sfrench/cifs-2.6.git
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86
Linus Torvalds [Fri, 1 Feb 2008 21:27:50 +0000 (08:27 +1100)]
Merge git://git./linux/kernel/git/x86/linux-2.6-x86

* git://git.kernel.org/pub/scm/linux/kernel/git/x86/linux-2.6-x86:
  x86: avoid section mismatch involving arch_register_cpu
  x86: fixes for lookup_address args
  x86: fix sparse warnings in cpu/common.c
  x86: make early_console static in early_printk.c
  x86: remove unneeded round_up
  x86: fix section mismatch warning in kernel/pci-calgary
  x86: fix section mismatch warning in acpi/boot.c
  x86: fix section mismatch warnings when referencing notifiers
  x86: silence section mismatch warning in smpboot_64.c
  x86: fix comments in vmlinux_64.lds
  x86_64: make bootmap_start page align v6
  x86_64: add debug name for early_res

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched
Linus Torvalds [Fri, 1 Feb 2008 21:27:00 +0000 (08:27 +1100)]
Merge git://git./linux/kernel/git/mingo/linux-2.6-sched

* git://git.kernel.org/pub/scm/linux/kernel/git/mingo/linux-2.6-sched:
  latencytop: Change Kconfig dependency.
  futex: Add bitset conditional wait/wakeup functionality
  futex: Remove warn on in return fixup path
  x86: replace LOCK_PREFIX in futex.h
  tick-sched: add more debug information
  timekeeping: update xtime_cache when time(zone) changes
  hrtimer: fix hrtimer_init_sleeper() users

14 years agox86: avoid section mismatch involving arch_register_cpu
Alexander van Heukelum [Fri, 1 Feb 2008 16:49:43 +0000 (17:49 +0100)]
x86: avoid section mismatch involving arch_register_cpu

Avoid section mismatch involving arch_register_cpu.

Marking arch_register_cpu as __init and removing the export
for non-hotplug-cpu configurations makes the following warning
go away:

Section mismatch in reference from the function
arch_register_cpu() to the function .devinit.text:register_cpu()
The function  arch_register_cpu() references
the function __devinit register_cpu().
This is often because arch_register_cpu lacks a __devinit
annotation or the annotation of register_cpu is wrong.

The only external user of arch_register_cpu in the tree is
in drivers/acpi/processor_core.c where it is guarded by
ACPI_HOTPLUG_CPU (which depends on HOTPLUG_CPU).

Signed-off-by: Alexander van Heukelum <heukelum@fastmail.fm>
CC: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fixes for lookup_address args
Harvey Harrison [Fri, 1 Feb 2008 16:49:43 +0000 (17:49 +0100)]
x86: fixes for lookup_address args

Signedness mismatches in level argument.

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fix sparse warnings in cpu/common.c
Harvey Harrison [Fri, 1 Feb 2008 16:49:43 +0000 (17:49 +0100)]
x86: fix sparse warnings in cpu/common.c

The casts will always be needed, may as well make them the right
signedness.  The ebx variables can easily be unsigned, may as well.

arch/x86/kernel/cpu/common.c:261:21: warning: incorrect type in argument 2 (different signedness)
arch/x86/kernel/cpu/common.c:261:21:    expected unsigned int *eax
arch/x86/kernel/cpu/common.c:261:21:    got int *<noident>
arch/x86/kernel/cpu/common.c:262:9: warning: incorrect type in argument 3 (different signedness)
arch/x86/kernel/cpu/common.c:262:9:    expected unsigned int *ebx
arch/x86/kernel/cpu/common.c:262:9:    got int *<noident>
arch/x86/kernel/cpu/common.c:263:9: warning: incorrect type in argument 4 (different signedness)
arch/x86/kernel/cpu/common.c:263:9:    expected unsigned int *ecx
arch/x86/kernel/cpu/common.c:263:9:    got int *<noident>
arch/x86/kernel/cpu/common.c:264:9: warning: incorrect type in argument 5 (different signedness)
arch/x86/kernel/cpu/common.c:264:9:    expected unsigned int *edx
arch/x86/kernel/cpu/common.c:264:9:    got int *<noident>
arch/x86/kernel/cpu/common.c:293:30: warning: incorrect type in argument 3 (different signedness)
arch/x86/kernel/cpu/common.c:293:30:    expected unsigned int *ebx
arch/x86/kernel/cpu/common.c:293:30:    got int *<noident>
arch/x86/kernel/cpu/common.c:350:22: warning: incorrect type in argument 2 (different signedness)
arch/x86/kernel/cpu/common.c:350:22:    expected unsigned int *eax
arch/x86/kernel/cpu/common.c:350:22:    got int *<noident>
arch/x86/kernel/cpu/common.c:351:10: warning: incorrect type in argument 3 (different signedness)
arch/x86/kernel/cpu/common.c:351:10:    expected unsigned int *ebx
arch/x86/kernel/cpu/common.c:351:10:    got int *<noident>
arch/x86/kernel/cpu/common.c:352:10: warning: incorrect type in argument 4 (different signedness)
arch/x86/kernel/cpu/common.c:352:10:    expected unsigned int *ecx
arch/x86/kernel/cpu/common.c:352:10:    got int *<noident>
arch/x86/kernel/cpu/common.c:353:10: warning: incorrect type in argument 5 (different signedness)
arch/x86/kernel/cpu/common.c:353:10:    expected unsigned int *edx
arch/x86/kernel/cpu/common.c:353:10:    got int *<noident>
arch/x86/kernel/cpu/common.c:362:30: warning: incorrect type in argument 3 (different signedness)
arch/x86/kernel/cpu/common.c:362:30:    expected unsigned int *ebx
arch/x86/kernel/cpu/common.c:362:30:    got int *<noident>

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: make early_console static in early_printk.c
Harvey Harrison [Fri, 1 Feb 2008 16:49:42 +0000 (17:49 +0100)]
x86: make early_console static in early_printk.c

Not necessary to expose it, also fixes sparse warning.

arch/x86/kernel/early_printk.c:196:16: warning: symbol 'early_console' was not declared. Should it be static?

Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: remove unneeded round_up
Yinghai Lu [Fri, 1 Feb 2008 16:49:42 +0000 (17:49 +0100)]
x86: remove unneeded round_up

Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fix section mismatch warning in kernel/pci-calgary
Sam Ravnborg [Fri, 1 Feb 2008 16:49:42 +0000 (17:49 +0100)]
x86: fix section mismatch warning in kernel/pci-calgary

Fix following warning:
WARNING: arch/x86/kernel/built-in.o(.text+0x1eb41): Section mismatch in reference from the function calgary_handle_quirks() to the function .init.text:calgary_set_split_completion_timeout()

calgary_handle_quirks() are only called at
__init time (in calgary_init_one() via handle_quirks ops).
So annotate this function and the sister function __init.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fix section mismatch warning in acpi/boot.c
Sam Ravnborg [Fri, 1 Feb 2008 16:49:42 +0000 (17:49 +0100)]
x86: fix section mismatch warning in acpi/boot.c

Fix following warning:
WARNING: o-x86_64/arch/x86/kernel/built-in.o(.text+0x13d15): Section mismatch in reference from the function acpi_map_lsapic() to the function .cpuinit.text:mp_register_lapic()

The function acpi_map_lsapic() is exported and thus not annotated.
But the sole user is acpi/processor_core.c in a __cpuinit path.
So create a small wrapper and put back the annotation thus
avoiding the warning.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fix section mismatch warnings when referencing notifiers
Sam Ravnborg [Fri, 1 Feb 2008 16:49:42 +0000 (17:49 +0100)]
x86: fix section mismatch warnings when referencing notifiers

Fix the following warnings:
WARNING: arch/x86/kernel/built-in.o(.exit.text+0xf8): Section mismatch in reference from the function msr_exit() to the variable .cpuinit.data:msr_class_cpu_notifier
WARNING: arch/x86/kernel/built-in.o(.exit.text+0x158): Section mismatch in reference from the function cpuid_exit() to the variable .cpuinit.data:cpuid_class_cpu_notifier
WARNING: arch/x86/kernel/built-in.o(.exit.text+0x171): Section mismatch in reference from the function microcode_exit() to the variable .cpuinit.data:mc_cpu_notifier

In all three cases there were a function annotated __exit
that referenced a variable annotated __cpuinitdata.

The fix was to replace the annotation of the notifier
with __refdata to tell modpost that the reference to
a _cpuinit function in the notifier are OK.
The unregister call that references the notifier
variable will simple delete the function pointer
so there is no problem ignoring the reference.

Note: This looks like another case where __cpuinit
has been used as replacement for proper use
of CONFIG_HOTPLUG_CPU to decide what code are used for
HOTPLUG_CPU.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: silence section mismatch warning in smpboot_64.c
Sam Ravnborg [Fri, 1 Feb 2008 16:49:41 +0000 (17:49 +0100)]
x86: silence section mismatch warning in smpboot_64.c

Silence the following warning:
WARNING: o-x86_64/arch/x86/kernel/built-in.o(.text+0x17cd3): Section mismatch in reference from the function remove_cpu_from_maps() to the variable .cpuinit.data:cpu_initialized

remove_cpu:maps() had a single user: __cpu_disable() so
mark it static and annotate it with __ref to silence the
warning from modpost.

_cpu_disable() has a single user in kernel/cpu.c:
 => take_cpu_down()
    which again has a single user in the following call:
    => __stop_machine_run(take_cpu_down, &tcd_param, cpu);
Here a kthread is created.

So maybe the warning is correct and the right fix is to
remove the __cpuinitdata annotation of cpu_initialized?

Note: The analysis were disturbed by the fact that we had a variable
      with the same name in cpu/common.c - but this is 32 bit only]
Note: Should smpboot_64 use cpu_clear()?

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: fix comments in vmlinux_64.lds
Yinghai Lu [Fri, 1 Feb 2008 16:49:41 +0000 (17:49 +0100)]
x86: fix comments in vmlinux_64.lds

for bzImage, the vmlinux_64.lds still have s32 bit code, and startup_32
should be 0. fix the comment.

Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86_64: make bootmap_start page align v6
Yinghai Lu [Fri, 1 Feb 2008 16:49:41 +0000 (17:49 +0100)]
x86_64: make bootmap_start page align v6

boot oopses when a system has 64 or 128 GB of RAM installed:

Calling initcall 0xffffffff80bc33b6: sctp_init+0x0/0x711()
BUG: unable to handle kernel NULL pointer dereference at 000000000000005f
IP: [<ffffffff802bfe55>] proc_register+0xe7/0x10f
PGD 0
Oops: 0000 [1] SMP
CPU 0
Modules linked in:
Pid: 1, comm: swapper Not tainted 2.6.24-smp-g5a514e21-dirty #6
RIP: 0010:[<ffffffff802bfe55>]  [<ffffffff802bfe55>] proc_register+0xe7/0x10f
RSP: 0000:ffff810824c57e60  EFLAGS: 00010246
RAX: 000000000000d7d7 RBX: ffff811024c5fa80 RCX: ffff810824c57e08
RDX: 0000000000000000 RSI: 0000000000000195 RDI: ffffffff80cc2460
RBP: ffffffffffffffff R08: 0000000000000000 R09: ffff811024c5fa80
R10: 0000000000000000 R11: 0000000000000002 R12: ffff810824c57e6c
R13: 0000000000000000 R14: ffff810824c57ee0 R15: 00000006abd25bee
FS:  0000000000000000(0000) GS:ffffffff80b4d000(0000) knlGS:0000000000000000
CS:  0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 000000000000005f CR3: 0000000000201000 CR4: 00000000000006e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process swapper (pid: 1, threadinfo ffff810824c56000, task ffff812024c52000)
Stack:  ffffffff80a57348 0000019500000000 ffff811024c5fa80 0000000000000000
 00000000ffffff97 ffffffff802bfef0 0000000000000000 ffffffffffffffff
 0000000000000000 ffffffff80bc3b4b ffff810824c57ee0 ffffffff80bc34a5
Call Trace:
 [<ffffffff802bfef0>] ? create_proc_entry+0x73/0x8a
 [<ffffffff80bc3b4b>] ? sctp_snmp_proc_init+0x1c/0x34
 [<ffffffff80bc34a5>] ? sctp_init+0xef/0x711
 [<ffffffff80b976e3>] ? kernel_init+0x175/0x2e1
 [<ffffffff8020ccf8>] ? child_rip+0xa/0x12
 [<ffffffff80b9756e>] ? kernel_init+0x0/0x2e1
 [<ffffffff8020ccee>] ? child_rip+0x0/0x12

Code: 1e 48 83 7b 38 00 75 08 48 c7 43 38 f0 e8 82 80 48 83 7b 30 00 75 08 48 c7 43 30 d0 e9 82 80 48 c7 c7 60 24 cc 80 e8 bd 5a 54 00 <48> 8b 45 60 48 89 6b 58 48 89 5d 60 48 89 43 50 fe 05 f5 25 a0
RIP  [<ffffffff802bfe55>] proc_register+0xe7/0x10f
 RSP <ffff810824c57e60>
CR2: 000000000000005f
---[ end trace 02c2d78def82877a ]---
Kernel panic - not syncing: Attempted to kill init!

it turns out some variables near end of bss are corrupted already.

in System.map we have
ffffffff80d40420 b rsi_table
ffffffff80d40620 B krb5_seq_lock
ffffffff80d40628 b i.20437
ffffffff80d40630 b xprt_rdma_inline_write_padding
ffffffff80d40638 b sunrpc_table_header
ffffffff80d40640 b zero
ffffffff80d40644 b min_memreg
ffffffff80d40648 b rpcrdma_tk_lock_g
ffffffff80d40650 B sctp_assocs_id_lock
ffffffff80d40658 B proc_net_sctp
ffffffff80d40660 B sctp_assocs_id
ffffffff80d40680 B sysctl_sctp_mem
ffffffff80d40690 B sysctl_sctp_rmem
ffffffff80d406a0 B sysctl_sctp_wmem
ffffffff80d406b0 b sctp_ctl_socket
ffffffff80d406b8 b sctp_pf_inet6_specific
ffffffff80d406c0 b sctp_pf_inet_specific
ffffffff80d406c8 b sctp_af_v4_specific
ffffffff80d406d0 b sctp_af_v6_specific
ffffffff80d406d8 b sctp_rand.33270
ffffffff80d406dc b sctp_memory_pressure
ffffffff80d406e0 b sctp_sockets_allocated
ffffffff80d406e4 b sctp_memory_allocated
ffffffff80d406e8 b sctp_sysctl_header
ffffffff80d406f0 b zero
ffffffff80d406f4 A __bss_stop
ffffffff80d406f4 A _end

and setup_node_bootmem() will use that page 0xd40000 for bootmap
Bootmem setup node 0 0000000000000000-0000000828000000
  NODE_DATA [000000000008a485 - 0000000000091484]
  bootmap [0000000000d406f4 -  0000000000e456f3] pages 105
Bootmem setup node 1 0000000828000000-0000001028000000
  NODE_DATA [0000000828000000 - 0000000828006fff]
  bootmap [0000000828007000 -  0000000828106fff] pages 100
Bootmem setup node 2 0000001028000000-0000001828000000
  NODE_DATA [0000001028000000 - 0000001028006fff]
  bootmap [0000001028007000 -  0000001028106fff] pages 100
Bootmem setup node 3 0000001828000000-0000002028000000
  NODE_DATA [0000001828000000 - 0000001828006fff]
  bootmap [0000001828007000 -  0000001828106fff] pages 100

setup_node_bootmem() makes NODE_DATA cacheline aligned,
and bootmap is page-aligned.

the patch updates find_e820_area() to make sure we can meet
the alignment constraints.

Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86_64: add debug name for early_res
Yinghai Lu [Fri, 1 Feb 2008 16:49:41 +0000 (17:49 +0100)]
x86_64: add debug name for early_res

helps debugging problems in this rather murky area of code.

Signed-off-by: Yinghai Lu <yinghai.lu@sun.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agolatencytop: Change Kconfig dependency.
Heiko Carstens [Fri, 1 Feb 2008 16:45:14 +0000 (17:45 +0100)]
latencytop: Change Kconfig dependency.

Change latencytop Kconfig entry so it doesn't list the archictectures
that support it. Instead introduce HAVE_LATENCY_SUPPORT which any
architecture can set. Should reduce patch conflicts.

Cc: Arjan van de Ven <arjan@linux.intel.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Holger Wolf <wolf@linux.vnet.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agofutex: Add bitset conditional wait/wakeup functionality
Thomas Gleixner [Fri, 1 Feb 2008 16:45:14 +0000 (17:45 +0100)]
futex: Add bitset conditional wait/wakeup functionality

To allow the implementation of optimized rw-locks in user space, glibc
needs a possibility to select waiters for wakeup depending on a bitset
mask.

This requires two new futex OPs: FUTEX_WAIT_BITS and FUTEX_WAKE_BITS
These OPs are basically the same as FUTEX_WAIT and FUTEX_WAKE plus an
additional argument - a bitset. Further the FUTEX_WAIT_BITS OP is
expecting an absolute timeout value instead of the relative one, which
is used for the FUTEX_WAIT OP.

FUTEX_WAIT_BITS calls into the kernel with a bitset. The bitset is
stored in the futex_q structure, which is used to enqueue the waiter
into the hashed futex waitqueue.

FUTEX_WAKE_BITS also calls into the kernel with a bitset. The wakeup
function logically ANDs the bitset with the bitset stored in each
waiters futex_q structure. If the result is zero (i.e. none of the set
bits in the bitsets is matching), then the waiter is not woken up. If
the result is not zero (i.e. one of the set bits in the bitsets is
matching), then the waiter is woken.

The bitset provided by the caller must be non zero. In case the
provided bitset is zero the kernel returns EINVAL.

Internaly the new OPs are only extensions to the existing FUTEX_WAIT
and FUTEX_WAKE functions. The existing OPs hand a bitset with all bits
set into the futex_wait() and futex_wake() functions.

Signed-off-by: Thomas Gleixner <tgxl@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agofutex: Remove warn on in return fixup path
Thomas Gleixner [Fri, 1 Feb 2008 16:45:14 +0000 (17:45 +0100)]
futex: Remove warn on in return fixup path

The WARN_ON() in the fixup return path of futex_lock_pi() can
trigger with false positives.

The following scenario happens:
t1 holds the futex and t2 and t3 are blocked on the kernel side rt_mutex.
t1 releases the futex (and the rt_mutex) and assigned t2 to be the next
owner of the futex.

t2 is interrupted and returns w/o acquiring the rt_mutex, before t1 can
release the rtmutex.

t1 releases the rtmutex and t3 becomes the pending owner of the rtmutex.

t2 notices that it is the designated owner (user space variable) and
fails to acquire the rt_mutex via trylock, because it is not allowed to
steal the rt_mutex from t3. Now it looks at the rt_mutex pending owner (t3)
and assigns the futex and the pi_state to it.

During the fixup t4 steals the rtmutex from t3.

t2 returns from the fixup and the owner of the rt_mutex has changed from
t3 to t4.

There is no need to do another round of fixups from t2. The important
part (t2 is not returning as the user space visible owner) is
done. The further fixups are done, before either t3 or t4 return to
user space.

For the user space it is not relevant which task (t3 or t4) is the real
owner, as long as those are both in the kernel, which is guaranteed by
the serialization of the hash bucket lock. Both tasks (which ever returns
first to userspace - t4 because it locked the rt_mutex or t3 due to a signal)
are going through the lock_futex_pi() return path where the ownership is
fixed before the return to user space.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agox86: replace LOCK_PREFIX in futex.h
Thomas Gleixner [Fri, 1 Feb 2008 16:45:14 +0000 (17:45 +0100)]
x86: replace LOCK_PREFIX in futex.h

The exception fixup for the futex macros __futex_atomic_op1/2 and
futex_atomic_cmpxchg_inatomic() is missing an entry when the lock
prefix is replaced by a NOP via SMP alternatives.

Chuck Ebert tracked this down from the information provided in:
https://bugzilla.redhat.com/show_bug.cgi?id=429412

A possible solution would be to add another fixup after the
LOCK_PREFIX, so both the LOCK and NOP case have their own entry in the
exception table, but it's not really worth the trouble.

Simply replace LOCK_PREFIX with lock and keep those untouched by SMP
alternatives.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agotick-sched: add more debug information
Thomas Gleixner [Fri, 1 Feb 2008 16:45:14 +0000 (17:45 +0100)]
tick-sched: add more debug information

To allow better diagnosis of tick-sched related, especially NOHZ
related problems, we need to know when the last wakeup via an irq
happened and when the CPU left the idle state.

Add two fields (idle_waketime, idle_exittime) to the tick_sched
structure and add them to the timer_list output.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agotimekeeping: update xtime_cache when time(zone) changes
Thomas Gleixner [Fri, 1 Feb 2008 16:45:13 +0000 (17:45 +0100)]
timekeeping: update xtime_cache when time(zone) changes

xtime_cache needs to be updated whenever xtime and or wall_to_monotic
are changed. Otherwise users of xtime_cache might see a stale (and in
the case of timezone changes utterly wrong) value until the next
update happens.

Fixup the obvious places, which miss this update.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: John Stultz <johnstul@us.ibm.com>
Tested-by: Dhaval Giani <dhaval@linux.vnet.ibm.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years agohrtimer: fix hrtimer_init_sleeper() users
Peter Zijlstra [Fri, 1 Feb 2008 16:45:13 +0000 (17:45 +0100)]
hrtimer: fix hrtimer_init_sleeper() users

this patch:

 commit 37bb6cb4097e29ffee970065b74499cbf10603a3
 Author: Peter Zijlstra <a.p.zijlstra@chello.nl>
 Date:   Fri Jan 25 21:08:32 2008 +0100

     hrtimer: unlock hrtimer_wakeup

Broke hrtimer_init_sleeper() users. It forgot to fix up the futex
caller of this function to detect the failed queueing and messed up
the do_nanosleep() caller in that it could leak a TASK_INTERRUPTIBLE
state.

Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
14 years ago[MIPS] Split the micro-assembler from tlbex.c.
Thiemo Seufer [Mon, 28 Jan 2008 20:05:38 +0000 (20:05 +0000)]
[MIPS] Split the micro-assembler from tlbex.c.

This patch moves the micro-assembler in a separate implementation, as
it is useful for further run-time optimizations. The only change in
behaviour is cutting down printk noise at kernel startup time.

Checkpatch complains about macro parameters which aren't protected by
parentheses. I believe this is a flaw in checkpatch, the paste operator
used in those macros won't work with parenthesised parameters.

Signed-off-by: Thiemo Seufer <ths@networkno.de>
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years ago[MIPS] Fix build after b0rked changes to <linux/init.h>.
Ralf Baechle [Wed, 30 Jan 2008 12:14:59 +0000 (12:14 +0000)]
[MIPS] Fix build after b0rked changes to <linux/init.h>.

Commit 312b1485fb509c9bc32eda28ad29537896658cb8 made __INIT_REFOK expand
into .section .section ".ref.text", "ax".  Since the assembler doesn't
tolerate stuttering in the source that broke all MIPS builds.

Since with this change Sam downgraded __INIT_REFOK the best fix is to
get replace it by the modern days operator.  With MIPS the only user
of __INIT_REFOK and __INITDATA_REFOK (which was equally broken) being
unused anyway these can be deleted but that's subject of a separate
commit.

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
14 years agoMerge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block
Linus Torvalds [Fri, 1 Feb 2008 10:48:45 +0000 (21:48 +1100)]
Merge branch 'for-linus' of git://git.kernel.dk/linux-2.6-block

* 'for-linus' of git://git.kernel.dk/linux-2.6-block:
  block: kill swap_io_context()
  as-iosched: fix inconsistent ioc->lock context
  ide-cd: fix leftover data BUG
  block: make elevator lib checkpatch compliant
  cfq-iosched: make checkpatch compliant
  block: make core bits checkpatch compliant
  block: new end request handling interface should take unsigned byte counts
  unexport add_disk_randomness
  block/sunvdc.c:print_version() must be __devinit
  splice: always updated atime in direct splice

14 years agoblock: kill swap_io_context()
Jens Axboe [Fri, 1 Feb 2008 10:34:49 +0000 (11:34 +0100)]
block: kill swap_io_context()

It blindly copies everything in the io_context, including the lock.
That doesn't work so well for either lock ordering or lockdep.

There seems zero point in swapping io contexts on a request to request
merge, so the best point of action is to just remove it.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agodocbook: rapidio: fix fatal filename error
Randy Dunlap [Fri, 1 Feb 2008 06:09:08 +0000 (22:09 -0800)]
docbook: rapidio: fix fatal filename error

Fix docbook fatal error (files were renamed):
docproc: linux-2.6.24-git9/arch/ppc/kernel/rio.c: No such file or directory

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agofix directory entry in arch-x86-Makefile
Steven Rostedt [Fri, 1 Feb 2008 05:13:58 +0000 (00:13 -0500)]
fix directory entry in arch-x86-Makefile

Doing a make randconfig I came across this error in the Makefile.

This patch makes a directory out of arch/x86/mach-default for
CONFIG_X86_RDC321X

Signed-off-by: Steven Rostedt <srostedt@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6
Linus Torvalds [Fri, 1 Feb 2008 10:06:56 +0000 (21:06 +1100)]
Merge git://git./linux/kernel/git/davem/sparc-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/sparc-2.6:
  [SPARC64]: Fix inconsistent .section usage in lib/
  [SPARC/SPARC64]: Fix usage of .section .sched.text in assembler code.

14 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Fri, 1 Feb 2008 10:06:29 +0000 (21:06 +1100)]
Merge git://git./linux/kernel/git/davem/net-2.6

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6: (173 commits)
  [NETNS]: Lookup in FIB semantic hashes taking into account the namespace.
  [NETNS]: Add a namespace mark to fib_info.
  [IPV4]: fib_sync_down rework.
  [NETNS]: Process interface address manipulation routines in the namespace.
  [IPV4]: Small style cleanup of the error path in rtm_to_ifaddr.
  [IPV4]: Fix memory leak on error path during FIB initialization.
  [NETFILTER]: Ipv6-related xt_hashlimit compilation fix.
  [NET_SCHED]: Add flow classifier
  [NET_SCHED]: sch_sfq: make internal queues visible as classes
  [NET_SCHED]: sch_sfq: add support for external classifiers
  [NET_SCHED]: Constify struct tcf_ext_map
  [BLUETOOTH]: Fix bugs in previous conn add/del workqueue changes.
  [TCP]: Unexport sysctl_tcp_tso_win_divisor
  [IPV4]: Make struct ipv4_devconf static.
  [TR] net/802/tr.c: sysctl_tr_rif_timeout static
  [XFRM]: Fix statistics.
  [XFRM]: Remove unused exports.
  [PKT_SCHED] sch_teql.c: Duplicate IFF_BROADCAST in FMASK, remove 2nd.
  [BNX2]: Fix ASYM PAUSE advertisement for remote PHY.
  [IPV4] route cache: Introduce rt_genid for smooth cache invalidation
  ...

14 years agopasemi: Fix thinko in dma_direct_ops setup
Olof Johansson [Thu, 31 Jan 2008 23:50:02 +0000 (17:50 -0600)]
pasemi: Fix thinko in dma_direct_ops setup

[POWERPC] pasemi: Fix thinko in dma_direct_ops setup

The first patch will just fall through and still set dma_data to a bad
value, make it return directly instead.

Signed-off-by: Olof Johansson <olof@lixom.net>
Acked-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: remove dead timer int pending code
Greg Ungerer [Fri, 1 Feb 2008 07:40:52 +0000 (17:40 +1000)]
m68knommu: remove dead timer int pending code

Remove all the dead timer interrupt checking functions for the ColdFire
CPU "timers" hardware that are not used after switching to GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: remove empty timer offset function for 68360
Greg Ungerer [Fri, 1 Feb 2008 07:40:49 +0000 (17:40 +1000)]
m68knommu: remove empty timer offset function for 68360

Switching to GENERIC_TIME means we no longer need the empty timer offset
function for the 68360 CPU.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: remove local gettimeofday code
Greg Ungerer [Fri, 1 Feb 2008 07:40:39 +0000 (17:40 +1000)]
m68knommu: remove local gettimeofday code

Remove unused local gettimeofday functions, now that we are using
GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: set config to use GENERIC_TIME
Greg Ungerer [Fri, 1 Feb 2008 07:40:30 +0000 (17:40 +1000)]
m68knommu: set config to use GENERIC_TIME

Switch m68knommu arch to using GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: switch ColdFire TIMER to use GENERIC_TIME
Greg Ungerer [Fri, 1 Feb 2008 07:40:26 +0000 (17:40 +1000)]
m68knommu: switch ColdFire TIMER to use GENERIC_TIME

Switch the ColdFire TIMER timer to use GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: switch ColdFire PIT timer to use GENERIC_TIME
Greg Ungerer [Fri, 1 Feb 2008 07:40:21 +0000 (17:40 +1000)]
m68knommu: switch ColdFire PIT timer to use GENERIC_TIME

Switch the ColdFire PIT timer to using GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: switch 68328 timer to use GENERIC_TIME
Greg Ungerer [Fri, 1 Feb 2008 07:40:17 +0000 (17:40 +1000)]
m68knommu: switch 68328 timer to use GENERIC_TIME

Switch the 68328 CPU timer code to using GENERIC_TIME.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: remove unused CONFIG_DISKtel symbol
Greg Ungerer [Fri, 1 Feb 2008 07:38:26 +0000 (17:38 +1000)]
m68knommu: remove unused CONFIG_DISKtel symbol

Remove unused CONFIG_DISKtel define.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: fix 528x ColdFire cache settings
Greg Ungerer [Fri, 1 Feb 2008 07:38:24 +0000 (17:38 +1000)]
m68knommu: fix 528x ColdFire cache settings

Fix problems with the 528x ColdFire CPU cache setup.
Do not cache the flash region (if present), and make the runtime
settings consistent with the init setting.

Problems pointed out by Bernd Buttner <b.buettner@mkc-gmbh.de>

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: use container_of in mcf.c
Greg Ungerer [Fri, 1 Feb 2008 07:38:34 +0000 (17:38 +1000)]
m68knommu: use container_of in mcf.c

Use container_of() instead of assuming local uart struct is first member
of uart_port struct.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: remove dead code from m68knommu timer code
Greg Ungerer [Fri, 1 Feb 2008 07:38:20 +0000 (17:38 +1000)]
m68knommu: remove dead code from m68knommu timer code

Remove unused LED heartbeat code from m68knommu timer code.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: use IRQF_DISABLE in m68328serial.c
Greg Ungerer [Fri, 1 Feb 2008 07:38:32 +0000 (17:38 +1000)]
m68knommu: use IRQF_DISABLE in m68328serial.c

Use IRQF_DISABLED instead of obsolete IRQ_FLG_STD for request_irq() flags
in m68328serial.c driver.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: create common DMA table for ColdFire CPUs
Greg Ungerer [Fri, 1 Feb 2008 07:38:03 +0000 (17:38 +1000)]
m68knommu: create common DMA table for ColdFire CPUs

Move the ColdFire DMA address table into its own file, and out
of each of the different CPU config files. No need to have a copy
of it in each of the config setup files.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: modify Makefiles to support common coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:56 +0000 (17:37 +1000)]
m68knommu: modify Makefiles to support common coldfire directory

Modify Makefiles to support separate coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: move ColdFire vectors.c to its own coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:46 +0000 (17:37 +1000)]
m68knommu: move ColdFire vectors.c to its own coldfire directory

Move common ColdFire CPU vectors.c to common coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: move ColdFire timers.c to its own coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:39 +0000 (17:37 +1000)]
m68knommu: move ColdFire timers.c to its own coldfire directory

Move common ColdFire CPU timers.c to common coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: move ColdFire pit.c to its own coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:36 +0000 (17:37 +1000)]
m68knommu: move ColdFire pit.c to its own coldfire directory

Move common ColdFire CPU pit.c to common coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: move ColdFire head.S to its own coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:28 +0000 (17:37 +1000)]
m68knommu: move ColdFire head.S to its own coldfire directory

Move common ColdFire CPU head.S to common coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: move ColdFire entry.S to its own coldfire directory
Greg Ungerer [Fri, 1 Feb 2008 07:37:18 +0000 (17:37 +1000)]
m68knommu: move ColdFire entry.S to its own coldfire directory

Move common ColdFire CPU entry.S to common coldfire platform directory.

Currently the common ColdFire CPU family code sits in the
arch/m68knommu/platform/5307 directory. This is confusing, the files
containing this common code are in no way specific to the 5307 ColdFire.
Create an arch/m68knommu/platform/coldfire directory to contain this
common code. Other m68knommu CPU varients do not need use this code
though, so it doesn't make sense to move it to arch/m68knommu/kernel.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5407 ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:35:03 +0000 (17:35 +1000)]
m68knommu: platform setup for 5407 ColdFire parts

Switch to platform style configuration for 5407 ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 532x ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:58 +0000 (17:34 +1000)]
m68knommu: platform setup for 532x ColdFire parts

Switch to platform style configuration for 532x ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 527x ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:40 +0000 (17:34 +1000)]
m68knommu: platform setup for 527x ColdFire parts

Switch to platform style configuration for 527x ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5307 ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:55 +0000 (17:34 +1000)]
m68knommu: platform setup for 5307 ColdFire parts

Switch to platform style configuration for 5307 ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 528x ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:50 +0000 (17:34 +1000)]
m68knommu: platform setup for 528x ColdFire parts

Switch to platform style configuration for 528x ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5249 ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:30 +0000 (17:34 +1000)]
m68knommu: platform setup for 5249 ColdFire parts

Switch to platform style configuration for 5249 ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5272 ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:32 +0000 (17:34 +1000)]
m68knommu: platform setup for 5272 ColdFire parts

Switch to platform style configuration for 5272 ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 520x ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:10 +0000 (17:34 +1000)]
m68knommu: platform setup for 520x ColdFire parts

Switch to platform style configuration for 520x ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 523x ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:15 +0000 (17:34 +1000)]
m68knommu: platform setup for 523x ColdFire parts

Switch to platform style configuration for 523x ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5206e ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:08 +0000 (17:34 +1000)]
m68knommu: platform setup for 5206e ColdFire parts

Switch to platform style configuration for 5206e ColdFire parts.
Initial support is for the UARTs. DMA support is moved to common code.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: platform setup for 5206 ColdFire parts
Greg Ungerer [Fri, 1 Feb 2008 07:34:04 +0000 (17:34 +1000)]
m68knommu: platform setup for 5206 ColdFire parts

Switch to platform style configuration for 5206 ColdFire parts.
Initial support is for the UARTs. DMA support moved to common code
for all ColdFire parts.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: configure support for new ColdFire serial driver
Greg Ungerer [Fri, 1 Feb 2008 07:32:00 +0000 (17:32 +1000)]
m68knommu: configure support for new ColdFire serial driver

Add configure support for new ColdFire serial driver.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agom68knommu: build support for new ColdFire serial driver
Greg Ungerer [Fri, 1 Feb 2008 07:31:58 +0000 (17:31 +1000)]
m68knommu: build support for new ColdFire serial driver

Add build support for new ColdFire serial driver.

Signed-off-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoFix dl2k constants
Al Viro [Fri, 1 Feb 2008 07:05:44 +0000 (07:05 +0000)]
Fix dl2k constants

The MSSR constants didn't match the reality - bitfield declarations used
to be correct (1000BT_FD - bit 11, 1000BT_HD - bit 10), but enum had
them the other way round.  Went unnoticed until the switch from the
bitfields use to the explicit arithmetics and I hadn't caught that one
when verifying correctness of change...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
14 years agoas-iosched: fix inconsistent ioc->lock context
Jens Axboe [Fri, 1 Feb 2008 08:44:28 +0000 (09:44 +0100)]
as-iosched: fix inconsistent ioc->lock context

Since it's acquired from irq context, all locking must be of the
irq safe variant. Most are already inside the queue lock (which
already disables interrupts), but the io scheduler rmmod path
always has irqs enabled and the put_io_context() path may legally
be called with irqs enabled (even if it isn't usually). So fixup
those two.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agoide-cd: fix leftover data BUG
Jens Axboe [Thu, 31 Jan 2008 12:57:51 +0000 (13:57 +0100)]
ide-cd: fix leftover data BUG

It's perfectly legal to have data leftovers when a drive signals
completion, it happens all the time when a user issues a command
with a transfer count that is too large.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agoblock: make elevator lib checkpatch compliant
Jens Axboe [Thu, 31 Jan 2008 23:37:27 +0000 (00:37 +0100)]
block: make elevator lib checkpatch compliant

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agocfq-iosched: make checkpatch compliant
Jens Axboe [Thu, 31 Jan 2008 12:08:54 +0000 (13:08 +0100)]
cfq-iosched: make checkpatch compliant

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agoblock: make core bits checkpatch compliant
Jens Axboe [Thu, 31 Jan 2008 12:03:55 +0000 (13:03 +0100)]
block: make core bits checkpatch compliant

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agoblock: new end request handling interface should take unsigned byte counts
Jens Axboe [Thu, 31 Jan 2008 11:36:19 +0000 (12:36 +0100)]
block: new end request handling interface should take unsigned byte counts

No point in passing signed integers as the byte count, they can never
be negative.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agounexport add_disk_randomness
Adrian Bunk [Wed, 30 Jan 2008 20:17:52 +0000 (21:17 +0100)]
unexport add_disk_randomness

This patch removes the no longer used EXPORT_SYMBOL(add_disk_randomness).

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Acked-by: Matt Mackall <mpm@selenic.com>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agoblock/sunvdc.c:print_version() must be __devinit
Adrian Bunk [Wed, 30 Jan 2008 20:17:23 +0000 (21:17 +0100)]
block/sunvdc.c:print_version() must be __devinit

This patch fixes the following section mismatches:

<--  snip  -->

...
WARNING: drivers/block/sunvdc.o(.text+0xf0): Section mismatch in reference from the function print_version() to the variable .devinit.data:version
WARNING: drivers/block/sunvdc.o(.text+0xf8): Section mismatch in reference from the function print_version() to the variable .devinit.data:version
...

<--  snip  -->

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years agosplice: always updated atime in direct splice
Jens Axboe [Wed, 30 Jan 2008 11:24:48 +0000 (12:24 +0100)]
splice: always updated atime in direct splice

Andre Majorel <aym-xunil@teaser.fr> points out that if we only updated
the atime when we transfer some data, we deviate from the standard
of always updating the atime. So change splice to always call
file_accessed() even if splice_direct_to_actor() didn't transfer
any data.

Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
14 years ago[SPARC64]: Fix inconsistent .section usage in lib/
Sam Ravnborg [Fri, 1 Feb 2008 00:48:49 +0000 (16:48 -0800)]
[SPARC64]: Fix inconsistent .section usage in lib/

A few places missed the "a" specifier for the __ex_table section. Add
these so we avoid generation an additional section at link time.

Latest modpost would otherwise complain like this:

WARNING: vmlinux.o (__ex_table.2): section name inconsistency.
(.[number]+) following section name.
Did you forget to use "ax"/"aw" in a .S file?
Note that for example <linux/init.h> contains
section definitions for use in .S files.
WARNING: vmlinux.o (__ex_table.4): section name inconsistency.
(.[number]+) following section name.
Did you forget to use "ax"/"aw" in a .S file?
Note that for example <linux/init.h> contains
section definitions for use in .S files.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[SPARC/SPARC64]: Fix usage of .section .sched.text in assembler code.
Sam Ravnborg [Tue, 29 Jan 2008 05:06:23 +0000 (21:06 -0800)]
[SPARC/SPARC64]: Fix usage of .section .sched.text in assembler code.

ld will generate an unique named section when assembler do not use
"ax" but gcc does. Add the missing annotation.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NETNS]: Lookup in FIB semantic hashes taking into account the namespace.
Denis V. Lunev [Fri, 1 Feb 2008 02:50:07 +0000 (18:50 -0800)]
[NETNS]: Lookup in FIB semantic hashes taking into account the namespace.

The namespace is not available in the fib_sync_down_addr, add it as a
parameter.

Looking up a device by the pointer to it is OK. Looking up using a
result from fib_trie/fib_hash table lookup is also safe. No need to
fix that at all.  So, just fix lookup by address and insertion to the
hash table path.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NETNS]: Add a namespace mark to fib_info.
Denis V. Lunev [Fri, 1 Feb 2008 02:49:32 +0000 (18:49 -0800)]
[NETNS]: Add a namespace mark to fib_info.

This is required to make fib_info lookups namespace aware. In the
other case initial namespace devices are marked as dead in the local
routing table during other namespace stop.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4]: fib_sync_down rework.
Denis V. Lunev [Fri, 1 Feb 2008 02:48:47 +0000 (18:48 -0800)]
[IPV4]: fib_sync_down rework.

fib_sync_down can be called with an address and with a device. In
reality it is called either with address OR with a device. The
codepath inside is completely different, so lets separate it into two
calls for these two cases.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NETNS]: Process interface address manipulation routines in the namespace.
Denis V. Lunev [Fri, 1 Feb 2008 02:47:40 +0000 (18:47 -0800)]
[NETNS]: Process interface address manipulation routines in the namespace.

The namespace is available when required except rtm_to_ifaddr. Add
namespace argument to it.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4]: Small style cleanup of the error path in rtm_to_ifaddr.
Denis V. Lunev [Fri, 1 Feb 2008 02:47:00 +0000 (18:47 -0800)]
[IPV4]: Small style cleanup of the error path in rtm_to_ifaddr.

Remove error code assignment inside brackets on failure. The code
looks better if the error is assigned before condition check. Also,
the compiler treats this better.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4]: Fix memory leak on error path during FIB initialization.
Denis V. Lunev [Fri, 1 Feb 2008 02:44:53 +0000 (18:44 -0800)]
[IPV4]: Fix memory leak on error path during FIB initialization.

net->ipv4.fib_table_hash is not freed when fib4_rules_init failed.

Signed-off-by: Denis V. Lunev <den@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NETFILTER]: Ipv6-related xt_hashlimit compilation fix.
Pavel Emelyanov [Fri, 1 Feb 2008 02:42:26 +0000 (18:42 -0800)]
[NETFILTER]: Ipv6-related xt_hashlimit compilation fix.

The hashlimit_ipv6_mask() is called from under IP6_NF_IPTABLES config
option, but is not under it by itself.

gcc warns us about it :) :
net/netfilter/xt_hashlimit.c:473: warning: "hashlimit_ipv6_mask" defined but not used

Signed-off-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NET_SCHED]: Add flow classifier
Patrick McHardy [Fri, 1 Feb 2008 02:37:42 +0000 (18:37 -0800)]
[NET_SCHED]: Add flow classifier

Add new "flow" classifier, which is meant to extend the SFQ hashing
capabilities without hard-coding new hash functions and also allows
deterministic mappings of keys to classes, replacing some out of tree
iptables patches like IPCLASSIFY (maps IPs to classes), IPMARK (maps
IPs to marks, with fw filters to classes), ...

Some examples:

- Classic SFQ hash:

  tc filter add ... flow hash \
   keys src,dst,proto,proto-src,proto-dst divisor 1024

- Classic SFQ hash, but using information from conntrack to work properly in
  combination with NAT:

  tc filter add ... flow hash \
   keys nfct-src,nfct-dst,proto,nfct-proto-src,nfct-proto-dst divisor 1024

- Map destination IPs of 192.168.0.0/24 to classids 1-257:

  tc filter add ... flow map \
   key dst addend -192.168.0.0 divisor 256

- alternatively:

  tc filter add ... flow map \
   key dst and 0xff

- similar, but reverse ordered:

  tc filter add ... flow map \
   key dst and 0xff xor 0xff

Perturbation is currently not supported because we can't reliable kill the
timer on destruction.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NET_SCHED]: sch_sfq: make internal queues visible as classes
Patrick McHardy [Fri, 1 Feb 2008 02:37:16 +0000 (18:37 -0800)]
[NET_SCHED]: sch_sfq: make internal queues visible as classes

Add support for dumping statistics and make internal queues visible as
classes.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NET_SCHED]: sch_sfq: add support for external classifiers
Patrick McHardy [Fri, 1 Feb 2008 02:36:52 +0000 (18:36 -0800)]
[NET_SCHED]: sch_sfq: add support for external classifiers

Add support for external classifiers to allow using different flow
hash functions similar to ESFQ. When no classifier is attached the
built-in hash is used as before.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NET_SCHED]: Constify struct tcf_ext_map
Patrick McHardy [Fri, 1 Feb 2008 02:36:18 +0000 (18:36 -0800)]
[NET_SCHED]: Constify struct tcf_ext_map

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[BLUETOOTH]: Fix bugs in previous conn add/del workqueue changes.
Dave Young [Fri, 1 Feb 2008 02:33:10 +0000 (18:33 -0800)]
[BLUETOOTH]: Fix bugs in previous conn add/del workqueue changes.

Jens Axboe noticed that we were queueing &conn->work on both btaddconn
and keventd_wq.

Signed-off-by: Dave Young <hidave.darkstar@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[TCP]: Unexport sysctl_tcp_tso_win_divisor
Adrian Bunk [Fri, 1 Feb 2008 01:18:50 +0000 (17:18 -0800)]
[TCP]: Unexport sysctl_tcp_tso_win_divisor

This patch removes the no longer used
EXPORT_SYMBOL(sysctl_tcp_tso_win_divisor).

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4]: Make struct ipv4_devconf static.
Adrian Bunk [Fri, 1 Feb 2008 01:17:31 +0000 (17:17 -0800)]
[IPV4]: Make struct ipv4_devconf static.

struct ipv4_devconf can now become static.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Acked-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[TR] net/802/tr.c: sysctl_tr_rif_timeout static
Adrian Bunk [Fri, 1 Feb 2008 01:16:32 +0000 (17:16 -0800)]
[TR] net/802/tr.c: sysctl_tr_rif_timeout static

sysctl_tr_rif_timeout can now become static.

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Acked-by: Pavel Emelyanov <xemul@openvz.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[XFRM]: Fix statistics.
Masahide NAKAMURA [Fri, 1 Feb 2008 01:14:58 +0000 (17:14 -0800)]
[XFRM]: Fix statistics.

o Outbound sequence number overflow error status
  is counted as XfrmOutStateSeqError.
o Additionaly, it changes inbound sequence number replay
  error name from XfrmInSeqOutOfWindow to XfrmInStateSeqError
  to apply name scheme above.
o Inbound IPv4 UDP encapsuling type mismatch error is wrongly
  mapped to XfrmInStateInvalid then this patch fiex the error
  to XfrmInStateMismatch.

Signed-off-by: Masahide NAKAMURA <nakam@linux-ipv6.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[XFRM]: Remove unused exports.
Adrian Bunk [Fri, 1 Feb 2008 01:10:30 +0000 (17:10 -0800)]
[XFRM]: Remove unused exports.

This patch removes the following no longer used EXPORT_SYMBOL's:
- xfrm_input.c: xfrm_parse_spi
- xfrm_state.c: xfrm_replay_check
- xfrm_state.c: xfrm_replay_advance

Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[PKT_SCHED] sch_teql.c: Duplicate IFF_BROADCAST in FMASK, remove 2nd.
Roel Kluin [Fri, 1 Feb 2008 01:08:47 +0000 (17:08 -0800)]
[PKT_SCHED] sch_teql.c: Duplicate IFF_BROADCAST in FMASK, remove 2nd.

Signed-off-by: Roel Kluin <12o3l@tiscali.nl>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[BNX2]: Fix ASYM PAUSE advertisement for remote PHY.
Michael Chan [Fri, 1 Feb 2008 01:07:21 +0000 (17:07 -0800)]
[BNX2]: Fix ASYM PAUSE advertisement for remote PHY.

We were checking for the ASYM_PAUSE bit for 1000Base-X twice instead
checking for both the 1000Base-X bit and the 10/100/1000Base-T bit.
The purpose of the logic is to tell the firmware that ASYM_PAUSE is
set on either the Serdes or Copper interface.

Problem was discovered by Roel Kluin <12o3l@tiscali.nl>

Signed-off-by: Michael Chan <mchan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4] route cache: Introduce rt_genid for smooth cache invalidation
Eric Dumazet [Fri, 1 Feb 2008 01:05:09 +0000 (17:05 -0800)]
[IPV4] route cache: Introduce rt_genid for smooth cache invalidation

Current ip route cache implementation is not suited to large caches.

We can consume a lot of CPU when cache must be invalidated, since we
currently need to evict all cache entries, and this eviction is
sometimes asynchronous. min_delay & max_delay can somewhat control this
asynchronism behavior, but whole thing is a kludge, regularly triggering
infamous soft lockup messages. When entries are still in use, this also
consumes a lot of ram, filling dst_garbage.list.

A better scheme is to use a generation identifier on each entry,
so that cache invalidation can be performed by changing the table
identifier, without having to scan all entries.
No more delayed flushing, no more stalling when secret_interval expires.

Invalidated entries will then be freed at GC time (controled by
ip_rt_gc_timeout or stress), or when an invalidated entry is found
in a chain when an insert is done.
Thus we keep a normal equilibrium.

This patch :
- renames rt_hash_rnd to rt_genid (and makes it an atomic_t)
- Adds a new rt_genid field to 'struct rtable' (filling a hole on 64bit)
- Checks entry->rt_genid at appropriate places :

14 years ago[PKTGEN]: pktgen should not print info that it is spinning
Jesse Brandeburg [Fri, 1 Feb 2008 00:59:47 +0000 (16:59 -0800)]
[PKTGEN]: pktgen should not print info that it is spinning

when using pktgen to send delay packets the module prints repeatedly
to the kernel log:

sleeping for X
sleeping for X
...

This is probably just a debugging item left in and should not be
enabled for regular use of the module.

Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[NET_SCHED]: sch_ingress: remove netfilter support
Patrick McHardy [Fri, 1 Feb 2008 00:57:15 +0000 (16:57 -0800)]
[NET_SCHED]: sch_ingress: remove netfilter support

Since the old policer code is gone, TC actions are needed for policing.
The ingress qdisc can get packets directly from netif_receive_skb()
in case TC actions are enabled or through netfilter otherwise, but
since without TC actions there is no policer the only thing it actually
does is count packets.

Remove the netfilter support and always require TC actions.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Acked-by: Jamal Hadi Salim <hadi@cyberus.ca>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[MACVLAN]: Setting macvlan_handle_frame_hook to NULL when rtnl_link_register() fails.
Rami Rosen [Fri, 1 Feb 2008 00:56:03 +0000 (16:56 -0800)]
[MACVLAN]: Setting macvlan_handle_frame_hook to NULL when rtnl_link_register() fails.

In drivers/net/macvlan.c, when rtnl_link_register() fails in
macvlan_init_module(), there is no point to set it (second time in
this method) to macvlan_handle_frame; macvlan_init_module() will
return a negative number, so instead this patch sets
macvlan_handle_frame_hook to NULL.

Signed-off-by: Rami Rosen <ramirose@gmail.com>
Acked-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[VLAN]: set_rx_mode support for unicast address list
Chris Leech [Fri, 1 Feb 2008 00:53:23 +0000 (16:53 -0800)]
[VLAN]: set_rx_mode support for unicast address list

Reuse the existing logic for multicast list synchronization for the
unicast address list. The core of dev_mc_sync/unsync are split out as
__dev_addr_sync/unsync and moved from dev_mcast.c to dev.c.  These are
then used to implement dev_unicast_sync/unsync as well.

I'm working on cleaning up Intel's FCoE stack, which generates new MAC
addresses from the fibre channel device id assigned by the fabric as
per the current draft specification in T11.  When using such a
protocol in a VLAN environment it would be nice to not always be
forced into promiscuous mode, assuming the underlying Ethernet driver
supports multiple unicast addresses as well.

Signed-off-by: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Patrick McHardy <kaber@trash.net>
14 years ago[TCP]: Fix a bug in strategy_allowed_congestion_control
Shan Wei [Fri, 1 Feb 2008 00:47:27 +0000 (16:47 -0800)]
[TCP]: Fix a bug in strategy_allowed_congestion_control

In strategy_allowed_congestion_control of the 2.6.24 kernel, when
sysctl_string return 1 on success,it should call
tcp_set_allowed_congestion_control to set the allowed congestion
control.But, it don't.  the sysctl_string return 1 on success,
otherwise return negative, never return 0.The patch fix the problem.

Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Acked-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
14 years ago[IPV4] fib_trie: rescan if key is lost during dump
Stephen Hemminger [Fri, 1 Feb 2008 00:45:47 +0000 (16:45 -0800)]
[IPV4] fib_trie: rescan if key is lost during dump

Normally during a dump the key of the last dumped entry is used for
continuation, but since lock is dropped it might be lost. In that case
fallback to the old counter based N^2 behaviour.  This means the dump
will end up skipping some routes which matches what FIB_HASH does.

Signed-off-by: Stephen Hemminger <shemminger@vyatta.com>
Signed-off-by: David S. Miller <davem@davemloft.net>