From patchwork Thu Aug 29 08:52:48 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexey Kardashevskiy X-Patchwork-Id: 1155106 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from lists.ozlabs.org (lists.ozlabs.org [203.11.71.2]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 46JxYN0CtBz9sMr for ; Thu, 29 Aug 2019 19:06:16 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from bilbo.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxYK6PD6zDr6V for ; Thu, 29 Aug 2019 19:06:13 +1000 (AEST) X-Original-To: linuxppc-dev@lists.ozlabs.org Delivered-To: linuxppc-dev@lists.ozlabs.org Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=ozlabs.ru (client-ip=107.173.13.209; helo=ozlabs.ru; envelope-from=aik@ozlabs.ru; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from ozlabs.ru (ozlabs.ru [107.173.13.209]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxGd60gJzDqQs for ; Thu, 29 Aug 2019 18:53:29 +1000 (AEST) Received: from fstn1-p1.ozlabs.ibm.com (localhost [IPv6:::1]) by ozlabs.ru (Postfix) with ESMTP id ADD2FAE800AD; Thu, 29 Aug 2019 04:52:34 -0400 (EDT) From: Alexey Kardashevskiy To: linuxppc-dev@lists.ozlabs.org Subject: [PATCH kernel v3 1/5] powerpc/powernv/ioda: Split out TCE invalidation from TCE updates Date: Thu, 29 Aug 2019 18:52:48 +1000 Message-Id: <20190829085252.72370-2-aik@ozlabs.ru> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190829085252.72370-1-aik@ozlabs.ru> References: <20190829085252.72370-1-aik@ozlabs.ru> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, Alexey Kardashevskiy , Alistair Popple , kvm-ppc@vger.kernel.org, Alex Williamson , David Gibson Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Sender: "Linuxppc-dev" At the moment updates in a TCE table are made by iommu_table_ops::exchange which update one TCE and invalidates an entry in the PHB/NPU TCE cache via set of registers called "TCE Kill" (hence the naming). Writing a TCE is a simple xchg() but invalidating the TCE cache is a relatively expensive OPAL call. Mapping a 100GB guest with PCI+NPU passed through devices takes about 20s. Thankfully we can do better. Since such big mappings happen at the boot time and when memory is plugged/onlined (i.e. not often), these requests come in 512 pages so we call call OPAL 512 times less which brings 20s from the above to less than 10s. Also, since TCE caches can be flushed entirely, calling OPAL for 512 TCEs helps skiboot [1] to decide whether to flush the entire cache or not. This implements 2 new iommu_table_ops callbacks: - xchg_no_kill() to update a single TCE with no TCE invalidation; - tce_kill() to invalidate multiple TCEs. This uses the same xchg_no_kill() callback for IODA1/2. This implements 2 new wrappers on top of the new callbacks similar to the existing iommu_tce_xchg(). This does not use the new callbacks yet, the next patches will; so this should not cause any behavioral change. Signed-off-by: Alexey Kardashevskiy --- arch/powerpc/include/asm/iommu.h | 17 ++++++++++++++ arch/powerpc/kernel/iommu.c | 27 +++++++++++++++++++++++ arch/powerpc/platforms/powernv/pci-ioda.c | 12 ++++++++++ 3 files changed, 56 insertions(+) diff --git a/arch/powerpc/include/asm/iommu.h b/arch/powerpc/include/asm/iommu.h index d7bf1f104c15..837b5122f257 100644 --- a/arch/powerpc/include/asm/iommu.h +++ b/arch/powerpc/include/asm/iommu.h @@ -58,6 +58,17 @@ struct iommu_table_ops { unsigned long *hpa, enum dma_data_direction *direction); + int (*xchg_no_kill)(struct iommu_table *tbl, + long index, + unsigned long *hpa, + enum dma_data_direction *direction, + bool realmode); + + void (*tce_kill)(struct iommu_table *tbl, + unsigned long index, + unsigned long pages, + bool realmode); + __be64 *(*useraddrptr)(struct iommu_table *tbl, long index, bool alloc); #endif void (*clear)(struct iommu_table *tbl, @@ -209,6 +220,12 @@ extern void iommu_del_device(struct device *dev); extern long iommu_tce_xchg(struct mm_struct *mm, struct iommu_table *tbl, unsigned long entry, unsigned long *hpa, enum dma_data_direction *direction); +extern long iommu_tce_xchg_no_kill(struct mm_struct *mm, + struct iommu_table *tbl, + unsigned long entry, unsigned long *hpa, + enum dma_data_direction *direction); +extern void iommu_tce_kill(struct iommu_table *tbl, + unsigned long entry, unsigned long pages); #else static inline void iommu_register_group(struct iommu_table_group *table_group, int pci_domain_number, diff --git a/arch/powerpc/kernel/iommu.c b/arch/powerpc/kernel/iommu.c index e7a2b160d4c6..070492f9b46e 100644 --- a/arch/powerpc/kernel/iommu.c +++ b/arch/powerpc/kernel/iommu.c @@ -1037,6 +1037,33 @@ long iommu_tce_xchg(struct mm_struct *mm, struct iommu_table *tbl, } EXPORT_SYMBOL_GPL(iommu_tce_xchg); +extern long iommu_tce_xchg_no_kill(struct mm_struct *mm, + struct iommu_table *tbl, + unsigned long entry, unsigned long *hpa, + enum dma_data_direction *direction) +{ + long ret; + unsigned long size = 0; + + ret = tbl->it_ops->xchg_no_kill(tbl, entry, hpa, direction, false); + if (!ret && ((*direction == DMA_FROM_DEVICE) || + (*direction == DMA_BIDIRECTIONAL)) && + !mm_iommu_is_devmem(mm, *hpa, tbl->it_page_shift, + &size)) + SetPageDirty(pfn_to_page(*hpa >> PAGE_SHIFT)); + + return ret; +} +EXPORT_SYMBOL_GPL(iommu_tce_xchg_no_kill); + +void iommu_tce_kill(struct iommu_table *tbl, + unsigned long entry, unsigned long pages) +{ + if (tbl->it_ops->tce_kill) + tbl->it_ops->tce_kill(tbl, entry, pages, false); +} +EXPORT_SYMBOL_GPL(iommu_tce_kill); + int iommu_take_ownership(struct iommu_table *tbl) { unsigned long flags, i, sz = (tbl->it_size + 7) >> 3; diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c index ec48ea25a674..4e56b2c620ec 100644 --- a/arch/powerpc/platforms/powernv/pci-ioda.c +++ b/arch/powerpc/platforms/powernv/pci-ioda.c @@ -1939,6 +1939,14 @@ static int pnv_ioda1_tce_build(struct iommu_table *tbl, long index, } #ifdef CONFIG_IOMMU_API +/* Common for IODA1 and IODA2 */ +static int pnv_ioda_tce_xchg_no_kill(struct iommu_table *tbl, long index, + unsigned long *hpa, enum dma_data_direction *direction, + bool realmode) +{ + return pnv_tce_xchg(tbl, index, hpa, direction, !realmode); +} + static int pnv_ioda1_tce_xchg(struct iommu_table *tbl, long index, unsigned long *hpa, enum dma_data_direction *direction) { @@ -1975,6 +1983,8 @@ static struct iommu_table_ops pnv_ioda1_iommu_ops = { #ifdef CONFIG_IOMMU_API .exchange = pnv_ioda1_tce_xchg, .exchange_rm = pnv_ioda1_tce_xchg_rm, + .xchg_no_kill = pnv_ioda_tce_xchg_no_kill, + .tce_kill = pnv_pci_p7ioc_tce_invalidate, .useraddrptr = pnv_tce_useraddrptr, #endif .clear = pnv_ioda1_tce_free, @@ -2140,6 +2150,8 @@ static struct iommu_table_ops pnv_ioda2_iommu_ops = { #ifdef CONFIG_IOMMU_API .exchange = pnv_ioda2_tce_xchg, .exchange_rm = pnv_ioda2_tce_xchg_rm, + .xchg_no_kill = pnv_ioda_tce_xchg_no_kill, + .tce_kill = pnv_pci_ioda2_tce_invalidate, .useraddrptr = pnv_tce_useraddrptr, #endif .clear = pnv_ioda2_tce_free, From patchwork Thu Aug 29 08:52:49 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexey Kardashevskiy X-Patchwork-Id: 1155097 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from lists.ozlabs.org (lists.ozlabs.org [203.11.71.2]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 46JxMD69Txz9s5b for ; Thu, 29 Aug 2019 18:57:28 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from bilbo.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxMD54r8zDrQ9 for ; Thu, 29 Aug 2019 18:57:28 +1000 (AEST) X-Original-To: linuxppc-dev@lists.ozlabs.org Delivered-To: linuxppc-dev@lists.ozlabs.org Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=ozlabs.ru (client-ip=107.173.13.209; helo=ozlabs.ru; envelope-from=aik@ozlabs.ru; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from ozlabs.ru (ozlabs.ru [107.173.13.209]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxG63958zDqTD for ; Thu, 29 Aug 2019 18:53:02 +1000 (AEST) Received: from fstn1-p1.ozlabs.ibm.com (localhost [IPv6:::1]) by ozlabs.ru (Postfix) with ESMTP id E807AAE802FA; Thu, 29 Aug 2019 04:52:36 -0400 (EDT) From: Alexey Kardashevskiy To: linuxppc-dev@lists.ozlabs.org Subject: [PATCH kernel v3 2/5] KVM: PPC: Book3S: Invalidate multiple TCEs at once Date: Thu, 29 Aug 2019 18:52:49 +1000 Message-Id: <20190829085252.72370-3-aik@ozlabs.ru> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190829085252.72370-1-aik@ozlabs.ru> References: <20190829085252.72370-1-aik@ozlabs.ru> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, Alexey Kardashevskiy , Alistair Popple , kvm-ppc@vger.kernel.org, Alex Williamson , David Gibson Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Sender: "Linuxppc-dev" Invalidating a TCE cache entry for each updated TCE is quite expensive. This makes use of the new iommu_table_ops::xchg_no_kill()/tce_kill() callbacks to bring down the time spent in mapping a huge guest DMA window; roughly 20s to 10s for each guest's 100GB of DMA space. Signed-off-by: Alexey Kardashevskiy Acked-by: Paul Mackerras --- Changes: v3: * fixed the subject line to include "Book3S" * added "ack" from Paul --- arch/powerpc/kvm/book3s_64_vio.c | 29 +++++++++++++++------- arch/powerpc/kvm/book3s_64_vio_hv.c | 38 +++++++++++++++++++++-------- 2 files changed, 48 insertions(+), 19 deletions(-) diff --git a/arch/powerpc/kvm/book3s_64_vio.c b/arch/powerpc/kvm/book3s_64_vio.c index c4b606fe73eb..5834db0a54c6 100644 --- a/arch/powerpc/kvm/book3s_64_vio.c +++ b/arch/powerpc/kvm/book3s_64_vio.c @@ -416,7 +416,7 @@ static void kvmppc_clear_tce(struct mm_struct *mm, struct iommu_table *tbl, unsigned long hpa = 0; enum dma_data_direction dir = DMA_NONE; - iommu_tce_xchg(mm, tbl, entry, &hpa, &dir); + iommu_tce_xchg_no_kill(mm, tbl, entry, &hpa, &dir); } static long kvmppc_tce_iommu_mapped_dec(struct kvm *kvm, @@ -447,7 +447,8 @@ static long kvmppc_tce_iommu_do_unmap(struct kvm *kvm, unsigned long hpa = 0; long ret; - if (WARN_ON_ONCE(iommu_tce_xchg(kvm->mm, tbl, entry, &hpa, &dir))) + if (WARN_ON_ONCE(iommu_tce_xchg_no_kill(kvm->mm, tbl, entry, &hpa, + &dir))) return H_TOO_HARD; if (dir == DMA_NONE) @@ -455,7 +456,7 @@ static long kvmppc_tce_iommu_do_unmap(struct kvm *kvm, ret = kvmppc_tce_iommu_mapped_dec(kvm, tbl, entry); if (ret != H_SUCCESS) - iommu_tce_xchg(kvm->mm, tbl, entry, &hpa, &dir); + iommu_tce_xchg_no_kill(kvm->mm, tbl, entry, &hpa, &dir); return ret; } @@ -501,7 +502,7 @@ long kvmppc_tce_iommu_do_map(struct kvm *kvm, struct iommu_table *tbl, if (mm_iommu_mapped_inc(mem)) return H_TOO_HARD; - ret = iommu_tce_xchg(kvm->mm, tbl, entry, &hpa, &dir); + ret = iommu_tce_xchg_no_kill(kvm->mm, tbl, entry, &hpa, &dir); if (WARN_ON_ONCE(ret)) { mm_iommu_mapped_dec(mem); return H_TOO_HARD; @@ -579,6 +580,8 @@ long kvmppc_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn, ret = kvmppc_tce_iommu_map(vcpu->kvm, stt, stit->tbl, entry, ua, dir); + iommu_tce_kill(stit->tbl, entry, 1); + if (ret != H_SUCCESS) { kvmppc_clear_tce(vcpu->kvm->mm, stit->tbl, entry); goto unlock_exit; @@ -656,13 +659,13 @@ long kvmppc_h_put_tce_indirect(struct kvm_vcpu *vcpu, */ if (get_user(tce, tces + i)) { ret = H_TOO_HARD; - goto unlock_exit; + goto invalidate_exit; } tce = be64_to_cpu(tce); if (kvmppc_tce_to_ua(vcpu->kvm, tce, &ua)) { ret = H_PARAMETER; - goto unlock_exit; + goto invalidate_exit; } list_for_each_entry_lockless(stit, &stt->iommu_tables, next) { @@ -673,13 +676,17 @@ long kvmppc_h_put_tce_indirect(struct kvm_vcpu *vcpu, if (ret != H_SUCCESS) { kvmppc_clear_tce(vcpu->kvm->mm, stit->tbl, entry); - goto unlock_exit; + goto invalidate_exit; } } kvmppc_tce_put(stt, entry + i, tce); } +invalidate_exit: + list_for_each_entry_lockless(stit, &stt->iommu_tables, next) + iommu_tce_kill(stit->tbl, entry, npages); + unlock_exit: srcu_read_unlock(&vcpu->kvm->srcu, idx); @@ -718,7 +725,7 @@ long kvmppc_h_stuff_tce(struct kvm_vcpu *vcpu, continue; if (ret == H_TOO_HARD) - return ret; + goto invalidate_exit; WARN_ON_ONCE(1); kvmppc_clear_tce(vcpu->kvm->mm, stit->tbl, entry); @@ -728,6 +735,10 @@ long kvmppc_h_stuff_tce(struct kvm_vcpu *vcpu, for (i = 0; i < npages; ++i, ioba += (1ULL << stt->page_shift)) kvmppc_tce_put(stt, ioba >> stt->page_shift, tce_value); - return H_SUCCESS; +invalidate_exit: + list_for_each_entry_lockless(stit, &stt->iommu_tables, next) + iommu_tce_kill(stit->tbl, ioba >> stt->page_shift, npages); + + return ret; } EXPORT_SYMBOL_GPL(kvmppc_h_stuff_tce); diff --git a/arch/powerpc/kvm/book3s_64_vio_hv.c b/arch/powerpc/kvm/book3s_64_vio_hv.c index b4f20f13b860..ab6eeb8e753e 100644 --- a/arch/powerpc/kvm/book3s_64_vio_hv.c +++ b/arch/powerpc/kvm/book3s_64_vio_hv.c @@ -218,13 +218,14 @@ static long kvmppc_rm_ioba_validate(struct kvmppc_spapr_tce_table *stt, return H_SUCCESS; } -static long iommu_tce_xchg_rm(struct mm_struct *mm, struct iommu_table *tbl, +static long iommu_tce_xchg_no_kill_rm(struct mm_struct *mm, + struct iommu_table *tbl, unsigned long entry, unsigned long *hpa, enum dma_data_direction *direction) { long ret; - ret = tbl->it_ops->exchange_rm(tbl, entry, hpa, direction); + ret = tbl->it_ops->xchg_no_kill(tbl, entry, hpa, direction, true); if (!ret && ((*direction == DMA_FROM_DEVICE) || (*direction == DMA_BIDIRECTIONAL))) { @@ -240,13 +241,20 @@ static long iommu_tce_xchg_rm(struct mm_struct *mm, struct iommu_table *tbl, return ret; } +extern void iommu_tce_kill_rm(struct iommu_table *tbl, + unsigned long entry, unsigned long pages) +{ + if (tbl->it_ops->tce_kill) + tbl->it_ops->tce_kill(tbl, entry, pages, true); +} + static void kvmppc_rm_clear_tce(struct kvm *kvm, struct iommu_table *tbl, unsigned long entry) { unsigned long hpa = 0; enum dma_data_direction dir = DMA_NONE; - iommu_tce_xchg_rm(kvm->mm, tbl, entry, &hpa, &dir); + iommu_tce_xchg_no_kill_rm(kvm->mm, tbl, entry, &hpa, &dir); } static long kvmppc_rm_tce_iommu_mapped_dec(struct kvm *kvm, @@ -278,7 +286,7 @@ static long kvmppc_rm_tce_iommu_do_unmap(struct kvm *kvm, unsigned long hpa = 0; long ret; - if (iommu_tce_xchg_rm(kvm->mm, tbl, entry, &hpa, &dir)) + if (iommu_tce_xchg_no_kill_rm(kvm->mm, tbl, entry, &hpa, &dir)) /* * real mode xchg can fail if struct page crosses * a page boundary @@ -290,7 +298,7 @@ static long kvmppc_rm_tce_iommu_do_unmap(struct kvm *kvm, ret = kvmppc_rm_tce_iommu_mapped_dec(kvm, tbl, entry); if (ret) - iommu_tce_xchg_rm(kvm->mm, tbl, entry, &hpa, &dir); + iommu_tce_xchg_no_kill_rm(kvm->mm, tbl, entry, &hpa, &dir); return ret; } @@ -336,7 +344,7 @@ static long kvmppc_rm_tce_iommu_do_map(struct kvm *kvm, struct iommu_table *tbl, if (WARN_ON_ONCE_RM(mm_iommu_mapped_inc(mem))) return H_TOO_HARD; - ret = iommu_tce_xchg_rm(kvm->mm, tbl, entry, &hpa, &dir); + ret = iommu_tce_xchg_no_kill_rm(kvm->mm, tbl, entry, &hpa, &dir); if (ret) { mm_iommu_mapped_dec(mem); /* @@ -417,6 +425,8 @@ long kvmppc_rm_h_put_tce(struct kvm_vcpu *vcpu, unsigned long liobn, ret = kvmppc_rm_tce_iommu_map(vcpu->kvm, stt, stit->tbl, entry, ua, dir); + iommu_tce_kill_rm(stit->tbl, entry, 1); + if (ret != H_SUCCESS) { kvmppc_rm_clear_tce(vcpu->kvm, stit->tbl, entry); return ret; @@ -558,7 +568,7 @@ long kvmppc_rm_h_put_tce_indirect(struct kvm_vcpu *vcpu, ua = 0; if (kvmppc_rm_tce_to_ua(vcpu->kvm, tce, &ua, NULL)) { ret = H_PARAMETER; - goto unlock_exit; + goto invalidate_exit; } list_for_each_entry_lockless(stit, &stt->iommu_tables, next) { @@ -569,13 +579,17 @@ long kvmppc_rm_h_put_tce_indirect(struct kvm_vcpu *vcpu, if (ret != H_SUCCESS) { kvmppc_rm_clear_tce(vcpu->kvm, stit->tbl, entry); - goto unlock_exit; + goto invalidate_exit; } } kvmppc_rm_tce_put(stt, entry + i, tce); } +invalidate_exit: + list_for_each_entry_lockless(stit, &stt->iommu_tables, next) + iommu_tce_kill_rm(stit->tbl, entry, npages); + unlock_exit: if (rmap) unlock_rmap(rmap); @@ -618,7 +632,7 @@ long kvmppc_rm_h_stuff_tce(struct kvm_vcpu *vcpu, continue; if (ret == H_TOO_HARD) - return ret; + goto invalidate_exit; WARN_ON_ONCE_RM(1); kvmppc_rm_clear_tce(vcpu->kvm, stit->tbl, entry); @@ -628,7 +642,11 @@ long kvmppc_rm_h_stuff_tce(struct kvm_vcpu *vcpu, for (i = 0; i < npages; ++i, ioba += (1ULL << stt->page_shift)) kvmppc_rm_tce_put(stt, ioba >> stt->page_shift, tce_value); - return H_SUCCESS; +invalidate_exit: + list_for_each_entry_lockless(stit, &stt->iommu_tables, next) + iommu_tce_kill_rm(stit->tbl, ioba >> stt->page_shift, npages); + + return ret; } /* This can be called in either virtual mode or real mode */ From patchwork Thu Aug 29 08:52:50 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexey Kardashevskiy X-Patchwork-Id: 1155099 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from lists.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 46JxPn1pr2z9s5b for ; Thu, 29 Aug 2019 18:59:41 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from bilbo.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxPn0QHjzDrYf for ; Thu, 29 Aug 2019 18:59:41 +1000 (AEST) X-Original-To: linuxppc-dev@lists.ozlabs.org Delivered-To: linuxppc-dev@lists.ozlabs.org Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=ozlabs.ru (client-ip=107.173.13.209; helo=ozlabs.ru; envelope-from=aik@ozlabs.ru; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from ozlabs.ru (ozlabs.ru [107.173.13.209]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxG80p6BzDqLv for ; Thu, 29 Aug 2019 18:53:04 +1000 (AEST) Received: from fstn1-p1.ozlabs.ibm.com (localhost [IPv6:::1]) by ozlabs.ru (Postfix) with ESMTP id 318D5AE80382; Thu, 29 Aug 2019 04:52:39 -0400 (EDT) From: Alexey Kardashevskiy To: linuxppc-dev@lists.ozlabs.org Subject: [PATCH kernel v3 3/5] vfio/spapr_tce: Invalidate multiple TCEs at once Date: Thu, 29 Aug 2019 18:52:50 +1000 Message-Id: <20190829085252.72370-4-aik@ozlabs.ru> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190829085252.72370-1-aik@ozlabs.ru> References: <20190829085252.72370-1-aik@ozlabs.ru> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, Alexey Kardashevskiy , Alistair Popple , kvm-ppc@vger.kernel.org, Alex Williamson , David Gibson Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Sender: "Linuxppc-dev" Invalidating a TCE cache entry for each updated TCE is quite expensive. This makes use of the new iommu_table_ops::xchg_no_kill()/tce_kill() callbacks to bring down the time spent in mapping a huge guest DMA window. Signed-off-by: Alexey Kardashevskiy --- drivers/vfio/vfio_iommu_spapr_tce.c | 18 ++++++++++++------ 1 file changed, 12 insertions(+), 6 deletions(-) diff --git a/drivers/vfio/vfio_iommu_spapr_tce.c b/drivers/vfio/vfio_iommu_spapr_tce.c index babef8b00daf..3b18fa4d090a 100644 --- a/drivers/vfio/vfio_iommu_spapr_tce.c +++ b/drivers/vfio/vfio_iommu_spapr_tce.c @@ -435,7 +435,7 @@ static int tce_iommu_clear(struct tce_container *container, unsigned long oldhpa; long ret; enum dma_data_direction direction; - unsigned long lastentry = entry + pages; + unsigned long lastentry = entry + pages, firstentry = entry; for ( ; entry < lastentry; ++entry) { if (tbl->it_indirect_levels && tbl->it_userspace) { @@ -460,7 +460,7 @@ static int tce_iommu_clear(struct tce_container *container, direction = DMA_NONE; oldhpa = 0; - ret = iommu_tce_xchg(container->mm, tbl, entry, &oldhpa, + ret = iommu_tce_xchg_no_kill(container->mm, tbl, entry, &oldhpa, &direction); if (ret) continue; @@ -476,6 +476,8 @@ static int tce_iommu_clear(struct tce_container *container, tce_iommu_unuse_page(container, oldhpa); } + iommu_tce_kill(tbl, firstentry, pages); + return 0; } @@ -518,8 +520,8 @@ static long tce_iommu_build(struct tce_container *container, hpa |= offset; dirtmp = direction; - ret = iommu_tce_xchg(container->mm, tbl, entry + i, &hpa, - &dirtmp); + ret = iommu_tce_xchg_no_kill(container->mm, tbl, entry + i, + &hpa, &dirtmp); if (ret) { tce_iommu_unuse_page(container, hpa); pr_err("iommu_tce: %s failed ioba=%lx, tce=%lx, ret=%ld\n", @@ -536,6 +538,8 @@ static long tce_iommu_build(struct tce_container *container, if (ret) tce_iommu_clear(container, tbl, entry, i); + else + iommu_tce_kill(tbl, entry, pages); return ret; } @@ -572,8 +576,8 @@ static long tce_iommu_build_v2(struct tce_container *container, if (mm_iommu_mapped_inc(mem)) break; - ret = iommu_tce_xchg(container->mm, tbl, entry + i, &hpa, - &dirtmp); + ret = iommu_tce_xchg_no_kill(container->mm, tbl, entry + i, + &hpa, &dirtmp); if (ret) { /* dirtmp cannot be DMA_NONE here */ tce_iommu_unuse_page_v2(container, tbl, entry + i); @@ -593,6 +597,8 @@ static long tce_iommu_build_v2(struct tce_container *container, if (ret) tce_iommu_clear(container, tbl, entry, i); + else + iommu_tce_kill(tbl, entry, pages); return ret; } From patchwork Thu Aug 29 08:52:51 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexey Kardashevskiy X-Patchwork-Id: 1155100 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from lists.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 46JxSC2KPJz9s5b for ; Thu, 29 Aug 2019 19:01:47 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from bilbo.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxSC1098zDqjn for ; Thu, 29 Aug 2019 19:01:47 +1000 (AEST) X-Original-To: linuxppc-dev@lists.ozlabs.org Delivered-To: linuxppc-dev@lists.ozlabs.org Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=ozlabs.ru (client-ip=107.173.13.209; helo=ozlabs.ru; envelope-from=aik@ozlabs.ru; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from ozlabs.ru (ozlabs.ru [107.173.13.209]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxGB0jNLzDqLv for ; Thu, 29 Aug 2019 18:53:06 +1000 (AEST) Received: from fstn1-p1.ozlabs.ibm.com (localhost [IPv6:::1]) by ozlabs.ru (Postfix) with ESMTP id 46E3BAE80570; Thu, 29 Aug 2019 04:52:41 -0400 (EDT) From: Alexey Kardashevskiy To: linuxppc-dev@lists.ozlabs.org Subject: [PATCH kernel v3 4/5] powerpc/pseries/iommu: Switch to xchg_no_kill Date: Thu, 29 Aug 2019 18:52:51 +1000 Message-Id: <20190829085252.72370-5-aik@ozlabs.ru> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190829085252.72370-1-aik@ozlabs.ru> References: <20190829085252.72370-1-aik@ozlabs.ru> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, Alexey Kardashevskiy , Alistair Popple , kvm-ppc@vger.kernel.org, Alex Williamson , David Gibson Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Sender: "Linuxppc-dev" This is the last implementation of iommu_table_ops::exchange() which we are about to remove. This implements xchg_no_kill() for pseries. Since it is paravirtual platform, the hypervisor does TCE invalidations and we do not have to deal with it here, hence no tce_kill() hook. Signed-off-by: Alexey Kardashevskiy --- Changes: v3: * new to the series and it fixes compile error from v2 --- arch/powerpc/platforms/pseries/iommu.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/platforms/pseries/iommu.c b/arch/powerpc/platforms/pseries/iommu.c index 42fb03253334..df7db33ca93b 100644 --- a/arch/powerpc/platforms/pseries/iommu.c +++ b/arch/powerpc/platforms/pseries/iommu.c @@ -621,7 +621,8 @@ static void pci_dma_bus_setup_pSeries(struct pci_bus *bus) #ifdef CONFIG_IOMMU_API static int tce_exchange_pseries(struct iommu_table *tbl, long index, unsigned - long *tce, enum dma_data_direction *direction) + long *tce, enum dma_data_direction *direction, + bool realmode) { long rc; unsigned long ioba = (unsigned long) index << tbl->it_page_shift; @@ -649,7 +650,7 @@ static int tce_exchange_pseries(struct iommu_table *tbl, long index, unsigned struct iommu_table_ops iommu_table_lpar_multi_ops = { .set = tce_buildmulti_pSeriesLP, #ifdef CONFIG_IOMMU_API - .exchange = tce_exchange_pseries, + .xchg_no_kill = tce_exchange_pseries, #endif .clear = tce_freemulti_pSeriesLP, .get = tce_get_pSeriesLP From patchwork Thu Aug 29 08:52:52 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alexey Kardashevskiy X-Patchwork-Id: 1155104 Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from lists.ozlabs.org (lists.ozlabs.org [203.11.71.2]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature RSA-PSS (4096 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 46JxVk2YPXz9sBp for ; Thu, 29 Aug 2019 19:03:58 +1000 (AEST) Authentication-Results: ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from bilbo.ozlabs.org (lists.ozlabs.org [IPv6:2401:3900:2:1::3]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxVk0p8YzDrSL for ; Thu, 29 Aug 2019 19:03:58 +1000 (AEST) X-Original-To: linuxppc-dev@lists.ozlabs.org Delivered-To: linuxppc-dev@lists.ozlabs.org Authentication-Results: lists.ozlabs.org; spf=pass (mailfrom) smtp.mailfrom=ozlabs.ru (client-ip=107.173.13.209; helo=ozlabs.ru; envelope-from=aik@ozlabs.ru; receiver=) Authentication-Results: lists.ozlabs.org; dmarc=none (p=none dis=none) header.from=ozlabs.ru Received: from ozlabs.ru (ozlabs.ru [107.173.13.209]) by lists.ozlabs.org (Postfix) with ESMTP id 46JxGC3shszDqRS for ; Thu, 29 Aug 2019 18:53:07 +1000 (AEST) Received: from fstn1-p1.ozlabs.ibm.com (localhost [IPv6:::1]) by ozlabs.ru (Postfix) with ESMTP id 56F47AE80597; Thu, 29 Aug 2019 04:52:43 -0400 (EDT) From: Alexey Kardashevskiy To: linuxppc-dev@lists.ozlabs.org Subject: [PATCH kernel v3 5/5] powerpc/powernv/ioda: Remove obsolete iommu_table_ops::exchange callbacks Date: Thu, 29 Aug 2019 18:52:52 +1000 Message-Id: <20190829085252.72370-6-aik@ozlabs.ru> X-Mailer: git-send-email 2.17.1 In-Reply-To: <20190829085252.72370-1-aik@ozlabs.ru> References: <20190829085252.72370-1-aik@ozlabs.ru> X-BeenThere: linuxppc-dev@lists.ozlabs.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: kvm@vger.kernel.org, Alexey Kardashevskiy , Alistair Popple , kvm-ppc@vger.kernel.org, Alex Williamson , David Gibson Errors-To: linuxppc-dev-bounces+patchwork-incoming=ozlabs.org@lists.ozlabs.org Sender: "Linuxppc-dev" As now we have xchg_no_kill/tce_kill, these are not used anymore so remove them. Signed-off-by: Alexey Kardashevskiy --- arch/powerpc/include/asm/iommu.h | 10 ----- arch/powerpc/kernel/iommu.c | 26 +----------- arch/powerpc/platforms/powernv/pci-ioda.c | 50 ----------------------- 3 files changed, 1 insertion(+), 85 deletions(-) diff --git a/arch/powerpc/include/asm/iommu.h b/arch/powerpc/include/asm/iommu.h index 837b5122f257..350101e11ddb 100644 --- a/arch/powerpc/include/asm/iommu.h +++ b/arch/powerpc/include/asm/iommu.h @@ -48,16 +48,6 @@ struct iommu_table_ops { * returns old TCE and DMA direction mask. * @tce is a physical address. */ - int (*exchange)(struct iommu_table *tbl, - long index, - unsigned long *hpa, - enum dma_data_direction *direction); - /* Real mode */ - int (*exchange_rm)(struct iommu_table *tbl, - long index, - unsigned long *hpa, - enum dma_data_direction *direction); - int (*xchg_no_kill)(struct iommu_table *tbl, long index, unsigned long *hpa, diff --git a/arch/powerpc/kernel/iommu.c b/arch/powerpc/kernel/iommu.c index 070492f9b46e..9704f3f76e63 100644 --- a/arch/powerpc/kernel/iommu.c +++ b/arch/powerpc/kernel/iommu.c @@ -1013,30 +1013,6 @@ int iommu_tce_check_gpa(unsigned long page_shift, unsigned long gpa) } EXPORT_SYMBOL_GPL(iommu_tce_check_gpa); -long iommu_tce_xchg(struct mm_struct *mm, struct iommu_table *tbl, - unsigned long entry, unsigned long *hpa, - enum dma_data_direction *direction) -{ - long ret; - unsigned long size = 0; - - ret = tbl->it_ops->exchange(tbl, entry, hpa, direction); - - if (!ret && ((*direction == DMA_FROM_DEVICE) || - (*direction == DMA_BIDIRECTIONAL)) && - !mm_iommu_is_devmem(mm, *hpa, tbl->it_page_shift, - &size)) - SetPageDirty(pfn_to_page(*hpa >> PAGE_SHIFT)); - - /* if (unlikely(ret)) - pr_err("iommu_tce: %s failed on hwaddr=%lx ioba=%lx kva=%lx ret=%d\n", - __func__, hwaddr, entry << tbl->it_page_shift, - hwaddr, ret); */ - - return ret; -} -EXPORT_SYMBOL_GPL(iommu_tce_xchg); - extern long iommu_tce_xchg_no_kill(struct mm_struct *mm, struct iommu_table *tbl, unsigned long entry, unsigned long *hpa, @@ -1076,7 +1052,7 @@ int iommu_take_ownership(struct iommu_table *tbl) * requires exchange() callback defined so if it is not * implemented, we disallow taking ownership over the table. */ - if (!tbl->it_ops->exchange) + if (!tbl->it_ops->xchg_no_kill) return -EINVAL; spin_lock_irqsave(&tbl->large_pool.lock, flags); diff --git a/arch/powerpc/platforms/powernv/pci-ioda.c b/arch/powerpc/platforms/powernv/pci-ioda.c index 4e56b2c620ec..c28d0d9b7ee0 100644 --- a/arch/powerpc/platforms/powernv/pci-ioda.c +++ b/arch/powerpc/platforms/powernv/pci-ioda.c @@ -1946,28 +1946,6 @@ static int pnv_ioda_tce_xchg_no_kill(struct iommu_table *tbl, long index, { return pnv_tce_xchg(tbl, index, hpa, direction, !realmode); } - -static int pnv_ioda1_tce_xchg(struct iommu_table *tbl, long index, - unsigned long *hpa, enum dma_data_direction *direction) -{ - long ret = pnv_tce_xchg(tbl, index, hpa, direction, true); - - if (!ret) - pnv_pci_p7ioc_tce_invalidate(tbl, index, 1, false); - - return ret; -} - -static int pnv_ioda1_tce_xchg_rm(struct iommu_table *tbl, long index, - unsigned long *hpa, enum dma_data_direction *direction) -{ - long ret = pnv_tce_xchg(tbl, index, hpa, direction, false); - - if (!ret) - pnv_pci_p7ioc_tce_invalidate(tbl, index, 1, true); - - return ret; -} #endif static void pnv_ioda1_tce_free(struct iommu_table *tbl, long index, @@ -1981,8 +1959,6 @@ static void pnv_ioda1_tce_free(struct iommu_table *tbl, long index, static struct iommu_table_ops pnv_ioda1_iommu_ops = { .set = pnv_ioda1_tce_build, #ifdef CONFIG_IOMMU_API - .exchange = pnv_ioda1_tce_xchg, - .exchange_rm = pnv_ioda1_tce_xchg_rm, .xchg_no_kill = pnv_ioda_tce_xchg_no_kill, .tce_kill = pnv_pci_p7ioc_tce_invalidate, .useraddrptr = pnv_tce_useraddrptr, @@ -2113,30 +2089,6 @@ static int pnv_ioda2_tce_build(struct iommu_table *tbl, long index, return ret; } -#ifdef CONFIG_IOMMU_API -static int pnv_ioda2_tce_xchg(struct iommu_table *tbl, long index, - unsigned long *hpa, enum dma_data_direction *direction) -{ - long ret = pnv_tce_xchg(tbl, index, hpa, direction, true); - - if (!ret) - pnv_pci_ioda2_tce_invalidate(tbl, index, 1, false); - - return ret; -} - -static int pnv_ioda2_tce_xchg_rm(struct iommu_table *tbl, long index, - unsigned long *hpa, enum dma_data_direction *direction) -{ - long ret = pnv_tce_xchg(tbl, index, hpa, direction, false); - - if (!ret) - pnv_pci_ioda2_tce_invalidate(tbl, index, 1, true); - - return ret; -} -#endif - static void pnv_ioda2_tce_free(struct iommu_table *tbl, long index, long npages) { @@ -2148,8 +2100,6 @@ static void pnv_ioda2_tce_free(struct iommu_table *tbl, long index, static struct iommu_table_ops pnv_ioda2_iommu_ops = { .set = pnv_ioda2_tce_build, #ifdef CONFIG_IOMMU_API - .exchange = pnv_ioda2_tce_xchg, - .exchange_rm = pnv_ioda2_tce_xchg_rm, .xchg_no_kill = pnv_ioda_tce_xchg_no_kill, .tce_kill = pnv_pci_ioda2_tce_invalidate, .useraddrptr = pnv_tce_useraddrptr,