From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1756697Ab1EPQCE (ORCPT ); Mon, 16 May 2011 12:02:04 -0400 Received: from DMZ-MAILSEC-SCANNER-5.MIT.EDU ([18.7.68.34]:53528 "EHLO dmz-mailsec-scanner-5.mit.edu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756423Ab1EPQCB (ORCPT ); Mon, 16 May 2011 12:02:01 -0400 X-AuditID: 12074422-b7b0eae000007f48-db-4dd14a832229 From: Andy Lutomirski To: x86@kernel.org Cc: linux-kernel@vger.kernel.org, Ingo Molnar , Andi Kleen , Linus Torvalds , "David S. Miller" , Eric Dumazet , Peter Zijlstra , Thomas Gleixner , Borislav Petkov , Andy Lutomirski Subject: [PATCH v4 5/6] x86-64: Move vread_tsc into a new file with sensible options Date: Mon, 16 May 2011 12:01:02 -0400 Message-Id: <76102e8ad58c0931d0016da189c6901182379f28.1305560561.git.luto@mit.edu> X-Mailer: git-send-email 1.7.5.1 In-Reply-To: References: In-Reply-To: References: X-Brightmail-Tracker: H4sIAAAAAAAAA+NgFlrKKsWRmVeSWpSXmKPExsUixG6nrtvsddHXYMFnVYuLbRfZLPquHGW3 OHLtO7vFxcZtLBZzzrewWOx7f5bN4vKuOWwWWy41s1ps3jSV2eJR31t2ix8bHrM6cHt0P/rE 6HHlKYfHlpU3mTxutf1h9pi/8yOjx85Zd9k9Nq3qZPN4d+4cu8eJGb9ZPD5vkgvgiuKySUnN ySxLLdK3S+DKWLp7HnvBJK2KfYePsDUwrlHqYuTkkBAwkZjdNJERwhaTuHBvPVsXIxeHkMA+ RolDHX/YQBJCAhsYJSYul4CwnzFJnP5cBmKzCahIdCx9wNTFyMEhIiAksfRuHUgvs8BaZol9 3YuZQGqEBUIkOvefB5vDIqAq8XDjG2aQel6BIInlfdkQexUkrlyZxwJicwoYSGw/vIUZYpW+ RNea/4y4xCcwCixgZFjFKJuSW6Wbm5iZU5yarFucnJiXl1qka6qXm1mil5pSuokRHAMuSjsY fx5UOsQowMGoxMP7x+mirxBrYllxZe4hRkkOJiVRXgNXoBBfUn5KZUZicUZ8UWlOavEhRgkO ZiUR3jP1F3yFeFMSK6tSi/JhUtIcLErivHMk1X2FBNITS1KzU1MLUotgsjIcHEoSvKs9gYYK FqWmp1akZeaUIKSZODhBhvMADb8AUsNbXJCYW5yZDpE/xajL8fnRoQOMQix5+XmpUuK8O0CK BECKMkrz4ObAUtcrRnGgt4R5j4FU8QDTHtykV0BLmICWrDoF8kFxSSJCSqqBMeN99DyDn0/a nKL5l11O4TJas+XmI5Uji6W+OK/ateOqnNPDc4yd8w91OaVs1Ir3MMjoWnA3O1OGISbmwaK0 Is71jH5eJaa2nrucfk3rm7iNeXEEj8mVxuV58S/FWuQyLwReMuxV/pictZznnzRbi4DYMQ8F hRnPRR7l5V8yCGUVrTrqXKCnxFKckWioxVxUnAgABaxiEDgDAAA= Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org vread_tsc is short and hot, and it's userspace code so the usual reasons to enable -pg and turn off sibling calls don't apply. (OK, turning off sibling calls has no effect. But it might someday...) As an added benefit, tsc.c is profilable now. Signed-off-by: Andy Lutomirski --- arch/x86/include/asm/tsc.h | 4 ++++ arch/x86/kernel/Makefile | 8 +++++--- arch/x86/kernel/tsc.c | 34 ---------------------------------- arch/x86/kernel/vread_tsc_64.c | 36 ++++++++++++++++++++++++++++++++++++ 4 files changed, 45 insertions(+), 37 deletions(-) create mode 100644 arch/x86/kernel/vread_tsc_64.c diff --git a/arch/x86/include/asm/tsc.h b/arch/x86/include/asm/tsc.h index 1ca132f..8f2b3c6 100644 --- a/arch/x86/include/asm/tsc.h +++ b/arch/x86/include/asm/tsc.h @@ -51,6 +51,10 @@ extern int unsynchronized_tsc(void); extern int check_tsc_unstable(void); extern unsigned long native_calibrate_tsc(void); +#ifdef CONFIG_X86_64 +extern cycles_t vread_tsc(void); +#endif + /* * Boot-time check whether the TSCs are synchronized across * all CPUs/cores: diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile index 34244b2..7ea1aa2 100644 --- a/arch/x86/kernel/Makefile +++ b/arch/x86/kernel/Makefile @@ -8,7 +8,6 @@ CPPFLAGS_vmlinux.lds += -U$(UTS_MACHINE) ifdef CONFIG_FUNCTION_TRACER # Do not profile debug and lowlevel utilities -CFLAGS_REMOVE_tsc.o = -pg CFLAGS_REMOVE_rtc.o = -pg CFLAGS_REMOVE_paravirt-spinlocks.o = -pg CFLAGS_REMOVE_pvclock.o = -pg @@ -24,13 +23,16 @@ endif nostackp := $(call cc-option, -fno-stack-protector) CFLAGS_vsyscall_64.o := $(PROFILING) -g0 $(nostackp) CFLAGS_hpet.o := $(nostackp) -CFLAGS_tsc.o := $(nostackp) +CFLAGS_vread_tsc_64.o := $(nostackp) CFLAGS_paravirt.o := $(nostackp) GCOV_PROFILE_vsyscall_64.o := n GCOV_PROFILE_hpet.o := n GCOV_PROFILE_tsc.o := n GCOV_PROFILE_paravirt.o := n +# vread_tsc_64 is hot and should be fully optimized: +CFLAGS_REMOVE_vread_tsc_64.o = -pg -fno-optimize-sibling-calls + obj-y := process_$(BITS).o signal.o entry_$(BITS).o obj-y += traps.o irq.o irq_$(BITS).o dumpstack_$(BITS).o obj-y += time.o ioport.o ldt.o dumpstack.o @@ -39,7 +41,7 @@ obj-$(CONFIG_IRQ_WORK) += irq_work.o obj-$(CONFIG_X86_32) += probe_roms_32.o obj-$(CONFIG_X86_32) += sys_i386_32.o i386_ksyms_32.o obj-$(CONFIG_X86_64) += sys_x86_64.o x8664_ksyms_64.o -obj-$(CONFIG_X86_64) += syscall_64.o vsyscall_64.o +obj-$(CONFIG_X86_64) += syscall_64.o vsyscall_64.o vread_tsc_64.o obj-y += bootflag.o e820.o obj-y += pci-dma.o quirks.o i8237.o topology.o kdebugfs.o obj-y += alternative.o i8253.o pci-nommu.o hw_breakpoint.o diff --git a/arch/x86/kernel/tsc.c b/arch/x86/kernel/tsc.c index db4c6e6..5346381 100644 --- a/arch/x86/kernel/tsc.c +++ b/arch/x86/kernel/tsc.c @@ -763,40 +763,6 @@ static cycle_t read_tsc(struct clocksource *cs) ret : clocksource_tsc.cycle_last; } -#ifdef CONFIG_X86_64 -static cycle_t __vsyscall_fn vread_tsc(void) -{ - cycle_t ret; - u64 last; - - /* - * Empirically, a fence (of type that depends on the CPU) - * before rdtsc is enough to ensure that rdtsc is ordered - * with respect to loads. The various CPU manuals are unclear - * as to whether rdtsc can be reordered with later loads, - * but no one has ever seen it happen. - */ - rdtsc_barrier(); - ret = (cycle_t)vget_cycles(); - - last = VVAR(vsyscall_gtod_data).clock.cycle_last; - - if (likely(ret >= last)) - return ret; - - /* - * GCC likes to generate cmov here, but this branch is extremely - * predictable (it's just a funciton of time and the likely is - * very likely) and there's a data dependence, so force GCC - * to generate a branch instead. I don't barrier() because - * we don't actually need a barrier, and if this function - * ever gets inlined it will generate worse code. - */ - asm volatile (""); - return last; -} -#endif - static void resume_tsc(struct clocksource *cs) { clocksource_tsc.cycle_last = 0; diff --git a/arch/x86/kernel/vread_tsc_64.c b/arch/x86/kernel/vread_tsc_64.c new file mode 100644 index 0000000..a81aa9e --- /dev/null +++ b/arch/x86/kernel/vread_tsc_64.c @@ -0,0 +1,36 @@ +/* This code runs in userspace. */ + +#define DISABLE_BRANCH_PROFILING +#include + +notrace cycle_t __vsyscall_fn vread_tsc(void) +{ + cycle_t ret; + u64 last; + + /* + * Empirically, a fence (of type that depends on the CPU) + * before rdtsc is enough to ensure that rdtsc is ordered + * with respect to loads. The various CPU manuals are unclear + * as to whether rdtsc can be reordered with later loads, + * but no one has ever seen it happen. + */ + rdtsc_barrier(); + ret = (cycle_t)vget_cycles(); + + last = VVAR(vsyscall_gtod_data).clock.cycle_last; + + if (likely(ret >= last)) + return ret; + + /* + * GCC likes to generate cmov here, but this branch is extremely + * predictable (it's just a funciton of time and the likely is + * very likely) and there's a data dependence, so force GCC + * to generate a branch instead. I don't barrier() because + * we don't actually need a barrier, and if this function + * ever gets inlined it will generate worse code. + */ + asm volatile (""); + return last; +} -- 1.7.5.1