From mboxrd@z Thu Jan  1 00:00:00 1970
From: Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [PATCH 04/12] KVM: x86: Replace call-back set_tsc_khz() with a
 common function
Date: Tue, 6 Oct 2015 12:40:49 +0200
Message-ID: <5613A531.5050506@redhat.com>
References: <1443418691-24050-1-git-send-email-haozhong.zhang@intel.com>
 <1443418691-24050-5-git-send-email-haozhong.zhang@intel.com>
 <20151005195325.GA4508@potion.brq.redhat.com>
 <20151006040647.GD3798@hzzhang-OptiPlex-9020.sh.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
To: =?UTF-8?B?UmFkaW0gS3LEjW3DocWZ?= <rkrcmar@redhat.com>,
	David Matlack <dmatlack@google.com>, kvm@vger.kernel.org,
	Gleb Natapov <gleb@kernel.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	Ingo Molnar <mingo@redhat.com>,
	"H. Peter Anvin" <hpa@zytor.com>, x86@kernel.org,
	Joerg Roedel <joro@8bytes.org>,
	Wanpeng Li <wanpeng.li@linux.intel.com>,
	Xiao Guangrong <guangrong.xiao@linux.intel.com>,
	=?UTF-8?Q?Mihai_Don=c8=9bu?= <mdontu@bitdefender.com>,
	Andy Lutomirski <luto@kernel.org>,
	Kai Huang <kai.huang@linux.intel.com>,
	linux-kernel@vger.kernel.org
Return-path: <linux-kernel-owner@vger.kernel.org>
In-Reply-To: <20151006040647.GD3798@hzzhang-OptiPlex-9020.sh.intel.com>
Sender: linux-kernel-owner@vger.kernel.org
List-Id: kvm.vger.kernel.org



On 06/10/2015 06:06, Haozhong Zhang wrote:
> Alternatively, it's also possible to follow David's comment to use
> divq on x86_64 to keep both precision and safety. On i386, it just
> falls back to above truncating approach.

khz is just 32 bits, so we can do a 96/32 division.  And because this is
a slow path, we can code a generic u64*u32/u32 function and use it to do
(1 << kvm_tsc_scaling_ratio_frac_bits) * khz / tsc_khz:

diff --git a/include/linux/math64.h b/include/linux/math64.h
index c45c089bfdac..5b70af4fa386 100644
--- a/include/linux/math64.h
+++ b/include/linux/math64.h
@@ -142,6 +142,13 @@ static inline u64 mul_u64_u32_shr(u64 a, u32 mul,
unsigned int shift)
 }
 #endif /* mul_u64_u32_shr */

+#ifndef mul_u64_u32_div
+static inline u64 mul_u64_u32_div(u64 x, u32 num, u32 den)
+{
+	return (u64)(((unsigned __int128)a * mul) / den);
+}
+#endif
+
 #else

 #ifndef mul_u64_u32_shr
@@ -161,6 +168,35 @@ static inline u64 mul_u64_u32_shr(u64 a, u32 mul,
unsigned int shift)
 }
 #endif /* mul_u64_u32_shr */

+#ifndef mul_u64_u32_div
+static inline u64 mul_u64_u32_div(u64 a, u32 num, u32 den)
+{
+	union {
+		u64 ll;
+		struct {
+#ifdef __BIG_ENDIAN
+			u32 high, low;
+#else
+			u32 low, high;
+#endif
+		} l;
+	} u, rl, rh;
+
+	u.ll = a;
+	rl.ll = (u64)u.l.low * num;
+	rh.ll = (u64)u.l.high * num + rl.l.high;
+
+	/* Bits 32-63 of the result will be in rh.l.low.  */
+	rl.l.high = do_div(rh.ll, den);
+
+	/* Bits 0-31 of the result will be in rl.l.low.  */
+	do_div(rl.ll, den);
+
+	rl.l.high = rh.l.low;
+	return rl.ll;
+}
+#endif
+
 #endif

 #endif /* _LINUX_MATH64_H */