- irq handlers should be cache-affine to a large degree, without the
  explicit use of /proc/irq/*/smp_affinity.

- idle CPUs should be preferred over busy CPUs when directing IRQs towards
  them.

- the distribution of IRQs should be random, to avoid all IRQs going to
  the same CPU, and to avoid 'heavy' IRQs from loading certain CPUs
  unfairly over CPUs that handle 'light' IRQs. The IRQ system has no
  knowledge about how 'heavy' an IRQ handler is in terms of CPU cycles.

here is the design and implementation:

- we make per-irq decisions about where the IRQ will go to next. Right
  now it's a fastpath and a slowpath, the real stuff happens in the slow
  path. The fastpath is very lightweight.

- [ i decided not to measure IRQ handler overhead via RDTSC - it ends up
    being very messy, and if we want to be 100% fair then we also need to
    measure softirq overhead, and since there is no 1:1 relationship
    between softirq load and hardirq load, it's impossible to do
    correctly. So the IRQ balancer achieves fairness via randomness. ]

- we stay affine in the micro timescale, and we are loading the CPUs
  fairly in the macro timescale. The IO-APIC's lowest priority
  distribution method rotated IRQs between CPUs once per IRQ, which was
  the worst possible solution for good cache-affinity.

- to achieve fairness and to avoid lock-step situations some real
  randomness is needed. The IRQs will wander in the allowed CPU group
  randomly, in a brownean motion fashion. This is what the 'move()'
  function accomplishes. The IRQ moves one step forward or one step
  backwards in the allowed CPU mask. [ Note that this achieves a level of
  NUMA affinity as well, nearby CPUs are more likely to be NUMA-affine. ]

- the irq balancer has some knowledge about 'how idle' a single CPU is.
  The idle task updates the idle_timestamp. Since this update is in the
  idle-to-be codepath, it does not increase the latency of idle-wakeup,
  the overhead should be zero in all cases that matter. The idle-balancing
  happens the following way: when searching for the next target CPU after
  a 'IRQ tick' has expired, we first search 'idle enough' CPUs in the
  allowed set. If this does not succeed then we search all CPUs.

- the patch is fully compatible with the /proc/irq/*/smp_affinity
  interface as well, everything works as expected.

note that the current implementation can be expressed equivalently in
terms of timer-interrupt-driven IRQ redirection. But i wanted to get some
real feedback before removing the possibility to do finer grained
decisions - and the per-IRQ overhead is very small anyway.

This makes the cpqfc driver recognize the HP Tachyon.  I moved the
device list to an __initdata structure so the driver doesn't build it at
runtime and changed it to use the proper PCI_DEVICE_ID_* names. 

With this patch applied, the driver happily detects the disks attached
to my HP Tachyon.

- Kill unneded parameters to ide_cmd_ioctl() and ide_task_ioctl().

- Apply Petr Vendrovecs fix for 32bit ver 16bit transfers.

- Make CD-ROM usable again by guarding the generic routines against request
   field abuse found there. We will try to convert this driver to the just to be
   finished struct ata_request after the generic changes stabilize a bit.
   The strcut ata_taskfile and struct ata_request merge to be more preciese.

- Don't provide symbolic links in /proc/ide - they are redundant data.

- Try to use a more reasonable default capacity value in ata_capacity().

- Fix ata_put() ata_get() usage in ide_check_media_change().

- Small readability fixes to the option parsing code.

- Apply Vojtech Pavliks /proc PIIX output fix.

- Replace all occurrences of ide_wait_taskfile() with ide_raw_taskfile().  One
   duplicated code path fewer.

Included is the latest version of the code. Indenting it properly
actually reduced the size of the files by another 10K or so.

Removes an atavism in declaration of sys_nfsservctl() - sorry, I should've
remove that junk when cond_syscall() thing was done.

Found by Andrew Morton <akpm@zip.com.au>

 This patch fixes small typo in ikernel informational message.

 This patch fixes a problem that was created during inode structure
 cleanup/ private parts separation. This fix was made by Chris Mason.
 This is very critical bugfix. Without it, filesystem corruption
 happens on savelinks processing and possibly in some other cases.

This patch removes one tail_conversion object out of build list,
because it was specified twice. (noticed by Jeff Garzik)

This patch removes confusing warning about journal replay on readonly FS

This patch is to change comment of CONFIG_REISERFS_PROC_INFO config item,
to make it more clear.

This patch renames reiserfs debugging option in config output,
to make its meaning more clear.

This patch is to fix journal replay bug where old code would replay
transactions with mount_id != mount_id recorded in journal header.
Fixed by Chris Mason.

This patch solves a problem where separate journal device was not freed
if journal initialisation failed

This patch is to add forgotten metadata journaling for a case when
we free blocks after tail conversion failures. Found and fixed by Chris Mason

This patch is to fix a case where flag was not set at inode-read time which
prevented 32bit uid/gid to work correctly.

This patch is to convert pap14030 panic into warning. While doing this,
a bug was uncovered, that when get_block() returns a failure, buffer
is still marked as mapped, and on subsequent access to this buffer
get_block() was not called anymore. This is also fixed.

This patch is to fix a lookup problem on bigendian platforms

This patch is to fix a problem when directory's atime was not updated on
readdir(). Patch is written by Chris Mason.

- Integrate the TCQ stuff from Jens Axboe. Deal with the conflicts, apply some
   cosmetic changes. We are still not at a stage where we could immediately
   integrate ata_request and ata_taskfile but we are no longer far away.

- Clean up the data transfer function in ide-disk to use ata_request structures
   directly.

- Kill useless leading version information in ide-disk.c

- Replace the ATA_AR_INIT macro with inline ata_ar_init() function.

- Replace IDE_CLEAR_TAG with ata_clear_tag().

- Replace IDE_SET_TAG with ata_set_tag().

- Kill georgeous ide_dmafunc_verbose().

- Fix typo in ide_enable_queued() (ide-tcq.c!)

Apparently there still problems with a TCQ enabled device and a not enabled
device on the same channel, but let's first synchronize up with Jens.

- Eliminate ide_task_t and rename struct ide_task_s to struct ata_taskfile.
   This should become the entity which is holding all data for a request in the
   future. If this turns out to be the case, we will just rename it to
   ata_request.

- Reduce the number of arguments for the ata_taskfile() function. This helps to
   wipe quite a lot of code out as well.

This stage is not sensitive, so let's make a patch before we start to integrate
the last work of Jens Axboe.

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

and scheduling-in-interrupt problems we had, and also makes it
much faster when handling large numbers (100s or more) of PPP units.

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

and cacheflush.h in a few places where they are needed.

flushing code a little.

Patch to cciss driver in 2.4.8-pre2 to use pdev->irq
and other pci_dev structure elements only after calling
pci_enable_device.

Morten Helgesen <admin@nextframe.net> sent me this.

The latest ACPI merge accidentally clobbered another change in pci-irq.c.
Here's the original patch again (applies fine except for an offset)
Thanks -- Andy

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

More places where we want the size of block device and have relevant
struct block_device * available,

Fixes races in jffs2_get_sb() - current code has a window when two
mounts of the same mtd device can miss each other, resulting in two
active instances of jffs2 fighting over the same device.