mutex subsystem, -V6

this is verion -V6 of the generic mutex subsystem. It consists of the following patches:

add-atomic-xchg.patch
mutex-generic-asm-implementations.patch
mutex-arch-mutex-h.patch
mutex-core.patch
mutex-docs.patch
mutex-debug.patch
mutex-debug-more.patch
xfs-mutex-namespace-collision-fix.patch

the patches are against Linus' latest GIT tree, and they should work fine on every Linux architecture.

the delta since -V5:

  53 files changed, 718 insertions(+), 454 deletions(-)

this release picks up Arjan's asm/mutex.h implementation, which adds asm-generic/mutex-dec.h, asm-generic/mutex-xchg.h for architectures to pick up. i386 and x86_64 use their own optimized version already, the other architectures default to mutex-xchg.h. Architectures specify the following functions:

-------------------------------------------------------------------
* __mutex_fastpath_lock - try to take the lock by moving the count
* from 1 to a 0 value
* @count: pointer of type atomic_t
* @fn: function to call if the original value was not 1
-------------------------------------------------------------------
* __mutex_fastpath_lock_retval - try to take the lock by moving the count
* from 1 to a 0 value
* @count: pointer of type atomic_t
* @fn: function to call if the original value was not 1
-------------------------------------------------------------------

 *  __mutex_fastpath_unlock - try to promote the mutex from 0 to 1
 *  @count: pointer of type atomic_t
 *  @fn: function to call if the original value was not 1

-------------------------------------------------------------------

and __mutex_slowpath_needs_to_unlock(), to specify whether the fastpath has touched the count or not.

i have tested this on x86, and i have booted all 4 variants:
mutex-xchg.h, mutex-dec.h, asm-i386/mutex.h and the debug version. (in MUTEX_DEBUG_FULL mode, i.e. with the mutexes in real use.)

Nico, Christoph, does this approach work for you? Nico, you might want to try an ARM-specific mutex.h implementation.

Ingo