From patchwork Sat Sep 14 10:17:23 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Dapeng Mi X-Patchwork-Id: 13804292 Received: from mgamail.intel.com (mgamail.intel.com [198.175.65.11]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id A0CE21D049D; Sat, 14 Sep 2024 07:01:46 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=198.175.65.11 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1726297308; cv=none; b=gpNLKkVRlB92mO9FRx4lnI3CKnNjL8C0iraU4t+TJl8XmzbsKGPkFP9C+pPnf6g4UBElmxcHW5/kSvbiqe6yyJYm3f2emEZtY6Y+inB2Qj22HZNCgaTHt1j45iq5YBeaitAniyOnkhmn9e99ChlH4ws+iX/cmVbShRWdMZiCa14= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1726297308; c=relaxed/simple; bh=Ih04t7D/0iaLudr3E/C/IlUJ8oePrTKqbytNL2ob+RE=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=OeSR94SyjS7jqh6tlzu7kBWR/BR3aBFubYOnCF+2RwJT8POJV3eimuFubvz//SswT3106vrotenJhc3/4rHY1in5AORyyS2whAAztxydn4mxSKBFxsnRj+Pxbd2wlaqwifJHKl8WekFEGOmifLBCQUxa/t4+QQZ5GmrChge7wz0= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com; spf=none smtp.mailfrom=linux.intel.com; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b=HaxM0CtN; arc=none smtp.client-ip=198.175.65.11 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; spf=none smtp.mailfrom=linux.intel.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=intel.com header.i=@intel.com header.b="HaxM0CtN" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1726297307; x=1757833307; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=Ih04t7D/0iaLudr3E/C/IlUJ8oePrTKqbytNL2ob+RE=; b=HaxM0CtNveJljxjyw4+W+QF7fK+uFLPCYXiP8ut73hutqeuK7idtQKWs DVJ4DuZ9mWPDHgLajnx0gV1tCoXVJ/ZpPMu34Mut/3ui40JzVlWIkndee jBZKvEbErxQzyhGvIU2kKC9Ae+Mv9ArbIdlgkRkd+DdiBXlGSI4lTpyCy reF6f27BWR0IqokpKv88BctktvpKvFeDOdc8N6VTyFcu1gT39KWAVJ/SJ n6mx3fp+92dfAC02cYoiNnYKFxVOPe1WLOfWRThwdCBKa6SAoQ1PlNekx sKHJQ6AUTOrSGmzDI4M0iT2Ym+b0kpJ3X6gAOhHcPRzTj0oeS/IynPk17 g==; X-CSE-ConnectionGUID: wEFe2iz2SKe4mHDfdNT2CA== X-CSE-MsgGUID: tfKzfjC0R3aSx5kxW83TuQ== X-IronPort-AV: E=McAfee;i="6700,10204,11194"; a="35778844" X-IronPort-AV: E=Sophos;i="6.10,228,1719903600"; d="scan'208";a="35778844" Received: from fmviesa006.fm.intel.com ([10.60.135.146]) by orvoesa103.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 14 Sep 2024 00:01:46 -0700 X-CSE-ConnectionGUID: DqgFYJEGQj+a1W0wwMSUuQ== X-CSE-MsgGUID: XZVUL+iFStWvfLybBSp6LQ== X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="6.10,228,1719903600"; d="scan'208";a="67950986" Received: from emr.sh.intel.com ([10.112.229.56]) by fmviesa006.fm.intel.com with ESMTP; 14 Sep 2024 00:01:43 -0700 From: Dapeng Mi To: Sean Christopherson , Paolo Bonzini Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org, Jim Mattson , Mingwei Zhang , Xiong Zhang , Zhenyu Wang , Like Xu , Jinrong Liang , Yongwei Ma , Dapeng Mi , Dapeng Mi Subject: [kvm-unit-tests patch v6 13/18] x86: pmu: Improve instruction and branches events verification Date: Sat, 14 Sep 2024 10:17:23 +0000 Message-Id: <20240914101728.33148-14-dapeng1.mi@linux.intel.com> X-Mailer: git-send-email 2.40.1 In-Reply-To: <20240914101728.33148-1-dapeng1.mi@linux.intel.com> References: <20240914101728.33148-1-dapeng1.mi@linux.intel.com> Precedence: bulk X-Mailing-List: kvm@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 If HW supports GLOBAL_CTRL MSR, enabling and disabling PMCs are moved in __precise_count_loop(). Thus, instructions and branches events can be verified against a precise count instead of a rough range. BTW, some intermittent failures on AMD processors using PerfMonV2 is seen due to variance in counts. This probably has to do with the way instructions leading to a VM-Entry or VM-Exit are accounted when counting retired instructions and branches. https://lore.kernel.org/all/6d512a14-ace1-41a3-801e-0beb41425734@amd.com/ So only enable this precise check for Intel processors. Signed-off-by: Dapeng Mi --- x86/pmu.c | 37 +++++++++++++++++++++++++++++++++++++ 1 file changed, 37 insertions(+) diff --git a/x86/pmu.c b/x86/pmu.c index 270f11b9..13c7c45d 100644 --- a/x86/pmu.c +++ b/x86/pmu.c @@ -19,6 +19,11 @@ #define EXPECTED_INSTR 17 #define EXPECTED_BRNCH 5 + +/* Enable GLOBAL_CTRL + disable GLOBAL_CTRL instructions */ +#define EXTRA_INSTRNS (3 + 3) +#define LOOP_INSTRNS (N * 10 + EXTRA_INSTRNS) +#define LOOP_BRANCHES (N) #define LOOP_ASM(_wrmsr) \ _wrmsr "\n\t" \ "mov %%ecx, %%edi; mov %%ebx, %%ecx;\n\t" \ @@ -123,6 +128,30 @@ static inline void loop(u64 cntrs) __precise_loop(cntrs); } +static void adjust_events_range(struct pmu_event *gp_events, + int instruction_idx, int branch_idx) +{ + /* + * If HW supports GLOBAL_CTRL MSR, enabling and disabling PMCs are + * moved in __precise_loop(). Thus, instructions and branches events + * can be verified against a precise count instead of a rough range. + * + * We see some intermittent failures on AMD processors using PerfMonV2 + * due to variance in counts. This probably has to do with the way + * instructions leading to a VM-Entry or VM-Exit are accounted when + * counting retired instructions and branches. Thus only enable the + * precise validation for Intel processors. + */ + if (pmu.is_intel && this_cpu_has_perf_global_ctrl()) { + /* instructions event */ + gp_events[instruction_idx].min = LOOP_INSTRNS; + gp_events[instruction_idx].max = LOOP_INSTRNS; + /* branches event */ + gp_events[branch_idx].min = LOOP_BRANCHES; + gp_events[branch_idx].max = LOOP_BRANCHES; + } +} + volatile uint64_t irq_received; static void cnt_overflow(isr_regs_t *regs) @@ -832,6 +861,9 @@ static void check_invalid_rdpmc_gp(void) int main(int ac, char **av) { + int instruction_idx; + int branch_idx; + setup_vm(); handle_irq(PMI_VECTOR, cnt_overflow); buf = malloc(N*64); @@ -845,13 +877,18 @@ int main(int ac, char **av) } gp_events = (struct pmu_event *)intel_gp_events; gp_events_size = sizeof(intel_gp_events)/sizeof(intel_gp_events[0]); + instruction_idx = INTEL_INSTRUCTIONS_IDX; + branch_idx = INTEL_BRANCHES_IDX; report_prefix_push("Intel"); set_ref_cycle_expectations(); } else { gp_events_size = sizeof(amd_gp_events)/sizeof(amd_gp_events[0]); gp_events = (struct pmu_event *)amd_gp_events; + instruction_idx = AMD_INSTRUCTIONS_IDX; + branch_idx = AMD_BRANCHES_IDX; report_prefix_push("AMD"); } + adjust_events_range(gp_events, instruction_idx, branch_idx); printf("PMU version: %d\n", pmu.version); printf("GP counters: %d\n", pmu.nr_gp_counters);