From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1423999AbdD3LpK (ORCPT ); Sun, 30 Apr 2017 07:45:10 -0400 Received: from Galois.linutronix.de ([146.0.238.70]:37479 "EHLO Galois.linutronix.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S968342AbdD3LpA (ORCPT ); Sun, 30 Apr 2017 07:45:00 -0400 Date: Sun, 30 Apr 2017 13:44:58 +0200 (CEST) From: Thomas Gleixner To: Linus Torvalds cc: LKML , Andrew Morton , Ingo Molnar , "H. Peter Anvin" Subject: [GIT pull] x86 fixes for 4.11 Message-ID: User-Agent: Alpine 2.20 (DEB 67 2015-01-07) MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Linus, please pull the latest x86-urgent-for-linus git tree from: git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip.git x86-urgent-for-linus The final fixes for 4.11: - Prevent a triple fault with function graph tracing triggered via suspend to ram. - Prevent optimizing for size when function graph tracing is enabled and the compiler does not support -mfentry - Prevent mwaitx() being called with a zero timeout as mwaitx() might never return. Observed on the new Ryzen CPUs. Thanks, tglx ------------------> Janakarajan Natarajan (1): Prevent timer value 0 for MWAITX Josh Poimboeuf (2): ftrace/x86: Fix triple fault with graph tracing and suspend-to-ram x86/build: convert function graph '-Os' error to warning arch/x86/Makefile | 8 ++++++++ arch/x86/kernel/ftrace.c | 18 ++++++++++++------ arch/x86/lib/delay.c | 7 +++++++ 3 files changed, 27 insertions(+), 6 deletions(-) diff --git a/arch/x86/Makefile b/arch/x86/Makefile index a94a4d10f2df..49d160b781f0 100644 --- a/arch/x86/Makefile +++ b/arch/x86/Makefile @@ -154,6 +154,14 @@ ifdef CONFIG_FUNCTION_GRAPH_TRACER else ifeq ($(call cc-option-yn, -mfentry), n) ACCUMULATE_OUTGOING_ARGS := 1 + + # GCC ignores '-maccumulate-outgoing-args' when used with '-Os'. + # If '-Os' is enabled, disable it and print a warning. + ifdef CONFIG_CC_OPTIMIZE_FOR_SIZE + undefine CONFIG_CC_OPTIMIZE_FOR_SIZE + $(warning Disabling CONFIG_CC_OPTIMIZE_FOR_SIZE. Your compiler does not have -mfentry so you cannot optimize for size with CONFIG_FUNCTION_GRAPH_TRACER.) + endif + endif endif endif diff --git a/arch/x86/kernel/ftrace.c b/arch/x86/kernel/ftrace.c index cbd73eb42170..5b7153540727 100644 --- a/arch/x86/kernel/ftrace.c +++ b/arch/x86/kernel/ftrace.c @@ -29,12 +29,6 @@ #include #include -#if defined(CONFIG_FUNCTION_GRAPH_TRACER) && \ - !defined(CC_USING_FENTRY) && \ - !defined(CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE) -# error The following combination is not supported: ((compiler missing -mfentry) || (CONFIG_X86_32 and !CONFIG_DYNAMIC_FTRACE)) && CONFIG_FUNCTION_GRAPH_TRACER && CONFIG_CC_OPTIMIZE_FOR_SIZE -#endif - #ifdef CONFIG_DYNAMIC_FTRACE int ftrace_arch_code_modify_prepare(void) @@ -989,6 +983,18 @@ void prepare_ftrace_return(unsigned long self_addr, unsigned long *parent, unsigned long return_hooker = (unsigned long) &return_to_handler; + /* + * When resuming from suspend-to-ram, this function can be indirectly + * called from early CPU startup code while the CPU is in real mode, + * which would fail miserably. Make sure the stack pointer is a + * virtual address. + * + * This check isn't as accurate as virt_addr_valid(), but it should be + * good enough for this purpose, and it's fast. + */ + if (unlikely((long)__builtin_frame_address(0) >= 0)) + return; + if (unlikely(ftrace_graph_is_dead())) return; diff --git a/arch/x86/lib/delay.c b/arch/x86/lib/delay.c index a8e91ae89fb3..29df077cb089 100644 --- a/arch/x86/lib/delay.c +++ b/arch/x86/lib/delay.c @@ -93,6 +93,13 @@ static void delay_mwaitx(unsigned long __loops) { u64 start, end, delay, loops = __loops; + /* + * Timer value of 0 causes MWAITX to wait indefinitely, unless there + * is a store on the memory monitored by MONITORX. + */ + if (loops == 0) + return; + start = rdtsc_ordered(); for (;;) {