- 02 Apr, 2010 7 commits
-
-
Oleg Nesterov authored
sched_exec()->select_task_rq() reads/updates ->cpus_allowed lockless. This can race with other CPUs updating our ->cpus_allowed, and this looks meaningless to me. The task is current and running, it must have online cpus in ->cpus_allowed, the fallback mode is bogus. And, if ->sched_class returns the "wrong" cpu, this likely means we raced with set_cpus_allowed() which was called for reason, why should sched_exec() retry and call ->select_task_rq() again? Change the code to call sched_class->select_task_rq() directly and do nothing if the returned cpu is wrong after re-checking under rq->lock. From now task_struct->cpus_allowed is always stable under TASK_WAKING, select_fallback_rq() is always called under rq-lock or the caller or the caller owns TASK_WAKING (select_task_rq). Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <20100315091019.GA9141@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Oleg Nesterov authored
The previous patch preserved the retry logic, but it looks unneeded. __migrate_task() can only fail if we raced with migration after we dropped the lock, but in this case the caller of set_cpus_allowed/etc must initiate migration itself if ->on_rq == T. We already fixed p->cpus_allowed, the changes in active/online masks must be visible to racer, it should migrate the task to online cpu correctly. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <20100315091014.GA9138@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Oleg Nesterov authored
move_task_off_dead_cpu()->select_fallback_rq() reads/updates ->cpus_allowed lockless. We can race with set_cpus_allowed() running in parallel. Change it to take rq->lock around select_fallback_rq(). Note that it is not trivial to move this spin_lock() into select_fallback_rq(), we must recheck the task was not migrated after we take the lock and other callers do not need this lock. To avoid the races with other callers of select_fallback_rq() which rely on TASK_WAKING, we also check p->state != TASK_WAKING and do nothing otherwise. The owner of TASK_WAKING must update ->cpus_allowed and choose the correct CPU anyway, and the subsequent __migrate_task() is just meaningless because p->se.on_rq must be false. Alternatively, we could change select_task_rq() to take rq->lock right after it calls sched_class->select_task_rq(), but this looks a bit ugly. Also, change it to not assume irqs are disabled and absorb __migrate_task_irq(). Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <20100315091010.GA9131@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Oleg Nesterov authored
This patch just states the fact the cpusets/cpuhotplug interaction is broken and removes the deadlockable code which only pretends to work. - cpuset_lock() doesn't really work. It is needed for cpuset_cpus_allowed_locked() but we can't take this lock in try_to_wake_up()->select_fallback_rq() path. - cpuset_lock() is deadlockable. Suppose that a task T bound to CPU takes callback_mutex. If cpu_down(CPU) happens before T drops callback_mutex stop_machine() preempts T, then migration_call(CPU_DEAD) tries to take cpuset_lock() and hangs forever because CPU is already dead and thus T can't be scheduled. - cpuset_cpus_allowed_locked() is deadlockable too. It takes task_lock() which is not irq-safe, but try_to_wake_up() can be called from irq. Kill them, and change select_fallback_rq() to use cpu_possible_mask, like we currently do without CONFIG_CPUSETS. Also, with or without this patch, with or without CONFIG_CPUSETS, the callers of select_fallback_rq() can race with each other or with set_cpus_allowed() pathes. The subsequent patches try to to fix these problems. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> LKML-Reference: <20100315091003.GA9123@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Li Zefan authored
USER_SCHED has been removed, so update the documentation accordingly. Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Acked-by: Serge E. Hallyn <serue@us.ibm.com> LKML-Reference: <4BA9A07E.8070508@cn.fujitsu.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Li Zefan authored
This is left over from commit 7c941438 ("sched: Remove USER_SCHED"") Signed-off-by: Li Zefan <lizf@cn.fujitsu.com> Acked-by: Dhaval Giani <dhaval.giani@gmail.com> Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: David Howells <dhowells@redhat.com> LKML-Reference: <4BA9A05F.7010407@cn.fujitsu.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
Ingo Molnar authored
Merge reason: update to latest upstream Signed-off-by: Ingo Molnar <mingo@elte.hu>
-
- 01 Apr, 2010 33 commits
-
-
git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6Linus Torvalds authored
* 'drm-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6: (76 commits) drm/radeon/kms: enable ACPI powermanagement mode on radeon gpus. drm/radeon/kms: rs400/480 should set common registers. drm/radeon/kms: add sanity check to wptr. drm/radeon/kms/evergreen: get DP working drm/radeon/kms: add hw_i2c module option drm/radeon/kms: use new pre/post_xfer i2c bit algo hooks drm/radeon/kms: disable MSI on IGP chips drm/radeon/kms: display watermark updates (v2) drm/radeon/kms/dp: disable training pattern on the sink at the end of link training drm/radeon/kms: minor fixes for eDP with LCD* device tags (v2) drm/radeon/kms/dp: remove extraneous training complete call drm/radeon/kms/atom: minor fixes to transmitter setup drm/radeon/kms: Only restrict BO to visible VRAM size when pinning to VRAM. drm: fix build error when SYSRQ is disabled drm/radeon/kms: fix macbookpro connector quirk drm/radeon/r6xx/r7xx: further safe reg clean up drm/radeon: bump the UMS driver version for r6xx/r7xx const buffer support drm/radeon/kms: bump the version for r6xx/r7xx const buffer support drm/radeon/r6xx/r7xx: CS parser fixes drm/radeon/kms: fix some typos in r6xx/r7xx hpd setup ... Fix up MSI-related conflicts in drivers/gpu/drm/radeon/radeon_irq_kms.c
-
git://git.monstr.eu/linux-2.6-microblazeLinus Torvalds authored
* 'for-linus' of git://git.monstr.eu/linux-2.6-microblaze: (35 commits) microblaze: Support word copying in copy_tofrom_user microblaze: Print early printk information to log buffer microblaze: head.S typo fix microblaze: Use MICROBLAZE_TLB_SIZE in asm code microblaze: Kconfig Fix - pci microblaze: Adding likely macros microblaze: Add .type and .size to ASM functions microblaze: Fix TLB macros microblaze: Use instruction with delay slot microblaze: Remove additional resr and rear loading microblaze: Change register usage for ESR and EAR microblaze: Prepare work for optimization in exception code microblaze: Add DEBUG option microblaze: Support systems without lmb bram microblaze: uaccess: Sync strlen, strnlen, copy_to/from_user microblaze: uaccess: Unify __copy_tofrom_user microblaze: uaccess: Move functions to generic location microblaze: uaccess: Fix put_user for noMMU microblaze: uaccess: Fix get_user macro for noMMU microblaze: uaccess: fix clear_user for noMMU kernel ...
-
Linus Torvalds authored
Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86 * 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mjg59/platform-drivers-x86: eeepc-wmi: new driver for WMI based hotkeys on Eee PC laptops asus-laptop: fix warning in asus_handle_init
-
Oleg Nesterov authored
proc_oom_score(task) has a reference to task_struct, but that is all. If this task was already released before we take tasklist_lock - we can't use task->group_leader, it points to nowhere - it is not safe to call badness() even if this task is ->group_leader, has_intersects_mems_allowed() assumes it is safe to iterate over ->thread_group list. - even worse, badness() can hit ->signal == NULL Add the pid_alive() check to ensure __unhash_process() was not called. Also, use "task" instead of task->group_leader. badness() should return the same result for any sub-thread. Currently this is not true, but this should be changed anyway. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
-
Michal Simek authored
Word copying is used only for aligned addresses. Here is space for improving to use any better copying technique. Look at memcpy implementation. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
If early printk console is not enabled then all messages are written to log buffer. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
I forget to change register name in comments. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
TLB size was hardcoded in asm code. This patch brings ability to change TLB size only in one place. (mmu.h). Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
I forget to remove pci Kconfig option. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
On the base on GCOV analytics is helpful to add likely/unlikely macros. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Cachegrind analysis need this fix to be able to log asm functions. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
To be able to do trace TLB operations. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Sync labels. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
RESR and REAR uses the same regs in whole file. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
This change synchronize register usage in code. ESR = R4 EAR = R3 Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Any sync branch must follow mts instructions not mfs. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Disable debug option in asm code. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
When the system has no lmb bram, main memory should be start from zero because of microblaze vectors. DTS fragment could look like: DDR2_SDRAM: memory@0 { device_type = "memory"; reg = < 0x0 0x10000000 >; } ; Then you have to setup CONFIG_KERNEL_BASE_ADDR=0 which caused that kernel physical start address will be zero. On reset vector place will be jump to 0x100 and on 0x100 starts kernel text. You have to solve how to load the kernel before cpu starts. Tested with XMD. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Last sync. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Move to generic location. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
noMMU and MMU use them. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Here is small regression on dhrystone tests and I think that on all benchmarking tests. It is due to better checking mechanism in put_user macro Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Use unified version. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Previous patches fixed only MMU version and this is the first patch for noMMU kernel Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Generic implementation for noMMU and MMU version Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
copy_from_user macro also use copy_tofrom_user function Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
noMMU and MMU kernel will use copy copy_tofrom_user asm implementation. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Add macro description and resort. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Use FIXUP macros and resort them. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
It is used __FIXUP_SECTION and __EX_TABLE_SECTION macros. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
This is the first patch which does uaccess unification. I choosed to do several patches to be able to use bisect in future if any fault happens. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
The same noMMU and MMU functions should be placed in one file. Signed-off-by: Michal Simek <monstr@monstr.eu>
-
Michal Simek authored
Just sort to be able remove whole block. Signed-off-by: Michal Simek <monstr@monstr.eu>
-