]> git.hungrycats.org Git - linux/log
linux
22 years agoLinux v2.5.50 v2.5.50
Linus Torvalds [Wed, 27 Nov 2002 06:33:58 +0000 (22:33 -0800)]
Linux v2.5.50

22 years agoMerge kroah.com:/home/linux/linux/BK/bleeding-2.5
Greg Kroah-Hartman [Wed, 27 Nov 2002 03:24:48 +0000 (19:24 -0800)]
Merge kroah.com:/home/linux/linux/BK/bleeding-2.5
into kroah.com:/home/linux/linux/BK/gregkh-2.5

22 years ago[PATCH] WhiteHEAT update
Stuart MacDonald [Wed, 27 Nov 2002 02:56:01 +0000 (18:56 -0800)]
[PATCH] WhiteHEAT update

1-fix-lowlat:

QA found that running all four ports at 460800 would drop data. I
traced it to data being dropped in the read callback because the flip
buffers were full. Turning on the low latency flag fixed things.

2-fix-taint

A side-effect of turning on low latency; the interrupt context from
the callback is now passed through to the tty layer, passing it on to
calls back into usb-serial.c. Which causes deadlocks when trying to
re-acquire the per-port semaphore. We've already talked about this.
This patch is my work-around for the usb-serial.c brokenness.
Basically, implemement a buffering scheme, and schedule a software
interrupt to handle the data handoff to the tty layer sometime later.
urb_pool_size defaults to 8, but is a module parameter and can be
modified at runtime.
The buffering is needed so that the driver can run while data is
waiting to be processed, but I could have used the tty layer
scheduling instead of doing my own by turning off low latency.
However, I looked at the tty layer and it seems to me that there's
nothing preventing a really fast device from flipping one buffer,
flipping the next, and flipping back to the still full buffer from
before (actually, the flip just gets scheduled for later), so my
driver needs to be able to hold onto buffered data and schedule them
for processing later anyway. So, might as well leave low_latency on.

diff -Naur linux-2.5.49-0-virgin/drivers/usb/serial/whiteheat.c linux-2.5.49-1-fix-
lowlat/drivers/usb/serial/whiteheat.c

22 years ago[PATCH] usb-serial.c disconnect race
Stuart MacDonald [Wed, 27 Nov 2002 02:27:41 +0000 (18:27 -0800)]
[PATCH] usb-serial.c disconnect race

Attached is a patch that changes the 2.5.x disconnect to be similar to
2.4.x disconnect. This doesn't fix the race, but does shrink the
window such that I've never seen it trigger, even under testing
designed to do that.

There doesn't seem to be a good way to fix the race. The fix should be
to have _disconnect force any sleeping semaphore holders to run to
completion between the end of the loop in the patch below and the spot
where the underlying memory is freed, but I don't see a way to do
that.

diff -Naur linux-2.5.49-2-fix-taint/drivers/usb/serial/usb-serial.c linux-2.5.49-3-fix-drvdata/drivers/usb/serial/usb-
serial.c

22 years ago[PATCH] [patch] speed/clean up vicam_decode_color
John Tyner [Wed, 27 Nov 2002 01:54:32 +0000 (17:54 -0800)]
[PATCH] [patch] speed/clean up vicam_decode_color

This patch cleans up the vicam_decode_color function by removing
unused/useless variables and combining the two "x" loops inside the
y loop into one. It also reduces the number of times that the "x"
loop occurs from 512 to 320 which should provide a decent speed
increase. It also fixes a bug in the y loop that wrote beyond its bound.

22 years ago[PATCH] [PATCH] PWC 8.9
Nemosoft Unv. [Wed, 27 Nov 2002 01:46:14 +0000 (17:46 -0800)]
[PATCH] [PATCH] PWC 8.9

After a little absence, here's a patch to bring the Philips Webcam driver up
to version 8.9 (skipping 8.8 which has been available as a download on my
website for a while).

This patch is against 2.5.49, and includes some of the following:
* New USB IDs for Logitech and Visionite webcams.
* Better URB link/unlink sequence when opening/closing device
  and switching resolutions.
* Adding probe for CCD/CMOS sensor type.
* Removed remnants of YUV420 palette stuff.

Also updated the description in 'Kconfig'.

22 years ago[PATCH] USB ov511 driver: Update to version 1.63
Mark W. McClelland [Wed, 27 Nov 2002 01:40:14 +0000 (17:40 -0800)]
[PATCH] USB ov511 driver: Update to version 1.63

22 years ago[PATCH] added support for insmod options to specify vendor/product id. this
Ganesh Varadarajan [Wed, 27 Nov 2002 01:06:12 +0000 (17:06 -0800)]
[PATCH] added support for insmod options to specify vendor/product id. this

22 years ago[PATCH] Add `unlikely' to error-return path in v850 __syscall_return macro
Miles Bader [Tue, 26 Nov 2002 13:21:40 +0000 (05:21 -0800)]
[PATCH] Add `unlikely' to error-return path in v850 __syscall_return macro

Add `unlikely' to error-return path in v850 __syscall_return macro

22 years ago[PATCH] Shrink v850 exception-trap handling code a bit
Miles Bader [Tue, 26 Nov 2002 13:21:35 +0000 (05:21 -0800)]
[PATCH] Shrink v850 exception-trap handling code a bit

Shrink v850 exception-trap handling code a bit

22 years ago[PATCH] Change type of v850 function `gbus_int_disable_irqs' to void
Miles Bader [Tue, 26 Nov 2002 13:21:31 +0000 (05:21 -0800)]
[PATCH] Change type of v850 function `gbus_int_disable_irqs' to void

Change type of v850 function `gbus_int_disable_irqs' to void

22 years ago[PATCH] Make v850 syscall6 macro support both old and new gcc versions
Miles Bader [Tue, 26 Nov 2002 13:21:25 +0000 (05:21 -0800)]
[PATCH] Make v850 syscall6 macro support both old and new gcc versions

Make v850 syscall6 macro support both old and new gcc versions

22 years ago[PATCH] Add id for v850 `nb85e_uart' to serial_core.h
Miles Bader [Tue, 26 Nov 2002 13:21:19 +0000 (05:21 -0800)]
[PATCH] Add id for v850 `nb85e_uart' to serial_core.h

Add id for v850 `nb85e_uart' to serial_core.h

22 years ago[PATCH] Give a compile-time error on the v850 if MAX_ORDER is too large
Miles Bader [Tue, 26 Nov 2002 13:21:13 +0000 (05:21 -0800)]
[PATCH] Give a compile-time error on the v850 if MAX_ORDER is too large

Give a compile-time error on the v850 if MAX_ORDER is too large

22 years ago[PATCH] Make the v850 leds driver's seek routine always return a value
Miles Bader [Tue, 26 Nov 2002 13:21:06 +0000 (05:21 -0800)]
[PATCH] Make the v850 leds driver's seek routine always return a value

Make the v850 leds driver's seek routine always return a value

22 years ago[PATCH] Update make variable used by initramfs `binary blob' creation on v850
Miles Bader [Tue, 26 Nov 2002 13:21:00 +0000 (05:21 -0800)]
[PATCH] Update make variable used by initramfs `binary blob' creation on v850

Update make variable used by initramfs `binary blob' creation on v850

22 years ago[PATCH] Change the default baud rate of the `nb85e_uart' driver to 115200
Miles Bader [Tue, 26 Nov 2002 13:20:55 +0000 (05:20 -0800)]
[PATCH] Change the default baud rate of the `nb85e_uart' driver to 115200

Change the default baud rate of the `nb85e_uart' driver to 115200

22 years ago[PATCH] Tweak some v850 name strings
Miles Bader [Tue, 26 Nov 2002 13:20:51 +0000 (05:20 -0800)]
[PATCH] Tweak some v850 name strings

Tweak some v850 name strings
[basically only affects printed output]

22 years ago[PATCH] Random whitespace tweaks in v850 files
Miles Bader [Tue, 26 Nov 2002 13:20:46 +0000 (05:20 -0800)]
[PATCH] Random whitespace tweaks in v850 files

Random whitespace tweaks in v850 files

22 years ago[PATCH] Add my name to CREDITS file
Miles Bader [Tue, 26 Nov 2002 13:20:40 +0000 (05:20 -0800)]
[PATCH] Add my name to CREDITS file

Add my name to CREDITS file

22 years ago[PATCH] Update includes in v850 files to reflect recent header changes
Miles Bader [Tue, 26 Nov 2002 13:20:35 +0000 (05:20 -0800)]
[PATCH] Update includes in v850 files to reflect recent header changes

Update includes in v850 files to reflect recent header changes

22 years ago[PATCH] Update v850 to use kstat_cpu in irqs.c
Miles Bader [Tue, 26 Nov 2002 13:20:30 +0000 (05:20 -0800)]
[PATCH] Update v850 to use kstat_cpu in irqs.c

Update v850 to use kstat_cpu in irqs.c

22 years ago[PATCH] The v850 doesn't need sys_old_getrlimit
Miles Bader [Tue, 26 Nov 2002 13:20:25 +0000 (05:20 -0800)]
[PATCH] The v850 doesn't need sys_old_getrlimit

The v850 doesn't need sys_old_getrlimit

22 years ago[PATCH] Add TASK_UNMAPPED_BASE for v850
Miles Bader [Tue, 26 Nov 2002 13:20:20 +0000 (05:20 -0800)]
[PATCH] Add TASK_UNMAPPED_BASE for v850

Add TASK_UNMAPPED_BASE for v850

22 years ago[PATCH] Update v850 usage of do_fork to supply new args
Miles Bader [Tue, 26 Nov 2002 13:20:15 +0000 (05:20 -0800)]
[PATCH] Update v850 usage of do_fork to supply new args

Update v850 usage of do_fork to supply new args

22 years ago[PATCH] Add v850 support for initramfs
Miles Bader [Tue, 26 Nov 2002 13:20:10 +0000 (05:20 -0800)]
[PATCH] Add v850 support for initramfs

Add v850 support for initramfs

22 years ago[PATCH] Random minor fixes for v850 `anna' platform
Miles Bader [Tue, 26 Nov 2002 13:20:05 +0000 (05:20 -0800)]
[PATCH] Random minor fixes for v850 `anna' platform

Random minor fixes for v850 `anna' platform

22 years ago[PATCH] Update `nb85e_uart' driver for recent serial framework changes
Miles Bader [Tue, 26 Nov 2002 13:20:00 +0000 (05:20 -0800)]
[PATCH] Update `nb85e_uart' driver for recent serial framework changes

Update `nb85e_uart' driver for recent serial framework changes

22 years ago[PATCH] Make NB85E_UART_CKSR_MAX_FREQ (in nb85e_uart driver) overridable
Miles Bader [Tue, 26 Nov 2002 13:19:56 +0000 (05:19 -0800)]
[PATCH] Make NB85E_UART_CKSR_MAX_FREQ (in nb85e_uart driver) overridable

Make NB85E_UART_CKSR_MAX_FREQ (in nb85e_uart driver) overridable
[The proper value is CPU-dependent]

22 years ago[PATCH] Add v850 nanosecond stat fields
Miles Bader [Tue, 26 Nov 2002 13:19:51 +0000 (05:19 -0800)]
[PATCH] Add v850 nanosecond stat fields

Add v850 nanosecond stat fields

22 years ago[PATCH] Add support for AS85EP1 platform to v850 arch
Miles Bader [Tue, 26 Nov 2002 13:19:46 +0000 (05:19 -0800)]
[PATCH] Add support for AS85EP1 platform to v850 arch

Add support for AS85EP1 platform to v850 arch

22 years ago[PATCH] Add v850 versions of dump_stack and show_stack
Miles Bader [Tue, 26 Nov 2002 13:19:40 +0000 (05:19 -0800)]
[PATCH] Add v850 versions of dump_stack and show_stack

Add v850 versions of dump_stack and show_stack
Also move v850 show_regs to same source file (bug.c)

22 years ago[PATCH] Add v850 support for new module loader
Miles Bader [Tue, 26 Nov 2002 13:19:36 +0000 (05:19 -0800)]
[PATCH] Add v850 support for new module loader

Add v850 support for new module loader

22 years ago[PATCH] Add asm macros on v850 so the size of asm data objects can be recorded
Miles Bader [Tue, 26 Nov 2002 13:19:30 +0000 (05:19 -0800)]
[PATCH] Add asm macros on v850 so the size of asm data objects can be recorded

Add asm macros on v850 so the size of asm data objects can be recorded
[This just makes it easier to see the sizes using readelf or whatever]

22 years ago[PATCH] Fix some minor type problems in v850 code
Miles Bader [Tue, 26 Nov 2002 13:19:25 +0000 (05:19 -0800)]
[PATCH] Fix some minor type problems in v850 code

Fix some minor type problems in v850 code
[Mostly signed/unsigned mismatches]

22 years ago[PATCH] Consolidate various v850 platform banner printks
Miles Bader [Tue, 26 Nov 2002 13:19:19 +0000 (05:19 -0800)]
[PATCH] Consolidate various v850 platform banner printks

Consolidate various v850 platform banner printks

22 years ago[PATCH] oprofile build fix
Andrew Morton [Tue, 26 Nov 2002 11:47:50 +0000 (03:47 -0800)]
[PATCH] oprofile build fix

cpu_buffer.c needs errno.h for ENOMEM.

22 years agoMerge bk://bk.arm.linux.org.uk
Linus Torvalds [Tue, 26 Nov 2002 10:45:05 +0000 (02:45 -0800)]
Merge bk://bk.arm.linux.org.uk
into home.transmeta.com:/home/torvalds/v2.5/linux

22 years agoMerge http://linux-acpi.bkbits.net/linux-acpi
Linus Torvalds [Tue, 26 Nov 2002 10:43:05 +0000 (02:43 -0800)]
Merge http://linux-acpi.bkbits.net/linux-acpi
into home.transmeta.com:/home/torvalds/v2.5/linux

22 years ago[PATCH] USB serial: split the generic functions out into their own file
Greg Kroah-Hartman [Tue, 26 Nov 2002 07:33:15 +0000 (23:33 -0800)]
[PATCH] USB serial: split the generic functions out into their own file

22 years ago[PATCH] tiglusb timeouts
Randy Dunlap [Tue, 26 Nov 2002 06:27:06 +0000 (22:27 -0800)]
[PATCH] tiglusb timeouts

It addresses the timeout parameter in the tiglusb driver.

1.  timeout could be 0, causing a divide-by-zero.
The patch prevents this.

2.  The timeout value to usb_bulk_msg() could be rounded
down to cause a divide-by-zero if timeout was < 10, e.g. 9,
in:
result = usb_bulk_msg (s->dev, pipe, buffer, bytes_to_read,
       &bytes_read, HZ / (timeout / 10));
9 / 10 == 0 => divide-by-zero !!

3.  The timeout value above doesn't do very well on converting
timeout to tenths of seconds.  Even for the default timeout
value of 15 (1.5 seconds), it becomes:
HZ / (15 / 10) == HZ / 1 == HZ, or 1 second.
The patch corrects this formula to use:
(HZ * 10) / timeout

22 years ago[PATCH] USB serial: move the ezusb functions into their own file.
Greg Kroah-Hartman [Tue, 26 Nov 2002 06:16:49 +0000 (22:16 -0800)]
[PATCH] USB serial: move the ezusb functions into their own file.

22 years agoMerge bk://lsm.bkbits.net/linus-2.5
Linus Torvalds [Tue, 26 Nov 2002 06:15:41 +0000 (22:15 -0800)]
Merge bk://lsm.bkbits.net/linus-2.5
into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/lsm-tree

22 years ago[PATCH] Silence debugging message
Andries E. Brouwer [Tue, 26 Nov 2002 05:51:09 +0000 (21:51 -0800)]
[PATCH] Silence debugging message

That annoying ldm message is just debugging output somebody forgot to
remove.

22 years agoMerge kroah.com:/home/greg/linux/BK/bleeding_edge-2.5
Greg Kroah-Hartman [Tue, 26 Nov 2002 03:29:49 +0000 (19:29 -0800)]
Merge kroah.com:/home/greg/linux/BK/bleeding_edge-2.5
into kroah.com:/home/greg/linux/BK/lsm-2.5

22 years agoMerge groveronline.com:/root/bk/linux-2.5
Andy Grover [Tue, 26 Nov 2002 02:41:07 +0000 (18:41 -0800)]
Merge groveronline.com:/root/bk/linux-2.5
into groveronline.com:/root/bk/linux-acpi

22 years agoMerge http://groveronline.com:5555
Andy Grover [Tue, 26 Nov 2002 01:59:06 +0000 (17:59 -0800)]
Merge http://groveronline.com:5555
into groveronline.com:/root/bk/linux-acpi

22 years ago[PATCH] reduced latency in dentry and inode cache shrinking
Andrew Morton [Tue, 26 Nov 2002 01:57:32 +0000 (17:57 -0800)]
[PATCH] reduced latency in dentry and inode cache shrinking

Shrinking a huge number of dentries or inodes can hold dcache_lock or
inode_lock for a long time.  Not only does this hold off preemption -
holding those locks basically shuts down the whole VFS.

A neat fix for all such caches is to chunk the work up at the
shrink_slab() level.

I made the chunksize pretty small, for scalability reasons - avoid
holding the lock for too long so another CPU can come in, acquire it
and go off to do some work.

22 years ago[PATCH] Don't hold BKL across sync_blockdev() in blkdev_put()
Andrew Morton [Tue, 26 Nov 2002 01:57:27 +0000 (17:57 -0800)]
[PATCH] Don't hold BKL across sync_blockdev() in blkdev_put()

Running sync_blockdev() inside lock_kernel causes long scheduling
stalls under some conditions because the lock_kernel() disables
preemption.

We don't need the BKL across sync_blockdev(), so move it down a bit.

This is a minimal fix - we probably don't need BKL for much at all in
there.

22 years ago[PATCH] Add a scheduling point to page reclaim
Andrew Morton [Tue, 26 Nov 2002 01:57:22 +0000 (17:57 -0800)]
[PATCH] Add a scheduling point to page reclaim

When the VM is under heavy load and a page allocator caller is holding
the BKL there can be several-millisecond scheduling stalls.

So add a cond_resched() at every 32nd page (approx).

22 years ago[PATCH] swapoff accounting cleanup
Andrew Morton [Tue, 26 Nov 2002 01:57:17 +0000 (17:57 -0800)]
[PATCH] swapoff accounting cleanup

From Hugh.  Remove some strangeness in the swapoff path.

"it dates from the days when that block set *swap_map to 0, but in
 2.4.13 I changed it to set *swap_map to 1 and delete_from_swap_cache
 afterwards: it's been wrong ever since, hasn't it?  swap_list_locking
 can go, it was there to guard nr_swap_pages for si_swapinfo; the
 swap_device_locking has to stay because swap_map is an array of
 unsigned _shorts_."

22 years ago[PATCH] realtime swapspace accounting
Andrew Morton [Tue, 26 Nov 2002 01:57:13 +0000 (17:57 -0800)]
[PATCH] realtime swapspace accounting

There are a couple of statistical functions which scan the entire swap
map counting things up, to display in /proc.

On my machine, these hold spinlocks for 19 milliseconds which is
unacceptable from a scheduling latency point of view.

And an application which sits in a loop reading /proc/swaps on a large
machine is probably a decent denial-of-service attack - it will limit
swap allocations to tens of pages per second.

So add a counter to swap_info_struct and use it to track how many pages
are currently in use, so those reporting functions don't need to add
them all up.

22 years ago[PATCH] Add some low-latency scheduling points
Andrew Morton [Tue, 26 Nov 2002 01:57:08 +0000 (17:57 -0800)]
[PATCH] Add some low-latency scheduling points

This is the first in a little batch of patches which address long-held
locks in the VFS/MM layer which are affecting our worst-case scheduling
latency, and are making CONFIG_PREEMPT not very useful.

We end up with a worst-case of 500 microseconds at 500MHz, which is
very good.  Unless you do an exit with lots of mmapped memory.
unmap_page_range() needs work.

Some of these patches also add rescheduling points for non-preemptible
kernels - where I felt that the code path could be long enough to be
perceptible.

Three places in the generic pagecache functions need manual
rescheduling points even for non-preemptible kernels:

- generic_file_read()  (Can hold the CPU for seconds)

- generic_file_write() (ditto)

- filemap_fdatawait().  This won't hold the CPU for so long, but it
  can walk many thousands of pages under the lock.  It needs a lock
  drop and scheduling point for both preemptible and non-preemptible
  kernels.  (This one's a bit ugly...)

22 years ago[PATCH] reduced context switch rate in writeback
Andrew Morton [Tue, 26 Nov 2002 01:57:03 +0000 (17:57 -0800)]
[PATCH] reduced context switch rate in writeback

pdflush writes back chunks of ~1000 pages.  It currently takes a short
nap if it writes back no pages at all.  That could cause it to write
back lots of small batches of pages, as it bounces against a congested
queue.

Change it to sleep if it failed to write back the entire batch against
a congested queue.  Ths reduces the context switch rate a little.

The context switch rate is still fairly high (150/sec) - this appears
to be due to add_disk_randomness() scheduling a work function.

22 years ago[PATCH] shrink task_struct by removing per_cpu utime and stime
Andrew Morton [Tue, 26 Nov 2002 01:56:58 +0000 (17:56 -0800)]
[PATCH] shrink task_struct by removing per_cpu utime and stime

Patch from Bill Irwin.  It has the potential to break userspace
monitoring tools a little bit, and I'm a rater uncertain about
how useful the per-process per-cpu accounting is.

Bill sent this out as an RFC on July 29:

"These statistics severely bloat the task_struct and nothing in
 userspace can rely on them as they're conditional on CONFIG_SMP.  If
 anyone is using them (or just wants them around), please speak up."

And nobody spoke up.

If we apply this, the contents of /proc/783/cpu will go from

cpu  1 1
cpu0 0 0
cpu1 0 0
cpu2 1 1
cpu3 0 0

to

cpu  1 1

And we shall save 256 bytes from the ia32 task_struct.

On my SMP build with NR_CPUS=32:

Without this patch, sizeof(task_struct) is 1824, slab uses a 1-order
allocation and we are getting 2 task_structs per page.

With this patch, sizeof(task_struct) is 1568, slab uses a 2-order
allocation and we are getting 2.5 task_structs per page.

So it seems worthwhile.

(Maybe this highlights a shortcoming in slab.  For the 1824-byte case
it could have used a 0-order allocation)

22 years ago[PATCH] Updated Documentation/kernel-parameters.txt
Andrew Morton [Tue, 26 Nov 2002 01:56:52 +0000 (17:56 -0800)]
[PATCH] Updated Documentation/kernel-parameters.txt

Patch from Petr Baudis <pasky@ucw.cz>

This patch (against 2.5.49) updates Documentation/kernel-parameters.txt
to the current state of kernel.  It was somehow abadonded lately, so I
did my best, but it's possible that I still missed some of the options
- thus, if you will notice your favourite boot option missing there,
please speak up.  Note also that I will probably send up another update
after few further kernel releases..

Also, I attempted to introduce some uniform format to the entries, I
added the format description where I was able to find it out and
decypher it, and I also added gross amount of external links to the
headers of the source files or to the README-like files, where the
options are described into more degree.  This way, hopefully this file
has a chance to be actually usable for the users ;-).

There are almost certainly some entries which I missed, it was really a
huge number and the main reason is that some of the boot options don't
use the __setup macro, which I grep'd for.

I hope the patch is ok, there should be no problems with it. Please apply.

Note that this is the fourth submission of the patch - I took the
opportunity and updated the patch from 2.5.48 to 2.5.49.  AFAIK mutt
shouldn't mangle the patch in any way, so it should apply cleanly to
your tree, Linus.

22 years ago[PATCH] kernel_stat cleanup
Andrew Morton [Tue, 26 Nov 2002 01:56:47 +0000 (17:56 -0800)]
[PATCH] kernel_stat cleanup

Patch from Dipankar Sarma <dipankar@gamebox.net>

This is a trivial cleanup removing two old unused macros from
kernel_stat.h that made no sense with the new per-CPU kstat.
Also included a few finicky coding style changes. Please apply.

22 years ago[PATCH] Allow for profile_buf size = kernel text size
Andrew Morton [Tue, 26 Nov 2002 01:56:43 +0000 (17:56 -0800)]
[PATCH] Allow for profile_buf size = kernel text size

Patch from Ravikiran G Thirumalai <kiran@in.ibm.com>

Allows the kernel profiler to use a shift value of zero for
high-resolution profiling.

So instead of overloading a shift of zero to mean "not profiling", use
a separate flag for that.

This means that anyone who is currently using "profile=0" to disable
profiling has accidentally turned profiling on, so I added a printk.

22 years ago[PATCH] misc fixes
Andrew Morton [Tue, 26 Nov 2002 01:56:37 +0000 (17:56 -0800)]
[PATCH] misc fixes

- Missing parentheses around macro arg in __pte_offest()

- struct page_state needs to be cacheline aligned (the first seven
  fields should be in the same cacheline)

22 years ago[PATCH] blk_run_queues() locking fix
Andrew Morton [Tue, 26 Nov 2002 01:56:31 +0000 (17:56 -0800)]
[PATCH] blk_run_queues() locking fix

blk_run_queues() places queues onto a local list without locking.  But
interrupt-time activity in scsi_lib.c will replug these queues, which
involves accessing that list which blk_run_queues() is walking.

Net effect: list corruption and an oops.  Very hard to reproduce...

So hold the lock while blk_run_queues() is walking the local list.

The patch was acked by Jens.  It also uninlines a few largeish
functions and adds a couple of WARN_ON()s in code which worried me.

22 years agoRevert bad PCI bridge resource handling from -dj tree
Linus Torvalds [Tue, 26 Nov 2002 01:45:14 +0000 (17:45 -0800)]
Revert bad PCI bridge resource handling from -dj tree
Cset exclude: davej@codemonkey.org.uk|ChangeSet|20021126023731|33210

22 years agoRevert over-eager memory leak "fix" from the -dj merge
Linus Torvalds [Tue, 26 Nov 2002 01:41:45 +0000 (17:41 -0800)]
Revert over-eager memory leak "fix" from the -dj merge

22 years agoMerge master.kernel.org:/home/davem/BK/sparc-2.5
Linus Torvalds [Tue, 26 Nov 2002 00:55:56 +0000 (16:55 -0800)]
Merge master.kernel.org:/home/davem/BK/sparc-2.5
into home.transmeta.com:/home/torvalds/v2.5/linux

22 years agoRevert duplicate initialization from -ac merge
Linus Torvalds [Tue, 26 Nov 2002 00:53:27 +0000 (16:53 -0800)]
Revert duplicate initialization from -ac merge
Cset exclude: alan@lxorguk.ukuu.org.uk|ChangeSet|20021126021252|43411

22 years agoMerge bk://ppc.bkbits.net/for-linus-ppc
Linus Torvalds [Tue, 26 Nov 2002 00:50:30 +0000 (16:50 -0800)]
Merge bk://ppc.bkbits.net/for-linus-ppc
into home.transmeta.com:/home/torvalds/v2.5/linux

22 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Wed, 27 Nov 2002 08:09:11 +0000 (19:09 +1100)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

22 years agoUndo bogon from the -dj merge.
Linus Torvalds [Tue, 26 Nov 2002 00:48:51 +0000 (16:48 -0800)]
Undo bogon from the -dj merge.

22 years ago[PATCH] PATCH: fix wrong check in BUG() test for UDMA on serverworks
Alan Cox [Tue, 26 Nov 2002 00:44:28 +0000 (16:44 -0800)]
[PATCH] PATCH: fix wrong check in BUG() test for UDMA on serverworks

Otherwise a UDMA CD-ROM is correctly set up and then panics the kernel 8(
The UDMA rule is on disks not on CD

22 years agoMerge tetrachloride.(none):/home/davej/src/bk-linus
Dave Jones [Tue, 26 Nov 2002 10:26:47 +0000 (09:26 -0100)]
Merge tetrachloride.(none):/home/davej/src/bk-linus
into tetrachloride.(none):/home/davej/src/watchdog

22 years agowatchdog compile fixes
Dave Jones [Mon, 25 Nov 2002 16:03:15 +0000 (15:03 -0100)]
watchdog compile fixes

22 years ago[SPARC ESP]: Convert to slave_{attach,destroy}
David S. Miller [Mon, 25 Nov 2002 15:13:09 +0000 (07:13 -0800)]
[SPARC ESP]: Convert to slave_{attach,destroy}

22 years ago[AIC7XXX]: aic7xxx_osm.h needs asm/io.h
David S. Miller [Mon, 25 Nov 2002 15:12:06 +0000 (07:12 -0800)]
[AIC7XXX]: aic7xxx_osm.h needs asm/io.h

22 years ago[PATCH] fix build with IDE disabled
Anton Blanchard [Mon, 25 Nov 2002 13:47:00 +0000 (05:47 -0800)]
[PATCH] fix build with IDE disabled

More fall-out from the -ac merge.

22 years ago[PATCH] IDE fix for -ac merge
Andrew Morton [Mon, 25 Nov 2002 13:35:18 +0000 (05:35 -0800)]
[PATCH] IDE fix for -ac merge

do_ide_setup_pci_device() is returning an uninitialised
ata_index_t causing an oops at bootup.

22 years ago[SPARC]: Add DEBUG_SPINLOCK_SLEEP config option.
David S. Miller [Mon, 25 Nov 2002 13:06:37 +0000 (05:06 -0800)]
[SPARC]: Add DEBUG_SPINLOCK_SLEEP config option.

22 years agoMerge nuts.ninka.net:/home/davem/src/BK/sparcwork-2.5
David S. Miller [Mon, 25 Nov 2002 12:54:27 +0000 (04:54 -0800)]
Merge nuts.ninka.net:/home/davem/src/BK/sparcwork-2.5
into nuts.ninka.net:/home/davem/src/BK/sparc-2.5

22 years ago[SPARC]: Ignore SIGURG if not caught.
David S. Miller [Mon, 25 Nov 2002 12:33:04 +0000 (04:33 -0800)]
[SPARC]: Ignore SIGURG if not caught.

22 years agoLSM: put CONFIG_SECURITY back into the Kconfig file (was lost in the merge)
Greg Kroah-Hartman [Mon, 25 Nov 2002 12:10:47 +0000 (04:10 -0800)]
LSM: put CONFIG_SECURITY back into the Kconfig file (was lost in the merge)

22 years agoLSM: Move the definition of capable() into sched.h if CONFIG_SECURITY is set to help
Greg Kroah-Hartman [Mon, 25 Nov 2002 12:09:46 +0000 (04:09 -0800)]
LSM: Move the definition of capable() into sched.h if CONFIG_SECURITY is set to help
make the #include nightmare more managable.

22 years agoLSM: fix up some needed header file #includes
Greg Kroah-Hartman [Mon, 25 Nov 2002 11:57:47 +0000 (03:57 -0800)]
LSM: fix up some needed header file #includes

22 years agoMerge
Linus Torvalds [Mon, 25 Nov 2002 11:18:45 +0000 (03:18 -0800)]
Merge

22 years agoMerge
Linus Torvalds [Mon, 25 Nov 2002 11:16:12 +0000 (03:16 -0800)]
Merge

22 years agoFix up Alan's huge patch set. It couldn't have compiled for
Linus Torvalds [Mon, 25 Nov 2002 11:10:06 +0000 (03:10 -0800)]
Fix up Alan's huge patch set. It couldn't have compiled for
him either. Tssk, tssk.

22 years ago[PATCH] region handling cleanup for de600
Dave Jones [Mon, 25 Nov 2002 10:42:01 +0000 (02:42 -0800)]
[PATCH] region handling cleanup for de600

22 years ago[PATCH] memleak in macsonic
Dave Jones [Mon, 25 Nov 2002 10:41:51 +0000 (02:41 -0800)]
[PATCH] memleak in macsonic

22 years ago[PATCH] si1_isa board addition to sx driver
Dave Jones [Mon, 25 Nov 2002 10:41:41 +0000 (02:41 -0800)]
[PATCH] si1_isa board addition to sx driver

From 2.4

22 years ago[PATCH] extra parport IDs
Dave Jones [Mon, 25 Nov 2002 10:41:31 +0000 (02:41 -0800)]
[PATCH] extra parport IDs

22 years ago[PATCH] region handling cleanups for appletalk ltpc
Dave Jones [Mon, 25 Nov 2002 10:41:21 +0000 (02:41 -0800)]
[PATCH] region handling cleanups for appletalk ltpc

22 years ago[PATCH] Remove unneeded verify_area from sg.c
Dave Jones [Mon, 25 Nov 2002 10:41:12 +0000 (02:41 -0800)]
[PATCH] Remove unneeded verify_area from sg.c

Makes the code simpler..

22 years ago[PATCH] remove unused EISA_bus__is_a_macro macro
Dave Jones [Mon, 25 Nov 2002 10:41:02 +0000 (02:41 -0800)]
[PATCH] remove unused EISA_bus__is_a_macro macro

22 years ago[PATCH] region cleanups etc in ips driver
Dave Jones [Mon, 25 Nov 2002 10:40:51 +0000 (02:40 -0800)]
[PATCH] region cleanups etc in ips driver

22 years ago[PATCH] Check request_region() in inia100
Dave Jones [Mon, 25 Nov 2002 10:40:42 +0000 (02:40 -0800)]
[PATCH] Check request_region() in inia100

22 years ago[PATCH] add missing clipping for zr36067
Dave Jones [Mon, 25 Nov 2002 10:40:31 +0000 (02:40 -0800)]
[PATCH] add missing clipping for zr36067

from 2.4

22 years ago[PATCH] named struct initialiser updates
Dave Jones [Mon, 25 Nov 2002 10:40:22 +0000 (02:40 -0800)]
[PATCH] named struct initialiser updates

22 years ago[PATCH] Make various bits of synclink static
Dave Jones [Mon, 25 Nov 2002 10:40:12 +0000 (02:40 -0800)]
[PATCH] Make various bits of synclink static

22 years ago[PATCH] unwrapped var usage in unbz64wrap
Dave Jones [Mon, 25 Nov 2002 10:40:01 +0000 (02:40 -0800)]
[PATCH] unwrapped var usage in unbz64wrap

22 years ago[PATCH] extern inline -> static inline
Dave Jones [Mon, 25 Nov 2002 10:39:51 +0000 (02:39 -0800)]
[PATCH] extern inline -> static inline

As per CodingStyle

22 years ago[PATCH] scsi scan blacklist update
Dave Jones [Mon, 25 Nov 2002 10:39:41 +0000 (02:39 -0800)]
[PATCH] scsi scan blacklist update

22 years ago[PATCH] request_region fix in madgemc.c
Dave Jones [Mon, 25 Nov 2002 10:39:31 +0000 (02:39 -0800)]
[PATCH] request_region fix in madgemc.c

22 years ago[PATCH] pcmcia tcic region cleanups
Dave Jones [Mon, 25 Nov 2002 10:39:21 +0000 (02:39 -0800)]
[PATCH] pcmcia tcic region cleanups