core will create the "dev" file containing the major/minor
numbers automatically.
Signed-off-by: Kay Sievers <kay.sievers@vrfy.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Separate platform device name from platform device number such that
names ending with numbers aren't confusing.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Even though these 2.4. interfaces are already gone in Dave Jones' cpufreq
bitkeeper tree, here's a patch which properly announces it in
Documentation/feature-removal-schedule.txt:


Add meaningful content concerning the removal of deprecated interfaces to
the cpufreq core.
Signed-off-by: Dominik Brodowski <linux@brodo.de>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

Add 2.4.x cpufreq /proc and sysctl interface removal
to the feature-removal-schedule.
Signed-off-by: Randy Dunlap <rddunlap@osdl.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

kobject_get_path and kobject_rename are only used by the sysfs core code
and not aren't really driver-ish code. Remove the unused exports
Signed-off-by: Arjan van de Ven <arjan@infradead.org>
Signed-off-by: Greg Kroah-Hartman <greg@kroah.com>

 - Move the bits_wide parameter from txn_alloc_data to txn_alloc_irq.
   It's too late by the time it hits txn_alloc_data(), we can only panic.
   In txn_alloc_irq, we can fail the allocation and continue.
 - Also fix a bug where we'd only allow up to half the interrupts to be
   allocated.  Not a problem for iosapic machines, but might have sucked
   on a really big GSC-based machine.
 - Add missing irq_enter() / irq_exit()
 - Allow interrupt processing to be interrupted by the timer tick so we
   actually account hard interrupts.  Also speed up the handling of CPU
   interrupts by not masking with cpu_eiem again.
 - Remove sufficiently obsolete DEBUG_IRQ code
 - Remove limit on times around the loop.  If we exit the loop while
   interrupts are still pending, we'll only be re-interrupted as soon
   as we exit the function.
 - Remove unnecessary includes
Signed-off-by: Matthew Wilcox <willy@parisc-linux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

Fix "make configcheck | fgrep 'not needed' | fgrep parisc"
Signed-off-by: Grant Grundler <grundler@parisc-linux.org>
Signed-off-by: Matthew Wilcox <willy@parisc-linux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

 - The entire silence buffer is now being filled, instead of just 1/8th of it.
 - The silence buffer is now prepared during initialisation of the driver.
 - snd_pcm_format_size() is missing from the ALSA headers:
   using snd_pcm_format_physical_width() instead.
Signed-off-by: Stuart Brady <sdbrady@ntlworld.com>
Signed-off-by: Matthew Wilcox <willy@parisc-linux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

 - Missing UL on large integers
 - 0/NULL confusion
 - __user annotations
 - C99-style array elements
Signed-off-by: Randolph Chung <tausq@parisc-linux.org>
Signed-off-by: Matthew Wilcox <willy@parisc-linux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

 - Convert ldil/ldo to using the load32 macro.
Signed-Off-By: Kyle McMartin <kyle@parisc-linux.org>

 - loop on signal delivery like other archs
 - unify VM initialization code for 32/64-bit
Signed-off-by: Randolph Chung <tausq@parisc-linux.org>
Signed-off-by: Matthew Wilcox <willy@parisc-linux.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Resolve a problem where a Sony Ericsson GC79 Cardbus device was not
being correctly resumed across a S3 suspend, as reported by Hendrik
Hoeth.
Signed-off-by: Russell King <rmk@arm.linux.org.uk>

Patch from Petr Vandrovec

XScale detection needs access to Interrupt Enable Register on UART.
But this register shares position with high byte clock divisor, and
previous detection steps were leaving clock divisor and not IER
selected, causing misdetection of all 16550A chips as XScale.
 
Fix this by disabling access to clock divisor at the end of previous
detection step, so chip is in same mode after each detection step.
 
Signed-off-by: Petr Vandrovec

into ppc970.osdl.org:/home/torvalds/v2.6/linux

Patch from Tony Lindgren

This patch adds support for 32k timer on OMAP 16xx, and 64-bit
sched clock to the MPU timer.
This is an update version of ARM patch 2337/1. The 32k timer
modulo code has been left out, and the dynamic tick (VST) timer
will be submitted in a separate patch.

Signed-off-by: Tony Lindgren
Signed-off-by: Russell King

Oops, it broke.  Glue the bits back together, replacing yrs with
tm->tm_year + 1900.

I will not merge untested changes into Linus' tree.
I will not merge untested changes into Linus' tree.
I will not merge untested changes into Linus' tree.
I will not ...
Signed-off-by: Russell King <rmk@arm.linux.org.uk>

Patch from Richard Purdie

Sharp SCOOP: Devices with multiple scoop interfaces are now
available so:
* add support for mutliple device support to the driver
* Update corgi, collie and poodle to share the scoop
  device structure so a device can be selected in drivers
* Update drivers to use the device structures

Signed-off-by: Richard Purdie
Signed-off-by: Russell King

Patch from Lennert Buytenhek

Another round of ixp2000 typo fixes.

Signed-off-by: Lennert Buytenhek
Signed-off-by: Russell King

Patch from Ben Dooks

Add the Acer N30 machine, from Christer Weinigel

Signed-off-by: Christer Weinigel

Signed-off-by: Ben Dooks
Signed-off-by: Russell King

into suse.de:/home/greg/linux/BK/usb-2.6

into ppc970.osdl.org:/home/torvalds/v2.6/linux

into pobox.com:/garz/repo/net-drivers-2.6

This patch is from Nathan Lynch <ntl@pobox.com>.

Use static inlines instead of #defines for stub functions when
CONFIG_EEH=n, to eliminate "statement with no effect" warnings with
some toolchains.
Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

This patch is from Nathan Lynch <ntl@pobox.com>.

When offlining a cpu, any device interrupts which are bound to the cpu
have their affinity forcibly reset to all cpus (the default).
However, the value in /proc/irq/XXX/smp_affinity remains unchanged.
Since we're doing this while all the other cpus are stopped, it should
be safe to just call desc->handler->set_affinity and manually update
the irq_affinity array.
Signed-off-by: Nathan Lynch <ntl@pobox.com>
Signed-off-by: Paul Mackerras <paulus@samba.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>