From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S965889AbbJVWHp (ORCPT ); Thu, 22 Oct 2015 18:07:45 -0400 Received: from mga03.intel.com ([134.134.136.65]:22339 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752153AbbJVWHn (ORCPT ); Thu, 22 Oct 2015 18:07:43 -0400 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.20,184,1444719600"; d="scan'208";a="586267597" From: Andi Kleen To: peterz@infradead.org Cc: mingo@kernel.org, linux-kernel@vger.kernel.org, Andi Kleen Subject: [PATCH 2/2] x86, perf: Optimize stack walk user accesses Date: Thu, 22 Oct 2015 15:07:21 -0700 Message-Id: <1445551641-13379-2-git-send-email-andi@firstfloor.org> X-Mailer: git-send-email 2.4.3 In-Reply-To: <1445551641-13379-1-git-send-email-andi@firstfloor.org> References: <1445551641-13379-1-git-send-email-andi@firstfloor.org> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Andi Kleen Change the perf user stack walking to use the new __copy_from_user_nmi, and split each access into word sized transfer sizes. This allows to inline the complete access and optimize it all into a single load. The main advantage is that this avoids the overhead of double page faults. When normal copy_from_user fails it reexecutes the copy to compute an accurate number of non copied bytes. This leads to executing the expensive page fault twice. While walking stacks having a fault at some point is relatively common (typically when some part of the program isn't compiled with frame pointers), so this is a large overhead. With the optimized copies we avoid this problem because they only do all accesses once. And of course they're much faster too when the access does not fault because they're just single instructions instead of complex function calls. While profiling a kernel build with -g, the patch brings down the average time of the PMI handler from 966ns to 552ns (-43%) v2: Disable page faults explicitly to handle software trace points. Fix sparse warning Signed-off-by: Andi Kleen --- arch/x86/kernel/cpu/perf_event.c | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) diff --git a/arch/x86/kernel/cpu/perf_event.c b/arch/x86/kernel/cpu/perf_event.c index a74fab5..c80740a 100644 --- a/arch/x86/kernel/cpu/perf_event.c +++ b/arch/x86/kernel/cpu/perf_event.c @@ -2253,12 +2253,18 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) ss_base = get_segment_base(regs->ss); fp = compat_ptr(ss_base + regs->bp); + pagefault_disable(); while (entry->nr < PERF_MAX_STACK_DEPTH) { unsigned long bytes; frame.next_frame = 0; frame.return_address = 0; - bytes = copy_from_user_nmi(&frame, fp, sizeof(frame)); + if (!access_ok(VERIFY_READ, fp, 8)) + break; + bytes = __copy_from_user_nmi(&frame.next_frame, fp, 4); + if (bytes != 0) + break; + bytes = __copy_from_user_nmi(&frame.return_address, fp+4, 4); if (bytes != 0) break; @@ -2268,6 +2274,7 @@ perf_callchain_user32(struct pt_regs *regs, struct perf_callchain_entry *entry) perf_callchain_store(entry, cs_base + frame.return_address); fp = compat_ptr(ss_base + frame.next_frame); } + pagefault_enable(); return 1; } #else @@ -2305,12 +2312,19 @@ perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) if (perf_callchain_user32(regs, entry)) return; + pagefault_disable(); while (entry->nr < PERF_MAX_STACK_DEPTH) { unsigned long bytes; frame.next_frame = NULL; frame.return_address = 0; - bytes = copy_from_user_nmi(&frame, fp, sizeof(frame)); + if (!access_ok(VERIFY_READ, fp, 16)) + break; + + bytes = __copy_from_user_nmi(&frame.next_frame, fp, 8); + if (bytes != 0) + break; + bytes = __copy_from_user_nmi(&frame.return_address, fp+8, 8); if (bytes != 0) break; @@ -2318,8 +2332,9 @@ perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs) break; perf_callchain_store(entry, frame.return_address); - fp = frame.next_frame; + fp = (void __user *)frame.next_frame; } + pagefault_enable(); } /* -- 2.4.3