On Mon, Aug 31, 2015 at 6:37 PM, Andy Lutomirski <l...@amacapital.net> wrote: > I got random errors from perf kvm, but I think I found at least part > of the issue. The two irqs_disabled() calls in common.c are kind of > expensive. I should disable them on non-lockdep kernels. > > The context tracking hooks are also too expensive, even when disabled. > I should do something to optimize those. Hello, static keys? This > doesn't affect syscalls, though. > > With context tracking off and the irqs_disabled checks commented out, > we're probably doing well enough. We can always tweak the C code and > aggressively force inlining if we want a few cycles back.
Currently, a compat AT_SYSINFO syscall (getpid) is 171 cycles for me. With my patches, it's 196 cycles, so it's really not that bad. The impact will probably be slightly worse on native 32-bit because of increased register pressure and because one of the micro-optimizations I threw in are 64-bit specific. We could probably tune the C code a bit more to get a few of the cycles back. On the flip side, the rewrite is *far* faster in some of the slow path cases because the slow path no longer forces IRET. On 32-bit, there's the added benefit that we could drop asmlinkage from the syscall bodies on top of the rewrite. --Andy > > --Andy -- Andy Lutomirski AMA Capital Management, LLC -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/