linux-toradex.git/kernel/locking/rtmutex.c, branch v5.13-rc3

locking/rtmutex: Clean up signal handling in __rt_mutex_slowlock()

2021-03-29T13:57:05+00:00

The signal handling in __rt_mutex_slowlock() is open coded.

Use signal_pending_state() instead.

Aside of the cleanup this also prepares for the RT lock substituions which
require support for TASK_KILLABLE.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153944.533811987@linutronix.de

locking/rtmutex: Restrict the trylock WARN_ON() to debug

2021-03-29T13:57:04+00:00

The warning as written is expensive and not really required for a
production kernel. Make it depend on rt mutex debugging and use !in_task()
for the condition which generates far better code and gives the same
answer.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153944.436565064@linutronix.de

locking/rtmutex: Fix misleading comment in rt_mutex_postunlock()

2021-03-29T13:57:04+00:00

Preemption is disabled in mark_wakeup_next_waiter(,) not in
rt_mutex_slowunlock().

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153944.341734608@linutronix.de

locking/rtmutex: Consolidate the fast/slowpath invocation

2021-03-29T13:57:04+00:00

The indirection via a function pointer (which is at least optimized into a
tail call by the compiler) is making the code hard to read.

Clean it up and move the futex related trylock functions down to the futex
section.

Move the wake_q wakeup into rt_mutex_slowunlock(). No point in handing it
to the caller. The futex code uses a different function.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153944.247927548@linutronix.de

locking/rtmutex: Make text section and inlining consistent

2021-03-29T13:57:04+00:00

rtmutex is half __sched and the other half is not. If the compiler decides
to not inline larger static functions then part of the code ends up in the
regular text section.

There are also quite some performance related small helpers which are
either static or plain inline. Force inline those which make sense and mark
the rest __sched.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153944.152977820@linutronix.de

locking/rtmutex: Decrapify __rt_mutex_init()

2021-03-29T13:57:03+00:00

The conditional debug handling is just another layer of obfuscation. Split
the function so rt_mutex_init_proxy_locked() can invoke the inner init and
__rt_mutex_init() gets the full treatment.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153943.955697588@linutronix.de

locking/rtmutex: Inline chainwalk depth check

2021-03-29T13:57:03+00:00

There is no point for this wrapper at all.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153943.754254046@linutronix.de

locking/rtmutex: Move rt_mutex_debug_task_free() to rtmutex.c

2021-03-29T13:57:03+00:00

Prepare for removing the header maze.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153943.646359691@linutronix.de

locking/rtmutex: Remove empty and unused debug stubs

2021-03-29T13:57:03+00:00

No users or useless and therefore just ballast.

Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153943.549192485@linutronix.de

locking/rtmutex: Remove output from deadlock detector

2021-03-29T13:57:02+00:00

The rtmutex specific deadlock detector predates lockdep coverage of rtmutex
and since commit f5694788ad8da ("rt_mutex: Add lockdep annotations") it
contains a lot of redundant functionality:

 - lockdep will detect an potential deadlock before rtmutex-debug
   has a chance to do so

 - the deadlock debugging is restricted to rtmutexes which are not
   associated to futexes and have an active waiter, which is covered by
   lockdep already

Remove the redundant functionality and move actual deadlock WARN() into the
deadlock code path. The latter needs a seperate cleanup.

Signed-off-by: Sebastian Andrzej Siewior 
Signed-off-by: Thomas Gleixner 
Signed-off-by: Ingo Molnar 
Acked-by: Peter Zijlstra (Intel) 
Link: https://lore.kernel.org/r/20210326153943.320398604@linutronix.de