Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752225Ab0HTIJI (ORCPT ); Fri, 20 Aug 2010 04:09:08 -0400 Received: from mx1.redhat.com ([209.132.183.28]:2568 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752150Ab0HTIIv (ORCPT ); Fri, 20 Aug 2010 04:08:51 -0400 From: Zachary Amsden To: kvm@vger.kernel.org Cc: Zachary Amsden , Avi Kivity , Marcelo Tosatti , Glauber Costa , Thomas Gleixner , John Stultz , linux-kernel@vger.kernel.org Subject: [KVM timekeeping 20/35] Make math work for other scales Date: Thu, 19 Aug 2010 22:07:34 -1000 Message-Id: <1282291669-25709-21-git-send-email-zamsden@redhat.com> In-Reply-To: <1282291669-25709-1-git-send-email-zamsden@redhat.com> References: <1282291669-25709-1-git-send-email-zamsden@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2772 Lines: 86 The math in kvm_get_time_scale relies on the fact that NSEC_PER_SEC < 2^32. To use the same function to compute arbitrary time scales, we must extend the first reduction step to shrink the base rate to a 32-bit value, and possibly reduce the scaled rate into a 32-bit as well. Note we must take care to avoid an arithmetic overflow when scaling up the tps32 value (this could not happen with the fixed scaled value of NSEC_PER_SEC, but can happen with scaled rates above 2^31. Signed-off-by: Zachary Amsden --- arch/x86/kvm/x86.c | 30 ++++++++++++++++++------------ 1 files changed, 18 insertions(+), 12 deletions(-) diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index ce03c2c..718bed9 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -866,31 +866,35 @@ static uint32_t div_frac(uint32_t dividend, uint32_t divisor) return quotient; } -static void kvm_set_time_scale(uint32_t tsc_khz, struct pvclock_vcpu_time_info *hv_clock) +static void kvm_get_time_scale(uint32_t scaled_khz, uint32_t base_khz, + s8 *pshift, u32 *pmultiplier) { - uint64_t nsecs = 1000000000LL; + uint64_t scaled64; int32_t shift = 0; uint64_t tps64; uint32_t tps32; - tps64 = tsc_khz * 1000LL; - while (tps64 > nsecs*2) { + tps64 = base_khz * 1000LL; + scaled64 = scaled_khz * 1000LL; + while (tps64 > scaled64*2 || tps64 & 0xffffffff00000000UL) { tps64 >>= 1; shift--; } tps32 = (uint32_t)tps64; - while (tps32 <= (uint32_t)nsecs) { - tps32 <<= 1; + while (tps32 <= scaled64 || scaled64 & 0xffffffff00000000UL) { + if (scaled64 & 0xffffffff00000000UL || tps32 & 0x80000000) + scaled64 >>= 1; + else + tps32 <<= 1; shift++; } - hv_clock->tsc_shift = shift; - hv_clock->tsc_to_system_mul = div_frac(nsecs, tps32); + *pshift = shift; + *pmultiplier = div_frac(scaled64, tps32); - pr_debug("%s: tsc_khz %u, tsc_shift %d, tsc_mul %u\n", - __func__, tsc_khz, hv_clock->tsc_shift, - hv_clock->tsc_to_system_mul); + pr_debug("%s: base_khz %u => %u, shift %d, mul %u\n", + __func__, base_khz, scaled_khz, shift, *pmultiplier); } static DEFINE_PER_CPU(unsigned long, cpu_tsc_khz); @@ -1016,7 +1020,9 @@ static int kvm_write_guest_time(struct kvm_vcpu *v) } if (unlikely(vcpu->hw_tsc_khz != this_tsc_khz)) { - kvm_set_time_scale(this_tsc_khz, &vcpu->hv_clock); + kvm_get_time_scale(NSEC_PER_SEC / 1000, this_tsc_khz, + &vcpu->hv_clock.tsc_shift, + &vcpu->hv_clock.tsc_to_system_mul); vcpu->hw_tsc_khz = this_tsc_khz; } -- 1.7.1 -- 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/