From patchwork Mon Feb 12 18:33:35 2018 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Jean-Philippe Brucker X-Patchwork-Id: 10213977 Return-Path: Received: from mail.wl.linuxfoundation.org (pdx-wl-mail.web.codeaurora.org [172.30.200.125]) by pdx-korg-patchwork.web.codeaurora.org (Postfix) with ESMTP id B6A1E60153 for ; Mon, 12 Feb 2018 18:32:58 +0000 (UTC) Received: from mail.wl.linuxfoundation.org (localhost [127.0.0.1]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id ACF7A28AB7 for ; Mon, 12 Feb 2018 18:32:58 +0000 (UTC) Received: by mail.wl.linuxfoundation.org (Postfix, from userid 486) id A180B28AD3; Mon, 12 Feb 2018 18:32:58 +0000 (UTC) X-Spam-Checker-Version: SpamAssassin 3.3.1 (2010-03-16) on pdx-wl-mail.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.9 required=2.0 tests=BAYES_00,RCVD_IN_DNSWL_HI autolearn=unavailable version=3.3.1 Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.wl.linuxfoundation.org (Postfix) with ESMTP id 0C1DF28ADA for ; Mon, 12 Feb 2018 18:32:58 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753952AbeBLSc4 (ORCPT ); Mon, 12 Feb 2018 13:32:56 -0500 Received: from usa-sjc-mx-foss1.foss.arm.com ([217.140.101.70]:46026 "EHLO foss.arm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753932AbeBLScy (ORCPT ); Mon, 12 Feb 2018 13:32:54 -0500 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 122B21435; Mon, 12 Feb 2018 10:32:54 -0800 (PST) Received: from e106794-lin.cambridge.arm.com (e106794-lin.cambridge.arm.com [10.1.210.24]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 34F573F24D; Mon, 12 Feb 2018 10:32:48 -0800 (PST) From: Jean-Philippe Brucker To: linux-arm-kernel@lists.infradead.org, linux-pci@vger.kernel.org, linux-acpi@vger.kernel.org, devicetree@vger.kernel.org, iommu@lists.linux-foundation.org, kvm@vger.kernel.org Cc: joro@8bytes.org, robh+dt@kernel.org, mark.rutland@arm.com, catalin.marinas@arm.com, will.deacon@arm.com, lorenzo.pieralisi@arm.com, hanjun.guo@linaro.org, sudeep.holla@arm.com, rjw@rjwysocki.net, lenb@kernel.org, robin.murphy@arm.com, bhelgaas@google.com, alex.williamson@redhat.com, tn@semihalf.com, liubo95@huawei.com, thunder.leizhen@huawei.com, xieyisheng1@huawei.com, xuzaibo@huawei.com, ilias.apalodimas@linaro.org, jonathan.cameron@huawei.com, shunyong.yang@hxt-semitech.com, nwatters@codeaurora.org, okaya@codeaurora.org, jcrouse@codeaurora.org, rfranz@cavium.com, dwmw2@infradead.org, jacob.jun.pan@linux.intel.com, yi.l.liu@intel.com, ashok.raj@intel.com, robdclark@gmail.com, christian.koenig@amd.com, bharatku@xilinx.com Subject: [PATCH 20/37] iommu/arm-smmu-v3: Share process page tables Date: Mon, 12 Feb 2018 18:33:35 +0000 Message-Id: <20180212183352.22730-21-jean-philippe.brucker@arm.com> X-Mailer: git-send-email 2.15.1 In-Reply-To: <20180212183352.22730-1-jean-philippe.brucker@arm.com> References: <20180212183352.22730-1-jean-philippe.brucker@arm.com> Sender: kvm-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: kvm@vger.kernel.org X-Virus-Scanned: ClamAV using ClamSMTP With Shared Virtual Addressing (SVA), we need to mirror CPU TTBR, TCR, MAIR and ASIDs in SMMU contexts. Each SMMU has a single ASID space split into two sets, shared and private. Shared ASIDs correspond to those obtained from the arch ASID allocator, and private ASIDs are used for "classic" map/unmap DMA. Replace the ASID IDA with an IDR, allowing to keep information about each context. Initialize shared contexts with info obtained from the mm. Signed-off-by: Jean-Philippe Brucker --- drivers/iommu/arm-smmu-v3-context.c | 181 ++++++++++++++++++++++++++++++++++-- 1 file changed, 171 insertions(+), 10 deletions(-) diff --git a/drivers/iommu/arm-smmu-v3-context.c b/drivers/iommu/arm-smmu-v3-context.c index aaffc2071966..b7c90384ff56 100644 --- a/drivers/iommu/arm-smmu-v3-context.c +++ b/drivers/iommu/arm-smmu-v3-context.c @@ -10,9 +10,11 @@ #include #include #include +#include #include #include "iommu-pasid.h" +#include "io-pgtable-arm.h" /* * Linear: when less than 1024 SSIDs are supported @@ -87,6 +89,9 @@ struct arm_smmu_cd { u64 ttbr; u64 tcr; u64 mair; + + refcount_t refs; + struct mm_struct *mm; }; #define pasid_entry_to_cd(entry) \ @@ -118,7 +123,8 @@ struct arm_smmu_cd_tables { #define pasid_ops_to_tables(ops) \ pasid_to_cd_tables(iommu_pasid_table_ops_to_table(ops)) -static DEFINE_IDA(asid_ida); +static DEFINE_SPINLOCK(asid_lock); +static DEFINE_IDR(asid_idr); static int arm_smmu_alloc_cd_leaf_table(struct device *dev, struct arm_smmu_cd_table *desc, @@ -260,7 +266,8 @@ static int arm_smmu_write_ctx_desc(struct arm_smmu_cd_tables *tbl, int ssid, CTXDESC_CD_0_ENDI | #endif CTXDESC_CD_0_R | CTXDESC_CD_0_A | - CTXDESC_CD_0_ASET_PRIVATE | + (cd->mm ? CTXDESC_CD_0_ASET_SHARED : + CTXDESC_CD_0_ASET_PRIVATE) | CTXDESC_CD_0_AA64 | (cd->entry.tag & CTXDESC_CD_0_ASID_MASK) << CTXDESC_CD_0_ASID_SHIFT | @@ -277,10 +284,145 @@ static int arm_smmu_write_ctx_desc(struct arm_smmu_cd_tables *tbl, int ssid, return 0; } +static bool arm_smmu_free_asid(struct arm_smmu_cd *cd) +{ + bool free; + struct arm_smmu_cd *old_cd; + + spin_lock(&asid_lock); + free = refcount_dec_and_test(&cd->refs); + if (free) { + old_cd = idr_remove(&asid_idr, (u16)cd->entry.tag); + WARN_ON(old_cd != cd); + } + spin_unlock(&asid_lock); + + return free; +} + +static struct arm_smmu_cd *arm_smmu_alloc_cd(struct arm_smmu_cd_tables *tbl) +{ + struct arm_smmu_cd *cd; + + cd = kzalloc(sizeof(*cd), GFP_KERNEL); + if (!cd) + return NULL; + + refcount_set(&cd->refs, 1); + + return cd; +} + +static struct arm_smmu_cd *arm_smmu_share_asid(u16 asid) +{ + struct arm_smmu_cd *cd; + + cd = idr_find(&asid_idr, asid); + if (!cd) + return NULL; + + if (cd->mm) { + /* + * It's pretty common to find a stale CD when doing unbind-bind, + * given that the release happens after a RCU grace period. + * Simply reuse it. + */ + refcount_inc(&cd->refs); + return cd; + } + + /* + * Ouch, ASID is already in use for a private cd. + * TODO: seize it, for the common good. + */ + return ERR_PTR(-EEXIST); +} + static struct iommu_pasid_entry * arm_smmu_alloc_shared_cd(struct iommu_pasid_table_ops *ops, struct mm_struct *mm) { - return ERR_PTR(-ENODEV); + u16 asid; + u64 tcr, par, reg; + int ret = -ENOMEM; + struct arm_smmu_cd *cd; + struct arm_smmu_cd *old_cd = NULL; + struct arm_smmu_cd_tables *tbl = pasid_ops_to_tables(ops); + + asid = mm_context_get(mm); + if (!asid) + return ERR_PTR(-ESRCH); + + cd = arm_smmu_alloc_cd(tbl); + if (!cd) + goto err_put_context; + + idr_preload(GFP_KERNEL); + spin_lock(&asid_lock); + old_cd = arm_smmu_share_asid(asid); + if (!old_cd) + ret = idr_alloc(&asid_idr, cd, asid, asid + 1, GFP_ATOMIC); + spin_unlock(&asid_lock); + idr_preload_end(); + + if (!IS_ERR_OR_NULL(old_cd)) { + if (WARN_ON(old_cd->mm != mm)) { + ret = -EINVAL; + goto err_free_cd; + } + kfree(cd); + mm_context_put(mm); + return &old_cd->entry; + } else if (old_cd) { + ret = PTR_ERR(old_cd); + goto err_free_cd; + } + + tcr = TCR_T0SZ(VA_BITS) | TCR_IRGN0_WBWA | TCR_ORGN0_WBWA | + TCR_SH0_INNER | ARM_LPAE_TCR_EPD1; + + switch (PAGE_SIZE) { + case SZ_4K: + tcr |= TCR_TG0_4K; + break; + case SZ_16K: + tcr |= TCR_TG0_16K; + break; + case SZ_64K: + tcr |= TCR_TG0_64K; + break; + default: + WARN_ON(1); + ret = -EINVAL; + goto err_free_asid; + } + + reg = read_sanitised_ftr_reg(SYS_ID_AA64MMFR0_EL1); + par = cpuid_feature_extract_unsigned_field(reg, ID_AA64MMFR0_PARANGE_SHIFT); + tcr |= par << ARM_LPAE_TCR_IPS_SHIFT; + + cd->ttbr = virt_to_phys(mm->pgd); + cd->tcr = tcr; + /* + * MAIR value is pretty much constant and global, so we can just get it + * from the current CPU register + */ + cd->mair = read_sysreg(mair_el1); + + cd->mm = mm; + cd->entry.tag = asid; + + return &cd->entry; + +err_free_asid: + arm_smmu_free_asid(cd); + +err_free_cd: + kfree(cd); + +err_put_context: + mm_context_put(mm); + + return ERR_PTR(ret); } static struct iommu_pasid_entry * @@ -294,19 +436,23 @@ arm_smmu_alloc_priv_cd(struct iommu_pasid_table_ops *ops, struct arm_smmu_cd_tables *tbl = pasid_ops_to_tables(ops); struct arm_smmu_context_cfg *ctx_cfg = &tbl->pasid.cfg.arm_smmu; - cd = kzalloc(sizeof(*cd), GFP_KERNEL); + cd = arm_smmu_alloc_cd(tbl); if (!cd) return ERR_PTR(-ENOMEM); - asid = ida_simple_get(&asid_ida, 0, 1 << ctx_cfg->asid_bits, - GFP_KERNEL); + idr_preload(GFP_KERNEL); + spin_lock(&asid_lock); + asid = idr_alloc_cyclic(&asid_idr, cd, 0, 1 << ctx_cfg->asid_bits, + GFP_ATOMIC); + cd->entry.tag = asid; + spin_unlock(&asid_lock); + idr_preload_end(); + if (asid < 0) { kfree(cd); return ERR_PTR(asid); } - cd->entry.tag = asid; - switch (fmt) { case ARM_64_LPAE_S1: cd->ttbr = cfg->arm_lpae_s1_cfg.ttbr[0]; @@ -322,7 +468,7 @@ arm_smmu_alloc_priv_cd(struct iommu_pasid_table_ops *ops, return &cd->entry; err_free_asid: - ida_simple_remove(&asid_ida, asid); + arm_smmu_free_asid(cd); kfree(cd); @@ -334,7 +480,14 @@ static void arm_smmu_free_cd(struct iommu_pasid_table_ops *ops, { struct arm_smmu_cd *cd = pasid_entry_to_cd(entry); - ida_simple_remove(&asid_ida, (u16)entry->tag); + if (!arm_smmu_free_asid(cd)) + return; + + if (cd->mm) { + /* Unpin ASID */ + mm_context_put(cd->mm); + } + kfree(cd); } @@ -359,6 +512,14 @@ static void arm_smmu_clear_cd(struct iommu_pasid_table_ops *ops, int pasid, return; arm_smmu_write_ctx_desc(tbl, pasid, NULL); + + /* + * The ASID allocator won't broadcast the final TLB invalidations for + * this ASID, so we need to do it manually. For private contexts, + * freeing io-pgtable ops performs the invalidation. + */ + if (cd->mm) + iommu_pasid_flush_tlbs(&tbl->pasid, cd->pasid, entry); } static struct iommu_pasid_table *