Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756092AbbDJNHH (ORCPT ); Fri, 10 Apr 2015 09:07:07 -0400 Received: from mx1.redhat.com ([209.132.183.28]:56533 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755886AbbDJNHF (ORCPT ); Fri, 10 Apr 2015 09:07:05 -0400 From: Denys Vlasenko To: linux-kernel@vger.kernel.org Cc: Denys Vlasenko , Jiri Olsa , Peter Zijlstra , Paul Mackerras , Ingo Molnar , Arnaldo Carvalho de Melo , x86@kernel.org Subject: [PATCH 1/4] perf/x86_64: get_regs_user: Do not guess user_regs->cs,ss,sp Date: Fri, 10 Apr 2015 15:06:56 +0200 Message-Id: <1428671219-29341-1-git-send-email-dvlasenk@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 3537 Lines: 93 After recent changes to syscall entry points, user_regs->{cs,ss,sp} are always correct. (They used to be undefined while in syscalls). We can report them reliably, without guessing. Signed-off-by: Denys Vlasenko CC: Jiri Olsa CC: Peter Zijlstra CC: Paul Mackerras CC: Ingo Molnar CC: Arnaldo Carvalho de Melo CC: x86@kernel.org CC: linux-kernel@vger.kernel.org --- arch/x86/kernel/perf_regs.c | 23 +++++++++-------------- 1 file changed, 9 insertions(+), 14 deletions(-) diff --git a/arch/x86/kernel/perf_regs.c b/arch/x86/kernel/perf_regs.c index 02a8720..7ab198a 100644 --- a/arch/x86/kernel/perf_regs.c +++ b/arch/x86/kernel/perf_regs.c @@ -131,8 +131,8 @@ void perf_get_regs_user(struct perf_regs *regs_user, } /* - * RIP, flags, and the argument registers are usually saved. - * orig_ax is probably okay, too. + * These registers are always saved on 64-bit syscall entry. + * On 32-bit entry points, they are saved too except r8..r11. */ regs_user_copy->ip = user_regs->ip; regs_user_copy->cx = user_regs->cx; @@ -145,9 +145,12 @@ void perf_get_regs_user(struct perf_regs *regs_user, regs_user_copy->r11 = user_regs->r11; regs_user_copy->orig_ax = user_regs->orig_ax; regs_user_copy->flags = user_regs->flags; + regs_user_copy->sp = user_regs->sp; + regs_user_copy->cs = user_regs->cs; + regs_user_copy->ss = user_regs->ss; /* - * Don't even try to report the "rest" regs. + * Most system calls don't save these registers, don't report them. */ regs_user_copy->bx = -1; regs_user_copy->bp = -1; @@ -158,7 +161,7 @@ void perf_get_regs_user(struct perf_regs *regs_user, /* * For this to be at all useful, we need a reasonable guess for - * sp and the ABI. Be careful: we're in NMI context, and we're + * the ABI. Be careful: we're in NMI context, and we're * considering current to be the current task, so we should * be careful not to look at any other percpu variables that might * change during context switches. @@ -167,9 +170,6 @@ void perf_get_regs_user(struct perf_regs *regs_user, task_thread_info(current)->status & TS_COMPAT) { /* Easy case: we're in a compat syscall. */ regs_user->abi = PERF_SAMPLE_REGS_ABI_32; - regs_user_copy->sp = user_regs->sp; - regs_user_copy->cs = user_regs->cs; - regs_user_copy->ss = user_regs->ss; } else if (user_regs->orig_ax != -1) { /* * We're probably in a 64-bit syscall. @@ -177,17 +177,12 @@ void perf_get_regs_user(struct perf_regs *regs_user, * than just blindly copying user_regs. */ regs_user->abi = PERF_SAMPLE_REGS_ABI_64; - regs_user_copy->sp = user_regs->sp; - regs_user_copy->cs = __USER_CS; - regs_user_copy->ss = __USER_DS; - regs_user_copy->cx = -1; /* usually contains garbage */ + /* usually contains return address (same as ->ip) */ + regs_user_copy->cx = -1; } else { /* We're probably in an interrupt or exception. */ regs_user->abi = user_64bit_mode(user_regs) ? PERF_SAMPLE_REGS_ABI_64 : PERF_SAMPLE_REGS_ABI_32; - regs_user_copy->sp = user_regs->sp; - regs_user_copy->cs = user_regs->cs; - regs_user_copy->ss = user_regs->ss; } regs_user->regs = regs_user_copy; -- 1.8.1.4 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/