]> git.neil.brown.name Git - history.git/log
history.git
23 years agoMerge bk://68.0.152.218/linux-2.5-misc
Paul Mackerras [Thu, 9 Jan 2003 20:20:34 +0000 (07:20 +1100)]
Merge bk://68.0.152.218/linux-2.5-misc
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years agoPPC32: Default to 768MB of lowmem.
Tom Rini [Wed, 8 Jan 2003 07:24:41 +0000 (00:24 -0700)]
PPC32: Default to 768MB of lowmem.

23 years ago[PATCH] saa7134 driver update
Gerd Knorr [Wed, 8 Jan 2003 02:11:58 +0000 (18:11 -0800)]
[PATCH] saa7134 driver update

This updates the saa7134 driver.  It adds support for a few more cards
and includes adaptions to the video-buf.c changes sent earlier.

23 years ago[PATCH] i2c update for tuner.c
Gerd Knorr [Wed, 8 Jan 2003 02:11:50 +0000 (18:11 -0800)]
[PATCH] i2c update for tuner.c

This has some i2c adaptions and cleanups for the tv card tuner module.

23 years ago[PATCH] media/video i2c updates
Gerd Knorr [Wed, 8 Jan 2003 02:11:43 +0000 (18:11 -0800)]
[PATCH] media/video i2c updates

This updates a bunch of i2c modules in drivers/media/video.  Most of it
are adaptions to the recent i2c changes in the kernel.  While being at
it I also did some other cleanups like deleting unused+dead code, using
name-based initialization for some not-yet converted structs, ...

The also has a few small fixes here and there, but no major functional
changes.

23 years ago[PATCH] add bt832 module
Gerd Knorr [Wed, 8 Jan 2003 02:11:35 +0000 (18:11 -0800)]
[PATCH] add bt832 module

This adds a driver module for the bt832 chip.  It is needed by the bttv
driver to support the Pixelview Digital Camera.  The bt832 is connected
using the GPIO pins of the bt878 chip.

23 years ago[PATCH] update bttv documentation
Gerd Knorr [Wed, 8 Jan 2003 02:11:28 +0000 (18:11 -0800)]
[PATCH] update bttv documentation

23 years ago[PATCH] bttv driver update.
Gerd Knorr [Wed, 8 Jan 2003 02:11:20 +0000 (18:11 -0800)]
[PATCH] bttv driver update.

This updates the bttv driver.  Changes:

 * adaptions to the video-buf changes send in a previous mail.
 * adaptions to recent i2c changes in the kernel.
 * first code bits the pixelview digital camera support (not
   working yet).
 * lots of small fixes/changes for specific TV cards.

23 years ago[PATCH] add v4l1-compat module.
Gerd Knorr [Wed, 8 Jan 2003 02:11:13 +0000 (18:11 -0800)]
[PATCH] add v4l1-compat module.

This adds the v4l1-compat module.  This is a module which can translate
most (old) v4l1 ioctls into the new v4l2 API.  This makes it easier for
v4l2 drivers to present both old v4l and new v4l2 APIs to video4linux
applications.  The saa7134 driver uses this for example.

23 years ago[PATCH] video-buf.c update
Gerd Knorr [Wed, 8 Jan 2003 02:11:06 +0000 (18:11 -0800)]
[PATCH] video-buf.c update

This updates the video-buf helper module.  It changes the field handling
a bit and adds code do deal better with alternating field capture (=
capture even and odd fields to separate video buffers).

23 years ago[PATCH] add tda9887 module
Gerd Knorr [Wed, 8 Jan 2003 02:11:00 +0000 (18:11 -0800)]
[PATCH] add tda9887 module

This adds the tda9887 i2c module to the linux kernel.  This one is
needed for some TV cards.  It also adds the missing #define to
audiochip.h which currently breaks the bttv build.

23 years agoMerge bk://ppc.bkbits.net/for-linus-ppc
Linus Torvalds [Wed, 8 Jan 2003 02:03:40 +0000 (18:03 -0800)]
Merge bk://ppc.bkbits.net/for-linus-ppc
into home.transmeta.com:/home/torvalds/v2.5/linux

23 years agoPPC32: Add support for new IBM embedded PPC cpus.
Paul Mackerras [Thu, 9 Jan 2003 08:56:55 +0000 (19:56 +1100)]
PPC32: Add support for new IBM embedded PPC cpus.

This adds support for 405GPR, 405LP, STBx25, NP4GS3.

23 years agoPPC32: Add support for the "Redwood-6" STB03xxx-based eval board
Paul Mackerras [Thu, 9 Jan 2003 08:50:15 +0000 (19:50 +1100)]
PPC32: Add support for the "Redwood-6" STB03xxx-based eval board

23 years agoPPC32: Add support for the IBM405LP-based "Beech" board
Paul Mackerras [Thu, 9 Jan 2003 08:44:42 +0000 (19:44 +1100)]
PPC32: Add support for the IBM405LP-based "Beech" board

23 years agoPPC32: remove execute permission from some ppc source files.
Paul Mackerras [Thu, 9 Jan 2003 08:31:21 +0000 (19:31 +1100)]
PPC32: remove execute permission from some ppc source files.

23 years agoPPC32: Update the support for the IBM 40x embedded PowerPC chips and boards.
Paul Mackerras [Thu, 9 Jan 2003 04:11:56 +0000 (15:11 +1100)]
PPC32: Update the support for the IBM 40x embedded PowerPC chips and boards.

Some of this comes from recent work in the 2_4_devel tree, including the
conversion to using an array (core_ocp[]) to describe on-chip peripherals
in a standard format.  The other major change is to give the board file
control early on in the boot and let it call back to generic 4xx routines
rather than vice-versa.

23 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Wed, 8 Jan 2003 23:09:36 +0000 (10:09 +1100)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years agoMerge bk://68.0.152.218/linux-2.5-misc
Paul Mackerras [Wed, 8 Jan 2003 21:30:37 +0000 (08:30 +1100)]
Merge bk://68.0.152.218/linux-2.5-misc
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Wed, 8 Jan 2003 21:07:55 +0000 (08:07 +1100)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Wed, 8 Jan 2003 03:42:53 +0000 (14:42 +1100)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years ago[PATCH] {get,put}_compat_timspec: s390x
Stephen Rothwell [Wed, 8 Jan 2003 02:02:55 +0000 (18:02 -0800)]
[PATCH] {get,put}_compat_timspec: s390x

Here is the s390x part of the compat_timespec changes.

23 years ago[PATCH] {get,put}_compat_timspec: generic
Stephen Rothwell [Wed, 8 Jan 2003 02:02:49 +0000 (18:02 -0800)]
[PATCH] {get,put}_compat_timspec: generic

Andi Kleen asked for these two, so here they are.  This is the generic
part of creating compat functions for timespec get/put.

23 years ago[PATCH] Cleanup for SunRPC auth code
Trond Myklebust [Wed, 8 Jan 2003 01:59:44 +0000 (17:59 -0800)]
[PATCH] Cleanup for SunRPC auth code

Converts the RPC client auth code to use 'list_head' rather than a
custom pointer scheme.

Fixes a (relatively harmless) race which could cause several cred
entries to be created for the same user.

23 years ago[PATCH] allow arbitrary alignment of NFS read/write requests
Trond Myklebust [Wed, 8 Jan 2003 01:59:37 +0000 (17:59 -0800)]
[PATCH] allow arbitrary alignment of NFS read/write requests

Patch by Chuck Lever to add a new field called wb_pgbase to the
nfs_page struct. This separates the concept of file page offset from
buffer page offset, allowing NFS to specify read and write requests
into buffers at offsets that are independent of the file page
offset. It is a prequisite for direct I/O in NFS.

23 years ago[PATCH] Fix NFS 'off by one' bug
Trond Myklebust [Wed, 8 Jan 2003 01:59:30 +0000 (17:59 -0800)]
[PATCH] Fix NFS 'off by one' bug

 This fixes a bug that was found using the Stanford checker.

23 years ago[PATCH] dm: Remove redundant error checking
Joe Thornber [Wed, 8 Jan 2003 01:54:50 +0000 (17:54 -0800)]
[PATCH] dm: Remove redundant error checking

bio_alloc() shouldn't fail if GFP_NOIO is used, and the bvec count is
sensible.  So remove redundant error checking.

23 years ago[PATCH] dm: Export dm_table_get_mode()
Joe Thornber [Wed, 8 Jan 2003 01:54:43 +0000 (17:54 -0800)]
[PATCH] dm: Export dm_table_get_mode()

Export dm_table_get_mode()

23 years ago[PATCH] dm: Simplify error->map
Joe Thornber [Wed, 8 Jan 2003 01:54:36 +0000 (17:54 -0800)]
[PATCH] dm: Simplify error->map

Just return an error from the error targets map function, rather than
erroring the buffer_head.

23 years ago[PATCH] dm: printk tgt->error if dm_table_add_target() fails.
Joe Thornber [Wed, 8 Jan 2003 01:54:28 +0000 (17:54 -0800)]
[PATCH] dm: printk tgt->error if dm_table_add_target() fails.

printk tgt->error if dm_table_add_target() fails.

23 years ago[PATCH] dm: Remove explicit returns from void fns (fluff)
Joe Thornber [Wed, 8 Jan 2003 01:54:21 +0000 (17:54 -0800)]
[PATCH] dm: Remove explicit returns from void fns (fluff)

Remove explicit return at the end of a couple of void functions.

23 years ago[PATCH] dm: Call dm_put_target_type() *after* calling the destructor
Joe Thornber [Wed, 8 Jan 2003 01:54:15 +0000 (17:54 -0800)]
[PATCH] dm: Call dm_put_target_type() *after* calling the destructor

Call dm_put_target_type() *after* calling the destructor.

23 years ago[PATCH] dm: rwlock_t -> rw_semaphore (fluff)
Joe Thornber [Wed, 8 Jan 2003 01:54:09 +0000 (17:54 -0800)]
[PATCH] dm: rwlock_t -> rw_semaphore (fluff)

Use a rw_semaphore in dm_target.c rather than a rwlock_t, just to keep
in line with dm.c

23 years ago[PATCH] dm: Correct target_type reference counting
Joe Thornber [Wed, 8 Jan 2003 01:54:02 +0000 (17:54 -0800)]
[PATCH] dm: Correct target_type reference counting

ti->use was only getting incremented the first time a target type was
retrieved (bug introduced by recent hch patch).

23 years ago[PATCH] dm: Correct clone info initialisation
Joe Thornber [Wed, 8 Jan 2003 01:53:56 +0000 (17:53 -0800)]
[PATCH] dm: Correct clone info initialisation

Initialize the clone-info's index to the original bio's index.
Required to properly handle stacking DM devices. [Kevin Corry]

23 years ago[PATCH] dm: Don't let the ioctl interface drop a suspended device
Joe Thornber [Wed, 8 Jan 2003 01:53:48 +0000 (17:53 -0800)]
[PATCH] dm: Don't let the ioctl interface drop a suspended device

Don't let the ioctl interface drop a suspended device.

23 years ago[PATCH] Reduced wakeup rate in direct-io code
Andrew Morton [Wed, 8 Jan 2003 01:48:48 +0000 (17:48 -0800)]
[PATCH] Reduced wakeup rate in direct-io code

We don't need to wake up the synchronous waiter for every BIO
completion - just for the final BIO.

This reduces the context switch rate of a 180 megabyte/sec read from
3500/sec to 2200/sec when performing 1400 128k I/Os per second, so it
more than halves the overhead.

23 years ago[PATCH] AIO support for raw/O_DIRECT
Andrew Morton [Wed, 8 Jan 2003 01:48:41 +0000 (17:48 -0800)]
[PATCH] AIO support for raw/O_DIRECT

Patch from Badari Pulavarty <pbadari@us.ibm.com> and myself.

This patch adds AIO support for direct-IO - O_DIRECT files, O_DIRECT
blockdevs and /dev/raw/rawN.

Also, the controlling `struct dio' is now always kmalloced, saving ~400 bytes
of stack.

The best tool for testing AIO/DIO is the modified fsx-linux in ext3 CVS.
See http://www.zip.com.au/~akpm/linux/ext3/ for details.  For example:

fsx-linux -Z -A -R -W -r 512 -w 4096 foo

Running many instances of this against ext3 currently fails with incorrect
file data; some bug in the new ext3 O_DIRECT support.  ext2 is OK though.

23 years ago[PATCH] bio dirtying infrastructure
Andrew Morton [Wed, 8 Jan 2003 01:48:34 +0000 (17:48 -0800)]
[PATCH] bio dirtying infrastructure

In two cases (AIO-for-direct-IO and some CDROM DMA stuff which Jens
did), we need to run set_page_dirty() in interrupt context.  After DMA
hardware has altered userspace pages for direct-IO reads.

But mapping->page_lock, mapping->private_lock and inode_lock are not
irq-safe.  And really, we don't want to convert those locks just for this
problem.

So what we do is to dirty the pages *before* starting I/O.  Then, in
interrupt completion context, check to see that they are still dirty.
If so then there is nothing to do.  But if the pages were cleaned while
the direct-IO is in progress we need to redirty them.  The code uses
schedule_work() for that.

Of course, we could use schedule_work() for all BIOs and pages.  The
speculative dirty-before-starting-IO is just an optimisation.  It's
quite unlikely that any of the pages will be cleaned during the direct
IO.

This code is basically untestable under normal circumstances, because the
relevant pages are pinned via get_user_pages().  This makes
is_page_cache_freeable() evaluate false and the VM doesn't try to write them
out anyway.  But if the pages happen to be MAP_SHARED file pages, pdflush
could clean them.  The way I tested this code was to disable the call to
bio_set_pages_dirty() in fs/direct-io.c.

23 years ago[PATCH] AIO support for raw/O_DIRECT
Andrew Morton [Wed, 8 Jan 2003 01:48:27 +0000 (17:48 -0800)]
[PATCH] AIO support for raw/O_DIRECT

Patch from Badari Pulavarty <pbadari@us.ibm.com> and myself

This patch adds the infrastructure for performing asynchronous (AIO) blockdev
direct-IO.

- Adds generic_file_aio_write_nolock() and make other
  generic_file_*_write() to use it.

- Modify generic_file_direct_IO() and ->direct_IO() functions to take
  "kiocb *" instead of "file *".

- Renames generic_direct_IO() to blockdev_direct_IO().

- Move generic_file_direct_IO() to mm/filemap.c (it is not
  blockdev-specific, whereas the rest of fs/direct-io.c is).

- Add AIO read/write support to the raw driver.

23 years ago[PATCH] return the correct thing from direct-io
Andrew Morton [Wed, 8 Jan 2003 01:48:21 +0000 (17:48 -0800)]
[PATCH] return the correct thing from direct-io

If direct-io encounters an error, it currently returns that error even if it
performed some I/O.

Change it so that if there is an error and we read or wrote something, return
the number of bytes written/read.  Only return the error code if zero bytes
were written/read.

23 years ago[PATCH] AIO exit fix
Andrew Morton [Wed, 8 Jan 2003 01:48:13 +0000 (17:48 -0800)]
[PATCH] AIO exit fix

We need to run exit_aio() when the final user of the mm goes away, else lots
of things leak.

Also remove a printk which comes out when this change is made.

This patch was acked by bcrl.

23 years ago[PATCH] remove code for 2.0 kernels from drivers/char/ftape/*
Adrian Bunk [Wed, 8 Jan 2003 01:48:06 +0000 (17:48 -0800)]
[PATCH] remove code for 2.0 kernels from drivers/char/ftape/*

The patch below is large but trivial:

It removes old #if'd compatiblity code for 2.0 kernels from
drivers/char/ftape/* (this includes the removal of two header files
including only compatibility code and the #include's of these files).

23 years agoPPC32: Update the defconfigs
Paul Mackerras [Tue, 7 Jan 2003 23:22:29 +0000 (10:22 +1100)]
PPC32: Update the defconfigs

23 years agoMerge samba.org:/home/paulus/kernel/linux-2.5
Paul Mackerras [Tue, 7 Jan 2003 23:11:25 +0000 (10:11 +1100)]
Merge samba.org:/home/paulus/kernel/linux-2.5
into samba.org:/home/paulus/kernel/for-linus-ppc

23 years agoAutomerge
Linus Torvalds [Tue, 7 Jan 2003 12:02:27 +0000 (04:02 -0800)]
Automerge

23 years agoaic7xxx and aic79xx driver updates:
Justin T. Gibbs [Tue, 7 Jan 2003 12:58:41 +0000 (05:58 -0700)]
aic7xxx and aic79xx driver updates:

o Correct memory mapped I/O test for legacy controllers
  that do not have the "auto-access-pause" feature.
o Prevent NMIs from triggering should the MMAP I/O test fail.
o Fix aic7770 (EISA/VLB) controller regression.
o Kill a few compiler warnings.

23 years agoaic7xxx and aic79xx drivers Correct several DV issues:
Justin T. Gibbs [Tue, 7 Jan 2003 12:40:37 +0000 (05:40 -0700)]
aic7xxx and aic79xx drivers Correct several DV issues:

 o Do not fallback to a wide speed if the device does not support
   wide transfers.

 o Don't bother allocating target instances for wide IDs on narrow
   controllers.

 o Add a few additional diagnostics to aid in tracking down DV bugs.

23 years agoaic7xxx/aicasm:
Justin T. Gibbs [Tue, 7 Jan 2003 12:32:07 +0000 (05:32 -0700)]
aic7xxx/aicasm:
   Remove the numerical_value portion of the grammer which is no
   longer referenced.  This eliminates a yacc warning.

23 years agoUpdate the aic7xxx Makefile so that the register information tables
Justin T. Gibbs [Tue, 7 Jan 2003 12:29:46 +0000 (05:29 -0700)]
Update the aic7xxx Makefile so that the register information tables
are not rebuilt on every build.

Use better Kbuild rule style for building firmware.

Submitted by: Sam Ravnborg <sam@ravnborg.org>

23 years ago[PATCH] small module patch
Anton Blanchard [Tue, 7 Jan 2003 10:31:40 +0000 (02:31 -0800)]
[PATCH] small module patch

Add a missing inline, shows up when modules are turned off.

23 years agoPPC32: Move files for the "oak" 403-based platform in with the
Paul Mackerras [Tue, 7 Jan 2003 09:48:52 +0000 (20:48 +1100)]
PPC32: Move files for the "oak" 403-based platform in with the
other 4xx platforms.

23 years agoPCI hotplug: clean up the try_module_get() logic a bit.
Greg Kroah-Hartman [Tue, 7 Jan 2003 08:41:22 +0000 (00:41 -0800)]
PCI hotplug: clean up the try_module_get() logic a bit.

23 years agoPCI: properly unregister a PCI device if it is removed.
Greg Kroah-Hartman [Tue, 7 Jan 2003 08:29:23 +0000 (00:29 -0800)]
PCI: properly unregister a PCI device if it is removed.

This is only used by pci hotplug and cardbus systems.

23 years agoIBM PCI Hotplug: fix compile time error due to find_bus() function name.
Greg Kroah-Hartman [Tue, 7 Jan 2003 08:24:14 +0000 (00:24 -0800)]
IBM PCI Hotplug: fix compile time error due to find_bus() function name.

23 years agoMerge bk://kernel.bkbits.net/davem/sparc-2.5
Linus Torvalds [Tue, 7 Jan 2003 07:08:57 +0000 (23:08 -0800)]
Merge bk://kernel.bkbits.net/davem/sparc-2.5
into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

23 years agoMerge bk://kernel.bkbits.net/davem/net-2.5
Linus Torvalds [Tue, 7 Jan 2003 07:07:51 +0000 (23:07 -0800)]
Merge bk://kernel.bkbits.net/davem/net-2.5
into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

23 years agoMerge davem@nuts.ninka.net:/home/davem/src/BK/net-2.5
David S. Miller [Tue, 7 Jan 2003 06:33:41 +0000 (22:33 -0800)]
Merge davem@nuts.ninka.net:/home/davem/src/BK/net-2.5
into kernel.bkbits.net:/home/davem/net-2.5

23 years ago[IPV4]: Convert rt_cache_state to use kmalloc_percpu
Ravikiran G. Thirumalai [Tue, 7 Jan 2003 06:50:02 +0000 (22:50 -0800)]
[IPV4]: Convert rt_cache_state to use kmalloc_percpu

23 years agoMerge davem@nuts.ninka.net:/home/davem/src/BK/sparc-2.5
David S. Miller [Tue, 7 Jan 2003 06:33:05 +0000 (22:33 -0800)]
Merge davem@nuts.ninka.net:/home/davem/src/BK/sparc-2.5
into kernel.bkbits.net:/home/davem/sparc-2.5

23 years agoMerge penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/numaq
Linus Torvalds [Tue, 7 Jan 2003 05:50:10 +0000 (21:50 -0800)]
Merge penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/numaq
into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

23 years ago[PATCH] nuke clustered_apic_mode and friends
Martin J. Bligh [Tue, 7 Jan 2003 05:48:45 +0000 (21:48 -0800)]
[PATCH] nuke clustered_apic_mode and friends

OK, the grand finale ... NUMA-Q is now moved into subarch, so we can
kill off the last vestiges - CONFIG_CLUSTERED_APIC, clustered_apic_mode,
and smpboot.h (which only contains machine specific stuff now anyway).
the esr_disable switch was the last bit, that goes to subarch too.

If you end up with an empty smpboot.h due to patch / bitkeeper interactions,
please remove it after this ...

23 years ago[PATCH] remove clustered_apic_mode from smpboot.c
Martin J. Bligh [Tue, 7 Jan 2003 05:48:32 +0000 (21:48 -0800)]
[PATCH] remove clustered_apic_mode from smpboot.c

This removes clustered_apic_mode from smpboot.c into subarch,
creating a headerfile "mach_wakecpu" for all the cpu wakeup stuff.
This is pretty much the last of clustered_apic_mode ... ;-)

23 years ago[PATCH] cleanup apicid <-> cpu mapping
Martin J. Bligh [Tue, 7 Jan 2003 05:48:19 +0000 (21:48 -0800)]
[PATCH] cleanup apicid <-> cpu mapping

To be honest, I have no idea what I was smoking when I wrote this
originally, and it's kind of coincidental that it works at all currently.

We never use physical apicids after the cpus are all booted, so we should
just store the logical IDs which all subsequent things use. The only things
that were using the apicid->cpu mapping were hokey anyway, and it's hard to
maintain for machines that have a large apic addressing space (eg P4s
in clustered mode). Rips out everything except the mapping from
logical_apic_id -> cpu.

23 years ago[PATCH] move one more to subarch, general tidy up
Martin J. Bligh [Tue, 7 Jan 2003 05:48:08 +0000 (21:48 -0800)]
[PATCH] move one more to subarch, general tidy up

Moves check_phys_apicid_present() into subarch, and cleans up a
couple of stupid errors, and some bracketing issues in the macros.

23 years ago[PATCH] changes do_boot_cpu to return an error code
Martin J. Bligh [Tue, 7 Jan 2003 05:47:57 +0000 (21:47 -0800)]
[PATCH] changes do_boot_cpu to return an error code

Patch from James Cleverdon

Changes do_boot_cpu to return an error code, instead of trying to
work it out later by magic and voodoo. Removes the other usage
of apicid->cpu which is hard to maintain cleanly.

23 years ago[PATCH] make i386 topology caching
Martin J. Bligh [Tue, 7 Jan 2003 05:47:45 +0000 (21:47 -0800)]
[PATCH] make i386 topology caching

Patch originally by Matt Dobson. Reworked a little by me.

Stores the mappings between cpus and nodes in an array, instead of
working them out every time. Gives about 4% off systime for kernel
compile (we use these for every page allocation), and removes one
of the two only usages of apicid->cpu mapping, which is really awkward
to keep for systems with large apic spaces, and is genererally pretty
useless anyway (later patch removes).

23 years ago[PATCH] create generalised apic_to_node mapping
Martin J. Bligh [Tue, 7 Jan 2003 05:47:33 +0000 (21:47 -0800)]
[PATCH] create generalised apic_to_node mapping

Create a generalised apicid_to_node mapping, replacing the local
apicid_to_quad I had just locally in NUMA-Q subarch before.

23 years agoPPC32: Remove extra __KERNEL__ checks in some headers, as well
Tom Rini [Tue, 7 Jan 2003 05:45:22 +0000 (22:45 -0700)]
PPC32: Remove extra __KERNEL__ checks in some headers, as well
as adding /* __KERNEL__ */ to the #endif of others.

23 years agoMerge http://jfs.bkbits.net/linux-2.5
Linus Torvalds [Tue, 7 Jan 2003 05:43:23 +0000 (21:43 -0800)]
Merge http://jfs.bkbits.net/linux-2.5
into penguin.transmeta.com:/home/penguin/torvalds/repositories/kernel/linux

23 years agoJFS: define aio_read and aio_write file_operations.
Dave Kleikamp [Tue, 7 Jan 2003 09:24:04 +0000 (03:24 -0600)]
JFS: define aio_read and aio_write file_operations.

This simply uses generic_file_aio_read and generic_file_aio_write

23 years agoJFS: add sync_fs super_operation
Dave Kleikamp [Tue, 7 Jan 2003 06:30:20 +0000 (00:30 -0600)]
JFS: add sync_fs super_operation

23 years agoMerge jfs@jfs.bkbits.net:linux-2.5
Dave Kleikamp [Tue, 7 Jan 2003 04:07:34 +0000 (22:07 -0600)]
Merge jfs@jfs.bkbits.net:linux-2.5
into shaggy.austin.ibm.com:/shaggy/bk/jfs-2.5

23 years agoMerge bk://ldm.bkbits.net/linux-2.5-core
Linus Torvalds [Tue, 7 Jan 2003 03:33:55 +0000 (19:33 -0800)]
Merge bk://ldm.bkbits.net/linux-2.5-core
into home.transmeta.com:/home/torvalds/v2.5/linux

23 years agoMerge osdl.org:/home/mochel/src/kernel/devel/linux-2.5-virgin
Patrick Mochel [Tue, 7 Jan 2003 05:38:43 +0000 (23:38 -0600)]
Merge osdl.org:/home/mochel/src/kernel/devel/linux-2.5-virgin
into osdl.org:/home/mochel/src/kernel/devel/linux-2.5-core

23 years agoMerge bk://linux.bkbits.net/linux-2.5
Patrick Mochel [Tue, 7 Jan 2003 05:30:45 +0000 (23:30 -0600)]
Merge bk://linux.bkbits.net/linux-2.5
into osdl.org:/home/mochel/src/kernel/devel/linux-2.5-virgin

23 years agoUpdate kobject documentation.
Patrick Mochel [Tue, 7 Jan 2003 05:25:39 +0000 (23:25 -0600)]
Update kobject documentation.

23 years agoMerge bk://are.twiddle.net/axp-2.5
Linus Torvalds [Tue, 7 Jan 2003 03:27:09 +0000 (19:27 -0800)]
Merge bk://are.twiddle.net/axp-2.5
into home.transmeta.com:/home/torvalds/v2.5/linux

23 years ago[PATCH] x86_64 extable fixes
Andi Kleen [Tue, 7 Jan 2003 02:40:58 +0000 (18:40 -0800)]
[PATCH] x86_64 extable fixes

Make x86-64 compile again after recent extable changes.

23 years ago[PATCH] Re: unix_getname buglet - > 2.5.4(?)
Ray Lee [Tue, 7 Jan 2003 02:40:52 +0000 (18:40 -0800)]
[PATCH] Re: unix_getname buglet - > 2.5.4(?)

Micheal Meeks notes:

"Evolution is non-functioning on recent 2.5.X kernels, due to
    mal-performance in getpeername => net/unix/af_unix.c (unix_getname),
    where it seems we switch 'sk' on 'peer', but not the (previously)
    typecast pointer to it; this fixes it."

Whitespace fixes by Ray Lee.

23 years agoMove x86 signal handler return stub to the vsyscall page,
Linus Torvalds [Tue, 7 Jan 2003 02:08:55 +0000 (18:08 -0800)]
Move x86 signal handler return stub to the vsyscall page,
and stop honoring the SA_RESTORER information.

This will prepare us for alternate signal handler returns.

23 years agoPPC32: Fix a problem in the bootloader/wrapper where we might
Tom Rini [Tue, 7 Jan 2003 01:56:13 +0000 (18:56 -0700)]
PPC32: Fix a problem in the bootloader/wrapper where we might
overwrite part of the initrd.

23 years agoPPC32: Fix some 'prep' machines which are not true PRePs,
Tom Rini [Tue, 7 Jan 2003 01:52:38 +0000 (18:52 -0700)]
PPC32: Fix some 'prep' machines which are not true PRePs,
and can safely poll for interrupts on the i8259.

23 years agoPPC32: Add explicit parens to the _ALIGN macro.
Tom Rini [Tue, 7 Jan 2003 01:32:40 +0000 (18:32 -0700)]
PPC32: Add explicit parens to the _ALIGN macro.

23 years agoPPC32: Fix a delay which could occur when booting on machines
Tom Rini [Tue, 7 Jan 2003 01:29:52 +0000 (18:29 -0700)]
PPC32: Fix a delay which could occur when booting on machines
without an RTC.

23 years agoPPC32: Cleanup the questions under CONFIG_ADVANCED_OPTIONS
Tom Rini [Tue, 7 Jan 2003 01:06:29 +0000 (18:06 -0700)]
PPC32: Cleanup the questions under CONFIG_ADVANCED_OPTIONS

23 years ago[SCTP]: Convert mibstats to use kmalloc_percpu
Ravikiran G. Thirumalai [Mon, 6 Jan 2003 18:20:22 +0000 (10:20 -0800)]
[SCTP]: Convert mibstats to use kmalloc_percpu

23 years ago[IPV6]: Convert mibstats to use kmalloc_percpu
Ravikiran G. Thirumalai [Mon, 6 Jan 2003 18:19:53 +0000 (10:19 -0800)]
[IPV6]: Convert mibstats to use kmalloc_percpu

23 years ago[IPV4]: Convert mibstats to use kmalloc_percpu
Ravikiran G. Thirumalai [Mon, 6 Jan 2003 18:19:16 +0000 (10:19 -0800)]
[IPV4]: Convert mibstats to use kmalloc_percpu

23 years ago[NET]: Remove dup wireless.h include in socket.c.
David S. Miller [Mon, 6 Jan 2003 18:17:43 +0000 (10:17 -0800)]
[NET]: Remove dup wireless.h include in socket.c.

23 years ago[NET]: Protect secpath references in skbuff.c with CONFIG_INET.
Tomas Szepe [Mon, 6 Jan 2003 18:16:38 +0000 (10:16 -0800)]
[NET]: Protect secpath references in skbuff.c with CONFIG_INET.

23 years ago[NET]: Convert sockets_in_use to use per_cpu areas.
Ravikiran G. Thirumalai [Mon, 6 Jan 2003 17:25:03 +0000 (09:25 -0800)]
[NET]: Convert sockets_in_use to use per_cpu areas.

23 years ago[SUNSAB]: Bug fixes for new sunsab uart driver.
James Morris [Mon, 6 Jan 2003 17:13:49 +0000 (09:13 -0800)]
[SUNSAB]: Bug fixes for new sunsab uart driver.
- Ensure ports are detected in the right order (also fixes crash in
  interrupt handler).
- Make use of uart api port fields: port, line, irq, fifosize, mapbase.
- Remove some variable fifo size logic, driver always uses 32 byte fifos.
- Unmask XPR interrupt in transmit_chars() (per 2.4).
- Unmask XPR interrupt early in sunsab_start_tx(), as the user may have
  been put to sleep by the uart api with XPR masked, and will otherwise
  not be able to unmask XPR again.
- Do not enable RFDF, as the code is no longer checking for or
  even handling rx status bytes.
- Make sunsab_type() return more useful info.
- Increment instance_p in sab_attach_callback().
- Remove self-reporting, as the uart api now reports the ports.
- Handle sunserial_current_minor increment/decrement properly.

23 years ago[SUNGEM]: Add warning to ppc code wrt. MAX_ADDR_LEN change.
David S. Miller [Mon, 6 Jan 2003 17:07:51 +0000 (09:07 -0800)]
[SUNGEM]: Add warning to ppc code wrt. MAX_ADDR_LEN change.

23 years ago[NET]: Increase MAX_ADDR_LEN.
Roland Dreier [Mon, 6 Jan 2003 17:06:34 +0000 (09:06 -0800)]
[NET]: Increase MAX_ADDR_LEN.
- Add ARPHRD_INFINIBAND
- Increase MAX_ADDR_LEN to 32 from 8
- Add suitable length protection to SIOCGIFHWADDR and friends.
- Add RTM_SETLINK for portably setting larger hw addrs.

23 years ago[SPARC64]: Translate IPT_SO_SET_REPLACE socket option for 32-bit apps.
Laszlo Valko [Mon, 6 Jan 2003 17:00:22 +0000 (09:00 -0800)]
[SPARC64]: Translate IPT_SO_SET_REPLACE socket option for 32-bit apps.

23 years agoImplement find_bus() for finding a struct bus_type by name.
Patrick Mochel [Mon, 6 Jan 2003 15:28:43 +0000 (09:28 -0600)]
Implement find_bus() for finding a struct bus_type by name.

Simply, pass off to kset_find_obj() to iterate over list of buses and
search for bus with certain name.

23 years agoadd kset_find_obj() to search for object in a kset's list.
Patrick Mochel [Mon, 6 Jan 2003 15:20:40 +0000 (09:20 -0600)]
add kset_find_obj() to search for object in a kset's list.

The operation is simple:
- Take read lock for kset.
- Iterate over kset->list.
- Compare name to each kobject's name.
- Return kobject if found.

23 years agodriver model: allow manual binding of devices to drivers.
Patrick Mochel [Mon, 6 Jan 2003 14:57:46 +0000 (08:57 -0600)]
driver model: allow manual binding of devices to drivers.

This creates and exports device_bind_driver() and device_release_driver()
that allow a caller to manually bind a device to a driver. Apparantly, some
drivers use this functionality (like USB): a driver binds to other device
interfaces during the driver's probe() method call for the first interface.

Implementation is easy enough, since it simply involves renaming and exporting
the internal functions attach() and detach().

23 years agobus drivers: fix leaking refcounts.
Patrick Mochel [Mon, 6 Jan 2003 14:02:26 +0000 (08:02 -0600)]
bus drivers: fix leaking refcounts.

In drivers/base/bus.c, get_bus() was called without put_bus() in these
functions:

- bus_for_each_dev()
- bus_for_each_drv()