linux-toradex.git/include/linux/uaccess.h, branch v3.2.8-rt13

mm: raw_pagefault_disable

2012-02-15T16:32:32+00:00

Adding migrate_disable() to pagefault_disable() to preserve the
per-cpu thing for kmap_atomic might not have been the best of choices.
But short of adding preempt_disable/migrate_disable foo all over the
kmap code it still seems the best way.

It does however yield the below borkage as well as wreck !-rt builds
since !-rt does rely on pagefault_disable() not preempting. So fix all
that up by adding raw_pagefault_disable().

   [] warn_slowpath_common+0x85/0x9d
 [] warn_slowpath_fmt+0x46/0x48
 [] ? _raw_spin_lock+0x6c/0x73
 [] ? watchdog_overflow_callback+0x9b/0xd0
 [] watchdog_overflow_callback+0xb7/0xd0
 [] __perf_event_overflow+0x11c/0x1fe
 [] ? perf_event_update_userpage+0x149/0x151
 [] ? perf_event_task_disable+0x7c/0x7c
 [] perf_event_overflow+0x14/0x16
 [] x86_pmu_handle_irq+0xcb/0x108
 [] perf_event_nmi_handler+0x46/0x91
 [] notifier_call_chain+0x79/0xa6
 [] __atomic_notifier_call_chain+0x66/0x98
 [] ? notifier_call_chain+0xa6/0xa6
 [] atomic_notifier_call_chain+0x14/0x16
 [] notify_die+0x2e/0x30
 [] do_nmi+0x7e/0x22b
 [] nmi+0x1a/0x2c
 [] ? sub_preempt_count+0x4b/0xaa
 <>    [] delay_tsc+0xac/0xd1
 [] __delay+0xf/0x11
 [] do_raw_spin_lock+0xd2/0x13c
 [] _raw_spin_lock_irqsave+0x6b/0x85
 [] ? task_rq_lock+0x35/0x8d
 [] task_rq_lock+0x35/0x8d
 [] migrate_disable+0x65/0x12c
 [] pagefault_disable+0xe/0x1f
 [] dump_trace+0x21f/0x2e2
 [] show_trace_log_lvl+0x54/0x5d
 [] show_trace+0x15/0x17
 [] dump_stack+0x77/0x80
 [] spin_bug+0x9c/0xa3
 [] ? task_rq_lock+0x50/0x8d
 [] do_raw_spin_lock+0x47/0x13c
 [] _raw_spin_lock+0x60/0x73
 [] ? task_rq_lock+0x50/0x8d
 [] task_rq_lock+0x50/0x8d
 [] migrate_disable+0x65/0x12c
 [] pagefault_disable+0xe/0x1f
 [] dump_trace+0x21f/0x2e2
 [] save_stack_trace+0x2f/0x4c
 [] save_trace+0x3f/0xaf
 [] mark_lock+0x228/0x530
 [] __lock_acquire+0x662/0x1812
 [] ? native_sched_clock+0x37/0x6d
 [] ? trace_hardirqs_off_caller+0x1f/0x99
 [] ? sched_rt_period_timer+0xbd/0x218
 [] lock_acquire+0x145/0x18a
 [] ? sched_rt_period_timer+0xbd/0x218
 [] _raw_spin_lock+0x40/0x73
 [] ? sched_rt_period_timer+0xbd/0x218
 [] sched_rt_period_timer+0xbd/0x218
 [] __run_hrtimer+0x1e4/0x347
 [] ? can_migrate_task.clone.82+0x14a/0x14a
 [] hrtimer_interrupt+0xee/0x1d6
 [] ? add_preempt_count+0xae/0xb2
 [] smp_apic_timer_interrupt+0x85/0x98
 [] apic_timer_interrupt+0x13/0x20


Signed-off-by: Peter Zijlstra 
Link: http://lkml.kernel.org/n/tip-31keae8mkjiv8esq4rl76cib@git.kernel.org

mm: Prepare decoupling the page fault disabling logic

2012-02-15T16:32:32+00:00

Add a pagefault_disabled variable to task_struct to allow decoupling
the pagefault-disabled logic from the preempt count.

Signed-off-by: Ingo Molnar 
Signed-off-by: Thomas Gleixner

maccess,probe_kernel: Make write/read src const void *

2011-05-25T23:56:23+00:00

The functions probe_kernel_write() and probe_kernel_read() do not modify
the src pointer. Allow const pointers to be passed in without the need
of a typecast.

Acked-by: Mike Frysinger 
Acked-by: Heiko Carstens 
Acked-by: Martin Schwidefsky 
Signed-off-by: Steven Rostedt 
Link: http://lkml.kernel.org/r/1305824936.1465.4.camel@gandalf.stny.rr.com

maccess,probe_kernel: Allow arch specific override probe_kernel_(read|write)

2010-01-07T17:58:36+00:00

Some archs such as blackfin, would like to have an arch specific
probe_kernel_read() and probe_kernel_write() implementation which can
fall back to the generic implementation if no special operations are
needed.

CC: Thomas Gleixner 
CC: Ingo Molnar 
Signed-off-by: Jason Wessel 
Signed-off-by: Mike Frysinger

uaccess: fix parameters inversion for __copy_from_user_inatomic()

2008-09-16T05:44:46+00:00

The following patch changes to use __copy_from_user_inatomic(),
but the passing parameters incorrect:

    x86: some lock annotations for user copy paths, v3

This fixes the netfilter crash reported by Steven Noonan.

Reported-by: Steven Noonan 
Signed-off-by: Hiroshi Shimamoto 
Tested-by: Steven Noonan 
Signed-off-by: Ingo Molnar

x86: some lock annotations for user copy paths, v3

2008-09-11T19:42:59+00:00

- add annotation back to clear_user()
- change probe_kernel_address() to _inatomic*() method

Signed-off-by: Ingo Molnar

uaccess: add probe_kernel_write()

2008-04-17T18:05:36+00:00

add probe_kernel_read() and probe_kernel_write().

Uninlined and restricted to kernel range memory only, as suggested
by Linus.

Signed-off-by: Ingo Molnar 
Reviewed-by: Thomas Gleixner

Merge branch 'for-linus' of git://one.firstfloor.org/home/andi/git/linux-2.6

2006-12-07T16:59:11+00:00

* 'for-linus' of git://one.firstfloor.org/home/andi/git/linux-2.6: (156 commits)
  [PATCH] x86-64: Export smp_call_function_single
  [PATCH] i386: Clean up smp_tune_scheduling()
  [PATCH] unwinder: move .eh_frame to RODATA
  [PATCH] unwinder: fully support linker generated .eh_frame_hdr section
  [PATCH] x86-64: don't use set_irq_regs()
  [PATCH] x86-64: check vector in setup_ioapic_dest to verify if need setup_IO_APIC_irq
  [PATCH] x86-64: Make ix86 default to HIGHMEM4G instead of NOHIGHMEM
  [PATCH] i386: replace kmalloc+memset with kzalloc
  [PATCH] x86-64: remove remaining pc98 code
  [PATCH] x86-64: remove unused variable
  [PATCH] x86-64: Fix constraints in atomic_add_return()
  [PATCH] x86-64: fix asm constraints in i386 atomic_add_return
  [PATCH] x86-64: Correct documentation for bzImage protocol v2.05
  [PATCH] x86-64: replace kmalloc+memset with kzalloc in MTRR code
  [PATCH] x86-64: Fix numaq build error
  [PATCH] x86-64: include/asm-x86_64/cpufeature.h isn't a userspace header
  [PATCH] unwinder: Add debugging output to the Dwarf2 unwinder
  [PATCH] x86-64: Clarify error message in GART code
  [PATCH] x86-64: Fix interrupt race in idle callback (3rd try)
  [PATCH] x86-64: Remove unwind stack pointer alignment forcing again
  ...

Fixed conflict in include/linux/uaccess.h manually

Signed-off-by: Linus Torvalds

[PATCH] probe_kernel_address() needs to do set_fs()

2006-12-07T16:39:34+00:00

probe_kernel_address() purports to be generic, only it forgot to select
KERNEL_DS, so it presently won't work right on all architectures.

Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

[PATCH] mm: pagefault_{disable,enable}()

2006-12-07T16:39:21+00:00

Introduce pagefault_{disable,enable}() and use these where previously we did
manual preempt increments/decrements to make the pagefault handler do the
atomic thing.

Currently they still rely on the increased preempt count, but do not rely on
the disabled preemption, this might go away in the future.

(NOTE: the extra barrier() in pagefault_disable might fix some holes on
       machines which have too many registers for their own good)

[heiko.carstens@de.ibm.com: s390 fix]
Signed-off-by: Peter Zijlstra 
Acked-by: Nick Piggin 
Cc: Martin Schwidefsky 
Signed-off-by: Heiko Carstens 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds