linux-toradex.git/fs/select.c, branch v2.6.18

[PATCH] fs: sys_poll with timeout -1 bug fix

2006-06-25T17:01:22+00:00

If you do a poll() call with timeout -1, the wait will be a big number
(depending on HZ) instead of infinite wait, since -1 is passed to the
msecs_to_jiffies function.

Signed-off-by: Frode Isaksen 
Acked-by: Nishanth Aravamudan 
Cc: David Woodhouse 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] Poll cleanups/microoptimizations

2006-06-23T14:43:03+00:00

The "count" and "pt" variables are declared and modified by do_poll(), as
well as accessed and written indirectly in the do_pollfd() subroutine.

This patch pulls all handling of these variables into the do_poll()
function, thereby eliminating the odd use of indirection in do_pollfd().
This is done by pulling the "struct pollfd" traversal loop from do_pollfd()
into its only caller do_poll().  As an added bonus, the patch saves a few
clock cycles, and also adds comments to make the code easier to follow.

Signed-off-by: Vadim Lobanov 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] select: don't overflow if (SELECT_STACK_ALLOC % sizeof(long) != 0)

2006-04-11T13:18:41+00:00

If SELECT_STACK_ALLOC is not a multiple of sizeof(long) then stack_fds[]
would be shorter than SELECT_STACK_ALLOC bytes and could overflow later in
the function.  Fixed by simply rearranging the test later to work on
sizeof(stack_fds) Currently SELECT_STACK_ALLOC is 256 so this doesn't
happen, but it's nasty to have things like this hidden in the code.  What
if later someone decides to change SELECT_STACK_ALLOC to 300?

Signed-off-by: Mitchell Blank Jr 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] select() warning fixes

2006-04-11T13:18:30+00:00

fs/select.c: In function `core_sys_select':
fs/select.c:339: warning: assignment from incompatible pointer type
fs/select.c:376: warning: comparison of distinct pointer types lacks a cast

By using a void* we can remove lots of casts rather than adding more.

Cc: Jes Sorensen 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] avoid unaligned access when accessing poll stack

2006-03-31T20:30:48+00:00

Commit 70674f95c0a2ea694d5c39f4e514f538a09be36f:

  [PATCH] Optimize select/poll by putting small data sets on the stack

resulted in the poll stack being 4-byte aligned on 64-bit architectures,
causing misaligned accesses to elements in the array.

This patch fixes it by declaring the stack in terms of 'long' instead
of 'char'.

Force alignment of poll and select stacks to long to avoid unaligned
access on 64 bit architectures.

Signed-off-by: Jes Sorensen 
Signed-off-by: Linus Torvalds

[PATCH] mark f_ops const in the inode

2006-03-28T17:16:05+00:00

Mark the f_ops members of inodes as const, as well as fix the
ripple-through this causes by places that copy this f_ops and then "do
stuff" with it.

Signed-off-by: Arjan van de Ven 
Signed-off-by: Alexey Dobriyan 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] use fget_light() in select/poll

2006-03-28T17:16:04+00:00

Cc: Andi Kleen 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] Optimize select/poll by putting small data sets on the stack

2006-03-28T17:16:04+00:00

Optimize select and poll by a using stack space for small fd sets

This brings back an old optimization from Linux 2.0.  Using the stack is
faster than kmalloc.  On a Intel P4 system it speeds up a select of a
single pty fd by about 13% (~4000 cycles -> ~3500)

It also saves memory because a daemon hanging in select or poll will
usually save one or two less pages.  This can add up - e.g.  if you have 10
daemons blocking in poll/select you save 40KB of memory.

I did a patch for this long ago, but it was never applied.  This version is
a reimplementation of the old patch that tries to be less intrusive.  I
only did the minimal changes needed for the stack allocation.

The cut off point before external memory is allocated is currently at
832bytes.  The system calls always allocate this much memory on the stack.

These 832 bytes are divided into 256 bytes frontend data (for the select
bitmaps of the pollfds) and the rest of the space for the wait queues used
by the low level drivers.  There are some extreme cases where this won't
work out for select and it falls back to allocating memory too early -
especially with very sparse large select bitmaps - but the majority of
processes who only have a small number of file descriptors should be ok.
[TBD: 832/256 might not be the best split for select or poll]

I suspect more optimizations might be possible, but they would be more
complicated.  One way would be to cache the select/poll context over
multiple system calls because typically the input values should be similar.
 Problem is when to flush the file descriptors out though.

Signed-off-by: Andi Kleen 
Cc: Eric Dumazet 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] select: time comparison fixes

2006-02-17T21:59:28+00:00

I got all of these backwards.  We want to return

	min(input timeout, new timeout)

to userspace to prevent increasing the time-remaining value.

Thanks to Ernst Herzberg  for reporting and diagnosing.

Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] select: fix returned timeval

2006-02-12T05:41:11+00:00

With David Woodhouse 

select() presently has a habit of increasing the value of the user's
`timeout' argument on return.

We were writing back a timeout larger than the original.  We _deliberately_
round up, since we know we must wait at _least_ as long as the caller asks
us to.

The patch adds a couple of helper functions for magnitude comparison of
timespecs and of timevals, and uses them to prevent the various poll and
select functions from returning a timeout which is larger than the one which
was passed in.

The patch also fixes a bug in compat_sys_pselect7(): it was adding the new
timeout value to the old one and was returning that.  It should just return
the new timeout value.

(We have various handy timespec/timeval-to-from-nsec conversion functions in
time.h.  But this code open-codes it all).

Cc: "David S. Miller" 
Cc: Andi Kleen 
Cc: Ulrich Drepper 
Cc: Thomas Gleixner 
Cc: george anzinger 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds