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

This is the generic part of the start of the compatibility syscall
layer.  I think I have made it generic enough that each architecture can
define what compatibility means.

To use this, an architecture must create asm/compat.h and provide
typedefs for (currently) 'compat_time_t', 'struct compat_timeval' and
'struct compat_timespec'.

This fixes a lock and potential oops when accessing /proc/net/tr_rif is
the token ring interface is under heavy load.

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

SGI Modid: 2.5.x-xfs:slinx:134222a

SGI Modid: 2.5.x-xfs:slinx:134216a

SGI Modid: 2.5.x-xfs:slinx:134187a

SGI Modid: 2.5.x-xfs:slinx:134185a

SGI Modid: 2.5.x-xfs:slinx:134098a

around in active state for as long as possible. This allows us to
coalesce several transactions into one buffer and reduce the disk
traffic.

SGI Modid: 2.5.x-xfs:slinx:134077a

SGI Modid: 2.5.x-xfs:slinx:134179a

SGI Modid: 2.5.x-xfs:slinx:133408a

remove the callback processing from the log write path, we only
do callbacks on I/O completion now.

SGI Modid: 2.5.x-xfs:slinx:133285a

SGI Modid: 2.5.x-xfs:slinx:133254a

SGI Modid: 2.5.x-xfs:slinx:134176a

SGI Modid: 2.5.x-xfs:slinx:132911a

SGI Modid: 2.5.x-xfs:slinx:134172a

SGI Modid: 2.5.x-xfs:slinx:134113a

SGI Modid: 2.5.x-xfs:slinx:134107a

SGI Modid: 2.5.x-xfs:slinx:134068a

data structures (sb,agf,agi,agfl) are now sector size aware.  Cleaned up
the early mount code dealing with log devices and logsectsize.

SGI Modid: 2.5.x-xfs:slinx:134065a

SGI Modid: 2.5.x-xfs:slinx:134064a

SGI Modid: 2.5.x-xfs:slinx:134059a

SGI Modid: 2.5.x-xfs:slinx:134013a

SGI Modid: 2.5.x-xfs:slinx:133971a

Make old 32-bit getdents() look more like the updated
getdents64 for maintainability.

2.5 is 20% slower than 2.4 in an AIM9 test which is just running
readdir across /bin.  A lot of this is due to lots of tiny calls to
copy_to_user() in fs/readdir.c.  The patch speeds up that test by 50%,
so it's comfortably faster than 2.4.

Also, there were lots of unchecked copy_to_user() and put_user() calls
in there.  Fixed all that up as well.

The patch assumes that each arch has a working 64-bit put_user(), which
appears to be the case.

If a page is "freed" while in the deferred-lru-addition queue, the
final reference to it is the deferred lru addition queue.  When that
queue gets spilled onto the LRU, the page is actually freed.

Which is all expected and natural and works fine - it's a weird case.

But one of the AIM9 tests was taking a 20% performance hit (relative to
2.4) because it was going into the page allocator for new pages while
cache-hot pages were languishiung out in the deferred-addition queue.

So the patch changes things so that we spill the CPU's
deferred-lru-addition queue before starting to free pages.  This way,
the recently-used pages actually make it to the hot/cold lists and are
available for new allocations.

It gets back 15 of the lost 20%.  The other 5% is lost to the general
additional complexity of all this stuff.  (But we're 250% faster than
2.4 when running four instances of the test on 4-way).

This patch optimises the truncate of a zero-length file, which is a
sufficiently common case to justify the extra test-n-branch.

It does this by skipping the entire call into the fs if i_size is not
being altered.

The AIM9 `open_clo' test just loops, creating and unlinking a file.
This patch speeds it up 50% for ext2, 600% for reiserfs.

The buffer-stripping code gets upset when it sees a non-uptodate buffer
against an uptodate page.  This happens because the write end_io
handler clears BH_Uptodate.

Add a buffer_req() test to suppress these warnings.

Patch from Hugh Dickins and Robert Love.

Fixes up the PF_MEMDIE handling so that it actually works.

(PF_MEMDIE allows an oom-killed task to use the emergency memory
reserves so that it can actually get out of the page allocator and
die)

2.5's signal delivery is 20% slower than 2.4.  A signal send/handle
cycle is performing a total of 24 copy_*_user() calls, and
copy_*_user() got optimised for large copies.

The patch reduces that to six copy_*_user() calls, and gets us up to
about 5% slower than 2.4.  We'd have to go back to some additional
inlined copy_user() code to get the last 3% back.  And HZ=100 to get
the 2% back.

It is noteworthy that the benchmark is not using float at all during
the body of the test, yet the kernel is still doing all that floating
point stuff.

Patch from Mingming Cao <cmm@us.ibm.com>

- ipc_lock() need a read_barrier_depends() to prevent indexing
  uninitialized new array on the read side.  This is corresponding to
  the write memory barrier added in grow_ary() from Dipankar's patch to
  prevent indexing uninitialized array.

- Replaced "wmb()" in IPC code with "smp_wmb()"."wmb()" produces a
  full write memory barrier in both UP and SMP kernels, while
  "smp_wmb()" provides a full write memory barrier in an SMP kernel,
  but only a compiler directive in a UP kernel.  The same change are
  made for "rmb()".

- Removed rmb() in ipc_get().  We do not need a read memory barrier
  there since ipc_get() is protected by ipc_ids.sem semaphore.

- Added more comments about why write barriers and read barriers are
  needed (or not needed) here or there.

With some workloads a large number of pages coming off the LRU are
pinned blockdev pagecache - things like ext2 group descriptors, pages
which have buffers in the per-cpu buffer LRUs, etc.

They keep churning around the inactive list, reducing the overall page
reclaim effectiveness.

So move these pages onto the active list.

Pages from memory-backed filesystems are supposed to be moved up onto
the active list, but that's not working because fail_writepage() is
called when the page is not on the LRU.

So look for this case in page reclaim and handle it there.

And it's more efficient, the VM knows more about what is going on and
it later leads to the removal of fail_writepage().

The patch addresses some search complexity failures which occur when
there is a large amount of dirty data on the inactive list.

Normally we attempt to write out those pages and then move them to the
head of the inactive list.  But this goes against page aging, and means
that the page has to traverse the entire list again before it can be
reclaimed.

But the VM really wants to reclaim that page - it has reached the tail
of the LRU.

So what we do in this patch is to mark the page as needing reclamation,
and then start I/O.  In the IO completion handler we check to see if
the page is still probably reclaimable and if so, move it to the tail of
the inactive list, where it can be reclaimed immediately.

Under really heavy swap-intensive loads this increases the page reclaim
efficiency (pages reclaimed/pages scanned) from 10% to 25%.  Which is
OK for that sort of load.  Not great, but OK.

This code path takes the LRU lock once per page.  I didn't bother
playing games with batching up the locking work - it's a rare code
path, and the machine has plenty of CPU to spare when this is
happening.

This removes the last remnant of the 2.4 way of throttling page
allocators: the wait_on_page_writeback() against mapped-or-swapcache
pages.

I did this because:

a) It's not used much.
b) It's already causing big latencies
c) With Jens' large-queue stuff, it can cause huuuuuuuuge latencies.
   Like: ninety seconds.

So kill it, and rely on blk_congestion_wait() to slow the allocator
down to match the rate at which the IO system can retire writes.

These are the mount options which turn off and on the Orlov allocator.
ext2 supports them but Ted forgot to wire them up for ext3.