> Subject says it all.
The new lock_kernel/unlock_kernel implemenetation. Non SMP kernels don't
grow.
Just do the math: 187 system calls in 2.1.112, nearly all of them
do lock_kernel/unlock_kernel as inline code, lets say a lock/unlock pair
needs 150 bytes, then 170[1] * 150bytes = 24k [probably more because
zImage is compressed]
Linus, how about moving the slow part (that loops) of the lock out-of-line,
just to give the L1 Icache a bit more air to breath?
-Andi
[1] to account for the few calls that don't lock
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html