safe/jmp/linux-2.6
18 years ago[PATCH] uml: tls support: hack to make it compile on any host
Paolo 'Blaisorblade' Giarrusso [Fri, 31 Mar 2006 10:30:23 +0000 (02:30 -0800)]
[PATCH] uml: tls support: hack to make it compile on any host

Copy the definition of struct user_desc (with another name) for use by
userspace sources (where we use the host headers, and we can't be sure about
their content) to make sure UML compiles.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: implement {get,set}_thread_area for i386
Paolo 'Blaisorblade' Giarrusso [Fri, 31 Mar 2006 10:30:22 +0000 (02:30 -0800)]
[PATCH] uml: implement {get,set}_thread_area for i386

Implement sys_[gs]et_thread_area and the corresponding ptrace operations for
UML.  This is the main chunk, additional parts follow.  This implementation is
now well tested and has run reliably for some time, and we've understood all
the previously existing problems.

Their implementation saves the new GDT content and then forwards the call to
the host when appropriate, i.e.  immediately when the target process is
running or on context switch otherwise (i.e.  on fork and on ptrace() calls).

In SKAS mode, we must switch registers on each context switch (because SKAS
does not switches tls_array together with current->mm).

Also, added get_cpu() locking; this has been done for SKAS mode, since TT does
not need it (it does not use smp_processor_id()).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: clean arch_switch usage
Paolo 'Blaisorblade' Giarrusso [Fri, 31 Mar 2006 10:30:21 +0000 (02:30 -0800)]
[PATCH] uml: clean arch_switch usage

Call arch_switch also in switch_to_skas, even if it's, for now, a no-op for
that case (and mark this in the comment); this will change soon.

Also, arch_switch for TT mode is actually useless when the PT proxy (a
complicate debugging instrumentation for TT mode) is not enabled.  In fact, it
only calls update_debugregs, which checks debugregs_seq against seq (to check
if the registers are up-to-date - seq here means a "version number" of the
registers).

If the ptrace proxy is not enabled, debugregs_seq always stays 0 and
update_debugregs will be a no-op.  So, optimize this out (the compiler can't
do it).

Also, I've been disappointed by the fact that it would make a lot of sense if,
after calling a successful
update_debugregs(current->thread.arch.debugregs_seq),
current->thread.arch.debugregs_seq were updated with the new debugregs_seq.
But this is not done.  Is this a bug or a feature?  For all purposes, it seems
a bug (otherwise the whole mechanism does not make sense, which is also a
possibility to check), which causes some performance only problems (not
correctness), since we write_debugregs when not needed.

Also, as suggested by Jeff, remove a redundant enabling of SIGVTALRM,
comprised in the subsequent local_irq_enable().  I'm just a bit dubious if
ordering matters there...

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: split ldt.h in arch-independent and arch-dependant code
Paolo 'Blaisorblade' Giarrusso [Fri, 31 Mar 2006 10:30:19 +0000 (02:30 -0800)]
[PATCH] uml: split ldt.h in arch-independent and arch-dependant code

ldt-{i386,x86_64}.h is made of two different parts - some code for parsing of
LDT descriptors, which is arch-dependant, and the code to handle uml_ldt_t (an
LDT block inside UML), which is mostly arch-independant (among x86 and x86_64,
at least).

Join the common part in a single file (ldt.h) and split the rest away
(host_ldt-{i386,x86_64}.h).

This is needed because processor.h, with next patches, will start including
the LDT descriptor parsing macros in host_ldt.h, but it can't include ldt.h
because it uses semaphores (and to define semaphores one must first include
processor.h!).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: idle thread needn't take access to init_mm
Paolo 'Blaisorblade' Giarrusso [Fri, 31 Mar 2006 10:30:18 +0000 (02:30 -0800)]
[PATCH] uml: idle thread needn't take access to init_mm

Comparing this code which is the actual body of the arch-independent
cpu_idle(), it is clear that it's unnecessary to set ->mm and ->active_mm;
beyond that, a kernel thread is not supposed to have ->mm != NULL, only
active_mm.

This showed up because I used the assumption (which is IMHO valid) that kernel
thread have their ->mm == NULL, and it failed for this thread.

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Acked-by: Jeff Dike <jdike@addtoit.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: fix min usage
Al Viro [Fri, 31 Mar 2006 10:30:17 +0000 (02:30 -0800)]
[PATCH] uml: fix min usage

type-safe min() in arch/um/drivers/mconsole_kern.c

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: remove unused make variables
Al Viro [Fri, 31 Mar 2006 10:30:16 +0000 (02:30 -0800)]
[PATCH] uml: remove unused make variables

Removed assignments to unused variables in arch/um/os-Linux/Makefile

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: __user annotations
Al Viro [Fri, 31 Mar 2006 10:30:16 +0000 (02:30 -0800)]
[PATCH] uml: __user annotations

__user annotations (hppfs)

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: sparse cleanups
Al Viro [Fri, 31 Mar 2006 10:30:15 +0000 (02:30 -0800)]
[PATCH] uml: sparse cleanups

misc sparse annotations

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: kconfigs
Al Viro [Fri, 31 Mar 2006 10:30:14 +0000 (02:30 -0800)]
[PATCH] uml: kconfigs

kconfig sanitized around drivers/net

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: eliminate duplicate mrpropered files
Al Viro [Fri, 31 Mar 2006 10:30:13 +0000 (02:30 -0800)]
[PATCH] uml: eliminate duplicate mrpropered files

no need to add the same file twice to MRPROPER_FILES

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: clean up remapping code build magic
Al Viro [Fri, 31 Mar 2006 10:30:12 +0000 (02:30 -0800)]
[PATCH] uml: clean up remapping code build magic

kills unmap magic

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: eliminate symlinks to host arch
Al Viro [Fri, 31 Mar 2006 10:30:11 +0000 (02:30 -0800)]
[PATCH] uml: eliminate symlinks to host arch

kills symlinks in arch/um/sys-*

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: fix initcall return values
Jeff Dike [Fri, 31 Mar 2006 10:30:10 +0000 (02:30 -0800)]
[PATCH] uml: fix initcall return values

A number of UML initcalls were improperly returning 1.  Also removed any
nearby emacs formatting comments.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] uml: redeclare highmem
Jeff Dike [Fri, 31 Mar 2006 10:30:09 +0000 (02:30 -0800)]
[PATCH] uml: redeclare highmem

The earlier printf patch missed a corresponding change in the printed
variable.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] UML: Hotplug memory, take 2
Jeff Dike [Fri, 31 Mar 2006 10:30:08 +0000 (02:30 -0800)]
[PATCH] UML: Hotplug memory, take 2

Changes since first version
added check for MADV_REMOVE support on the host
fixed error return botch
shrunk sprintf array by one character

This adds hotplug memory support to UML.  The mconsole syntax is
  config mem=[+-]n[KMG]
In other words, add or subtract some number of kilobytes, megabytes, or
gigabytes.

Unplugged pages are allocated and then madvise(MADV_TRUNCATE), which is a
currently experimental madvise extension.  These pages are tracked so they
can be plugged back in later if the admin decides to give them back.  The
first page to be unplugged is used to keep track of about 4M of other
pages.  A list_head is the first thing on this page.  The rest is filled
with addresses of other unplugged pages.  This first page is not madvised,
obviously.

When this page is filled, the next page is used in a similar way and linked
onto a list with the first page.  Etc.  This whole process reverses when
pages are plugged back in.  When a tracking page no longer tracks any
unplugged pages, then it is next in line for plugging, which is done by
freeing pages back to the kernel.

Signed-off-by: Jeff Dike <jdike@addtoit.com>
Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix suspend with traced tasks
Pavel Machek [Fri, 31 Mar 2006 10:30:06 +0000 (02:30 -0800)]
[PATCH] Fix suspend with traced tasks

strace /bin/bash misbehaves after resume; this fixes it.

(akpm: it's scary calling refrigerator() in state TASK_TRACED, but it seems to
do the right thing).

Signed-off-by: Pavel Machek <pavel@suse.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] i386 kdump timer vector lockup fix
Vivek Goyal [Fri, 31 Mar 2006 10:30:05 +0000 (02:30 -0800)]
[PATCH] i386 kdump timer vector lockup fix

Porting the patch I posted for x86_64 to i386.

http://marc.theaimsgroup.com/?l=linux-kernel&m=114178139610707&w=2

o While using kdump, after a system crash when second kernel boots, timer
  vector gets (0x31) locked and CPU does not see timer interrupts
  travelling from IOAPIC to APIC.  Currently it does not lead to boot
  failure in second kernel as timer interrupts continues to come as ExtInt
  through LAPIC directly, but fixing it is good in case some boards do not
  support the other mode.

o After a system crash, it is not safe to service interrupts any more,
  hence interrupts are disabled.  This leads to pending interrupts at
  LAPIC.  LAPIC sends these interrupts to the CPU during early boot of
  second kernel.  Other pending interrupts are discarded saying unexpected
  trap but timer interrupt is serviced and CPU does not issue an LAPIC EOI
  because it think this interrupt came from i8259 and sends ack to 8259.
  This leads to vector 0x31 locking as LAPIC does not clear respective ISR
  and keeps on waiting for EOI.

o This patch issues extra EOI for the pending interrupts who have ISR set.

o Though today only timer seems to be the special case because in early
  boot it thinks interrupts are coming from i8259 and uses
  mask_and_ack_8259A() as ack handler and does not issue LAPIC EOI.  But
  probably doing it in generic manner for all vectors makes sense.

Signed-off-by: Vivek Goyal <vgoyal@in.ibm.com>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Cc: Andi Kleen <ak@muc.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Remove long dead i386 floppy asm code
Brian Gerst [Fri, 31 Mar 2006 10:30:04 +0000 (02:30 -0800)]
[PATCH] Remove long dead i386 floppy asm code

It's been disabled since v2.1.88

Signed-off-by: Brian Gerst <bgerst@didntduck.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] for_each_possible_cpu: sh
KAMEZAWA Hiroyuki [Fri, 31 Mar 2006 10:30:02 +0000 (02:30 -0800)]
[PATCH] for_each_possible_cpu: sh

for_each_cpu() actually iterates across all possible CPUs.  We've had mistakes
in the past where people were using for_each_cpu() where they should have been
iterating across only online or present CPUs.  This is inefficient and
possibly buggy.

We're renaming for_each_cpu() to for_each_possible_cpu() to avoid this in the
future.

This patch replaces for_each_cpu with for_each_possible_cpu.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Paul Mundt <lethal@linux-sh.org>
Cc: Kazumoto Kojima <kkojima@rr.iij4u.or.jp>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] hugetlb: don't allow free hugetlb count fall below reserved count
Chen, Kenneth W [Fri, 31 Mar 2006 10:30:01 +0000 (02:30 -0800)]
[PATCH] hugetlb: don't allow free hugetlb count fall below reserved count

With strict page reservation, I think kernel should enforce number of free
hugetlb page don't fall below reserved count.  Currently it is possible in
the sysctl path.  Add proper check in sysctl to disallow that.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] fix extra page ref count in follow_hugetlb_page
Chen, Kenneth W [Fri, 31 Mar 2006 10:29:57 +0000 (02:29 -0800)]
[PATCH] fix extra page ref count in follow_hugetlb_page

git-commit: d5d4b0aa4e1430d73050babba999365593bdb9d2
"[PATCH] optimize follow_hugetlb_page" breaks mlock on hugepage areas.

I mis-interpret pages argument and made get_page() unconditional.  It
should only get a ref count when "pages" argument is non-null.

Credit goes to Adam Litke who spotted the bug.

Signed-off-by: Ken Chen <kenneth.w.chen@intel.com>
Acked-by: Adam Litke <agl@us.ibm.com>
Cc: David Gibson <david@gibson.dropbear.id.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mm: schedule find_trylock_page() removal
Nick Piggin [Fri, 31 Mar 2006 10:29:56 +0000 (02:29 -0800)]
[PATCH] mm: schedule find_trylock_page() removal

find_trylock_page() is an odd interface in that it doesn't take a reference
like the others.  Now that XFS no longer uses it, and its last remaining
caller actually wants an elevated refcount, opencode that callsite and
schedule find_trylock_page() for removal.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Acked-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] migrate_pages_to() must be defined for the no swap case
Christoph Lameter [Fri, 31 Mar 2006 10:29:56 +0000 (02:29 -0800)]
[PATCH] migrate_pages_to() must be defined for the no swap case

Fix migrate_pages_to() definition.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/mtd/maps/vmax301.c: fix off by one vmax_mtd
Petri T. Koistinen [Fri, 31 Mar 2006 10:29:54 +0000 (02:29 -0800)]
[PATCH] drivers/mtd/maps/vmax301.c: fix off by one vmax_mtd

Fix an obvious off-by-one error (vmax_mtd[] contains two elements).

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mtd: fix broken name_to_dev_t() declaration
Herbert Poetzl [Fri, 31 Mar 2006 10:29:53 +0000 (02:29 -0800)]
[PATCH] mtd: fix broken name_to_dev_t() declaration

drivers/mtd/devices/blkmtd.c uses a local declaration of name_to_dev_t()
which is inconsistant with the real one.  the following patch fixes this by
removing the local declaration and including mount.h instead

this patch was originally done by Micah Anderson.

Signed-off-by: Herbert Poetzl <herbert@13thfloor.at>
Acked-by: Micah Anderson <micah@riseup.net>
Acked-by: Daniel Hokka Zakrisson <daniel@hozac.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mtd/redboot: Handle holes in fis table
Peter Korsgaard [Fri, 31 Mar 2006 10:29:52 +0000 (02:29 -0800)]
[PATCH] mtd/redboot: Handle holes in fis table

Redboot simply sets the first character of a fis entry to 0xff on "fis
delete".  The kernel redboot parser stops parsing on such an entry, and
without this patch any entries after a deleted image would not be detected.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Add chip used in collie to jedec_probe
Pavel Machek [Fri, 31 Mar 2006 10:29:51 +0000 (02:29 -0800)]
[PATCH] Add chip used in collie to jedec_probe

This adds flash chip used in Sharp Zaurus sl5500 (collie) to jedec_probe.
Values work for read-only access, but I have not figured out how to do
read-write.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: Richard Purdie <rpurdie@rpsys.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Dead code in mtd/maps/pci.c
Eric Sesterhenn [Fri, 31 Mar 2006 10:29:50 +0000 (02:29 -0800)]
[PATCH] Dead code in mtd/maps/pci.c

This fixes coverity bug #12.  The first two gotos in the function still
have the initial value for mtd set.  And the third goto just triggers for
!mtd

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Kill ifdefs in mtdcore.c
Pavel Machek [Fri, 31 Mar 2006 10:29:49 +0000 (02:29 -0800)]
[PATCH] Kill ifdefs in mtdcore.c

Kill unneccessary ifdefs in mtdcore.c.

Signed-off-by: Pavel Machek <pavel@suse.cz>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Fix debug statement in inftlcore.c
Eric Sesterhenn / snakebyte [Fri, 31 Mar 2006 10:29:47 +0000 (02:29 -0800)]
[PATCH] Fix debug statement in inftlcore.c

Fix a copy\a/paste bug found by cpminer inside the inftlcore.c file

Signed-off-by: Eric Sesterhenn <snakebyte@gmx.de>
Acked-by: Greg Ungerer <gerg@snapgear.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] mtd cmdlinepart: allow zero offset value
Atsushi Nemoto [Fri, 31 Mar 2006 10:29:46 +0000 (02:29 -0800)]
[PATCH] mtd cmdlinepart: allow zero offset value

Current cmdlinepart.c uses offset value 0 to specify a continuous
partition.  This prevents creating a second partition starting at 0.

For example, I can split 4MB device using "mtdparts=id:2M,2M", but I can
not do "mtdparts=id:2M@2M,2M@0" to swap mtd0 and mtd1.

This patch introduces special OFFSET_CONTINUOUS value for a continuous
partition and allows 0 for offset value.

Also this patch replaces 0xffffffff with UINT_MAX for SIZE_REMAINING.

Signed-off-by: Atsushi Nemoto <anemo@mba.ocn.ne.jp>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/mtd: Use ARRAY_SIZE macro
Tobias Klauser [Fri, 31 Mar 2006 10:29:45 +0000 (02:29 -0800)]
[PATCH] drivers/mtd: Use ARRAY_SIZE macro

Use ARRAY_SIZE macro instead of sizeof(x)/sizeof(x[0]) and remove
duplicates of the macro.

Signed-off-by: Tobias Klauser <tklauser@nuerscht.ch>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] MTD_NAND_SHARPSL and MTD_NAND_NANDSIM should be tristate's
Adrian Bunk [Fri, 31 Mar 2006 10:29:43 +0000 (02:29 -0800)]
[PATCH] MTD_NAND_SHARPSL and MTD_NAND_NANDSIM should be tristate's

MTD_NAND=m and MTD_NAND_SHARPSL=y or MTD_NAND_NANDSIM=y are illegal
combinations that mustn't be allowed.

This patch fixes this bug by making MTD_NAND_SHARPSL and MTD_NAND_NANDSIM
tristate's.

Additionally, it fixes some whitespace damage at these options.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Richard Purdie <rpurdie@rpsys.net>
Acked-by: "Artem B. Bityutskiy" <dedekind@yandex.ru>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] drivers/mtd/: small cleanups
Adrian Bunk [Fri, 31 Mar 2006 10:29:42 +0000 (02:29 -0800)]
[PATCH] drivers/mtd/: small cleanups

- chips/sharp.c: make two needlessly global functions static

- move some declarations to a header file where they belong to

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sem2mutex: drivers/mtd/
Ingo Molnar [Fri, 31 Mar 2006 10:29:41 +0000 (02:29 -0800)]
[PATCH] sem2mutex: drivers/mtd/

Semaphore to mutex conversion.

The conversion was generated via scripts, and the result was validated
automatically via a script as well.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] sem2mutex: mtd/doc2000.c
Ingo Molnar [Fri, 31 Mar 2006 10:29:40 +0000 (02:29 -0800)]
[PATCH] sem2mutex: mtd/doc2000.c

Semaphore to mutex conversion.

The conversion was generated via scripts, and the result was validated
automatically via a script as well.

Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cc: Thomas Gleixner <tglx@linutronix.de>
Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] m25p80: printk warning fix
Andrew Morton [Fri, 31 Mar 2006 10:29:39 +0000 (02:29 -0800)]
[PATCH] m25p80: printk warning fix

drivers/mtd/devices/m25p80.c: In function `m25p80_erase':
drivers/mtd/devices/m25p80.c:189: warning: signed size_t format, different type arg (arg 6)

Acked-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] pcmcia: declare pccard_iodyn_ops (fix m8xx_pcmcia.c compilation error)
Marcelo Tosatti [Fri, 31 Mar 2006 22:37:25 +0000 (16:37 -0600)]
[PATCH] pcmcia: declare pccard_iodyn_ops (fix m8xx_pcmcia.c compilation error)

Apparently the pccard_iodyn_ops declaration has been forgotten, which
results in a compilation error for m8xx_pcmcia.c

Signed-off-by: Marcelo Tosatti <marcelo.tosatti@cyclades.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[IA64] Avoid "u64 foo : 32;" for gcc3 vs. gcc4 compatibility
Tony Luck [Fri, 31 Mar 2006 18:28:29 +0000 (10:28 -0800)]
[IA64] Avoid "u64 foo : 32;" for gcc3 vs. gcc4 compatibility

gcc3 thinks that a 32-bit field of a u64 type is itself a u64, so
should be printed with "%ld".  gcc4 thinks it needs just "%d".
Make both versions happy by avoiding this construct.

Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoMerge with /pub/scm/linux/kernel/git/torvalds/linux-2.6.git
Steve French [Fri, 31 Mar 2006 15:43:35 +0000 (15:43 +0000)]
Merge ... /linux/kernel/git/torvalds/linux-2.6.git

Signed-off-by: Steve French <sfrench@us.ibm.com>
18 years ago[PATCH] pcmcia: fix pcmcia_device_remove oops
Hugh Dickins [Fri, 10 Mar 2006 21:15:19 +0000 (21:15 +0000)]
[PATCH] pcmcia: fix pcmcia_device_remove oops

Fix pcmcia_device_remove NULL pointer dereference at shutdown.

Signed-off-by: Hugh Dickins <hugh@veritas.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: Add support for Possio GCC AKA PCMCIA Siemens MC45
Petr Vandrovec [Fri, 10 Feb 2006 10:04:00 +0000 (02:04 -0800)]
[PATCH] pcmcia: Add support for Possio GCC AKA PCMCIA Siemens MC45

This ugly hack add support for Siemens MC45 PCMCIA GPRS card (which is
identical to Possio GCC, and which is offered by one of our local GPRS
providers).  Card has unfortunate feature that after poweron oxcf950 chip
is fully powered and works, but attached MC45 modem is powered down :-(
There is a special sequence (which takes 1 sec :-( ) to poweron MC45 (and
after MC45 powers on, it takes more than 2 secs until firmware fully
boots...) which needs to be executed after all powerons.

I'm really not familiar with PCMCIA subsystem, so I have no idea whether I
should issue request_region() on rest of oxcf950 address range (0-7 is
UART, 8-F are special configuration registers), or how this should be
better integrated with PM system and so on - I just put it in same place
where another hack already lived...

Card uses 18.432MHz XTAL, so to get it to work you must add lines below to
the /etc/pcmcia/serial.opts.

case "$MANFID-$FUNCID-$PRODID_1-$PRODID_2-$PRODID_3-$PRODID_4" in
'030c,0003-2-GPRS-CARD--')
    SERIAL_OPTS="baud_base 1152000"
    ;;
esac

Cc: Russell King <rmk@arm.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: pseudo device handling update
Dominik Brodowski [Sun, 5 Feb 2006 08:51:34 +0000 (09:51 +0100)]
[PATCH] pcmcia: pseudo device handling update

If the driver for the primary pseudo device is removed from the device,
the secondary driver must be removed as well -- it cannot exist on its own.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: convert DEV_OK to pcmcia_dev_present
Dominik Brodowski [Sun, 5 Mar 2006 10:04:33 +0000 (11:04 +0100)]
[PATCH] pcmcia: convert DEV_OK to pcmcia_dev_present

Instead of the DEV_OK macro, drivers should use pcmcia_dev_present().

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: use bitfield instead of p_state and state
Dominik Brodowski [Wed, 1 Mar 2006 23:09:29 +0000 (00:09 +0100)]
[PATCH] pcmcia: use bitfield instead of p_state and state

Instead of the two status values struct pcmcia_device->p_state and state,
use descriptive bitfields. Most value-checking in drivers was invalid, as
the core now only calls the ->remove() (a.k.a. detach) function in case the
attachement _and_ configuration was successful.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove unused p_dev->state flags
Dominik Brodowski [Fri, 20 Jan 2006 14:10:23 +0000 (15:10 +0100)]
[PATCH] pcmcia: remove unused p_dev->state flags

Remove the unused DEV_RELEASE_PENDING flag, and move the DEV_SUSPEND flag
into the p_dev structure, and make use of it at the core level.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: make pcmcia_release_{io,irq} static
Adrian Bunk [Thu, 19 Jan 2006 07:53:13 +0000 (23:53 -0800)]
[PATCH] pcmcia: make pcmcia_release_{io,irq} static

We can now make pcmcia_release_{io,irq} static.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: add return value to _config() functions
Dominik Brodowski [Fri, 31 Mar 2006 15:26:06 +0000 (17:26 +0200)]
[PATCH] pcmcia: add return value to _config() functions

Most of the driver initialization isn't done in the .probe function, but in
the internal _config() functions. Make them return a value, so that .probe
can properly report whether the probing of the device succeeded or not.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection
Dominik Brodowski [Fri, 31 Mar 2006 15:21:06 +0000 (17:21 +0200)]
[PATCH] pcmcia: remove dev_link_t and client_handle_t indirection

dev_link_t * and client_handle_t both mean struct pcmcai_device * by now.
Therefore, remove all such indirections.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device
Dominik Brodowski [Sun, 5 Mar 2006 09:45:09 +0000 (10:45 +0100)]
[PATCH] pcmcia: embed dev_link_t into struct pcmcia_device

Embed dev_link_t into struct pcmcia_device(), as they basically address the
same entity. The actual contents of dev_link_t will be cleaned up step by step.
This patch includes a bugfix from and signed-off-by Andrew Morton.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: rename pcmcia_device.state
Dominik Brodowski [Sun, 15 Jan 2006 15:26:00 +0000 (16:26 +0100)]
[PATCH] pcmcia: rename pcmcia_device.state

Rename pcmcia_device.state (which is used in very few places) to p_state
in order to avoid a namespace collision when moving the deprecated
dev_link_t into struct pcmcia_device

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove unneeded Vcc pseudo setting
Dominik Brodowski [Sun, 15 Jan 2006 11:43:16 +0000 (12:43 +0100)]
[PATCH] pcmcia: remove unneeded Vcc pseudo setting

As we do not allow setting Vcc in the pcmcia core, and Vpp1 and
Vpp2 can only be set to the same value, a lot of code can be
streamlined.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove export of pcmcia_release_configuration
Dominik Brodowski [Sun, 15 Jan 2006 10:18:12 +0000 (11:18 +0100)]
[PATCH] pcmcia: remove export of pcmcia_release_configuration

Handle the _modifying_ operation sm91c92_cs requires in
pcmcia_modify_configuration, so that the only remaining users
of pcmcia_release_configuration() are within the pcmcia core
module.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: default suspend and resume handling
Dominik Brodowski [Wed, 1 Mar 2006 23:02:33 +0000 (00:02 +0100)]
[PATCH] pcmcia: default suspend and resume handling

In all but one case, the suspend and resume functions of PCMCIA drivers
contain mostly of calls to pcmcia_release_configuration() and
pcmcia_request_configuration(). Therefore, move this code out of the
drivers and into the core.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: convert remaining users of pcmcia_release_io and _irq
Dominik Brodowski [Sun, 15 Jan 2006 09:05:19 +0000 (10:05 +0100)]
[PATCH] pcmcia: convert remaining users of pcmcia_release_io and _irq

Convert the remaining drivers which use pcmcia_release_io or
pcmcia_release_irq, and remove the EXPORT of these symbols.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: add pcmcia_disable_device
Dominik Brodowski [Sun, 15 Jan 2006 08:32:39 +0000 (09:32 +0100)]
[PATCH] pcmcia: add pcmcia_disable_device

pcmcia_disable_device(struct pcmcia_device *p_dev) performs the necessary
cleanups upon device or driver removal: it calls the appropriate
pcmcia_release_* functions, and can replace (most) of the current drivers'
_release() functions.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] libata: fix ata_xfer_tbl termination
Tejun Heo [Fri, 31 Mar 2006 13:48:52 +0000 (22:48 +0900)]
[PATCH] libata: fix ata_xfer_tbl termination

ata_xfer_tbl is terminated by entry with -1 as ->shift.  However,
->shift was unsigned int making the termination condition bogus.  This
patch converts ->shift and ->bits to int.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: make ata_qc_issue complete failed qcs
Tejun Heo [Fri, 31 Mar 2006 11:41:11 +0000 (20:41 +0900)]
[PATCH] libata: make ata_qc_issue complete failed qcs

There is no reason for the issuer to diddle with a failed qc as the
issuer has complete control over when a qc gets freed (usually in
->complete_fn).  Make ata_qc_issue() responsible for completing qcs
which failed to issue.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] libata: fix ata_qc_issue failure path
Tejun Heo [Fri, 31 Mar 2006 11:36:47 +0000 (20:36 +0900)]
[PATCH] libata: fix ata_qc_issue failure path

On sg_err failure path, ata_qc_issue() doesn't mark the qc active
before returning.  This triggers WARN_ON() in __ata_qc_complete() when
the qc gets completed.  This patch moves ap->active_tag and
QCFLAG_ACTIVE setting to the top of the function.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] ata_piix: fix ich6/m_map_db
Tejun Heo [Fri, 31 Mar 2006 11:01:50 +0000 (20:01 +0900)]
[PATCH] ata_piix: fix ich6/m_map_db

MAP tables of ich6 and ich6m are wrong.  Depending on port usage,
ata_piix may fail to initialize attached devices.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years ago[PATCH] serial_cs: add Merlin U630 IDs
Domen Puncer [Mon, 6 Mar 2006 09:25:53 +0000 (10:25 +0100)]
[PATCH] serial_cs: add Merlin U630 IDs

Add Merlin U630 IDs.

Signed-off-by: Domen Puncer <domen.puncer@ultra.si>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: AT91RM9200 Compact Flash driver
Andrew Victor [Tue, 21 Feb 2006 10:56:52 +0000 (12:56 +0200)]
[PATCH] pcmcia: AT91RM9200 Compact Flash driver

This patch adds support for the Compact Flash controller integrated in
the Atmel AT91RM9200 processor.

Signed-off-by: Andrew Victor <andrew@sanpeople.com
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: socket.functions starts with 1
Dominik Brodowski [Wed, 1 Mar 2006 13:04:52 +0000 (14:04 +0100)]
[PATCH] pcmcia: socket.functions starts with 1

socket.functions is the number of functions, and so must be one larger
than the maximum function number.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove wrong comment in ciscode.h
Komuro [Sun, 5 Feb 2006 08:56:59 +0000 (09:56 +0100)]
[PATCH] pcmcia: remove wrong comment in ciscode.h

Remove misleading comment.

Signed-off-by: Komuro <komurojun-mbn@nifty.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years agoMerge branch 'master'
Jeff Garzik [Fri, 31 Mar 2006 15:03:19 +0000 (10:03 -0500)]
Merge branch 'master'

18 years ago[PATCH] parport_cs: don't play games with resources
Dominik Brodowski [Sun, 15 Jan 2006 07:30:40 +0000 (08:30 +0100)]
[PATCH] parport_cs: don't play games with resources

pcmcia_request_io() doesn't mark the resource as busy in 2.6., therefore
there's no need to work around the registration of the resources into the
resource tree.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove duplicate fields in io_window_t
Dominik Brodowski [Sun, 15 Jan 2006 07:04:43 +0000 (08:04 +0100)]
[PATCH] pcmcia: remove duplicate fields in io_window_t

BasePort, NumPorts and Attributes are or can be embedded in
struct resource, so remove them.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: size reduction if ioctl isn't compiled
Dominik Brodowski [Sun, 15 Jan 2006 00:14:31 +0000 (01:14 +0100)]
[PATCH] pcmcia: size reduction if ioctl isn't compiled

If the kernel is configured to not include the deprecated PCMCIA ioctl,
some code doesn't need to be built.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove pcmcia_compat.c
Dominik Brodowski [Sat, 14 Jan 2006 23:51:53 +0000 (00:51 +0100)]
[PATCH] pcmcia: remove pcmcia_compat.c

Remove the compatibility wrappers, as they can (now) also be implemented
using macros. Please continue using these wrappers instead of new functions
until a new API has stabilized.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove include of config.h
Dominik Brodowski [Tue, 10 Jan 2006 20:35:49 +0000 (21:35 +0100)]
[PATCH] pcmcia: remove include of config.h

Remove the inclusion of include/config.h as it isn't needed any longer.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: use mutexes instead of semaphores
Dominik Brodowski [Tue, 10 Jan 2006 20:20:36 +0000 (21:20 +0100)]
[PATCH] pcmcia: use mutexes instead of semaphores

Use mutexes in the PCMICA core, as they suffice for what needs to be done.
Includes a bugfix from and Signed-off-by Andrew Morton.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove unused defines
Dominik Brodowski [Tue, 10 Jan 2006 19:54:46 +0000 (20:54 +0100)]
[PATCH] pcmcia: remove unused defines

Remove unused fields and declarations.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: make config_t independent, add reference counting
Dominik Brodowski [Tue, 10 Jan 2006 19:50:39 +0000 (20:50 +0100)]
[PATCH] pcmcia: make config_t independent, add reference counting

Handle config_t structs independent of struct pcmcia_socket, and add
reference counting for them.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: always use device pointer to config_t
Dominik Brodowski [Tue, 10 Jan 2006 19:48:59 +0000 (20:48 +0100)]
[PATCH] pcmcia: always use device pointer to config_t

Update the remaining users using the static lookup table of the PCMCIA
function configuration to use the struct pcmcia_device-contained pointer.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: access config_t using pointer instead of array
Dominik Brodowski [Tue, 10 Jan 2006 19:41:27 +0000 (20:41 +0100)]
[PATCH] pcmcia: access config_t using pointer instead of array

Access the PCMCIA config_t struct (one per device function) using
a pointer in struct pcmcia_device, instead of looking them up in
an array.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: remove unused field Present from config_t
Dominik Brodowski [Tue, 10 Jan 2006 19:40:40 +0000 (20:40 +0100)]
[PATCH] pcmcia: remove unused field Present from config_t

config_t.Present is set to the same value as CardValues, which isn't modified
anywhere. Therefore, we can use only one of these two objects.

Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[PATCH] pcmcia: permit single-character identifiers
Janos Farkas [Thu, 30 Mar 2006 23:04:57 +0000 (01:04 +0200)]
[PATCH] pcmcia: permit single-character identifiers

For some time, the core pcmcia drivers seem not to think single
character prod_ids are valid, thus preventing the "cleverly" named

  "D" "Link DWL-650 11Mbps WLAN Card"

Before (as in 2.6.16):
PRODID_1=""
PRODID_2="Link DWL-650 11Mbps WLAN Card"
PRODID_3="Version 01.02"
PRODID_4=""
MANFID=0156,0002
FUNCID=6

After (with the patch)
PRODID_1="D"
PRODID_2="Link DWL-650 11Mbps WLAN Card"
PRODID_3="Version 01.02"
PRODID_4=""
MANFID=0156,0002
FUNCID=6

Signed-off-by: Janos Farkas <chexum@gmail.com>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
18 years ago[NET]: Allow skb headroom to be overridden
Anton Blanchard [Fri, 31 Mar 2006 10:27:06 +0000 (02:27 -0800)]
[NET]: Allow skb headroom to be overridden

Previously we added NET_IP_ALIGN so an architecture can override the
padding done to align headers. The next step is to allow the skb
headroom to be overridden.

We currently always reserve 16 bytes to grow into, meaning all DMAs
start 16 bytes into a cacheline. On ppc64 we really want DMA writes to
start on a cacheline boundary, so we increase that headroom to one
cacheline.

Signed-off-by: Anton Blanchard <anton@samba.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[TCP]: Kill unused extern decl for tcp_v4_hash_connecting()
David S. Miller [Fri, 31 Mar 2006 10:25:46 +0000 (02:25 -0800)]
[TCP]: Kill unused extern decl for tcp_v4_hash_connecting()

Noticed by Alan Menegotto.

Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[NET]: add SO_RCVBUF comment
Andrew Morton [Fri, 31 Mar 2006 10:09:36 +0000 (02:09 -0800)]
[NET]: add SO_RCVBUF comment

Put a comment in there explaining why we double the setsockopt()
caller's SO_RCVBUF.  People keep wondering.

Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
18 years ago[PATCH] splice exports
Jeff Garzik [Fri, 31 Mar 2006 04:06:13 +0000 (23:06 -0500)]
[PATCH] splice exports

Woe be unto he who builds their filesystems as modules.

Signed-off-by: Jeff Garzik <jeff@garzik.org>
[ Obscure quote from the infamous geek bible? ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years agopowerpc: converted embedded platforms to use new define_machine support
Kumar Gala [Fri, 31 Mar 2006 05:39:57 +0000 (23:39 -0600)]
powerpc: converted embedded platforms to use new define_machine support

Removed platform_init usage on 83xx and 85xx and use define_machine and
probe().  For now we always return true in the problem since you can only
build for one specific board at a time.  This is an artificial constraint.
When we get ride of it we will need to update the Kconfig's for these
sub-arch's and make the board's probe() functions actually do something.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
18 years ago[CIFS] Fix unlink oops when indirectly called in rename error path
Steve French [Fri, 31 Mar 2006 03:37:08 +0000 (03:37 +0000)]
[CIFS] Fix unlink oops when indirectly called in rename error path
under heavy stress.

Signed-off-by: Steve French <sfrench@us.ibm.com>
18 years agoMerge with /pub/scm/linux/kernel/git/torvalds/linux-2.6.git
Steve French [Fri, 31 Mar 2006 03:35:56 +0000 (03:35 +0000)]
Merge ... /linux/kernel/git/torvalds/linux-2.6.git

Signed-off-by: Steve French <sfrench@us.ibm.com>
18 years agopowerpc: merge machine_check_exception between ppc32 & ppc64
Kumar Gala [Fri, 31 Mar 2006 03:11:15 +0000 (21:11 -0600)]
powerpc: merge machine_check_exception between ppc32 & ppc64

Make machine_check_exception handling code path the same on ppc32 & ppc64.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
18 years ago[IA64] Export cpu cache info by sysfs
Zhang, Yanmin [Mon, 27 Feb 2006 03:37:45 +0000 (11:37 +0800)]
[IA64] Export cpu cache info by sysfs

The patch exports 8 attributes of cpu cache info under
/sys/devices/system/cpu/cpuX/cache/indexX:
1) level
2) type
3) coherency_line_size
4) ways_of_associativity
5) size
6) shared_cpu_map
7) attributes
8) number_of_sets: number_of_sets=size/ways_of_associativity/coherency_line_size.

Signed-off-by: Zhang Yanmin <yanmin.zhang@intel.com>
Signed-off-by: Tony Luck <tony.luck@intel.com>
18 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6
Linus Torvalds [Thu, 30 Mar 2006 23:04:13 +0000 (15:04 -0800)]
Merge git://git./linux/kernel/git/kyle/parisc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/kyle/parisc-2.6: (24 commits)
  [PARISC] Fix double free when removing HIL drivers
  [PARISC] Add atomic_sub_and_test
  [PARISC] Enabled some NLS modules in a500, b180 and c3000 defconfigs
  [PARISC] Kill duplicated EXPORT_SYMBOL warnings
  [PARISC] Move ioremap EXPORT_SYMBOL from parisc_ksyms.c
  [PARISC] Make local_t use atomic_long_t
  [PARISC] Update defconfigs
  [PARISC] Add PREEMPT support
  [PARISC] More useful readwrite lock helpers
  [PARISC] Convert HIL drivers to use input_allocate_device
  [PARISC] Fixup CONFIG_EISA a bit
  [PARISC] getsockopt should be ENTRY_COMP
  [PARISC] Remove obsolete CONFIG_DEBUG_IOREMAP
  [PARISC] Temporary FIXME for ioremapping EISA regions
  [PARISC] Enable ioremap functionality unconditionally
  [PARISC] Fix stifb with IOREMAP and a 64-bit kernel
  [PARISC] Add CONFIG_HPPA_IOREMAP to conditionally enable ioremap
  [PARISC] Add STRICT_MM_TYPECHECKS
  [PARISC] Fix IOREMAP with a 64-bit kernel
  [PARISC] Add parisc implementation of flush_kernel_dcache_page()
  ...

18 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland...
Linus Torvalds [Thu, 30 Mar 2006 22:32:38 +0000 (14:32 -0800)]
Merge branch 'for-linus' of git://git./linux/kernel/git/roland/infiniband

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/roland/infiniband:
  IB/mad: RMPP support for additional classes
  IB/mad: include GID/class when matching receives
  IB/mthca: Fix section mismatch problems
  IPoIB: Fix oops with raw sockets
  IB/mthca: Fix check of size in SRQ creation
  IB/srp: Fix unmapping of fake scatterlist

18 years agoMerge branch 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik...
Linus Torvalds [Thu, 30 Mar 2006 22:29:20 +0000 (14:29 -0800)]
Merge branch 'upstream-linus' of /linux/kernel/git/jgarzik/libata-dev

* 'upstream-linus' of master.kernel.org:/pub/scm/linux/kernel/git/jgarzik/libata-dev:
  [PATCH] sata_mv: three bug fixes
  [PATCH] libata: ata_dev_init_params() fixes
  [PATCH] libata: Fix interesting use of "extern" and also some bracketing
  [PATCH] libata: Simplex and other mode filtering logic
  [PATCH] libata - ATA is both ATA and CFA
  [PATCH] libata: Add ->set_mode hook for odd drivers
  [PATCH] libata: BMDMA handling updates
  [PATCH] libata: kill trailing whitespace
  [PATCH] libata: add FIXME above ata_dev_xfermask()
  [PATCH] libata: cosmetic changes in ata_bus_softreset()
  [PATCH] libata: kill E.D.D.

18 years agoMerge branch 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied...
Linus Torvalds [Thu, 30 Mar 2006 22:26:27 +0000 (14:26 -0800)]
Merge branch 'drm-patches' of git://git./linux/kernel/git/airlied/drm-2.6

* 'drm-patches' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6:
  drm: remove drm_{alloc,free}_pages
  drm: sis fix compile warning
  drm: add new radeon PCI ids..
  drm: read breadcrumb in IRQ handler
  drm: fixup i915 breadcrumb read/write
  drm:  remove pointless checks in radeon_state
  drm: fixup improper cast.
  drm: rationalise some pci ids
  drm: Add general-purpose packet for manipulating scratch registers (r300)
  drm: rework radeon memory map (radeon 1.23)
  drm: update r300 register names
  drm: fixup PCI DMA support

18 years ago[libata] ahci: add ATI SB600 PCI IDs
Jeff Garzik [Thu, 30 Mar 2006 22:07:32 +0000 (17:07 -0500)]
[libata] ahci: add ATI SB600 PCI IDs

From: Anatoli Antonovitch <antonovi@ati.com>

Signed-off-by: Felix Kuehling <fkuehlin@ati.com>
Signed-off-by: Jeff Garzik <jeff@garzik.org>
18 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Thu, 30 Mar 2006 20:38:18 +0000 (12:38 -0800)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] ioremap() should prefer WB over UC
  [IA64] Add __mca_table to the DISCARD list in gate.lds
  [IA64] Move __mca_table out of the __init section
  [IA64] simplify some condition checks in iosapic_check_gsi_range
  [IA64] correct some messages and fixes some minor things
  [IA64-SGI] fix for-loop in sn_hwperf_geoid_to_cnode()
  [IA64-SGI] sn_hwperf use of num_online_cpus()
  [IA64] optimize flush_tlb_range on large numa box
  [IA64] lazy_mmu_prot_update needs to be aware of huge pages

18 years ago[PATCH] splice: add support for SPLICE_F_MOVE flag
Jens Axboe [Thu, 30 Mar 2006 13:16:46 +0000 (15:16 +0200)]
[PATCH] splice: add support for SPLICE_F_MOVE flag

This enables the caller to migrate pages from one address space page
cache to another.  In buzz word marketing, you can do zero-copy file
copies!

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PATCH] Introduce sys_splice() system call
Jens Axboe [Thu, 30 Mar 2006 13:15:30 +0000 (15:15 +0200)]
[PATCH] Introduce sys_splice() system call

This adds support for the sys_splice system call. Using a pipe as a
transport, it can connect to files or sockets (latter as output only).

From the splice.c comments:

   "splice": joining two ropes together by interweaving their strands.

   This is the "extended pipe" functionality, where a pipe is used as
   an arbitrary in-memory buffer. Think of a pipe as a small kernel
   buffer that you can use to transfer data from one end to the other.

   The traditional unix read/write is extended with a "splice()" operation
   that transfers data buffers to or from a pipe buffer.

   Named by Larry McVoy, original implementation from Linus, extended by
   Jens to support splicing to files and fixing the initial implementation
   bugs.

Signed-off-by: Jens Axboe <axboe@suse.de>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
18 years ago[PARISC] Fix double free when removing HIL drivers
Kyle McMartin [Thu, 30 Mar 2006 16:47:32 +0000 (11:47 -0500)]
[PARISC] Fix double free when removing HIL drivers

On Thu, Mar 30, 2006 at 08:31:02AM -0500, Dmitry Torokhov wrote:
> Don't do that, its double free. input_unregister_device() normally
> causes release() to be called and free the device. input_free_device
> is only to be called when input_register_device has not been called or
> failed.
>
> Plus you might want to unregister device after closing serio port,
> otherwise your interrupt routine might be referencing already freed
> memory.

Signed-off-by: Kyle McMartin <kyle@parisc-linux.org>
18 years ago[PARISC] Add atomic_sub_and_test
Kyle McMartin [Thu, 30 Mar 2006 00:47:32 +0000 (19:47 -0500)]
[PARISC] Add atomic_sub_and_test

Define atomic_sub_and_test to fix build failures.

Signed-off-by: Kyle McMartin <kyle@parisc-linux.org>
18 years ago[PARISC] Enabled some NLS modules in a500, b180 and c3000 defconfigs
Stuart Brady [Wed, 29 Mar 2006 20:11:25 +0000 (13:11 -0700)]
[PARISC] Enabled some NLS modules in a500, b180 and c3000 defconfigs

With c3000_defconfig and b180_defconfig, FAT couldn't be used
because no NLS modules were built.

Signed-off-by: Stuart Brady <sdb@parisc-linux.org>
Signed-off-by: Kyle McMartin <kyle@parisc-linux.org>
18 years ago[PARISC] Kill duplicated EXPORT_SYMBOL warnings
Kyle McMartin [Wed, 29 Mar 2006 22:21:12 +0000 (15:21 -0700)]
[PARISC] Kill duplicated EXPORT_SYMBOL warnings

Some symbols are exported both in parisc_ksyms.c, and at their
definition site. Nuke the redundant EXPORT_SYMBOL in ksyms to quiet
warnings when vmlinux is linked.

Signed-off-by: Kyle McMartin <kyle@parisc-linux.org>
18 years ago[PARISC] Move ioremap EXPORT_SYMBOL from parisc_ksyms.c
Kyle McMartin [Wed, 29 Mar 2006 22:18:32 +0000 (15:18 -0700)]
[PARISC] Move ioremap EXPORT_SYMBOL from parisc_ksyms.c

Move ioremap/iounmap EXPORT_SYMBOL to ioremap.c where they belong.

Signed-off-by: Kyle McMartin <kyle@parisc-linux.org>