Optimize two cases in the MP locking code. First, it is not necessary
to use a locked cmpexg when unlocking a lock that we already hold, since nobody else can touch the lock while we hold it. Second, it is not necessary to use a locked cmpexg when locking a lock that we already hold, for the same reason. These changes will allow MP locks to be used recursively without impacting performance. Modify two procedures that are called only by assembly and are already NOPROF entries to pass a critical argument in %edx instead of on the stack, removing a significant amount of code from the critical path as a consequence. Reviewed by: Alfred Perlstein <bright@wintelcom.net>, Peter Wemm <peter@netplex.com.au>
This commit is contained in:
parent
2996376ab5
commit
f94f8efc61
@ -55,9 +55,8 @@
|
||||
add $4, %esp
|
||||
|
||||
#define ISR_RELLOCK \
|
||||
pushl $_mp_lock ; /* GIANT_LOCK */ \
|
||||
call _MPrellock ; \
|
||||
add $4, %esp
|
||||
movl $_mp_lock,%edx ; /* GIANT_LOCK */ \
|
||||
call _MPrellock_edx
|
||||
|
||||
/*
|
||||
* Protects the IO APIC and apic_imen as a critical region.
|
||||
|
Loading…
Reference in New Issue
Block a user