Merge tag 'dma-mapping-4.19' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping updates from Christoph Hellwig:

 - a series from Robin to fix bus imposed dma limits by adding a
   separate mask for them to struct device instead of trying to squeeze
   a second meaning out of the existing dma mask as we did before.

   This has ACKs from the various other subsystems touched

 - a small swiotlb cleanup from Kees (acked by Konrad)

 - conversion of nios2 and sh to the new generic dma-noncoherent code.

   Various other architecture conversions will come through the
   architectures maintainers trees.

* tag 'dma-mapping-4.19' of git://git.infradead.org/users/hch/dma-mapping:
  sh: use generic dma_noncoherent_ops
  sh: split arch/sh/mm/consistent.c
  sh: use dma_direct_ops for the CONFIG_DMA_COHERENT case
  sh: introduce a sh_cacheop_vaddr helper
  sh: simplify get_arch_dma_ops
  OF: Don't set default coherent DMA mask
  ACPI/IORT: Don't set default coherent DMA mask
  iommu/dma: Respect bus DMA limit for IOVAs
  of/device: Set bus DMA mask as appropriate
  ACPI/IORT: Set bus DMA mask as appropriate
  dma-mapping: Generalise dma_32bit_limit flag
  ACPI/IORT: Support address size limit for root complexes
  of/platform: Initialise default DMA masks
  nios2: use generic dma_noncoherent_ops
  swiotlb: clean up reporting
  dma-mapping: relax warning for per-device areas
This commit is contained in:
Linus Torvalds
2018-08-14 11:11:52 -07:00
23 changed files with 193 additions and 402 deletions

View File

@@ -51,7 +51,6 @@ config SUPERH
select HAVE_ARCH_AUDITSYSCALL
select HAVE_FUTEX_CMPXCHG if FUTEX
select HAVE_NMI
select NEED_DMA_MAP_STATE
select NEED_SG_DMA_LENGTH
help
@@ -159,10 +158,13 @@ config SWAP_IO_SPACE
bool
config DMA_COHERENT
select DMA_DIRECT_OPS
bool
config DMA_NONCOHERENT
def_bool !DMA_COHERENT
select ARCH_HAS_SYNC_DMA_FOR_DEVICE
select DMA_NONCOHERENT_OPS
config PGTABLE_LEVELS
default 3 if X2TLB

View File

@@ -2,6 +2,7 @@ generic-y += compat.h
generic-y += current.h
generic-y += delay.h
generic-y += div64.h
generic-y += dma-mapping.h
generic-y += emergency-restart.h
generic-y += exec.h
generic-y += irq_regs.h

View File

@@ -101,5 +101,12 @@ void kunmap_coherent(void *kvaddr);
void cpu_cache_init(void);
static inline void *sh_cacheop_vaddr(void *vaddr)
{
if (__in_29bit_mode())
vaddr = (void *)CAC_ADDR((unsigned long)vaddr);
return vaddr;
}
#endif /* __KERNEL__ */
#endif /* __ASM_SH_CACHEFLUSH_H */

View File

@@ -1,23 +0,0 @@
/* SPDX-License-Identifier: GPL-2.0 */
#ifndef __ASM_SH_DMA_MAPPING_H
#define __ASM_SH_DMA_MAPPING_H
extern const struct dma_map_ops *dma_ops;
extern void no_iommu_init(void);
static inline const struct dma_map_ops *get_arch_dma_ops(struct bus_type *bus)
{
return dma_ops;
}
extern void *dma_generic_alloc_coherent(struct device *dev, size_t size,
dma_addr_t *dma_addr, gfp_t flag,
unsigned long attrs);
extern void dma_generic_free_coherent(struct device *dev, size_t size,
void *vaddr, dma_addr_t dma_handle,
unsigned long attrs);
void sh_sync_dma_for_device(void *vaddr, size_t size,
enum dma_data_direction dir);
#endif /* __ASM_SH_DMA_MAPPING_H */

View File

@@ -12,7 +12,7 @@ endif
CFLAGS_REMOVE_return_address.o = -pg
obj-y := debugtraps.o dma-nommu.o dumpstack.o \
obj-y := debugtraps.o dumpstack.o \
idle.o io.o irq.o irq_$(BITS).o kdebugfs.o \
machvec.o nmi_debug.o process.o \
process_$(BITS).o ptrace.o ptrace_$(BITS).o \
@@ -45,7 +45,7 @@ obj-$(CONFIG_DUMP_CODE) += disassemble.o
obj-$(CONFIG_HIBERNATION) += swsusp.o
obj-$(CONFIG_DWARF_UNWINDER) += dwarf.o
obj-$(CONFIG_PERF_EVENTS) += perf_event.o perf_callchain.o
obj-$(CONFIG_DMA_NONCOHERENT) += dma-coherent.o
obj-$(CONFIG_HAVE_HW_BREAKPOINT) += hw_breakpoint.o
ccflags-y := -Werror

View File

@@ -0,0 +1,83 @@
/*
* Copyright (C) 2004 - 2007 Paul Mundt
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*/
#include <linux/mm.h>
#include <linux/init.h>
#include <linux/dma-noncoherent.h>
#include <linux/module.h>
#include <asm/cacheflush.h>
#include <asm/addrspace.h>
void *arch_dma_alloc(struct device *dev, size_t size, dma_addr_t *dma_handle,
gfp_t gfp, unsigned long attrs)
{
void *ret, *ret_nocache;
int order = get_order(size);
gfp |= __GFP_ZERO;
ret = (void *)__get_free_pages(gfp, order);
if (!ret)
return NULL;
/*
* Pages from the page allocator may have data present in
* cache. So flush the cache before using uncached memory.
*/
arch_sync_dma_for_device(dev, virt_to_phys(ret), size,
DMA_BIDIRECTIONAL);
ret_nocache = (void __force *)ioremap_nocache(virt_to_phys(ret), size);
if (!ret_nocache) {
free_pages((unsigned long)ret, order);
return NULL;
}
split_page(pfn_to_page(virt_to_phys(ret) >> PAGE_SHIFT), order);
*dma_handle = virt_to_phys(ret);
if (!WARN_ON(!dev))
*dma_handle -= PFN_PHYS(dev->dma_pfn_offset);
return ret_nocache;
}
void arch_dma_free(struct device *dev, size_t size, void *vaddr,
dma_addr_t dma_handle, unsigned long attrs)
{
int order = get_order(size);
unsigned long pfn = (dma_handle >> PAGE_SHIFT);
int k;
if (!WARN_ON(!dev))
pfn += dev->dma_pfn_offset;
for (k = 0; k < (1 << order); k++)
__free_pages(pfn_to_page(pfn + k), 0);
iounmap(vaddr);
}
void arch_sync_dma_for_device(struct device *dev, phys_addr_t paddr,
size_t size, enum dma_data_direction dir)
{
void *addr = sh_cacheop_vaddr(phys_to_virt(paddr));
switch (dir) {
case DMA_FROM_DEVICE: /* invalidate only */
__flush_invalidate_region(addr, size);
break;
case DMA_TO_DEVICE: /* writeback only */
__flush_wback_region(addr, size);
break;
case DMA_BIDIRECTIONAL: /* writeback and invalidate */
__flush_purge_region(addr, size);
break;
default:
BUG();
}
}

View File

@@ -1,88 +0,0 @@
/*
* DMA mapping support for platforms lacking IOMMUs.
*
* Copyright (C) 2009 Paul Mundt
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
*/
#include <linux/dma-mapping.h>
#include <linux/io.h>
#include <asm/cacheflush.h>
static dma_addr_t nommu_map_page(struct device *dev, struct page *page,
unsigned long offset, size_t size,
enum dma_data_direction dir,
unsigned long attrs)
{
dma_addr_t addr = page_to_phys(page) + offset
- PFN_PHYS(dev->dma_pfn_offset);
WARN_ON(size == 0);
if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
sh_sync_dma_for_device(page_address(page) + offset, size, dir);
return addr;
}
static int nommu_map_sg(struct device *dev, struct scatterlist *sg,
int nents, enum dma_data_direction dir,
unsigned long attrs)
{
struct scatterlist *s;
int i;
WARN_ON(nents == 0 || sg[0].length == 0);
for_each_sg(sg, s, nents, i) {
dma_addr_t offset = PFN_PHYS(dev->dma_pfn_offset);
BUG_ON(!sg_page(s));
if (!(attrs & DMA_ATTR_SKIP_CPU_SYNC))
sh_sync_dma_for_device(sg_virt(s), s->length, dir);
s->dma_address = sg_phys(s) - offset;
s->dma_length = s->length;
}
return nents;
}
#ifdef CONFIG_DMA_NONCOHERENT
static void nommu_sync_single_for_device(struct device *dev, dma_addr_t addr,
size_t size, enum dma_data_direction dir)
{
sh_sync_dma_for_device(phys_to_virt(addr), size, dir);
}
static void nommu_sync_sg_for_device(struct device *dev, struct scatterlist *sg,
int nelems, enum dma_data_direction dir)
{
struct scatterlist *s;
int i;
for_each_sg(sg, s, nelems, i)
sh_sync_dma_for_device(sg_virt(s), s->length, dir);
}
#endif
const struct dma_map_ops nommu_dma_ops = {
.alloc = dma_generic_alloc_coherent,
.free = dma_generic_free_coherent,
.map_page = nommu_map_page,
.map_sg = nommu_map_sg,
#ifdef CONFIG_DMA_NONCOHERENT
.sync_single_for_device = nommu_sync_single_for_device,
.sync_sg_for_device = nommu_sync_sg_for_device,
#endif
};
void __init no_iommu_init(void)
{
if (dma_ops)
return;
dma_ops = &nommu_dma_ops;
}

View File

@@ -1,10 +1,6 @@
/*
* arch/sh/mm/consistent.c
*
* Copyright (C) 2004 - 2007 Paul Mundt
*
* Declared coherent memory functions based on arch/x86/kernel/pci-dma_32.c
*
* This file is subject to the terms and conditions of the GNU General Public
* License. See the file "COPYING" in the main directory of this archive
* for more details.
@@ -13,90 +9,7 @@
#include <linux/init.h>
#include <linux/platform_device.h>
#include <linux/dma-mapping.h>
#include <linux/dma-debug.h>
#include <linux/io.h>
#include <linux/module.h>
#include <linux/gfp.h>
#include <asm/cacheflush.h>
#include <asm/addrspace.h>
const struct dma_map_ops *dma_ops;
EXPORT_SYMBOL(dma_ops);
void *dma_generic_alloc_coherent(struct device *dev, size_t size,
dma_addr_t *dma_handle, gfp_t gfp,
unsigned long attrs)
{
void *ret, *ret_nocache;
int order = get_order(size);
gfp |= __GFP_ZERO;
ret = (void *)__get_free_pages(gfp, order);
if (!ret)
return NULL;
/*
* Pages from the page allocator may have data present in
* cache. So flush the cache before using uncached memory.
*/
sh_sync_dma_for_device(ret, size, DMA_BIDIRECTIONAL);
ret_nocache = (void __force *)ioremap_nocache(virt_to_phys(ret), size);
if (!ret_nocache) {
free_pages((unsigned long)ret, order);
return NULL;
}
split_page(pfn_to_page(virt_to_phys(ret) >> PAGE_SHIFT), order);
*dma_handle = virt_to_phys(ret);
if (!WARN_ON(!dev))
*dma_handle -= PFN_PHYS(dev->dma_pfn_offset);
return ret_nocache;
}
void dma_generic_free_coherent(struct device *dev, size_t size,
void *vaddr, dma_addr_t dma_handle,
unsigned long attrs)
{
int order = get_order(size);
unsigned long pfn = dma_handle >> PAGE_SHIFT;
int k;
if (!WARN_ON(!dev))
pfn += dev->dma_pfn_offset;
for (k = 0; k < (1 << order); k++)
__free_pages(pfn_to_page(pfn + k), 0);
iounmap(vaddr);
}
void sh_sync_dma_for_device(void *vaddr, size_t size,
enum dma_data_direction direction)
{
void *addr;
addr = __in_29bit_mode() ?
(void *)CAC_ADDR((unsigned long)vaddr) : vaddr;
switch (direction) {
case DMA_FROM_DEVICE: /* invalidate only */
__flush_invalidate_region(addr, size);
break;
case DMA_TO_DEVICE: /* writeback only */
__flush_wback_region(addr, size);
break;
case DMA_BIDIRECTIONAL: /* writeback and invalidate */
__flush_purge_region(addr, size);
break;
default:
BUG();
}
}
EXPORT_SYMBOL(sh_sync_dma_for_device);
static int __init memchunk_setup(char *str)
{

View File

@@ -339,22 +339,12 @@ void __init paging_init(void)
free_area_init_nodes(max_zone_pfns);
}
/*
* Early initialization for any I/O MMUs we might have.
*/
static void __init iommu_init(void)
{
no_iommu_init();
}
unsigned int mem_init_done = 0;
void __init mem_init(void)
{
pg_data_t *pgdat;
iommu_init();
high_memory = NULL;
for_each_online_pgdat(pgdat)
high_memory = max_t(void *, high_memory,