]> git.hungrycats.org Git - linux/log
linux
23 years agoMerge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
Russell King [Mon, 6 Oct 2003 00:50:35 +0000 (01:50 +0100)]
Merge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-rmk

23 years ago[ARM] Ensure BK file modes allow others to read.
Russell King [Sun, 5 Oct 2003 12:22:49 +0000 (13:22 +0100)]
[ARM] Ensure BK file modes allow others to read.

23 years ago[ARM] Fix badly placed writeback/invalidation fixes.
Russell King [Sun, 5 Oct 2003 11:03:26 +0000 (12:03 +0100)]
[ARM] Fix badly placed writeback/invalidation fixes.

23 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-pcmcia
Linus Torvalds [Sun, 5 Oct 2003 07:11:05 +0000 (00:11 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-pcmcia
into home.osdl.org:/home/torvalds/v2.5/linux

23 years agoMerge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
Russell King [Sun, 5 Oct 2003 12:36:42 +0000 (13:36 +0100)]
Merge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-pcmcia

23 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-serial
Linus Torvalds [Sun, 5 Oct 2003 07:07:23 +0000 (00:07 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-serial
into home.osdl.org:/home/torvalds/v2.5/linux

23 years ago[SERIAL] removing legacy UART cruft
Bjorn Helgaas [Sun, 5 Oct 2003 22:58:04 +0000 (23:58 +0100)]
[SERIAL] removing legacy UART cruft

I'm trying to remove all the legacy UART cruft from ia64.  In theory
we should be able to discover all UARTs on ia64 via ACPI and PCI
enumeration, so we shouldn't need SERIAL_PORT_DFNS.

But SERIAL_PORT_DFNS currently does two things:
        1) it tells you about legacy devices you can't discover
           via a standard enumeration method, and
        2) it sizes old_serial_port[], which determines UART_NR,
           which is the maximum number of ports the driver will
           support

So here's a proposal to make SERIAL_PORT_DFNS optional and
provide another mechanism for configuring the number of ports
to support.

23 years ago[SERIAL] remove unused RS_TABLE definitions
Bjorn Helgaas [Sun, 5 Oct 2003 22:54:13 +0000 (23:54 +0100)]
[SERIAL] remove unused RS_TABLE definitions

Patch from Bjorn Helgaas.

RS_TABLE is defined by lots of architectures, but only referenced
in arch/{mips,ppc}.

This patch removes most of the unused definitions from 2.6.

23 years agoMerge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
Russell King [Sun, 5 Oct 2003 13:04:14 +0000 (14:04 +0100)]
Merge flint.arm.linux.org.uk:/usr/src/bk/linux-2.6
into flint.arm.linux.org.uk:/usr/src/bk/linux-2.6-serial

23 years agoMerge bk://kernel.bkbits.net/davem/sparc-2.5
Linus Torvalds [Sun, 5 Oct 2003 07:04:03 +0000 (00:04 -0700)]
Merge bk://kernel.bkbits.net/davem/sparc-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

23 years agoUse "select" instead of "depends on" to select GAMEPORT support
Linus Torvalds [Sun, 5 Oct 2003 06:57:58 +0000 (23:57 -0700)]
Use "select" instead of "depends on" to select GAMEPORT support
automatically for the sound drivers that require it.

This is a low less confusing than telling people that they
need to enable gameport support in order to listen to sound.

23 years ago[SPARC64]: Kill stray task->tty reference in Solaris module.
David S. Miller [Sun, 5 Oct 2003 05:39:42 +0000 (22:39 -0700)]
[SPARC64]: Kill stray task->tty reference in Solaris module.

23 years ago[PATCH] Update V4L2 "Hexium" driver
Michael Hunold [Sun, 5 Oct 2003 04:42:37 +0000 (21:42 -0700)]
[PATCH] Update V4L2 "Hexium" driver

This updates my two Video4Linux-2 drivers for the "Hexium Gemini" and
"Hexium Orion" cards.

It adds the long missing input handling for the "Gemini" card and
removes the annoying compile time warning about unused structures.

 - remove various hexium header files, put all driver relevant
   informations into the drivers.  these informations are not needed
   anywhere else.
 - fix driver names in information printk()s
 - fix device initialization for Hexium Gemini cards
 - add input switching for Hexium Gemini cards
 - fix all remaining "fixme"s

23 years ago[PATCH] Fix x86-64 compilation
Andi Kleen [Sun, 5 Oct 2003 04:42:27 +0000 (21:42 -0700)]
[PATCH] Fix x86-64 compilation

Make x86-64 compile again.

23 years ago[PATCH] Fix -Wdeclaration-after-statement warnings for x86-64
Andi Kleen [Sun, 5 Oct 2003 04:42:19 +0000 (21:42 -0700)]
[PATCH] Fix -Wdeclaration-after-statement warnings for x86-64

Fix warnings on compilers that support -Wdeclaration-after-statement
in the x86-64 specific code. x86-64 was happily using ISO-C99 code
before because it only uses gcc 3+ which is ISO-C99 compliant, but
unfortunately this warning was added in the top level Makefile.

From Bryan O'Sullivan.

23 years ago[PATCH] Reserve vserver syscall for x86-64
Andi Kleen [Sun, 5 Oct 2003 04:42:12 +0000 (21:42 -0700)]
[PATCH] Reserve vserver syscall for x86-64

Reserve vserver syscall on x86-64 too.

23 years ago[PATCH] Documentation_Changes visual cleanup
Rusty Russell [Sun, 5 Oct 2003 04:33:48 +0000 (21:33 -0700)]
[PATCH] Documentation_Changes visual cleanup

From:  Marcel Sebek <sebek64@post.cz>

23 years ago[PATCH] Christoph Hellwig no longer works at Caldera
Rusty Russell [Sun, 5 Oct 2003 04:33:39 +0000 (21:33 -0700)]
[PATCH] Christoph Hellwig no longer works at Caldera

From:  Adrian Bunk <bunk@fs.tum.de>

(Acked by Christoph Hellwig <hch@infradead.org>)

23 years ago[PATCH] unused variable in drivers_char_isicom.c
Rusty Russell [Sun, 5 Oct 2003 04:33:31 +0000 (21:33 -0700)]
[PATCH] unused variable in drivers_char_isicom.c

From:  krishnakumar@naturesoft.net

23 years ago[PATCH] Small cleanups for input
Rusty Russell [Sun, 5 Oct 2003 04:33:22 +0000 (21:33 -0700)]
[PATCH] Small cleanups for input

From:  Pavel Machek <pavel@suse.cz>

23 years ago[PATCH] 2 spelling patches in helps
Rusty Russell [Sun, 5 Oct 2003 04:33:15 +0000 (21:33 -0700)]
[PATCH] 2 spelling patches in helps

From:  Maciej Soltysiak <solt@dns.toxicfilms.tv>

23 years ago[PATCH] HISAX_SEDLBAUER_CS needs HISAX_SEDLBAUER
Rusty Russell [Sun, 5 Oct 2003 04:33:07 +0000 (21:33 -0700)]
[PATCH] HISAX_SEDLBAUER_CS needs HISAX_SEDLBAUER

From:  Adrian Bunk <bunk@fs.tum.de>

23 years ago[PATCH] Change list_emtpy() to take a const pointer
Rusty Russell [Sun, 5 Oct 2003 04:32:59 +0000 (21:32 -0700)]
[PATCH] Change list_emtpy() to take a const pointer

From:  "Perez-Gonzalez, Inaky" <inaky.perez-gonzalez@intel.com>

23 years ago[PATCH] ISDN PCBIT: fix nonmodular compile
Rusty Russell [Sun, 5 Oct 2003 04:32:49 +0000 (21:32 -0700)]
[PATCH] ISDN PCBIT: fix nonmodular compile

From:  Adrian Bunk <bunk@fs.tum.de>

23 years ago[PATCH] drivers_cpufreq_Kconfig URL update
Rusty Russell [Sun, 5 Oct 2003 04:32:41 +0000 (21:32 -0700)]
[PATCH] drivers_cpufreq_Kconfig URL update

From:  Petri Koistinen <petri.koistinen@iki.fi>

23 years ago[PATCH] unused variable in drivers_char_esp.c
Rusty Russell [Sun, 5 Oct 2003 04:32:33 +0000 (21:32 -0700)]
[PATCH] unused variable in drivers_char_esp.c

From:  krishnakumar@naturesoft.net

23 years ago[PATCH] Get rid of magic numbers in fs
Rusty Russell [Sun, 5 Oct 2003 04:32:24 +0000 (21:32 -0700)]
[PATCH] Get rid of magic numbers in fs

From:  Scharfe <l.s.r@web.de>

23 years ago[PATCH] update Kanoj Sarcar email address in docs
Rusty Russell [Sun, 5 Oct 2003 04:32:15 +0000 (21:32 -0700)]
[PATCH] update Kanoj Sarcar email address in docs

From:  Ed L Cashin <ecashin@uga.edu>

(Acked by Kanoj Sarcar <kanojsarcar@yahoo.com>)

23 years ago[PATCH] Bugzilla bug # 993 - Documenation_Changes still reads 2.5
Rusty Russell [Sun, 5 Oct 2003 04:32:07 +0000 (21:32 -0700)]
[PATCH] Bugzilla bug # 993 - Documenation_Changes still reads 2.5

From:  Josef Jeff Sipek <jeffpc@optonline.net>

23 years ago[PATCH] [PATCH ac97_plugin_ad1980.c: warning fix
Rusty Russell [Sun, 5 Oct 2003 04:31:58 +0000 (21:31 -0700)]
[PATCH] [PATCH ac97_plugin_ad1980.c: warning fix

From:  Vinay K Nallamothu <vinay-rc@naturesoft.net>

23 years ago[PATCH] More modules.txt removals
Rusty Russell [Sun, 5 Oct 2003 04:31:50 +0000 (21:31 -0700)]
[PATCH] More modules.txt removals

From:  Nicolas Kaiser <nikai@nikai.net>

23 years ago[PATCH] irq_affinity_write_proc no longer writes garbage into irq proc entries
Rusty Russell [Sun, 5 Oct 2003 04:31:42 +0000 (21:31 -0700)]
[PATCH] irq_affinity_write_proc no longer writes garbage into irq proc entries

From:  <adobriyan@mail.ru>

23 years ago[PATCH] Fix comment in parse_hex_value
Rusty Russell [Sun, 5 Oct 2003 04:31:34 +0000 (21:31 -0700)]
[PATCH] Fix comment in parse_hex_value

From:  <adobriyan@mail.ru>

(Acked by William Lee Irwin III <wli@holomorphy.com>)

23 years ago[PATCH] Update for Documentation_binfmt_misc-document
Rusty Russell [Sun, 5 Oct 2003 04:31:25 +0000 (21:31 -0700)]
[PATCH] Update for Documentation_binfmt_misc-document

From:  "ismail (cartman) donmez" <kde@myrealbox.com>

23 years ago[PATCH] drivers_ide_Kconfig URL updates
Rusty Russell [Sun, 5 Oct 2003 04:31:16 +0000 (21:31 -0700)]
[PATCH] drivers_ide_Kconfig URL updates

From:  Petri Koistinen <petri.koistinen@iki.fi>

23 years ago[PATCH] sysv_hash() is cleanups
Rusty Russell [Sun, 5 Oct 2003 04:31:09 +0000 (21:31 -0700)]
[PATCH] sysv_hash() is cleanups

From:  Scharfe <l.s.r@web.de>

The sysvfs code was confusing. Rewrite it to be readable, and
use the existing VFS helper for recalculating the hash instead
of doing it all by hand.

23 years ago[PATCH] drivers_media_Kconfig URL update
Rusty Russell [Sun, 5 Oct 2003 04:31:00 +0000 (21:31 -0700)]
[PATCH] drivers_media_Kconfig URL update

From:  Petri Koistinen <petri.koistinen@iki.fi>

23 years ago[PATCH] hlist constification
Rusty Russell [Sun, 5 Oct 2003 04:30:51 +0000 (21:30 -0700)]
[PATCH] hlist constification

From:  Mitchell Blank Jr <mitch@sfgoth.com>

Make some more of the hlist functions accept constant arguments.

23 years ago[PATCH] kconfig language doc r.e. ---help---
Rusty Russell [Sun, 5 Oct 2003 04:30:43 +0000 (21:30 -0700)]
[PATCH] kconfig language doc r.e. ---help---

From:  Stewart Smith <stewart@linux.org.au>

Document "---help---" tag in Kconfig

(Acked by Roman Zippel <zippel@linux-m68k.org>)

23 years ago[PATCH] Documentation_vm_hugetlbfs.txt cleanup
Rusty Russell [Sun, 5 Oct 2003 04:30:35 +0000 (21:30 -0700)]
[PATCH] Documentation_vm_hugetlbfs.txt cleanup

From:  Maciej Soltysiak <solt@dns.toxicfilms.tv>

  This changes the hugetlbfs examples actually being possible to
  compile. errno is being used while it's not included.

(Acked by William Lee Irwin III <wli@holomorphy.com>)

23 years ago[PATCH] correct number of CPUs in Kconfig help file
Rusty Russell [Sun, 5 Oct 2003 04:30:26 +0000 (21:30 -0700)]
[PATCH] correct number of CPUs in Kconfig help file

From:  Marcel Sebek <sebek64@post.cz>

(Acked by  William Lee Irwin III <wli@holomorphy.com>)

23 years ago[PATCH] Fix Linux 2.5 -> Linux 2.6
Rusty Russell [Sun, 5 Oct 2003 04:30:18 +0000 (21:30 -0700)]
[PATCH] Fix Linux 2.5 -> Linux 2.6

From:  Holger Schurig <h.schurig@mn-logistik.de>

23 years ago[PATCH] Add hint on sysrq on some keyboards
Rusty Russell [Sun, 5 Oct 2003 04:30:10 +0000 (21:30 -0700)]
[PATCH] Add hint on sysrq on some keyboards

From:  Pavel Machek <pavel@suse.cz>

23 years ago[PATCH] Bugzilla bug # 984 - 2.6 readme is still for 2.5
Rusty Russell [Sun, 5 Oct 2003 04:30:02 +0000 (21:30 -0700)]
[PATCH] Bugzilla bug # 984 - 2.6 readme is still for 2.5

From:  Josef Jeff Sipek <jeffpc@optonline.net>

23 years ago[PATCH] Rearrange error handling in fs_pipe.c a bit
Rusty Russell [Sun, 5 Oct 2003 04:29:54 +0000 (21:29 -0700)]
[PATCH] Rearrange error handling in fs_pipe.c a bit

From:  Scharfe <l.s.r@web.de>

  clean up the error handling in pipefs a little bit,
  shrinking pipe.o by whopping 4 bytes! :)

23 years ago[PATCH] Obvious sched doc fix
Rusty Russell [Sun, 5 Oct 2003 04:29:46 +0000 (21:29 -0700)]
[PATCH] Obvious sched doc fix

From:  Mitchell Blank Jr <mitch@sfgoth.com>

  Obvious typo fix for the scheduler docs

(Acked by Ingo)

23 years ago[PATCH] Remove racy check_mem_region() call from pcbit_drv.c
Rusty Russell [Sun, 5 Oct 2003 04:29:38 +0000 (21:29 -0700)]
[PATCH] Remove racy check_mem_region() call from pcbit_drv.c

From:  Bob Miller <rem@osdl.org>

  Removed the check_mem_region() call and replaced with request_mem_region().

23 years ago[PATCH] Remove extra #includes
Rusty Russell [Sun, 5 Oct 2003 04:29:29 +0000 (21:29 -0700)]
[PATCH] Remove extra #includes

From:  Tom Rini <trini@kernel.crashing.org>

  This removes two extra #includes of <linux/spinlock.h>.
  Nothing in either of these files require <linux/spinlock.h>.

23 years ago[PATCH] Use mod_timer in net_wanrouter_af_wanpipe.c
Rusty Russell [Sun, 5 Oct 2003 04:29:21 +0000 (21:29 -0700)]
[PATCH] Use mod_timer in net_wanrouter_af_wanpipe.c

From:  Vinay K Nallamothu <vinay-rc@naturesoft.net>

23 years ago[PATCH] Bugzilla bug # 267 - scripts_ver_linux fix
Rusty Russell [Sun, 5 Oct 2003 04:29:12 +0000 (21:29 -0700)]
[PATCH] Bugzilla bug # 267 - scripts_ver_linux fix

From:  Jeff Sipek <jeffpc@optonline.net>

23 years agoMerge bk://kernel.bkbits.net/davem/net-2.5
Linus Torvalds [Sun, 5 Oct 2003 04:11:08 +0000 (21:11 -0700)]
Merge bk://kernel.bkbits.net/davem/net-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

23 years ago[PATCH] fix pte_chain leak in do_no_page()
Andrew Morton [Sun, 5 Oct 2003 04:04:28 +0000 (21:04 -0700)]
[PATCH] fix pte_chain leak in do_no_page()

From: "V. Rajesh" <vrajesh@eecs.umich.edu>

Fix a rare pte_chain memory leak in do_no_page()

23 years ago[PATCH] fix "compat ioctl consolidation" for "move job
Andrew Morton [Sun, 5 Oct 2003 04:04:20 +0000 (21:04 -0700)]
[PATCH] fix "compat ioctl consolidation" for "move job

A little fix which is needed if both the "compat ioctl consolidation" and
"move job control fields from task_struct to signal_struct" patches are
applied.

23 years ago[PATCH] move job control fields from task_struct to
Andrew Morton [Sun, 5 Oct 2003 04:04:11 +0000 (21:04 -0700)]
[PATCH] move job control fields from task_struct to

From: Roland McGrath <roland@redhat.com>

This patch completes what was started with the `process_group' accessor
function, moving all the job control-related fields from task_struct into
signal_struct and using process_foo accessor functions to read them.  All
these things are per-process in POSIX, none per-thread.  Off hand it's hard
to come up with the hairy MT scenarios in which the existing code would do
insane things, but trust me, they're there.  At any rate, all the uses
being done via inline accessor functions now has got to be all good.

I did a "make allyesconfig" build and caught the few random drivers and
whatnot that referred to these fields.  I was surprised to find how few
references to ->tty there really were to fix up.  I'm sure there will be a
few more fixups needed in non-x86 code.  The only actual testing of a
running kernel with these patches I've done is on my normal minimal x86
config.  Everything works fine as it did before as far as I can tell.

One issue that may be of concern is the lack of any locking on multiple
threads diddling these fields.  I don't think it really matters, though
there might be some obscure races that could produce inconsistent job
control results.  Nothing shattering, I'm sure; probably only something
like a multi-threaded program calling setsid while its other threads do tty
i/o, which never happens in reality.  This is the same situation we get by
using ->group_leader->foo without other synchronization, which seemed to be
the trend and noone was worried about it.

23 years ago[PATCH] cpufreq sysfs oops fix
Andrew Morton [Sun, 5 Oct 2003 04:04:02 +0000 (21:04 -0700)]
[PATCH] cpufreq sysfs oops fix

From: Dominik Brodowski <linux@brodo.de>

Fix an ordering problem whcih was causing cpufreq oopses in sysfs.

23 years ago[PATCH] dscc4 driver fixes
Andrew Morton [Sun, 5 Oct 2003 04:03:53 +0000 (21:03 -0700)]
[PATCH] dscc4 driver fixes

From: Francois Romieu <romieu@fr.zoreil.com>

- dscc4_release_ring() must not appear in dscc4_{open/close} as the rings
  are allocated/freed in dscc4_{found1/free1};

- more elegant handling of return status in dscc4_found1().

23 years ago[PATCH] ext3 block allocator locking fix
Andrew Morton [Sun, 5 Oct 2003 04:03:45 +0000 (21:03 -0700)]
[PATCH] ext3 block allocator locking fix

When the BKL was removed from ext3 we lost locking coverage for
get_block()-versus-get_block().  Nobody seems to have hit the race because
get_block() almost always runs under i_sem: only memory pressure-based
writeout over a file hole runs outside i_sem.

ext2 uses the dedicated i_meta_lock spinlock in the inode to provide the
needed locking.  But ext3 already has an rwsem around all the get_block()
activity to protect it from truncate-related races.

So this patch just converts that rwsem into a semaphore, so concurrent
get_block() can never occur.  This will be more efficient than adding the new
spinlock.

We lose the ability to have two threads run get_block() against the same file
at the same time but again, that only happens during pageout over a hole
anyway.

(Kudos Alex Tomas for noticing the bug)

23 years ago[PATCH] EISA_bus cleanup
Andrew Morton [Sun, 5 Oct 2003 04:03:36 +0000 (21:03 -0700)]
[PATCH] EISA_bus cleanup

From: Marc Zyngier <mzyngier@freesurf.fr>, Christoph Hellwig

I'd like to kill willy's CONFIG_EISA_ALWAYS kludge.  So make EISA_bus a
variable always when CONFIG_EISA is set and initialize it to 1 for alpha.
We probably want to do that only if the system actually supports eisa, but
I keep the old behaviour for now.

23 years ago[PATCH] kernel documentation fixes
Andrew Morton [Sun, 5 Oct 2003 04:03:26 +0000 (21:03 -0700)]
[PATCH] kernel documentation fixes

From: Michael Still <mikal@stillhq.com>

The patch squelches build errors in the kernel-doc make targets by adding
documentation to arguements previously not documented, and updating the
argument names where they have changed.

23 years ago[PATCH] /proc/sys/auxv
Andrew Morton [Sun, 5 Oct 2003 04:03:17 +0000 (21:03 -0700)]
[PATCH] /proc/sys/auxv

From: Roland McGrath <roland@redhat.com>

This was supposed to be part of the recent mm_struct.saved_auxv[] patch.

The /proc addition is half the reason for the patch, and the more important
one (letting you debug live processes, while NT_AUXV in core dumps lets you
debug dead ones).  The patch below was supposed to be part of the original.

23 years ago[PATCH] document the macro for translating PROT_ to VM_ bits
Andrew Morton [Sun, 5 Oct 2003 04:03:09 +0000 (21:03 -0700)]
[PATCH] document the macro for translating PROT_ to VM_ bits

From: Muli Ben-Yehuda <mulix@mulix.org>

Shed some much-needed light.

23 years ago[PATCH] compat ioctl consolidation
Andrew Morton [Sun, 5 Oct 2003 04:03:00 +0000 (21:03 -0700)]
[PATCH] compat ioctl consolidation

From: Arun Sharma <arun.sharma@intel.com>, kevin.tian@intel.com

Move a whole bunch of filesystem ioctl conversion functions out of per-arch
files and into fs/compat_ioctl.c

It moves linux32_dirent to compat.h and renames it as compat_dirent.
linux32_dirent has been eliminated from ia64.  Other archs should do the
same.

We'll leave old_linux_dirent32 as is, since it seems to be arch specific
(ia64 doesn't use it for example).

23 years ago[PATCH] node enumeration fixes
Andrew Morton [Sun, 5 Oct 2003 04:02:51 +0000 (21:02 -0700)]
[PATCH] node enumeration fixes

From: Matthew Dobson <colpatch@us.ibm.com>

Here's a small update to the numnodes fix that went into -mm3.  The biggest
changes are:

1) move the actual NODES_SHIFT and MAX_NUMNODES definitions into
   linux/numa.h and include this in linux/mmzone.h, instead of being
   directly in linux/mmzone.h.  This allows other files to include *just*
   the NUMNODES stuff w/out grabbing all of mmzone.h.

2) pull NODE_SHIFT out of linux/mm.h.  This isn't used anywhere in the
   kernel, and it will only get confused with NODES_SHIFT.

3) Fix the IA64 patch.  The original patch I had sent out hadn't been
   tested on IA64.  It was mostly right, but there were circular
   dependencies.  All better now, and acked by Jesse.

4) In linux/mmzone.h, insert code to define MAX_NODES_SHIFT based on the
   size of unsigned long.  For 64-bit arches, we can have a much larger
   value.  This allows IA64 to have 100's or 1000's of nodes.
   MAX_NODES_SHIFT is defined as 10 (ie: 1024 nodes) for 64-bit for now,
   although it could likely be much larger.  For 32-bit it is 6 (ie: 64
   nodes).

5) Small cleanup in include/asm-arm/memory.h.  Mostly the result of the
   new linux/numa.h file.  Much cleaner and more readable now.

23 years ago[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [5/5]
Andrew Morton [Sun, 5 Oct 2003 04:02:42 +0000 (21:02 -0700)]
[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [5/5]

From: Matthew Dobson <colpatch@us.ibm.com>

Fix up the ia64 arch.

23 years ago[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [4/5]
Andrew Morton [Sun, 5 Oct 2003 04:02:34 +0000 (21:02 -0700)]
[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [4/5]

From: Matthew Dobson <colpatch@us.ibm.com>

Fix up the arm arch.  This needs to be reviewed.  Relatively
straightforward replacement of NR_NODES with standard MAX_NUMNODES.

23 years ago[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [3/5]
Andrew Morton [Sun, 5 Oct 2003 04:02:25 +0000 (21:02 -0700)]
[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [3/5]

From: Matthew Dobson <colpatch@us.ibm.com>

Fix up the sh arch.  sh defined NR_NODES, change sh to use standard
MAX_NUMNODES instead.

23 years ago[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [2/5]
Andrew Morton [Sun, 5 Oct 2003 04:02:01 +0000 (21:02 -0700)]
[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [2/5]

From: Matthew Dobson <colpatch@us.ibm.com>

Remove MAX_NR_NODES.  This value is only used in a couple of places, and
it's incorrectly used in all those places as far as I can tell.  Replace
with MAX_NUMNODES.  Create MAX_NODES_SHIFT and use this value to check
NODES_SHIFT is appropriate.  A possible future patch should make
MAX_NODES_SHIFT vary based on 32 vs.  64 bit archs.

23 years ago[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [1/5]
Andrew Morton [Sun, 5 Oct 2003 04:01:50 +0000 (21:01 -0700)]
[PATCH] Clean up MAX_NR_NODES/NUMNODES/etc. [1/5]

From: Matthew Dobson <colpatch@us.ibm.com>

This starts a series of cleanups against the way in which the various
architectures specify the number of nodes and memory zones.  We end up
supporting up to 1024 memory zones on ia64, which is a recent requirement.

Has been tested on ia32, ia64 (UMA), ppa64 (UMA) and NUMAQ.

Make sure MAX_NUMNODES is defined in one and only one place.  Remove
superfluous definitions.  Instead of defining MAX_NUMNODES in
asm/numnodes.h, we define NODES_SHIFT there.  Then in linux/mmzone.h we
turn that NODES_SHIFT value into MAX_NUMNODES.

23 years ago[PATCH] boot-time selectable log buffer size
Andrew Morton [Sun, 5 Oct 2003 04:01:16 +0000 (21:01 -0700)]
[PATCH] boot-time selectable log buffer size

From: Andrea Arcangeli <andrea@suse.de>

Allow the printk log buffer size to be selected with a __setup parameter,
`log_buf_len=N', where N must be a power-of-two.

The default, initial statically allocated buffer size is still determined via
kernel config.

23 years ago[NET]: asm-parisc/checksum.h needs linux/in6.h
Arnaldo Carvalho de Melo [Sun, 5 Oct 2003 00:54:02 +0000 (17:54 -0700)]
[NET]: asm-parisc/checksum.h needs linux/in6.h

23 years ago[IPVS] fix the unlocking bug in the ip_vs_conn_seq_stop
Wensong Zhang [Sun, 5 Oct 2003 00:51:11 +0000 (17:51 -0700)]
[IPVS] fix the unlocking bug in the ip_vs_conn_seq_stop

23 years ago[IPVS]: Properly handle non-linear skbs.
Julian Anastasov [Sat, 4 Oct 2003 22:20:59 +0000 (15:20 -0700)]
[IPVS]: Properly handle non-linear skbs.

Most of the changes come from Paul `Rusty' Russell. Now we
modify the skbs only for IPVS packets.

23 years ago[CRYPTO]: Kconfig URL updates.
Petri T. Koistinen [Sat, 4 Oct 2003 22:18:38 +0000 (15:18 -0700)]
[CRYPTO]: Kconfig URL updates.

23 years agoMerge bk://kernel.bkbits.net/davem/sparc-2.5
Linus Torvalds [Sat, 4 Oct 2003 11:13:55 +0000 (04:13 -0700)]
Merge bk://kernel.bkbits.net/davem/sparc-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

23 years agoMerge bk://kernel.bkbits.net/davem/net-2.5
Linus Torvalds [Sat, 4 Oct 2003 11:08:14 +0000 (04:08 -0700)]
Merge bk://kernel.bkbits.net/davem/net-2.5
into home.osdl.org:/home/torvalds/v2.5/linux

23 years agoMerge bk://bk.arm.linux.org.uk/linux-2.6-rmk
Linus Torvalds [Sat, 4 Oct 2003 11:02:52 +0000 (04:02 -0700)]
Merge bk://bk.arm.linux.org.uk/linux-2.6-rmk
into home.osdl.org:/home/torvalds/v2.5/linux

23 years agoMerge
Russell King [Sat, 4 Oct 2003 23:59:33 +0000 (00:59 +0100)]
Merge

23 years ago[ARM PATCH] 1660/1: misc PXA/Lubbock fixes
Nicolas Pitre [Sat, 4 Oct 2003 22:59:41 +0000 (23:59 +0100)]
[ARM PATCH] 1660/1: misc PXA/Lubbock fixes

Patch from Nicolas Pitre

23 years ago[ARM PATCH] 1675/1: remove definitions of non-implemented system calls from include...
Robin Farine [Sat, 4 Oct 2003 22:49:47 +0000 (23:49 +0100)]
[ARM PATCH] 1675/1: remove definitions of non-implemented system calls from include/asm-arm/unistd.h

Patch from Robin Farine

Remove the definitions of system call numbers from include/asm-arm/unistd.h for which the corresponding vector in arch/arm/kernel/calls.S points to sys_ni_syscall.

Since the new glibc (2.3.2) implements some of the system call wrapper functions based on the presence or absence of such definitions, only implemented system calls should have a corresponding __NR_xyz preprocessor symbol in unistd.h.

23 years ago[ARM PATCH] 1624/1: BE support for io-readsl-armv4.S, io-reads-armv4.S, io-writesw...
Deepak Saxena [Sat, 4 Oct 2003 22:46:52 +0000 (23:46 +0100)]
[ARM PATCH] 1624/1: BE support for io-readsl-armv4.S, io-reads-armv4.S, io-writesw-armv4.S

Patch from Deepak Saxena

23 years ago[ARM PATCH] 1674/1: better ARM division routines
Nicolas Pitre [Sat, 4 Oct 2003 22:35:39 +0000 (23:35 +0100)]
[ARM PATCH] 1674/1: better ARM division routines

Patch from Nicolas Pitre

Here's my rewrite of the ARM division routines:

 - take advantage of the clz instruction on ARMv5+
 - simpler and faster code for modulo routines
 - fast path for power of 2 divisors
 - other miscellaneous optimizations

This code has already been merged into gcc-3.4.

23 years ago[ARM] Update Integrator IRQ decoding.
Russell King [Sat, 4 Oct 2003 18:58:38 +0000 (19:58 +0100)]
[ARM] Update Integrator IRQ decoding.

This takes account of the other IRQ changes, and makes the interrupt
decoding more efficient.

23 years ago[ARM] Add sysdev model for Integrator/AP.
Russell King [Sat, 4 Oct 2003 18:03:11 +0000 (19:03 +0100)]
[ARM] Add sysdev model for Integrator/AP.

23 years ago[ARM] Move Integrator flash map to driver model.
Russell King [Sat, 4 Oct 2003 17:52:57 +0000 (18:52 +0100)]
[ARM] Move Integrator flash map to driver model.

Since we want to support multiple platforms, we also move the
Integrator flash driver over to the driver model.

23 years ago[ARM] Clean up Integrator interrupt number definitions.
Russell King [Sat, 4 Oct 2003 16:11:52 +0000 (17:11 +0100)]
[ARM] Clean up Integrator interrupt number definitions.

This allows us to expand the interrupt numbers on the primary
interrupt controller.

23 years ago[ARM] Prepare Integrator support code for multiple machine support.
Russell King [Sat, 4 Oct 2003 15:38:32 +0000 (16:38 +0100)]
[ARM] Prepare Integrator support code for multiple machine support.

Split Integrator/AP and Integrator/PP2 specific support from the
Integrator core code.

23 years ago[ARM] Make die() more correct.
Russell King [Sat, 4 Oct 2003 14:09:44 +0000 (15:09 +0100)]
[ARM] Make die() more correct.

Use 'tsk' when accessing comm and pid members.
Use regs->ARM_sp not regs + 1 for base of stack dump.

23 years ago[ARM] Make ARM use Pat's generic PM suspend to RAM support
Russell King [Sat, 4 Oct 2003 12:38:39 +0000 (13:38 +0100)]
[ARM] Make ARM use Pat's generic PM suspend to RAM support

Hook the PXA and SA11x0 platform core power into the generic PM core.
Add support for freezing userspace processes.  Make our apm emulation
call the generic PM suspend to memory infrastructure.

23 years ago[ARM] leds.c, being so basic, needs linux/compiler.h
Russell King [Sat, 4 Oct 2003 12:16:04 +0000 (13:16 +0100)]
[ARM] leds.c, being so basic, needs linux/compiler.h

linux/init.h doesn't include linux/compiler.h, but needs some
definitions (the __attribute_unused__ stuff) from compiler.h.
Since this affects only two files, we include linux/compiler.h
directly into the affected files.

23 years ago[ARM] Add system device for LEDs.
Russell King [Sat, 4 Oct 2003 11:50:21 +0000 (12:50 +0100)]
[ARM] Add system device for LEDs.

This allows us to make the suspend/resume handling local to the LEDs
code itself, rather than being scattered in several files.

23 years ago[ARM] Add ARMv5TEJ to processor architecture list.
Russell King [Sat, 4 Oct 2003 11:34:40 +0000 (12:34 +0100)]
[ARM] Add ARMv5TEJ to processor architecture list.

23 years ago[ARM] Remove redundant include of net/bluetooth/Kconfig
Russell King [Sat, 4 Oct 2003 11:27:57 +0000 (12:27 +0100)]
[ARM] Remove redundant include of net/bluetooth/Kconfig

23 years ago[ARM] Select decompressor mmu handling based upon the architecture IDs.
Russell King [Sat, 4 Oct 2003 11:22:54 +0000 (12:22 +0100)]
[ARM] Select decompressor mmu handling based upon the architecture IDs.

23 years ago[PATCH] remove PDC-ADMA placeholders
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 11:00:26 +0000 (04:00 -0700)]
[PATCH] remove PDC-ADMA placeholders

They do not contain any real code.

23 years ago[PATCH] slc90e66: kill dummy init_dma_slc90e66()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:58:52 +0000 (03:58 -0700)]
[PATCH] slc90e66: kill dummy init_dma_slc90e66()

23 years ago[PATCH] sis5513: kill dummy init_dma_sis5513()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:57:54 +0000 (03:57 -0700)]
[PATCH] sis5513: kill dummy init_dma_sis5513()

23 years ago[PATCH] siimage: kill dummy init_dma_siimage()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:57:07 +0000 (03:57 -0700)]
[PATCH] siimage: kill dummy init_dma_siimage()

23 years ago[PATCH] sc1200: kill dummy init_dma_sc1200()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:56:25 +0000 (03:56 -0700)]
[PATCH] sc1200: kill dummy init_dma_sc1200()

23 years ago[PATCH] piix: kill dummy init_dma_piix()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:55:48 +0000 (03:55 -0700)]
[PATCH] piix: kill dummy init_dma_piix()

23 years ago[PATCH] pdc202xx_new: kill dummy init_dma_pdc202new()
Bartlomiej Zolnierkiewicz [Sat, 4 Oct 2003 10:55:11 +0000 (03:55 -0700)]
[PATCH] pdc202xx_new: kill dummy init_dma_pdc202new()