RISC-V: Resurrect smp_mb__after_spinlock()

I removed this last week because of an incorrect comment:
smp_mb__after_spinlock() is actually still used, and is necessary on
RISC-V.  It's been resurrected, with a comment that describes what it
actually does this time.  Thanks to Andrea for finding the bug!

Fixes: 3343eb6806 ("RISC-V: Remove smb_mb__{before,after}_spinlock()")
CC: Andrea Parri <parri.andrea@gmail.com>
Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
This commit is contained in:
Palmer Dabbelt 2017-12-05 17:48:11 -08:00 committed by Palmer Dabbelt
parent 86ad5c97ce
commit 3cfa500808
No known key found for this signature in database
GPG Key ID: EF4CA1502CCBAB41
1 changed files with 19 additions and 0 deletions

View File

@ -38,6 +38,25 @@
#define smp_rmb() RISCV_FENCE(r,r)
#define smp_wmb() RISCV_FENCE(w,w)
/*
* This is a very specific barrier: it's currently only used in two places in
* the kernel, both in the scheduler. See include/linux/spinlock.h for the two
* orderings it guarantees, but the "critical section is RCsc" guarantee
* mandates a barrier on RISC-V. The sequence looks like:
*
* lr.aq lock
* sc lock <= LOCKED
* smp_mb__after_spinlock()
* // critical section
* lr lock
* sc.rl lock <= UNLOCKED
*
* The AQ/RL pair provides a RCpc critical section, but there's not really any
* way we can take advantage of that here because the ordering is only enforced
* on that one lock. Thus, we're just doing a full fence.
*/
#define smp_mb__after_spinlock() RISCV_FENCE(rw,rw)
#include <asm-generic/barrier.h>
#endif /* __ASSEMBLY__ */