Merge tag 'dma-mapping-4.18' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping updates from Christoph Hellwig:

 - replace the force_dma flag with a dma_configure bus method. (Nipun
   Gupta, although one patch is іncorrectly attributed to me due to a
   git rebase bug)

 - use GFP_DMA32 more agressively in dma-direct. (Takashi Iwai)

 - remove PCI_DMA_BUS_IS_PHYS and rely on the dma-mapping API to do the
   right thing for bounce buffering.

 - move dma-debug initialization to common code, and apply a few
   cleanups to the dma-debug code.

 - cleanup the Kconfig mess around swiotlb selection

 - swiotlb comment fixup (Yisheng Xie)

 - a trivial swiotlb fix. (Dan Carpenter)

 - support swiotlb on RISC-V. (based on a patch from Palmer Dabbelt)

 - add a new generic dma-noncoherent dma_map_ops implementation and use
   it for arc, c6x and nds32.

 - improve scatterlist validity checking in dma-debug. (Robin Murphy)

 - add a struct device quirk to limit the dma-mask to 32-bit due to
   bridge/system issues, and switch x86 to use it instead of a local
   hack for VIA bridges.

 - handle devices without a dma_mask more gracefully in the dma-direct
   code.

* tag 'dma-mapping-4.18' of git://git.infradead.org/users/hch/dma-mapping: (48 commits)
  dma-direct: don't crash on device without dma_mask
  nds32: use generic dma_noncoherent_ops
  nds32: implement the unmap_sg DMA operation
  nds32: consolidate DMA cache maintainance routines
  x86/pci-dma: switch the VIA 32-bit DMA quirk to use the struct device flag
  x86/pci-dma: remove the explicit nodac and allowdac option
  x86/pci-dma: remove the experimental forcesac boot option
  Documentation/x86: remove a stray reference to pci-nommu.c
  core, dma-direct: add a flag 32-bit dma limits
  dma-mapping: remove unused gfp_t parameter to arch_dma_alloc_attrs
  dma-debug: check scatterlist segments
  c6x: use generic dma_noncoherent_ops
  arc: use generic dma_noncoherent_ops
  arc: fix arc_dma_{map,unmap}_page
  arc: fix arc_dma_sync_sg_for_{cpu,device}
  arc: simplify arc_dma_sync_single_for_{cpu,device}
  dma-mapping: provide a generic dma-noncoherent implementation
  dma-mapping: simplify Kconfig dependencies
  riscv: add swiotlb support
  riscv: only enable ZONE_DMA32 for 64-bit
  ...

11 files changed, 325 insertions, 45 deletions

diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index 8767e45..435dbc0 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -25,7 +25,6 @@ config SPARC
 	select RTC_CLASS
 	select RTC_DRV_M48T59
 	select RTC_SYSTOHC
-	select HAVE_DMA_API_DEBUG
 	select HAVE_ARCH_JUMP_LABEL if SPARC64
 	select GENERIC_IRQ_SHOW
 	select ARCH_WANT_IPC_PARSE_VERSION
@@ -44,6 +43,8 @@ config SPARC
 	select ARCH_HAS_SG_CHAIN
 	select CPU_NO_EFFICIENT_FFS
 	select LOCKDEP_SMALL if LOCKDEP
+	select NEED_DMA_MAP_STATE
+	select NEED_SG_DMA_LENGTH
 
 config SPARC32
 	def_bool !64BIT
@@ -67,6 +68,7 @@ config SPARC64
 	select HAVE_SYSCALL_TRACEPOINTS
 	select HAVE_CONTEXT_TRACKING
 	select HAVE_DEBUG_KMEMLEAK
+	select IOMMU_HELPER
 	select SPARSE_IRQ
 	select RTC_DRV_CMOS
 	select RTC_DRV_BQ4802
@@ -102,14 +104,6 @@ config ARCH_ATU
 	bool
 	default y if SPARC64
 
-config ARCH_DMA_ADDR_T_64BIT
-	bool
-	default y if ARCH_ATU
-
-config IOMMU_HELPER
-	bool
-	default y if SPARC64
-
 config STACKTRACE_SUPPORT
 	bool
 	default y if SPARC64
@@ -146,12 +140,6 @@ config ZONE_DMA
 	bool
 	default y if SPARC32
 
-config NEED_DMA_MAP_STATE
-	def_bool y
-
-config NEED_SG_DMA_LENGTH
-	def_bool y
-
 config GENERIC_ISA_DMA
 	bool
 	default y if SPARC32
diff --git a/arch/sparc/include/asm/iommu-common.h b/arch/sparc/include/asm/iommu-common.h
new file mode 100644
index 0000000..802c90c
--- /dev/null
+++ b/arch/sparc/include/asm/iommu-common.h
@@ -0,0 +1,53 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef _LINUX_IOMMU_COMMON_H
+#define _LINUX_IOMMU_COMMON_H
+
+#include <linux/spinlock_types.h>
+#include <linux/device.h>
+#include <asm/page.h>
+
+#define IOMMU_POOL_HASHBITS     4
+#define IOMMU_NR_POOLS          (1 << IOMMU_POOL_HASHBITS)
+#define IOMMU_ERROR_CODE	(~(unsigned long) 0)
+
+struct iommu_pool {
+	unsigned long	start;
+	unsigned long	end;
+	unsigned long	hint;
+	spinlock_t	lock;
+};
+
+struct iommu_map_table {
+	unsigned long		table_map_base;
+	unsigned long		table_shift;
+	unsigned long		nr_pools;
+	void			(*lazy_flush)(struct iommu_map_table *);
+	unsigned long		poolsize;
+	struct iommu_pool	pools[IOMMU_NR_POOLS];
+	u32			flags;
+#define	IOMMU_HAS_LARGE_POOL	0x00000001
+#define	IOMMU_NO_SPAN_BOUND	0x00000002
+#define	IOMMU_NEED_FLUSH	0x00000004
+	struct iommu_pool	large_pool;
+	unsigned long		*map;
+};
+
+extern void iommu_tbl_pool_init(struct iommu_map_table *iommu,
+				unsigned long num_entries,
+				u32 table_shift,
+				void (*lazy_flush)(struct iommu_map_table *),
+				bool large_pool, u32 npools,
+				bool skip_span_boundary_check);
+
+extern unsigned long iommu_tbl_range_alloc(struct device *dev,
+					   struct iommu_map_table *iommu,
+					   unsigned long npages,
+					   unsigned long *handle,
+					   unsigned long mask,
+					   unsigned int align_order);
+
+extern void iommu_tbl_range_free(struct iommu_map_table *iommu,
+				 u64 dma_addr, unsigned long npages,
+				 unsigned long entry);
+
+#endif
diff --git a/arch/sparc/include/asm/iommu_64.h b/arch/sparc/include/asm/iommu_64.h
index 9ed6b54..0ef6ded 100644
--- a/arch/sparc/include/asm/iommu_64.h
+++ b/arch/sparc/include/asm/iommu_64.h
@@ -17,7 +17,7 @@
 #define IOPTE_WRITE   0x0000000000000002UL
 
 #define IOMMU_NUM_CTXS	4096
-#include <linux/iommu-common.h>
+#include <asm/iommu-common.h>
 
 struct iommu_arena {
 	unsigned long	*map;
diff --git a/arch/sparc/include/asm/pci_32.h b/arch/sparc/include/asm/pci_32.h
index 98917e4..cfc0ee9 100644
--- a/arch/sparc/include/asm/pci_32.h
+++ b/arch/sparc/include/asm/pci_32.h
@@ -17,10 +17,6 @@
 
 #define PCI_IRQ_NONE		0xffffffff
 
-/* Dynamic DMA mapping stuff.
- */
-#define PCI_DMA_BUS_IS_PHYS	(0)
-
 #endif /* __KERNEL__ */
 
 #ifndef CONFIG_LEON_PCI
diff --git a/arch/sparc/include/asm/pci_64.h b/arch/sparc/include/asm/pci_64.h
index 671274e..fac7781 100644
--- a/arch/sparc/include/asm/pci_64.h
+++ b/arch/sparc/include/asm/pci_64.h
@@ -17,12 +17,6 @@
 
 #define PCI_IRQ_NONE		0xffffffff
 
-/* The PCI address space does not equal the physical memory
- * address space.  The networking and block device layers use
- * this boolean for bounce buffer decisions.
- */
-#define PCI_DMA_BUS_IS_PHYS	(0)
-
 /* PCI IOMMU mapping bypass support. */
 
 /* PCI 64-bit addressing works for all slots on all controller
diff --git a/arch/sparc/kernel/Makefile b/arch/sparc/kernel/Makefile
index 76cb577..cf86408 100644
--- a/arch/sparc/kernel/Makefile
+++ b/arch/sparc/kernel/Makefile
@@ -59,7 +59,7 @@ obj-$(CONFIG_SPARC32)   += leon_pmc.o
 
 obj-$(CONFIG_SPARC64)   += reboot.o
 obj-$(CONFIG_SPARC64)   += sysfs.o
-obj-$(CONFIG_SPARC64)   += iommu.o
+obj-$(CONFIG_SPARC64)   += iommu.o iommu-common.o
 obj-$(CONFIG_SPARC64)   += central.o
 obj-$(CONFIG_SPARC64)   += starfire.o
 obj-$(CONFIG_SPARC64)   += power.o
@@ -74,8 +74,6 @@ obj-$(CONFIG_SPARC64)	+= pcr.o
 obj-$(CONFIG_SPARC64)	+= nmi.o
 obj-$(CONFIG_SPARC64_SMP) += cpumap.o
 
-obj-y                     += dma.o
-
 obj-$(CONFIG_PCIC_PCI)    += pcic.o
 obj-$(CONFIG_LEON_PCI)    += leon_pci.o
 obj-$(CONFIG_SPARC_GRPCI2)+= leon_pci_grpci2.o
diff --git a/arch/sparc/kernel/dma.c b/arch/sparc/kernel/dma.c
deleted file mode 100644
index f73e759..0000000
--- a/arch/sparc/kernel/dma.c
+++ /dev/null
@@ -1,13 +0,0 @@
-// SPDX-License-Identifier: GPL-2.0
-#include <linux/kernel.h>
-#include <linux/dma-mapping.h>
-#include <linux/dma-debug.h>
-
-#define PREALLOC_DMA_DEBUG_ENTRIES       (1 << 15)
-
-static int __init dma_init(void)
-{
-	dma_debug_init(PREALLOC_DMA_DEBUG_ENTRIES);
-	return 0;
-}
-fs_initcall(dma_init);
diff --git a/arch/sparc/kernel/iommu-common.c b/arch/sparc/kernel/iommu-common.c
new file mode 100644
index 0000000..59cb166
--- /dev/null
+++ b/arch/sparc/kernel/iommu-common.c
@@ -0,0 +1,264 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * IOMMU mmap management and range allocation functions.
+ * Based almost entirely upon the powerpc iommu allocator.
+ */
+
+#include <linux/export.h>
+#include <linux/bitmap.h>
+#include <linux/bug.h>
+#include <linux/iommu-helper.h>
+#include <linux/dma-mapping.h>
+#include <linux/hash.h>
+#include <asm/iommu-common.h>
+
+static unsigned long iommu_large_alloc = 15;
+
+static	DEFINE_PER_CPU(unsigned int, iommu_hash_common);
+
+static inline bool need_flush(struct iommu_map_table *iommu)
+{
+	return ((iommu->flags & IOMMU_NEED_FLUSH) != 0);
+}
+
+static inline void set_flush(struct iommu_map_table *iommu)
+{
+	iommu->flags |= IOMMU_NEED_FLUSH;
+}
+
+static inline void clear_flush(struct iommu_map_table *iommu)
+{
+	iommu->flags &= ~IOMMU_NEED_FLUSH;
+}
+
+static void setup_iommu_pool_hash(void)
+{
+	unsigned int i;
+	static bool do_once;
+
+	if (do_once)
+		return;
+	do_once = true;
+	for_each_possible_cpu(i)
+		per_cpu(iommu_hash_common, i) = hash_32(i, IOMMU_POOL_HASHBITS);
+}
+
+/*
+ * Initialize iommu_pool entries for the iommu_map_table. `num_entries'
+ * is the number of table entries. If `large_pool' is set to true,
+ * the top 1/4 of the table will be set aside for pool allocations
+ * of more than iommu_large_alloc pages.
+ */
+void iommu_tbl_pool_init(struct iommu_map_table *iommu,
+			 unsigned long num_entries,
+			 u32 table_shift,
+			 void (*lazy_flush)(struct iommu_map_table *),
+			 bool large_pool, u32 npools,
+			 bool skip_span_boundary_check)
+{
+	unsigned int start, i;
+	struct iommu_pool *p = &(iommu->large_pool);
+
+	setup_iommu_pool_hash();
+	if (npools == 0)
+		iommu->nr_pools = IOMMU_NR_POOLS;
+	else
+		iommu->nr_pools = npools;
+	BUG_ON(npools > IOMMU_NR_POOLS);
+
+	iommu->table_shift = table_shift;
+	iommu->lazy_flush = lazy_flush;
+	start = 0;
+	if (skip_span_boundary_check)
+		iommu->flags |= IOMMU_NO_SPAN_BOUND;
+	if (large_pool)
+		iommu->flags |= IOMMU_HAS_LARGE_POOL;
+
+	if (!large_pool)
+		iommu->poolsize = num_entries/iommu->nr_pools;
+	else
+		iommu->poolsize = (num_entries * 3 / 4)/iommu->nr_pools;
+	for (i = 0; i < iommu->nr_pools; i++) {
+		spin_lock_init(&(iommu->pools[i].lock));
+		iommu->pools[i].start = start;
+		iommu->pools[i].hint = start;
+		start += iommu->poolsize; /* start for next pool */
+		iommu->pools[i].end = start - 1;
+	}
+	if (!large_pool)
+		return;
+	/* initialize large_pool */
+	spin_lock_init(&(p->lock));
+	p->start = start;
+	p->hint = p->start;
+	p->end = num_entries;
+}
+
+unsigned long iommu_tbl_range_alloc(struct device *dev,
+				struct iommu_map_table *iommu,
+				unsigned long npages,
+				unsigned long *handle,
+				unsigned long mask,
+				unsigned int align_order)
+{
+	unsigned int pool_hash = __this_cpu_read(iommu_hash_common);
+	unsigned long n, end, start, limit, boundary_size;
+	struct iommu_pool *pool;
+	int pass = 0;
+	unsigned int pool_nr;
+	unsigned int npools = iommu->nr_pools;
+	unsigned long flags;
+	bool large_pool = ((iommu->flags & IOMMU_HAS_LARGE_POOL) != 0);
+	bool largealloc = (large_pool && npages > iommu_large_alloc);
+	unsigned long shift;
+	unsigned long align_mask = 0;
+
+	if (align_order > 0)
+		align_mask = ~0ul >> (BITS_PER_LONG - align_order);
+
+	/* Sanity check */
+	if (unlikely(npages == 0)) {
+		WARN_ON_ONCE(1);
+		return IOMMU_ERROR_CODE;
+	}
+
+	if (largealloc) {
+		pool = &(iommu->large_pool);
+		pool_nr = 0; /* to keep compiler happy */
+	} else {
+		/* pick out pool_nr */
+		pool_nr =  pool_hash & (npools - 1);
+		pool = &(iommu->pools[pool_nr]);
+	}
+	spin_lock_irqsave(&pool->lock, flags);
+
+ again:
+	if (pass == 0 && handle && *handle &&
+	    (*handle >= pool->start) && (*handle < pool->end))
+		start = *handle;
+	else
+		start = pool->hint;
+
+	limit = pool->end;
+
+	/* The case below can happen if we have a small segment appended
+	 * to a large, or when the previous alloc was at the very end of
+	 * the available space. If so, go back to the beginning. If a
+	 * flush is needed, it will get done based on the return value
+	 * from iommu_area_alloc() below.
+	 */
+	if (start >= limit)
+		start = pool->start;
+	shift = iommu->table_map_base >> iommu->table_shift;
+	if (limit + shift > mask) {
+		limit = mask - shift + 1;
+		/* If we're constrained on address range, first try
+		 * at the masked hint to avoid O(n) search complexity,
+		 * but on second pass, start at 0 in pool 0.
+		 */
+		if ((start & mask) >= limit || pass > 0) {
+			spin_unlock(&(pool->lock));
+			pool = &(iommu->pools[0]);
+			spin_lock(&(pool->lock));
+			start = pool->start;
+		} else {
+			start &= mask;
+		}
+	}
+
+	if (dev)
+		boundary_size = ALIGN(dma_get_seg_boundary(dev) + 1,
+				      1 << iommu->table_shift);
+	else
+		boundary_size = ALIGN(1ULL << 32, 1 << iommu->table_shift);
+
+	boundary_size = boundary_size >> iommu->table_shift;
+	/*
+	 * if the skip_span_boundary_check had been set during init, we set
+	 * things up so that iommu_is_span_boundary() merely checks if the
+	 * (index + npages) < num_tsb_entries
+	 */
+	if ((iommu->flags & IOMMU_NO_SPAN_BOUND) != 0) {
+		shift = 0;
+		boundary_size = iommu->poolsize * iommu->nr_pools;
+	}
+	n = iommu_area_alloc(iommu->map, limit, start, npages, shift,
+			     boundary_size, align_mask);
+	if (n == -1) {
+		if (likely(pass == 0)) {
+			/* First failure, rescan from the beginning.  */
+			pool->hint = pool->start;
+			set_flush(iommu);
+			pass++;
+			goto again;
+		} else if (!largealloc && pass <= iommu->nr_pools) {
+			spin_unlock(&(pool->lock));
+			pool_nr = (pool_nr + 1) & (iommu->nr_pools - 1);
+			pool = &(iommu->pools[pool_nr]);
+			spin_lock(&(pool->lock));
+			pool->hint = pool->start;
+			set_flush(iommu);
+			pass++;
+			goto again;
+		} else {
+			/* give up */
+			n = IOMMU_ERROR_CODE;
+			goto bail;
+		}
+	}
+	if (iommu->lazy_flush &&
+	    (n < pool->hint || need_flush(iommu))) {
+		clear_flush(iommu);
+		iommu->lazy_flush(iommu);
+	}
+
+	end = n + npages;
+	pool->hint = end;
+
+	/* Update handle for SG allocations */
+	if (handle)
+		*handle = end;
+bail:
+	spin_unlock_irqrestore(&(pool->lock), flags);
+
+	return n;
+}
+
+static struct iommu_pool *get_pool(struct iommu_map_table *tbl,
+				   unsigned long entry)
+{
+	struct iommu_pool *p;
+	unsigned long largepool_start = tbl->large_pool.start;
+	bool large_pool = ((tbl->flags & IOMMU_HAS_LARGE_POOL) != 0);
+
+	/* The large pool is the last pool at the top of the table */
+	if (large_pool && entry >= largepool_start) {
+		p = &tbl->large_pool;
+	} else {
+		unsigned int pool_nr = entry / tbl->poolsize;
+
+		BUG_ON(pool_nr >= tbl->nr_pools);
+		p = &tbl->pools[pool_nr];
+	}
+	return p;
+}
+
+/* Caller supplies the index of the entry into the iommu map table
+ * itself when the mapping from dma_addr to the entry is not the
+ * default addr->entry mapping below.
+ */
+void iommu_tbl_range_free(struct iommu_map_table *iommu, u64 dma_addr,
+			  unsigned long npages, unsigned long entry)
+{
+	struct iommu_pool *pool;
+	unsigned long flags;
+	unsigned long shift = iommu->table_shift;
+
+	if (entry == IOMMU_ERROR_CODE) /* use default addr->entry mapping */
+		entry = (dma_addr - iommu->table_map_base) >> shift;
+	pool = get_pool(iommu, entry);
+
+	spin_lock_irqsave(&(pool->lock), flags);
+	bitmap_clear(iommu->map, entry, npages);
+	spin_unlock_irqrestore(&(pool->lock), flags);
+}
diff --git a/arch/sparc/kernel/iommu.c b/arch/sparc/kernel/iommu.c
index b08dc34..40d008b 100644
--- a/arch/sparc/kernel/iommu.c
+++ b/arch/sparc/kernel/iommu.c
@@ -14,7 +14,7 @@
 #include <linux/errno.h>
 #include <linux/iommu-helper.h>
 #include <linux/bitmap.h>
-#include <linux/iommu-common.h>
+#include <asm/iommu-common.h>
 
 #ifdef CONFIG_PCI
 #include <linux/pci.h>
diff --git a/arch/sparc/kernel/ldc.c b/arch/sparc/kernel/ldc.c
index 86b625f..c0fa3ef 100644
--- a/arch/sparc/kernel/ldc.c
+++ b/arch/sparc/kernel/ldc.c
@@ -16,7 +16,7 @@
 #include <linux/list.h>
 #include <linux/init.h>
 #include <linux/bitmap.h>
-#include <linux/iommu-common.h>
+#include <asm/iommu-common.h>
 
 #include <asm/hypervisor.h>
 #include <asm/iommu.h>
diff --git a/arch/sparc/kernel/pci_sun4v.c b/arch/sparc/kernel/pci_sun4v.c
index 2493672..565d9ac 100644
--- a/arch/sparc/kernel/pci_sun4v.c
+++ b/arch/sparc/kernel/pci_sun4v.c
@@ -16,7 +16,7 @@
 #include <linux/export.h>
 #include <linux/log2.h>
 #include <linux/of_device.h>
-#include <linux/iommu-common.h>
+#include <asm/iommu-common.h>
 
 #include <asm/iommu.h>
 #include <asm/irq.h>