From patchwork Mon Jan 28 11:50:25 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Julien Grall X-Patchwork-Id: 156720 Delivered-To: patch@linaro.org Received: by 2002:a02:48:0:0:0:0:0 with SMTP id 69csp3333351jaa; Mon, 28 Jan 2019 03:52:32 -0800 (PST) X-Google-Smtp-Source: ALg8bN40wIu4XFCTH7sTIJp10tD437osVspYL+fjy0bEGOqejjKLRoodzxsfYaRyPF49OCIj+7Sz X-Received: by 2002:a25:7308:: with SMTP id o8mr20238370ybc.517.1548676352576; Mon, 28 Jan 2019 03:52:32 -0800 (PST) ARC-Seal: i=1; a=rsa-sha256; t=1548676352; cv=none; d=google.com; s=arc-20160816; b=gMvrca60115+dNtKUQ3pWfsq2K6Wtrewc1MEi36Dg2Az7B7nUBD+Li8AGrFNzEdi4u DRron1Kdfctgmt4zotyZP0UP93OLlsTCvvOdHd7tpA5dK4+T6iCZBrcjlsujgzwpF71U CclL8+ilIAl9PWfz69CbZPteH1GYvu31bdonsiNkUanDNpQPeGPjWE9Bbl9f2TM072SZ TNlmk4YQJAZcwlwDIliWjds8uIOguUg3OR8SM/qqChT7d5/5dERDFP6N0ou1lAwmHbh7 jTC8cwJLt1xy3H+sKsdE6YLiEDx3O/yCDDnIcV3TG0RYmdkrxSP8XQgk/s4qHeKm39SE f7sg== ARC-Message-Signature: i=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=arc-20160816; h=sender:errors-to:content-transfer-encoding:mime-version:cc :list-subscribe:list-help:list-post:list-unsubscribe:list-id :precedence:subject:references:in-reply-to:message-id:date:to:from; bh=ryWKBOolETSWduJrHH2rWe/D63aZ66WoB+lBgKTW090=; b=uTLF02yZE4ywWAIIceZiD4IVW5JADbGDdxEQ8Z2fgGKSKh83ZX+G7GarH8KOzU8/Yd l3EZQOkByz9YjNsvIjt8JPdC9SgJi7Nvhemj5Gz665PYIIrfZLlM8WkDOtIy9UIfQBoS xOhJi0dm6RqapuMGMmwEwapq07E//icRtYBZBPaq9NCxZC8G2OuM8HECjT2U5/XfrJV5 F2igT8Cw+w2F3AJnAeMOPda3e5b9jqahXrc8B+IbqShAgCzCfN5fcor96E+lcWZerfNf s5kEn/Io9b1CAntTWb8g0D/vfQKIC1U8ObSWnDvLF/HiJEdKLek1FZH3D8lC2GkiG6In uQMQ== ARC-Authentication-Results: i=1; mx.google.com; spf=pass (google.com: best guess record for domain of xen-devel-bounces@lists.xenproject.org designates 192.237.175.120 as permitted sender) smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Return-Path: Received: from lists.xenproject.org (lists.xenproject.org. [192.237.175.120]) by mx.google.com with ESMTPS id a188si19975535ywe.369.2019.01.28.03.52.32 (version=TLS1_2 cipher=ECDHE-RSA-CHACHA20-POLY1305 bits=256/256); Mon, 28 Jan 2019 03:52:32 -0800 (PST) Received-SPF: pass (google.com: best guess record for domain of xen-devel-bounces@lists.xenproject.org designates 192.237.175.120 as permitted sender) client-ip=192.237.175.120; Authentication-Results: mx.google.com; spf=pass (google.com: best guess record for domain of xen-devel-bounces@lists.xenproject.org designates 192.237.175.120 as permitted sender) smtp.mailfrom=xen-devel-bounces@lists.xenproject.org Received: from localhost ([127.0.0.1] helo=lists.xenproject.org) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1go5R4-00037g-H7; Mon, 28 Jan 2019 11:50:42 +0000 Received: from all-amaz-eas1.inumbo.com ([34.197.232.57] helo=us1-amaz-eas2.inumbo.com) by lists.xenproject.org with esmtp (Exim 4.89) (envelope-from ) id 1go5R2-000369-MI for xen-devel@lists.xenproject.org; Mon, 28 Jan 2019 11:50:40 +0000 X-Inumbo-ID: eec673ae-22f2-11e9-a5d2-f7d37ba7b507 Received: from foss.arm.com (unknown [217.140.101.70]) by us1-amaz-eas2.inumbo.com (Halon) with ESMTP id eec673ae-22f2-11e9-a5d2-f7d37ba7b507; Mon, 28 Jan 2019 11:50:39 +0000 (UTC) Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.72.51.249]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id 3FA81EBD; Mon, 28 Jan 2019 03:50:39 -0800 (PST) Received: from e108454-lin.cambridge.arm.com (e108454-lin.cambridge.arm.com [10.1.196.50]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPSA id 267203F59C; Mon, 28 Jan 2019 03:50:38 -0800 (PST) From: Julien Grall To: xen-devel@lists.xenproject.org Date: Mon, 28 Jan 2019 11:50:25 +0000 Message-Id: <20190128115026.3336-7-julien.grall@arm.com> X-Mailer: git-send-email 2.11.0 In-Reply-To: <20190128115026.3336-1-julien.grall@arm.com> References: <20190128115026.3336-1-julien.grall@arm.com> Subject: [Xen-devel] [PATCH for-4.12 v2 6/7] xen/arm: Implement workaround for Cortex-A76 erratum 1165522 X-BeenThere: xen-devel@lists.xenproject.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: Xen developer discussion List-Unsubscribe: , List-Post: List-Help: List-Subscribe: , Cc: Julien Grall , sstabellini@kernel.org, Andrii Anisov MIME-Version: 1.0 Errors-To: xen-devel-bounces@lists.xenproject.org Sender: "Xen-devel" Early version of Cortex-A76 can end-up with corrupt TLBs if they speculate an AT instruction while the S1/S2 system registers are in an inconsistent state. This can happen during guest context switch and when invalidating the TLBs for other than the current VMID. The workaround implemented in Xen will: - Use an empty stage-2 with a reserved VMID while context switching between 2 guests - Use an empty stage-2 with the VMID where TLBs need to be flushed Signed-off-by: Julien Grall Reviewed-by: Andrii Anisov --- Changes in v2: - Fix typo - Add Andrii's reviewed-by --- docs/misc/arm/silicon-errata.txt | 1 + xen/arch/arm/cpuerrata.c | 6 ++++ xen/arch/arm/domain.c | 8 +++-- xen/arch/arm/p2m.c | 78 ++++++++++++++++++++++++++++++++++++++-- xen/include/asm-arm/cpufeature.h | 3 +- xen/include/asm-arm/processor.h | 2 ++ 6 files changed, 93 insertions(+), 5 deletions(-) diff --git a/docs/misc/arm/silicon-errata.txt b/docs/misc/arm/silicon-errata.txt index 906bf5fd48..6cd1366f15 100644 --- a/docs/misc/arm/silicon-errata.txt +++ b/docs/misc/arm/silicon-errata.txt @@ -48,4 +48,5 @@ stable hypervisors. | ARM | Cortex-A57 | #852523 | N/A | | ARM | Cortex-A57 | #832075 | ARM64_ERRATUM_832075 | | ARM | Cortex-A57 | #834220 | ARM64_ERRATUM_834220 | +| ARM | Cortex-A76 | #1165522 | N/A | | ARM | MMU-500 | #842869 | N/A | diff --git a/xen/arch/arm/cpuerrata.c b/xen/arch/arm/cpuerrata.c index f4815cafb4..4431b244fd 100644 --- a/xen/arch/arm/cpuerrata.c +++ b/xen/arch/arm/cpuerrata.c @@ -489,6 +489,12 @@ static const struct arm_cpu_capabilities arm_errata[] = { .matches = has_ssbd_mitigation, }, #endif + { + /* Cortex-A76 r0p0 - r2p0 */ + .desc = "ARM erratum 116522", + .capability = ARM64_WORKAROUND_AT_SPECULATE, + MIDR_RANGE(MIDR_CORTEX_A76, 0, 2 << MIDR_VARIANT_SHIFT), + }, {}, }; diff --git a/xen/arch/arm/domain.c b/xen/arch/arm/domain.c index 41f101746e..6dc633ed50 100644 --- a/xen/arch/arm/domain.c +++ b/xen/arch/arm/domain.c @@ -181,8 +181,6 @@ static void ctxt_switch_to(struct vcpu *n) if ( is_idle_vcpu(n) ) return; - p2m_restore_state(n); - vpidr = READ_SYSREG32(MIDR_EL1); WRITE_SYSREG32(vpidr, VPIDR_EL2); WRITE_SYSREG(n->arch.vmpidr, VMPIDR_EL2); @@ -235,6 +233,12 @@ static void ctxt_switch_to(struct vcpu *n) #endif isb(); + /* + * ARM64_WORKAROUND_AT_SPECULATE: The P2M should be restored after + * the stage-1 MMU sysregs have been restored. + */ + p2m_restore_state(n); + /* Control Registers */ WRITE_SYSREG(n->arch.cpacr, CPACR_EL1); diff --git a/xen/arch/arm/p2m.c b/xen/arch/arm/p2m.c index 44391a5f8c..c38bd7e16e 100644 --- a/xen/arch/arm/p2m.c +++ b/xen/arch/arm/p2m.c @@ -5,6 +5,7 @@ #include #include +#include #include #include #include @@ -41,6 +42,8 @@ static const paddr_t level_masks[] = static const uint8_t level_orders[] = { ZEROETH_ORDER, FIRST_ORDER, SECOND_ORDER, THIRD_ORDER }; +static mfn_t __read_mostly empty_root_mfn; + static uint64_t generate_vttbr(uint16_t vmid, mfn_t root_mfn) { return (mfn_to_maddr(root_mfn) | ((uint64_t)vmid << 48)); @@ -92,9 +95,25 @@ void dump_p2m_lookup(struct domain *d, paddr_t addr) P2M_ROOT_LEVEL, P2M_ROOT_PAGES); } +/* + * p2m_save_state and p2m_restore_state work in pair to workaround + * ARM64_WORKAROUND_AT_SPECULATE. p2m_save_state will set-up VTTBR to + * point to the empty page-tables to stop allocating TLB entries. + */ void p2m_save_state(struct vcpu *p) { p->arch.sctlr = READ_SYSREG(SCTLR_EL1); + + if ( cpus_have_const_cap(ARM64_WORKAROUND_AT_SPECULATE) ) + { + WRITE_SYSREG64(generate_vttbr(INVALID_VMID, empty_root_mfn), VTTBR_EL2); + /* + * Ensure VTTBR_EL2 is correctly synchronized so we can restore + * the next vCPU context without worrying about AT instruction + * speculation. + */ + isb(); + } } void p2m_restore_state(struct vcpu *n) @@ -105,10 +124,17 @@ void p2m_restore_state(struct vcpu *n) if ( is_idle_vcpu(n) ) return; - WRITE_SYSREG64(p2m->vttbr, VTTBR_EL2); WRITE_SYSREG(n->arch.sctlr, SCTLR_EL1); WRITE_SYSREG(n->arch.hcr_el2, HCR_EL2); + /* + * ARM64_WORKAROUND_AT_SPECULATE: VTTBR_EL2 should be restored after all + * registers associated to EL1/EL0 translations regime have been + * synchronized. + */ + asm volatile(ALTERNATIVE("nop", "isb", ARM64_WORKAROUND_AT_SPECULATE)); + WRITE_SYSREG64(p2m->vttbr, VTTBR_EL2); + last_vcpu_ran = &p2m->last_vcpu_ran[smp_processor_id()]; /* @@ -149,8 +175,23 @@ static void p2m_force_tlb_flush_sync(struct p2m_domain *p2m) ovttbr = READ_SYSREG64(VTTBR_EL2); if ( ovttbr != p2m->vttbr ) { + uint64_t vttbr; + local_irq_save(flags); - WRITE_SYSREG64(p2m->vttbr, VTTBR_EL2); + + /* + * ARM64_WORKAROUND_AT_SPECULATE: We need to stop AT to allocate + * TLBs entries because the context is partially modified. We + * only need the VMID for flushing the TLBs, so we can generate + * a new VTTBR with the VMID to flush and the empty root table. + */ + if ( !cpus_have_const_cap(ARM64_WORKAROUND_AT_SPECULATE) ) + vttbr = p2m->vttbr; + else + vttbr = generate_vttbr(p2m->vmid, empty_root_mfn); + + WRITE_SYSREG64(vttbr, VTTBR_EL2); + /* Ensure VTTBR_EL2 is synchronized before flushing the TLBs */ isb(); } @@ -1913,6 +1954,23 @@ static uint32_t __read_mostly vtcr; static void setup_virt_paging_one(void *data) { WRITE_SYSREG32(vtcr, VTCR_EL2); + + /* + * ARM64_WORKAROUND_AT_SPECULATE: We want to keep the TLBs free from + * entries related to EL1/EL0 translation regime until a guest vCPU + * is running. For that, we need to set-up VTTBR to point to an empty + * page-table and turn on stage-2 translation. The TLB entries + * associated with EL1/EL0 translation regime will also be flushed in case + * an AT instruction was speculated before hand. + */ + if ( cpus_have_cap(ARM64_WORKAROUND_AT_SPECULATE) ) + { + WRITE_SYSREG64(generate_vttbr(INVALID_VMID, empty_root_mfn), VTTBR_EL2); + WRITE_SYSREG(READ_SYSREG(HCR_EL2) | HCR_VM, HCR_EL2); + isb(); + + flush_tlb_all_local(); + } } void __init setup_virt_paging(void) @@ -1996,6 +2054,22 @@ void __init setup_virt_paging(void) /* It is not allowed to concatenate a level zero root */ BUG_ON( P2M_ROOT_LEVEL == 0 && P2M_ROOT_ORDER > 0 ); vtcr = val; + + /* + * ARM64_WORKAROUND_AT_SPECULATE requires to allocate root table + * with all entries zeroed. + */ + if ( cpus_have_cap(ARM64_WORKAROUND_AT_SPECULATE) ) + { + struct page_info *root; + + root = p2m_allocate_root(); + if ( !root ) + panic("Unable to allocate root table for ARM64_WORKAROUND_AT_SPECULATE\n"); + + empty_root_mfn = page_to_mfn(root); + } + setup_virt_paging_one(NULL); smp_call_function(setup_virt_paging_one, NULL, 1); } diff --git a/xen/include/asm-arm/cpufeature.h b/xen/include/asm-arm/cpufeature.h index 17de928467..c2c8f3417c 100644 --- a/xen/include/asm-arm/cpufeature.h +++ b/xen/include/asm-arm/cpufeature.h @@ -45,8 +45,9 @@ #define ARM_HARDEN_BRANCH_PREDICTOR 7 #define ARM_SSBD 8 #define ARM_SMCCC_1_1 9 +#define ARM64_WORKAROUND_AT_SPECULATE 10 -#define ARM_NCAPS 10 +#define ARM_NCAPS 11 #ifndef __ASSEMBLY__ diff --git a/xen/include/asm-arm/processor.h b/xen/include/asm-arm/processor.h index cb781751a6..b5f515805d 100644 --- a/xen/include/asm-arm/processor.h +++ b/xen/include/asm-arm/processor.h @@ -52,6 +52,7 @@ #define ARM_CPU_PART_CORTEX_A72 0xD08 #define ARM_CPU_PART_CORTEX_A73 0xD09 #define ARM_CPU_PART_CORTEX_A75 0xD0A +#define ARM_CPU_PART_CORTEX_A76 0xD0B #define MIDR_CORTEX_A12 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A12) #define MIDR_CORTEX_A17 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A17) @@ -61,6 +62,7 @@ #define MIDR_CORTEX_A72 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A72) #define MIDR_CORTEX_A73 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A73) #define MIDR_CORTEX_A75 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A75) +#define MIDR_CORTEX_A76 MIDR_CPU_MODEL(ARM_CPU_IMP_ARM, ARM_CPU_PART_CORTEX_A76) /* MPIDR Multiprocessor Affinity Register */ #define _MPIDR_UP (30)