iommu.lists.linux-foundation.org archive mirror
 help / color / mirror / Atom feed
From: Tom Murphy <murphyt7@tcd.ie>
To: iommu@lists.linux-foundation.org
Cc: David Woodhouse <dwmw2@infradead.org>,
	linux-kernel@vger.kernel.org, Tom Murphy <murphyt7@tcd.ie>
Subject: [PATCH V2 1/5] iommu: Handle freelists when using deferred flushing in iommu drivers
Date: Thu,  3 Sep 2020 21:18:33 +0100	[thread overview]
Message-ID: <20200903201839.7327-2-murphyt7@tcd.ie> (raw)
In-Reply-To: <20200903201839.7327-1-murphyt7@tcd.ie>

Allow the iommu_unmap_fast to return newly freed page table pages and
pass the freelist to queue_iova in the dma-iommu ops path.

This is useful for iommu drivers (in this case the intel iommu driver)
which need to wait for the ioTLB to be flushed before newly
free/unmapped page table pages can be freed. This way we can still batch
ioTLB free operations and handle the freelists.

Signed-off-by: Tom Murphy <murphyt7@tcd.ie>
---
 drivers/iommu/dma-iommu.c   | 30 ++++++++++++++------
 drivers/iommu/intel/iommu.c | 55 ++++++++++++++++++++++++-------------
 include/linux/iommu.h       |  1 +
 3 files changed, 59 insertions(+), 27 deletions(-)

diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
index 4959f5df21bd..f69dc9467d71 100644
--- a/drivers/iommu/dma-iommu.c
+++ b/drivers/iommu/dma-iommu.c
@@ -50,6 +50,18 @@ struct iommu_dma_cookie {
 	struct iommu_domain		*fq_domain;
 };
 
+static void iommu_dma_entry_dtor(unsigned long data)
+{
+	struct page *freelist = (struct page *)data;
+
+	while (freelist != NULL) {
+		unsigned long p = (unsigned long)page_address(freelist);
+
+		freelist = freelist->freelist;
+		free_page(p);
+	}
+}
+
 static inline size_t cookie_msi_granule(struct iommu_dma_cookie *cookie)
 {
 	if (cookie->type == IOMMU_DMA_IOVA_COOKIE)
@@ -344,7 +356,8 @@ static int iommu_dma_init_domain(struct iommu_domain *domain, dma_addr_t base,
 	if (!cookie->fq_domain && !iommu_domain_get_attr(domain,
 			DOMAIN_ATTR_DMA_USE_FLUSH_QUEUE, &attr) && attr) {
 		cookie->fq_domain = domain;
-		init_iova_flush_queue(iovad, iommu_dma_flush_iotlb_all, NULL);
+		init_iova_flush_queue(iovad, iommu_dma_flush_iotlb_all,
+				iommu_dma_entry_dtor);
 	}
 
 	if (!dev)
@@ -438,7 +451,7 @@ static dma_addr_t iommu_dma_alloc_iova(struct iommu_domain *domain,
 }
 
 static void iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
-		dma_addr_t iova, size_t size)
+		dma_addr_t iova, size_t size, struct page *freelist)
 {
 	struct iova_domain *iovad = &cookie->iovad;
 
@@ -447,7 +460,8 @@ static void iommu_dma_free_iova(struct iommu_dma_cookie *cookie,
 		cookie->msi_iova -= size;
 	else if (cookie->fq_domain)	/* non-strict mode */
 		queue_iova(iovad, iova_pfn(iovad, iova),
-				size >> iova_shift(iovad), 0);
+				size >> iova_shift(iovad),
+				(unsigned long) freelist);
 	else
 		free_iova_fast(iovad, iova_pfn(iovad, iova),
 				size >> iova_shift(iovad));
@@ -472,7 +486,7 @@ static void __iommu_dma_unmap(struct device *dev, dma_addr_t dma_addr,
 
 	if (!cookie->fq_domain)
 		iommu_tlb_sync(domain, &iotlb_gather);
-	iommu_dma_free_iova(cookie, dma_addr, size);
+	iommu_dma_free_iova(cookie, dma_addr, size, iotlb_gather.freelist);
 }
 
 static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
@@ -494,7 +508,7 @@ static dma_addr_t __iommu_dma_map(struct device *dev, phys_addr_t phys,
 		return DMA_MAPPING_ERROR;
 
 	if (iommu_map_atomic(domain, iova, phys - iova_off, size, prot)) {
-		iommu_dma_free_iova(cookie, iova, size);
+		iommu_dma_free_iova(cookie, iova, size, NULL);
 		return DMA_MAPPING_ERROR;
 	}
 	return iova + iova_off;
@@ -649,7 +663,7 @@ static void *iommu_dma_alloc_remap(struct device *dev, size_t size,
 out_free_sg:
 	sg_free_table(&sgt);
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, size);
+	iommu_dma_free_iova(cookie, iova, size, NULL);
 out_free_pages:
 	__iommu_dma_free_pages(pages, count);
 	return NULL;
@@ -900,7 +914,7 @@ static int iommu_dma_map_sg(struct device *dev, struct scatterlist *sg,
 	return __finalise_sg(dev, sg, nents, iova);
 
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, iova_len);
+	iommu_dma_free_iova(cookie, iova, iova_len, NULL);
 out_restore_sg:
 	__invalidate_sg(sg, nents);
 	return 0;
@@ -1194,7 +1208,7 @@ static struct iommu_dma_msi_page *iommu_dma_get_msi_page(struct device *dev,
 	return msi_page;
 
 out_free_iova:
-	iommu_dma_free_iova(cookie, iova, size);
+	iommu_dma_free_iova(cookie, iova, size, NULL);
 out_free_page:
 	kfree(msi_page);
 	return NULL;
diff --git a/drivers/iommu/intel/iommu.c b/drivers/iommu/intel/iommu.c
index 237a470e1e9c..03699860880b 100644
--- a/drivers/iommu/intel/iommu.c
+++ b/drivers/iommu/intel/iommu.c
@@ -1160,17 +1160,17 @@ static struct page *dma_pte_clear_level(struct dmar_domain *domain, int level,
    pages can only be freed after the IOTLB flush has been done. */
 static struct page *domain_unmap(struct dmar_domain *domain,
 				 unsigned long start_pfn,
-				 unsigned long last_pfn)
+				 unsigned long last_pfn,
+				 struct page *freelist)
 {
-	struct page *freelist;
-
 	BUG_ON(!domain_pfn_supported(domain, start_pfn));
 	BUG_ON(!domain_pfn_supported(domain, last_pfn));
 	BUG_ON(start_pfn > last_pfn);
 
 	/* we don't need lock here; nobody else touches the iova range */
 	freelist = dma_pte_clear_level(domain, agaw_to_level(domain->agaw),
-				       domain->pgd, 0, start_pfn, last_pfn, NULL);
+				       domain->pgd, 0, start_pfn, last_pfn,
+				       freelist);
 
 	/* free pgd */
 	if (start_pfn == 0 && last_pfn == DOMAIN_MAX_PFN(domain->gaw)) {
@@ -1924,7 +1924,8 @@ static void domain_exit(struct dmar_domain *domain)
 	if (domain->pgd) {
 		struct page *freelist;
 
-		freelist = domain_unmap(domain, 0, DOMAIN_MAX_PFN(domain->gaw));
+		freelist = domain_unmap(domain, 0, DOMAIN_MAX_PFN(domain->gaw),
+				NULL);
 		dma_free_pagelist(freelist);
 	}
 
@@ -3480,7 +3481,7 @@ static void intel_unmap(struct device *dev, dma_addr_t dev_addr, size_t size)
 	if (dev_is_pci(dev))
 		pdev = to_pci_dev(dev);
 
-	freelist = domain_unmap(domain, start_pfn, last_pfn);
+	freelist = domain_unmap(domain, start_pfn, last_pfn, NULL);
 	if (intel_iommu_strict || (pdev && pdev->untrusted) ||
 			!has_iova_flush_queue(&domain->iovad)) {
 		iommu_flush_iotlb_psi(iommu, domain, start_pfn,
@@ -4575,7 +4576,8 @@ static int intel_iommu_memory_notifier(struct notifier_block *nb,
 			struct page *freelist;
 
 			freelist = domain_unmap(si_domain,
-						start_vpfn, last_vpfn);
+						start_vpfn, last_vpfn,
+						NULL);
 
 			rcu_read_lock();
 			for_each_active_iommu(iommu, drhd)
@@ -5543,10 +5545,8 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain,
 				struct iommu_iotlb_gather *gather)
 {
 	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
-	struct page *freelist = NULL;
 	unsigned long start_pfn, last_pfn;
-	unsigned int npages;
-	int iommu_id, level = 0;
+	int level = 0;
 
 	/* Cope with horrid API which requires us to unmap more than the
 	   size argument if it happens to be a large-page mapping. */
@@ -5558,22 +5558,38 @@ static size_t intel_iommu_unmap(struct iommu_domain *domain,
 	start_pfn = iova >> VTD_PAGE_SHIFT;
 	last_pfn = (iova + size - 1) >> VTD_PAGE_SHIFT;
 
-	freelist = domain_unmap(dmar_domain, start_pfn, last_pfn);
-
-	npages = last_pfn - start_pfn + 1;
-
-	for_each_domain_iommu(iommu_id, dmar_domain)
-		iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain,
-				      start_pfn, npages, !freelist, 0);
-
-	dma_free_pagelist(freelist);
+	gather->freelist = domain_unmap(dmar_domain, start_pfn, last_pfn,
+			gather->freelist);
 
 	if (dmar_domain->max_addr == iova + size)
 		dmar_domain->max_addr = iova;
 
+	iommu_iotlb_gather_add_page(domain, gather, iova, size);
+
 	return size;
 }
 
+static void intel_iommu_tlb_sync(struct iommu_domain *domain,
+		struct iommu_iotlb_gather *gather)
+{
+	struct dmar_domain *dmar_domain = to_dmar_domain(domain);
+	unsigned long iova_pfn = IOVA_PFN(gather->start);
+	size_t size = gather->end - gather->start;
+	unsigned long start_pfn, last_pfn;
+	unsigned long nrpages;
+	int iommu_id;
+
+	nrpages = aligned_nrpages(gather->start, size);
+	start_pfn = mm_to_dma_pfn(iova_pfn);
+	last_pfn = start_pfn + nrpages - 1;
+
+	for_each_domain_iommu(iommu_id, dmar_domain)
+		iommu_flush_iotlb_psi(g_iommus[iommu_id], dmar_domain,
+			start_pfn, nrpages, !gather->freelist, 0);
+
+	dma_free_pagelist(gather->freelist);
+}
+
 static phys_addr_t intel_iommu_iova_to_phys(struct iommu_domain *domain,
 					    dma_addr_t iova)
 {
@@ -6058,6 +6074,7 @@ const struct iommu_ops intel_iommu_ops = {
 	.aux_get_pasid		= intel_iommu_aux_get_pasid,
 	.map			= intel_iommu_map,
 	.unmap			= intel_iommu_unmap,
+	.iotlb_sync		= intel_iommu_tlb_sync,
 	.iova_to_phys		= intel_iommu_iova_to_phys,
 	.probe_device		= intel_iommu_probe_device,
 	.probe_finalize		= intel_iommu_probe_finalize,
diff --git a/include/linux/iommu.h b/include/linux/iommu.h
index 5f0b7859d2eb..e3eafb3cf4ba 100644
--- a/include/linux/iommu.h
+++ b/include/linux/iommu.h
@@ -186,6 +186,7 @@ struct iommu_iotlb_gather {
 	unsigned long		start;
 	unsigned long		end;
 	size_t			pgsize;
+	struct page		*freelist;
 };
 
 /**
-- 
2.20.1

_______________________________________________
iommu mailing list
iommu@lists.linux-foundation.org
https://lists.linuxfoundation.org/mailman/listinfo/iommu

  reply	other threads:[~2020-09-03 20:25 UTC|newest]

Thread overview: 22+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-09-03 20:18 [PATCH V2 0/5] Convert the intel iommu driver to the dma-iommu api Tom Murphy
2020-09-03 20:18 ` Tom Murphy [this message]
2020-09-03 20:18 ` [PATCH V2 2/5] iommu: Add iommu_dma_free_cpu_cached_iovas function Tom Murphy
2020-09-09  0:45   ` Lu Baolu
2020-09-09  7:05     ` Christoph Hellwig
2020-09-12  2:55       ` Lu Baolu
2020-09-03 20:18 ` [PATCH V2 3/5] iommu: allow the dma-iommu api to use bounce buffers Tom Murphy
2020-09-09  1:34   ` Lu Baolu
2020-09-03 20:18 ` [PATCH V2 4/5] iommu/vt-d: Convert intel iommu driver to the iommu ops Tom Murphy
2020-09-09  1:59   ` Lu Baolu
2020-09-03 20:18 ` [PATCH V2 5/5] DO NOT MERGE: iommu: disable list appending in dma-iommu Tom Murphy
2020-09-07  7:00   ` Christoph Hellwig
2020-09-07 20:18     ` Tom Murphy
2020-09-08  5:36       ` Christoph Hellwig
2020-09-08  5:55         ` Christoph Hellwig
2020-09-08  6:04           ` Lu Baolu
2020-09-08  6:23             ` Christoph Hellwig
2020-09-08  9:07               ` Lu Baolu
2020-09-09  1:43               ` Lu Baolu
2020-09-09  7:06                 ` Christoph Hellwig
2020-09-12  3:13                   ` Lu Baolu
2020-09-04 10:03 ` [PATCH V2 0/5] Convert the intel iommu driver to the dma-iommu api Joerg Roedel

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20200903201839.7327-2-murphyt7@tcd.ie \
    --to=murphyt7@tcd.ie \
    --cc=dwmw2@infradead.org \
    --cc=iommu@lists.linux-foundation.org \
    --cc=linux-kernel@vger.kernel.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).