Merge tag 'iommu-updates-v5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu

Pull iommu updates from Joerg Roedel:

 - Conversion of the AMD IOMMU driver to use the dma-iommu code for
   imlementing the DMA-API. This gets rid of quite some code in the
   driver itself, but also has some potential for regressions (non are
   known at the moment).

 - Support for the Qualcomm SMMUv2 implementation in the SDM845 SoC.
   This also includes some firmware interface changes, but those are
   acked by the respective maintainers.

 - Preparatory work to support two distinct page-tables per domain in
   the ARM-SMMU driver

 - Power management improvements for the ARM SMMUv2

 - Custom PASID allocator support

 - Multiple PCI DMA alias support for the AMD IOMMU driver

 - Adaption of the Mediatek driver to the changed IO/TLB flush interface
   of the IOMMU core code.

 - Preparatory patches for the Renesas IOMMU driver to support future
   hardware.

* tag 'iommu-updates-v5.5' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu: (62 commits)
  iommu/rockchip: Don't provoke WARN for harmless IRQs
  iommu/vt-d: Turn off translations at shutdown
  iommu/vt-d: Check VT-d RMRR region in BIOS is reported as reserved
  iommu/arm-smmu: Remove duplicate error message
  iommu/arm-smmu-v3: Don't display an error when IRQ lines are missing
  iommu/ipmmu-vmsa: Add utlb_offset_base
  iommu/ipmmu-vmsa: Add helper functions for "uTLB" registers
  iommu/ipmmu-vmsa: Calculate context registers' offset instead of a macro
  iommu/ipmmu-vmsa: Add helper functions for MMU "context" registers
  iommu/ipmmu-vmsa: tidyup register definitions
  iommu/ipmmu-vmsa: Remove all unused register definitions
  iommu/mediatek: Reduce the tlb flush timeout value
  iommu/mediatek: Get rid of the pgtlock
  iommu/mediatek: Move the tlb_sync into tlb_flush
  iommu/mediatek: Delete the leaf in the tlb_flush
  iommu/mediatek: Use gather to achieve the tlb range flush
  iommu/mediatek: Add a new tlb_lock for tlb_flush
  iommu/mediatek: Correct the flush_iotlb_all callback
  iommu/io-pgtable-arm: Rename IOMMU_QCOM_SYS_CACHE and improve doc
  iommu/io-pgtable-arm: Rationalise MAIR handling
  ...

1 files changed, 29 insertions, 61 deletions

diff --git a/drivers/iommu/mtk_iommu.c b/drivers/iommu/mtk_iommu.c
index 67a483c1a935..6fc1f5ecf91e 100644
--- a/drivers/iommu/mtk_iommu.c
+++ b/drivers/iommu/mtk_iommu.c
@@ -101,8 +101,6 @@
 #define MTK_M4U_TO_PORT(id)		((id) & 0x1f)
 
 struct mtk_iommu_domain {
-	spinlock_t			pgtlock; /* lock for page table */
-
 	struct io_pgtable_cfg		cfg;
 	struct io_pgtable_ops		*iop;
 
@@ -173,13 +171,16 @@ static void mtk_iommu_tlb_flush_all(void *cookie)
 	}
 }
 
-static void mtk_iommu_tlb_add_flush_nosync(unsigned long iova, size_t size,
-					   size_t granule, bool leaf,
-					   void *cookie)
+static void mtk_iommu_tlb_flush_range_sync(unsigned long iova, size_t size,
+					   size_t granule, void *cookie)
 {
 	struct mtk_iommu_data *data = cookie;
+	unsigned long flags;
+	int ret;
+	u32 tmp;
 
 	for_each_m4u(data) {
+		spin_lock_irqsave(&data->tlb_lock, flags);
 		writel_relaxed(F_INVLD_EN1 | F_INVLD_EN0,
 			       data->base + REG_MMU_INV_SEL);
 
@@ -188,23 +189,10 @@ static void mtk_iommu_tlb_add_flush_nosync(unsigned long iova, size_t size,
 			       data->base + REG_MMU_INVLD_END_A);
 		writel_relaxed(F_MMU_INV_RANGE,
 			       data->base + REG_MMU_INVALIDATE);
-		data->tlb_flush_active = true;
-	}
-}
-
-static void mtk_iommu_tlb_sync(void *cookie)
-{
-	struct mtk_iommu_data *data = cookie;
-	int ret;
-	u32 tmp;
-
-	for_each_m4u(data) {
-		/* Avoid timing out if there's nothing to wait for */
-		if (!data->tlb_flush_active)
-			return;
 
+		/* tlb sync */
 		ret = readl_poll_timeout_atomic(data->base + REG_MMU_CPE_DONE,
-						tmp, tmp != 0, 10, 100000);
+						tmp, tmp != 0, 10, 1000);
 		if (ret) {
 			dev_warn(data->dev,
 				 "Partial TLB flush timed out, falling back to full flush\n");
@@ -212,35 +200,24 @@ static void mtk_iommu_tlb_sync(void *cookie)
 		}
 		/* Clear the CPE status */
 		writel_relaxed(0, data->base + REG_MMU_CPE_DONE);
-		data->tlb_flush_active = false;
+		spin_unlock_irqrestore(&data->tlb_lock, flags);
 	}
 }
 
-static void mtk_iommu_tlb_flush_walk(unsigned long iova, size_t size,
-				     size_t granule, void *cookie)
-{
-	mtk_iommu_tlb_add_flush_nosync(iova, size, granule, false, cookie);
-	mtk_iommu_tlb_sync(cookie);
-}
-
-static void mtk_iommu_tlb_flush_leaf(unsigned long iova, size_t size,
-				     size_t granule, void *cookie)
-{
-	mtk_iommu_tlb_add_flush_nosync(iova, size, granule, true, cookie);
-	mtk_iommu_tlb_sync(cookie);
-}
-
 static void mtk_iommu_tlb_flush_page_nosync(struct iommu_iotlb_gather *gather,
 					    unsigned long iova, size_t granule,
 					    void *cookie)
 {
-	mtk_iommu_tlb_add_flush_nosync(iova, granule, granule, true, cookie);
+	struct mtk_iommu_data *data = cookie;
+	struct iommu_domain *domain = &data->m4u_dom->domain;
+
+	iommu_iotlb_gather_add_page(domain, gather, iova, granule);
 }
 
 static const struct iommu_flush_ops mtk_iommu_flush_ops = {
 	.tlb_flush_all = mtk_iommu_tlb_flush_all,
-	.tlb_flush_walk = mtk_iommu_tlb_flush_walk,
-	.tlb_flush_leaf = mtk_iommu_tlb_flush_leaf,
+	.tlb_flush_walk = mtk_iommu_tlb_flush_range_sync,
+	.tlb_flush_leaf = mtk_iommu_tlb_flush_range_sync,
 	.tlb_add_page = mtk_iommu_tlb_flush_page_nosync,
 };
 
@@ -316,8 +293,6 @@ static int mtk_iommu_domain_finalise(struct mtk_iommu_domain *dom)
 {
 	struct mtk_iommu_data *data = mtk_iommu_get_m4u_data();
 
-	spin_lock_init(&dom->pgtlock);
-
 	dom->cfg = (struct io_pgtable_cfg) {
 		.quirks = IO_PGTABLE_QUIRK_ARM_NS |
 			IO_PGTABLE_QUIRK_NO_PERMS |
@@ -412,22 +387,17 @@ static void mtk_iommu_detach_device(struct iommu_domain *domain,
 }
 
 static int mtk_iommu_map(struct iommu_domain *domain, unsigned long iova,
-			 phys_addr_t paddr, size_t size, int prot)
+			 phys_addr_t paddr, size_t size, int prot, gfp_t gfp)
 {
 	struct mtk_iommu_domain *dom = to_mtk_domain(domain);
 	struct mtk_iommu_data *data = mtk_iommu_get_m4u_data();
-	unsigned long flags;
-	int ret;
 
 	/* The "4GB mode" M4U physically can not use the lower remap of Dram. */
 	if (data->enable_4GB)
 		paddr |= BIT_ULL(32);
 
-	spin_lock_irqsave(&dom->pgtlock, flags);
-	ret = dom->iop->map(dom->iop, iova, paddr, size, prot);
-	spin_unlock_irqrestore(&dom->pgtlock, flags);
-
-	return ret;
+	/* Synchronize with the tlb_lock */
+	return dom->iop->map(dom->iop, iova, paddr, size, prot);
 }
 
 static size_t mtk_iommu_unmap(struct iommu_domain *domain,
@@ -435,25 +405,26 @@ static size_t mtk_iommu_unmap(struct iommu_domain *domain,
 			      struct iommu_iotlb_gather *gather)
 {
 	struct mtk_iommu_domain *dom = to_mtk_domain(domain);
-	unsigned long flags;
-	size_t unmapsz;
-
-	spin_lock_irqsave(&dom->pgtlock, flags);
-	unmapsz = dom->iop->unmap(dom->iop, iova, size, gather);
-	spin_unlock_irqrestore(&dom->pgtlock, flags);
 
-	return unmapsz;
+	return dom->iop->unmap(dom->iop, iova, size, gather);
 }
 
 static void mtk_iommu_flush_iotlb_all(struct iommu_domain *domain)
 {
-	mtk_iommu_tlb_sync(mtk_iommu_get_m4u_data());
+	mtk_iommu_tlb_flush_all(mtk_iommu_get_m4u_data());
 }
 
 static void mtk_iommu_iotlb_sync(struct iommu_domain *domain,
 				 struct iommu_iotlb_gather *gather)
 {
-	mtk_iommu_tlb_sync(mtk_iommu_get_m4u_data());
+	struct mtk_iommu_data *data = mtk_iommu_get_m4u_data();
+	size_t length = gather->end - gather->start;
+
+	if (gather->start == ULONG_MAX)
+		return;
+
+	mtk_iommu_tlb_flush_range_sync(gather->start, length, gather->pgsize,
+				       data);
 }
 
 static phys_addr_t mtk_iommu_iova_to_phys(struct iommu_domain *domain,
@@ -461,13 +432,9 @@ static phys_addr_t mtk_iommu_iova_to_phys(struct iommu_domain *domain,
 {
 	struct mtk_iommu_domain *dom = to_mtk_domain(domain);
 	struct mtk_iommu_data *data = mtk_iommu_get_m4u_data();
-	unsigned long flags;
 	phys_addr_t pa;
 
-	spin_lock_irqsave(&dom->pgtlock, flags);
 	pa = dom->iop->iova_to_phys(dom->iop, iova);
-	spin_unlock_irqrestore(&dom->pgtlock, flags);
-
 	if (data->enable_4GB && pa >= MTK_IOMMU_4GB_MODE_REMAP_BASE)
 		pa &= ~BIT_ULL(32);
 
@@ -733,6 +700,7 @@ static int mtk_iommu_probe(struct platform_device *pdev)
 	if (ret)
 		return ret;
 
+	spin_lock_init(&data->tlb_lock);
 	list_add_tail(&data->list, &m4ulist);
 
 	if (!iommu_present(&platform_bus_type))