2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Dynamic DMA mapping support.
|
|
|
|
*
|
2007-02-06 10:51:25 +08:00
|
|
|
* This implementation is a fallback for platforms that do not support
|
2005-04-17 06:20:36 +08:00
|
|
|
* I/O TLBs (aka DMA address translation hardware).
|
|
|
|
* Copyright (C) 2000 Asit Mallick <Asit.K.Mallick@intel.com>
|
|
|
|
* Copyright (C) 2000 Goutham Rao <goutham.rao@intel.com>
|
|
|
|
* Copyright (C) 2000, 2003 Hewlett-Packard Co
|
|
|
|
* David Mosberger-Tang <davidm@hpl.hp.com>
|
|
|
|
*
|
|
|
|
* 03/05/07 davidm Switch from PCI-DMA to generic device DMA API.
|
|
|
|
* 00/12/13 davidm Rename to swiotlb.c and add mark_clean() to avoid
|
|
|
|
* unnecessary i-cache flushing.
|
2005-09-30 05:45:24 +08:00
|
|
|
* 04/07/.. ak Better overflow handling. Assorted fixes.
|
|
|
|
* 05/09/10 linville Add support for syncing ranges, support syncing for
|
|
|
|
* DMA_BIDIRECTIONAL mappings, miscellaneous cleanup.
|
2008-12-23 02:26:09 +08:00
|
|
|
* 08/12/11 beckyb Add highmem support
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/cache.h>
|
2005-09-30 06:52:13 +08:00
|
|
|
#include <linux/dma-mapping.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/mm.h>
|
|
|
|
#include <linux/module.h>
|
|
|
|
#include <linux/spinlock.h>
|
|
|
|
#include <linux/string.h>
|
2008-12-17 04:17:27 +08:00
|
|
|
#include <linux/swiotlb.h>
|
2008-12-23 02:26:09 +08:00
|
|
|
#include <linux/pfn.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
#include <linux/types.h>
|
|
|
|
#include <linux/ctype.h>
|
2008-12-17 04:17:33 +08:00
|
|
|
#include <linux/highmem.h>
|
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 16:04:11 +08:00
|
|
|
#include <linux/gfp.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include <asm/io.h>
|
|
|
|
#include <asm/dma.h>
|
2005-09-30 06:52:13 +08:00
|
|
|
#include <asm/scatterlist.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/bootmem.h>
|
2008-04-29 15:59:36 +08:00
|
|
|
#include <linux/iommu-helper.h>
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
#define OFFSET(val,align) ((unsigned long) \
|
|
|
|
( (val) & ( (align) - 1)))
|
|
|
|
|
2005-09-07 01:20:49 +08:00
|
|
|
#define SLABS_PER_PAGE (1 << (PAGE_SHIFT - IO_TLB_SHIFT))
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Minimum IO TLB size to bother booting with. Systems with mainly
|
|
|
|
* 64bit capable cards will only lightly use the swiotlb. If we can't
|
|
|
|
* allocate a contiguous 1MB, we're probably in trouble anyway.
|
|
|
|
*/
|
|
|
|
#define IO_TLB_MIN_SLABS ((1<<20) >> IO_TLB_SHIFT)
|
|
|
|
|
2005-09-30 05:44:57 +08:00
|
|
|
/*
|
|
|
|
* Enumeration for sync targets
|
|
|
|
*/
|
|
|
|
enum dma_sync_target {
|
|
|
|
SYNC_FOR_CPU = 0,
|
|
|
|
SYNC_FOR_DEVICE = 1,
|
|
|
|
};
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
int swiotlb_force;
|
|
|
|
|
|
|
|
/*
|
2009-04-08 22:09:15 +08:00
|
|
|
* Used to do a quick range check in unmap_single and
|
|
|
|
* sync_single_*, to see if the memory was in fact allocated by this
|
2005-04-17 06:20:36 +08:00
|
|
|
* API.
|
|
|
|
*/
|
|
|
|
static char *io_tlb_start, *io_tlb_end;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The number of IO TLB blocks (in groups of 64) betweeen io_tlb_start and
|
|
|
|
* io_tlb_end. This is command line adjustable via setup_io_tlb_npages.
|
|
|
|
*/
|
|
|
|
static unsigned long io_tlb_nslabs;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* When the IOMMU overflows we return a fallback buffer. This sets the size.
|
|
|
|
*/
|
|
|
|
static unsigned long io_tlb_overflow = 32*1024;
|
|
|
|
|
|
|
|
void *io_tlb_overflow_buffer;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is a free list describing the number of free entries available from
|
|
|
|
* each index
|
|
|
|
*/
|
|
|
|
static unsigned int *io_tlb_list;
|
|
|
|
static unsigned int io_tlb_index;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* We need to save away the original address corresponding to a mapped entry
|
|
|
|
* for the sync operations.
|
|
|
|
*/
|
2008-12-23 02:26:08 +08:00
|
|
|
static phys_addr_t *io_tlb_orig_addr;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Protect the above data structures in the map and unmap calls
|
|
|
|
*/
|
|
|
|
static DEFINE_SPINLOCK(io_tlb_lock);
|
|
|
|
|
2009-11-10 18:46:18 +08:00
|
|
|
static int late_alloc;
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
static int __init
|
|
|
|
setup_io_tlb_npages(char *str)
|
|
|
|
{
|
|
|
|
if (isdigit(*str)) {
|
2005-08-05 04:06:00 +08:00
|
|
|
io_tlb_nslabs = simple_strtoul(str, &str, 0);
|
2005-04-17 06:20:36 +08:00
|
|
|
/* avoid tail segment of size < IO_TLB_SEGSIZE */
|
|
|
|
io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
|
|
|
|
}
|
|
|
|
if (*str == ',')
|
|
|
|
++str;
|
2009-11-11 23:03:28 +08:00
|
|
|
if (!strcmp(str, "force"))
|
2005-04-17 06:20:36 +08:00
|
|
|
swiotlb_force = 1;
|
2009-11-11 23:03:28 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
__setup("swiotlb=", setup_io_tlb_npages);
|
|
|
|
/* make io_tlb_overflow tunable too? */
|
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
/* Note that this doesn't work with highmem page */
|
2008-12-23 02:26:05 +08:00
|
|
|
static dma_addr_t swiotlb_virt_to_bus(struct device *hwdev,
|
|
|
|
volatile void *address)
|
2008-12-17 04:17:30 +08:00
|
|
|
{
|
2009-07-10 09:05:02 +08:00
|
|
|
return phys_to_dma(hwdev, virt_to_phys(address));
|
2008-12-17 04:17:30 +08:00
|
|
|
}
|
|
|
|
|
2009-11-10 18:46:19 +08:00
|
|
|
void swiotlb_print_info(void)
|
2008-12-17 04:17:34 +08:00
|
|
|
{
|
2009-11-10 18:46:19 +08:00
|
|
|
unsigned long bytes = io_tlb_nslabs << IO_TLB_SHIFT;
|
2008-12-17 04:17:34 +08:00
|
|
|
phys_addr_t pstart, pend;
|
|
|
|
|
|
|
|
pstart = virt_to_phys(io_tlb_start);
|
|
|
|
pend = virt_to_phys(io_tlb_end);
|
|
|
|
|
|
|
|
printk(KERN_INFO "Placing %luMB software IO TLB between %p - %p\n",
|
|
|
|
bytes >> 20, io_tlb_start, io_tlb_end);
|
2008-12-23 02:26:05 +08:00
|
|
|
printk(KERN_INFO "software IO TLB at phys %#llx - %#llx\n",
|
|
|
|
(unsigned long long)pstart,
|
|
|
|
(unsigned long long)pend);
|
2008-12-17 04:17:34 +08:00
|
|
|
}
|
|
|
|
|
2010-05-11 03:15:12 +08:00
|
|
|
void __init swiotlb_init_with_tbl(char *tlb, unsigned long nslabs, int verbose)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-02-06 10:51:25 +08:00
|
|
|
unsigned long i, bytes;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2010-05-11 03:15:12 +08:00
|
|
|
bytes = nslabs << IO_TLB_SHIFT;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2010-05-11 03:15:12 +08:00
|
|
|
io_tlb_nslabs = nslabs;
|
|
|
|
io_tlb_start = tlb;
|
2007-02-06 10:51:25 +08:00
|
|
|
io_tlb_end = io_tlb_start + bytes;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate and initialize the free list array. This array is used
|
|
|
|
* to find contiguous free memory regions of size up to IO_TLB_SEGSIZE
|
|
|
|
* between io_tlb_start and io_tlb_end.
|
|
|
|
*/
|
|
|
|
io_tlb_list = alloc_bootmem(io_tlb_nslabs * sizeof(int));
|
2007-03-07 05:31:45 +08:00
|
|
|
for (i = 0; i < io_tlb_nslabs; i++)
|
2005-04-17 06:20:36 +08:00
|
|
|
io_tlb_list[i] = IO_TLB_SEGSIZE - OFFSET(i, IO_TLB_SEGSIZE);
|
|
|
|
io_tlb_index = 0;
|
2008-12-23 02:26:08 +08:00
|
|
|
io_tlb_orig_addr = alloc_bootmem(io_tlb_nslabs * sizeof(phys_addr_t));
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Get the overflow emergency buffer
|
|
|
|
*/
|
|
|
|
io_tlb_overflow_buffer = alloc_bootmem_low(io_tlb_overflow);
|
2007-02-06 10:51:25 +08:00
|
|
|
if (!io_tlb_overflow_buffer)
|
|
|
|
panic("Cannot allocate SWIOTLB overflow buffer!\n");
|
2009-11-10 18:46:19 +08:00
|
|
|
if (verbose)
|
|
|
|
swiotlb_print_info();
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2010-05-11 03:15:12 +08:00
|
|
|
/*
|
|
|
|
* Statically reserve bounce buffer space and initialize bounce buffer data
|
|
|
|
* structures for the software IO TLB used to implement the DMA API.
|
|
|
|
*/
|
|
|
|
void __init
|
|
|
|
swiotlb_init_with_default_size(size_t default_size, int verbose)
|
|
|
|
{
|
|
|
|
unsigned long bytes;
|
|
|
|
|
|
|
|
if (!io_tlb_nslabs) {
|
|
|
|
io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
|
|
|
|
io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
|
|
|
|
}
|
|
|
|
|
|
|
|
bytes = io_tlb_nslabs << IO_TLB_SHIFT;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get IO TLB memory from the low pages
|
|
|
|
*/
|
|
|
|
io_tlb_start = alloc_bootmem_low_pages(bytes);
|
|
|
|
if (!io_tlb_start)
|
|
|
|
panic("Cannot allocate SWIOTLB buffer");
|
|
|
|
|
|
|
|
swiotlb_init_with_tbl(io_tlb_start, io_tlb_nslabs, verbose);
|
|
|
|
}
|
|
|
|
|
2007-02-06 10:51:25 +08:00
|
|
|
void __init
|
2009-11-10 18:46:19 +08:00
|
|
|
swiotlb_init(int verbose)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-11-10 18:46:19 +08:00
|
|
|
swiotlb_init_with_default_size(64 * (1<<20), verbose); /* default to 64MB */
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-09-07 01:20:49 +08:00
|
|
|
/*
|
|
|
|
* Systems with larger DMA zones (those that don't support ISA) can
|
|
|
|
* initialize the swiotlb later using the slab allocator if needed.
|
|
|
|
* This should be just like above, but with some error catching.
|
|
|
|
*/
|
|
|
|
int
|
2007-02-06 10:51:25 +08:00
|
|
|
swiotlb_late_init_with_default_size(size_t default_size)
|
2005-09-07 01:20:49 +08:00
|
|
|
{
|
2007-02-06 10:51:25 +08:00
|
|
|
unsigned long i, bytes, req_nslabs = io_tlb_nslabs;
|
2005-09-07 01:20:49 +08:00
|
|
|
unsigned int order;
|
|
|
|
|
|
|
|
if (!io_tlb_nslabs) {
|
|
|
|
io_tlb_nslabs = (default_size >> IO_TLB_SHIFT);
|
|
|
|
io_tlb_nslabs = ALIGN(io_tlb_nslabs, IO_TLB_SEGSIZE);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Get IO TLB memory from the low pages
|
|
|
|
*/
|
2007-02-06 10:51:25 +08:00
|
|
|
order = get_order(io_tlb_nslabs << IO_TLB_SHIFT);
|
2005-09-07 01:20:49 +08:00
|
|
|
io_tlb_nslabs = SLABS_PER_PAGE << order;
|
2007-02-06 10:51:25 +08:00
|
|
|
bytes = io_tlb_nslabs << IO_TLB_SHIFT;
|
2005-09-07 01:20:49 +08:00
|
|
|
|
|
|
|
while ((SLABS_PER_PAGE << order) > IO_TLB_MIN_SLABS) {
|
2009-07-10 09:04:51 +08:00
|
|
|
io_tlb_start = (void *)__get_free_pages(GFP_DMA | __GFP_NOWARN,
|
|
|
|
order);
|
2005-09-07 01:20:49 +08:00
|
|
|
if (io_tlb_start)
|
|
|
|
break;
|
|
|
|
order--;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!io_tlb_start)
|
|
|
|
goto cleanup1;
|
|
|
|
|
2007-02-06 10:51:25 +08:00
|
|
|
if (order != get_order(bytes)) {
|
2005-09-07 01:20:49 +08:00
|
|
|
printk(KERN_WARNING "Warning: only able to allocate %ld MB "
|
|
|
|
"for software IO TLB\n", (PAGE_SIZE << order) >> 20);
|
|
|
|
io_tlb_nslabs = SLABS_PER_PAGE << order;
|
2007-02-06 10:51:25 +08:00
|
|
|
bytes = io_tlb_nslabs << IO_TLB_SHIFT;
|
2005-09-07 01:20:49 +08:00
|
|
|
}
|
2007-02-06 10:51:25 +08:00
|
|
|
io_tlb_end = io_tlb_start + bytes;
|
|
|
|
memset(io_tlb_start, 0, bytes);
|
2005-09-07 01:20:49 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Allocate and initialize the free list array. This array is used
|
|
|
|
* to find contiguous free memory regions of size up to IO_TLB_SEGSIZE
|
|
|
|
* between io_tlb_start and io_tlb_end.
|
|
|
|
*/
|
|
|
|
io_tlb_list = (unsigned int *)__get_free_pages(GFP_KERNEL,
|
|
|
|
get_order(io_tlb_nslabs * sizeof(int)));
|
|
|
|
if (!io_tlb_list)
|
|
|
|
goto cleanup2;
|
|
|
|
|
|
|
|
for (i = 0; i < io_tlb_nslabs; i++)
|
|
|
|
io_tlb_list[i] = IO_TLB_SEGSIZE - OFFSET(i, IO_TLB_SEGSIZE);
|
|
|
|
io_tlb_index = 0;
|
|
|
|
|
2008-12-23 02:26:08 +08:00
|
|
|
io_tlb_orig_addr = (phys_addr_t *)
|
|
|
|
__get_free_pages(GFP_KERNEL,
|
|
|
|
get_order(io_tlb_nslabs *
|
|
|
|
sizeof(phys_addr_t)));
|
2005-09-07 01:20:49 +08:00
|
|
|
if (!io_tlb_orig_addr)
|
|
|
|
goto cleanup3;
|
|
|
|
|
2008-12-23 02:26:08 +08:00
|
|
|
memset(io_tlb_orig_addr, 0, io_tlb_nslabs * sizeof(phys_addr_t));
|
2005-09-07 01:20:49 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Get the overflow emergency buffer
|
|
|
|
*/
|
|
|
|
io_tlb_overflow_buffer = (void *)__get_free_pages(GFP_DMA,
|
|
|
|
get_order(io_tlb_overflow));
|
|
|
|
if (!io_tlb_overflow_buffer)
|
|
|
|
goto cleanup4;
|
|
|
|
|
2009-11-10 18:46:19 +08:00
|
|
|
swiotlb_print_info();
|
2005-09-07 01:20:49 +08:00
|
|
|
|
2009-11-10 18:46:18 +08:00
|
|
|
late_alloc = 1;
|
|
|
|
|
2005-09-07 01:20:49 +08:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
cleanup4:
|
2008-12-23 02:26:08 +08:00
|
|
|
free_pages((unsigned long)io_tlb_orig_addr,
|
|
|
|
get_order(io_tlb_nslabs * sizeof(phys_addr_t)));
|
2005-09-07 01:20:49 +08:00
|
|
|
io_tlb_orig_addr = NULL;
|
|
|
|
cleanup3:
|
2007-03-07 05:31:45 +08:00
|
|
|
free_pages((unsigned long)io_tlb_list, get_order(io_tlb_nslabs *
|
|
|
|
sizeof(int)));
|
2005-09-07 01:20:49 +08:00
|
|
|
io_tlb_list = NULL;
|
|
|
|
cleanup2:
|
2007-02-06 10:51:25 +08:00
|
|
|
io_tlb_end = NULL;
|
2005-09-07 01:20:49 +08:00
|
|
|
free_pages((unsigned long)io_tlb_start, order);
|
|
|
|
io_tlb_start = NULL;
|
|
|
|
cleanup1:
|
|
|
|
io_tlb_nslabs = req_nslabs;
|
|
|
|
return -ENOMEM;
|
|
|
|
}
|
|
|
|
|
2009-11-10 18:46:18 +08:00
|
|
|
void __init swiotlb_free(void)
|
|
|
|
{
|
|
|
|
if (!io_tlb_overflow_buffer)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (late_alloc) {
|
|
|
|
free_pages((unsigned long)io_tlb_overflow_buffer,
|
|
|
|
get_order(io_tlb_overflow));
|
|
|
|
free_pages((unsigned long)io_tlb_orig_addr,
|
|
|
|
get_order(io_tlb_nslabs * sizeof(phys_addr_t)));
|
|
|
|
free_pages((unsigned long)io_tlb_list, get_order(io_tlb_nslabs *
|
|
|
|
sizeof(int)));
|
|
|
|
free_pages((unsigned long)io_tlb_start,
|
|
|
|
get_order(io_tlb_nslabs << IO_TLB_SHIFT));
|
|
|
|
} else {
|
|
|
|
free_bootmem_late(__pa(io_tlb_overflow_buffer),
|
|
|
|
io_tlb_overflow);
|
|
|
|
free_bootmem_late(__pa(io_tlb_orig_addr),
|
|
|
|
io_tlb_nslabs * sizeof(phys_addr_t));
|
|
|
|
free_bootmem_late(__pa(io_tlb_list),
|
|
|
|
io_tlb_nslabs * sizeof(int));
|
|
|
|
free_bootmem_late(__pa(io_tlb_start),
|
|
|
|
io_tlb_nslabs << IO_TLB_SHIFT);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
static int is_swiotlb_buffer(phys_addr_t paddr)
|
2008-09-08 17:53:50 +08:00
|
|
|
{
|
2009-07-23 10:18:49 +08:00
|
|
|
return paddr >= virt_to_phys(io_tlb_start) &&
|
|
|
|
paddr < virt_to_phys(io_tlb_end);
|
2008-09-08 17:53:50 +08:00
|
|
|
}
|
|
|
|
|
2008-12-23 02:26:09 +08:00
|
|
|
/*
|
|
|
|
* Bounce: copy the swiotlb buffer back to the original dma location
|
|
|
|
*/
|
|
|
|
static void swiotlb_bounce(phys_addr_t phys, char *dma_addr, size_t size,
|
|
|
|
enum dma_data_direction dir)
|
|
|
|
{
|
|
|
|
unsigned long pfn = PFN_DOWN(phys);
|
|
|
|
|
|
|
|
if (PageHighMem(pfn_to_page(pfn))) {
|
|
|
|
/* The buffer does not have a mapping. Map it in and copy */
|
|
|
|
unsigned int offset = phys & ~PAGE_MASK;
|
|
|
|
char *buffer;
|
|
|
|
unsigned int sz = 0;
|
|
|
|
unsigned long flags;
|
|
|
|
|
|
|
|
while (size) {
|
2009-04-08 22:09:16 +08:00
|
|
|
sz = min_t(size_t, PAGE_SIZE - offset, size);
|
2008-12-23 02:26:09 +08:00
|
|
|
|
|
|
|
local_irq_save(flags);
|
|
|
|
buffer = kmap_atomic(pfn_to_page(pfn),
|
|
|
|
KM_BOUNCE_READ);
|
|
|
|
if (dir == DMA_TO_DEVICE)
|
|
|
|
memcpy(dma_addr, buffer + offset, sz);
|
2008-12-17 04:17:33 +08:00
|
|
|
else
|
2008-12-23 02:26:09 +08:00
|
|
|
memcpy(buffer + offset, dma_addr, sz);
|
|
|
|
kunmap_atomic(buffer, KM_BOUNCE_READ);
|
2008-12-17 04:17:33 +08:00
|
|
|
local_irq_restore(flags);
|
2008-12-23 02:26:09 +08:00
|
|
|
|
|
|
|
size -= sz;
|
|
|
|
pfn++;
|
|
|
|
dma_addr += sz;
|
|
|
|
offset = 0;
|
2008-12-17 04:17:33 +08:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
if (dir == DMA_TO_DEVICE)
|
2008-12-23 02:26:09 +08:00
|
|
|
memcpy(dma_addr, phys_to_virt(phys), size);
|
2008-12-17 04:17:33 +08:00
|
|
|
else
|
2008-12-23 02:26:09 +08:00
|
|
|
memcpy(phys_to_virt(phys), dma_addr, size);
|
2008-12-17 04:17:33 +08:00
|
|
|
}
|
2008-12-17 04:17:32 +08:00
|
|
|
}
|
|
|
|
|
swiotlb: add swiotlb_tbl_map_single library function
swiotlb_tbl_map_single() takes the dma address of iotlb instead of
using swiotlb_virt_to_bus().
[v2: changed swiotlb_tlb to swiotlb_tbl]
[v3: changed u64 to dma_addr_t]
This patch:
This is a set of patches that separate the address translation
(virt_to_phys, virt_to_bus, etc) and allocation of the SWIOTLB buffer
from the SWIOTLB library.
The idea behind this set of patches is to make it possible to have separate
mechanisms for translating virtual to physical or virtual to DMA addresses
on platforms which need an SWIOTLB, and where physical != PCI bus address
and also to allocate the core IOTLB memory outside SWIOTLB.
One customers of this is the pv-ops project, which can switch between
different modes of operation depending on the environment it is running in:
bare-metal or virtualized (Xen for now). Another is the Wii DMA - used to
implement the MEM2 DMA facility needed by its EHCI controller (for details:
http://lkml.org/lkml/2010/5/18/303)
On bare-metal SWIOTLB is used when there are no hardware IOMMU. In virtualized
environment it used when PCI pass-through is enabled for the guest. The problems
with PCI pass-through is that the guest's idea of PFN's is not the real thing.
To fix that, there is translation layer for PFN->machine frame number and vice-versa.
To bubble that up to the SWIOTLB layer there are two possible solutions.
One solution has been to wholesale copy the SWIOTLB, stick it in
arch/x86/xen/swiotlb.c and modify the virt_to_phys, phys_to_virt and others
to use the Xen address translation functions. Unfortunately, since the kernel can
run on bare-metal, there would be big code overlap with the real SWIOTLB.
(git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git xen/dom0/swiotlb-new)
Another approach, which this set of patches explores, is to abstract the
address translation and address determination functions away from the
SWIOTLB book-keeping functions. This way the core SWIOTLB library functions
are present in one place, while the address related functions are in
a separate library that can be loaded when running under non-bare-metal platform.
Changelog:
Since the last posting [v8.2] Konrad has done:
- Added this changelog in the patch and referenced in the other patches
this description.
- 'enum dma_data_direction direction' to 'enum dma.. dir' so to be
unified.
[v8-v8.2 changes:]
- Rolled-up the last two patches in one.
- Rebased against linus latest. That meant dealing with swiotlb_sync_single_range_* changes.
- added Acked-by: Fujita Tomonori and Tested-by: Albert Herranz
[v7-v8 changes:]
- Minimized the list of exported functions.
- Integrated Fujita's patches and changed "swiotlb_tlb" to "swiotlb_tbl" in them.
[v6-v7 changes:]
- Minimized the amount of exported functions/variable with a prefix of: "swiotbl_tbl".
- Made the usage of 'int dir' to be 'enum dma_data_direction'.
[v5-v6 changes:]
- Made the exported functions/variables have the 'swiotlb_bk' prefix.
- dropped the checkpatches/other reworks
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: Albert Herranz <albert_herranz@yahoo.es>
2010-05-11 03:14:54 +08:00
|
|
|
void *swiotlb_tbl_map_single(struct device *hwdev, dma_addr_t tbl_dma_addr,
|
|
|
|
phys_addr_t phys, size_t size, int dir)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
char *dma_addr;
|
|
|
|
unsigned int nslots, stride, index, wrap;
|
|
|
|
int i;
|
2008-02-05 14:28:16 +08:00
|
|
|
unsigned long mask;
|
|
|
|
unsigned long offset_slots;
|
|
|
|
unsigned long max_slots;
|
|
|
|
|
|
|
|
mask = dma_get_seg_boundary(hwdev);
|
|
|
|
|
swiotlb: add swiotlb_tbl_map_single library function
swiotlb_tbl_map_single() takes the dma address of iotlb instead of
using swiotlb_virt_to_bus().
[v2: changed swiotlb_tlb to swiotlb_tbl]
[v3: changed u64 to dma_addr_t]
This patch:
This is a set of patches that separate the address translation
(virt_to_phys, virt_to_bus, etc) and allocation of the SWIOTLB buffer
from the SWIOTLB library.
The idea behind this set of patches is to make it possible to have separate
mechanisms for translating virtual to physical or virtual to DMA addresses
on platforms which need an SWIOTLB, and where physical != PCI bus address
and also to allocate the core IOTLB memory outside SWIOTLB.
One customers of this is the pv-ops project, which can switch between
different modes of operation depending on the environment it is running in:
bare-metal or virtualized (Xen for now). Another is the Wii DMA - used to
implement the MEM2 DMA facility needed by its EHCI controller (for details:
http://lkml.org/lkml/2010/5/18/303)
On bare-metal SWIOTLB is used when there are no hardware IOMMU. In virtualized
environment it used when PCI pass-through is enabled for the guest. The problems
with PCI pass-through is that the guest's idea of PFN's is not the real thing.
To fix that, there is translation layer for PFN->machine frame number and vice-versa.
To bubble that up to the SWIOTLB layer there are two possible solutions.
One solution has been to wholesale copy the SWIOTLB, stick it in
arch/x86/xen/swiotlb.c and modify the virt_to_phys, phys_to_virt and others
to use the Xen address translation functions. Unfortunately, since the kernel can
run on bare-metal, there would be big code overlap with the real SWIOTLB.
(git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git xen/dom0/swiotlb-new)
Another approach, which this set of patches explores, is to abstract the
address translation and address determination functions away from the
SWIOTLB book-keeping functions. This way the core SWIOTLB library functions
are present in one place, while the address related functions are in
a separate library that can be loaded when running under non-bare-metal platform.
Changelog:
Since the last posting [v8.2] Konrad has done:
- Added this changelog in the patch and referenced in the other patches
this description.
- 'enum dma_data_direction direction' to 'enum dma.. dir' so to be
unified.
[v8-v8.2 changes:]
- Rolled-up the last two patches in one.
- Rebased against linus latest. That meant dealing with swiotlb_sync_single_range_* changes.
- added Acked-by: Fujita Tomonori and Tested-by: Albert Herranz
[v7-v8 changes:]
- Minimized the list of exported functions.
- Integrated Fujita's patches and changed "swiotlb_tlb" to "swiotlb_tbl" in them.
[v6-v7 changes:]
- Minimized the amount of exported functions/variable with a prefix of: "swiotbl_tbl".
- Made the usage of 'int dir' to be 'enum dma_data_direction'.
[v5-v6 changes:]
- Made the exported functions/variables have the 'swiotlb_bk' prefix.
- dropped the checkpatches/other reworks
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: Albert Herranz <albert_herranz@yahoo.es>
2010-05-11 03:14:54 +08:00
|
|
|
tbl_dma_addr &= mask;
|
|
|
|
|
|
|
|
offset_slots = ALIGN(tbl_dma_addr, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
|
2008-12-17 04:17:29 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Carefully handle integer overflow which can occur when mask == ~0UL.
|
|
|
|
*/
|
2008-03-13 17:13:30 +08:00
|
|
|
max_slots = mask + 1
|
|
|
|
? ALIGN(mask + 1, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT
|
|
|
|
: 1UL << (BITS_PER_LONG - IO_TLB_SHIFT);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* For mappings greater than a page, we limit the stride (and
|
|
|
|
* hence alignment) to a page size.
|
|
|
|
*/
|
|
|
|
nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
|
|
|
|
if (size > PAGE_SIZE)
|
|
|
|
stride = (1 << (PAGE_SHIFT - IO_TLB_SHIFT));
|
|
|
|
else
|
|
|
|
stride = 1;
|
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(!nslots);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Find suitable number of IO TLB entries size that will fit this
|
|
|
|
* request and allocate a buffer from that IO TLB pool.
|
|
|
|
*/
|
|
|
|
spin_lock_irqsave(&io_tlb_lock, flags);
|
2008-04-29 15:59:36 +08:00
|
|
|
index = ALIGN(io_tlb_index, stride);
|
|
|
|
if (index >= io_tlb_nslabs)
|
|
|
|
index = 0;
|
|
|
|
wrap = index;
|
|
|
|
|
|
|
|
do {
|
2008-04-29 15:59:36 +08:00
|
|
|
while (iommu_is_span_boundary(index, nslots, offset_slots,
|
|
|
|
max_slots)) {
|
2008-03-13 17:13:30 +08:00
|
|
|
index += stride;
|
|
|
|
if (index >= io_tlb_nslabs)
|
|
|
|
index = 0;
|
2008-04-29 15:59:36 +08:00
|
|
|
if (index == wrap)
|
|
|
|
goto not_found;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If we find a slot that indicates we have 'nslots' number of
|
|
|
|
* contiguous buffers, we allocate the buffers from that slot
|
|
|
|
* and mark the entries as '0' indicating unavailable.
|
|
|
|
*/
|
|
|
|
if (io_tlb_list[index] >= nslots) {
|
|
|
|
int count = 0;
|
|
|
|
|
|
|
|
for (i = index; i < (int) (index + nslots); i++)
|
|
|
|
io_tlb_list[i] = 0;
|
|
|
|
for (i = index - 1; (OFFSET(i, IO_TLB_SEGSIZE) != IO_TLB_SEGSIZE - 1) && io_tlb_list[i]; i--)
|
|
|
|
io_tlb_list[i] = ++count;
|
|
|
|
dma_addr = io_tlb_start + (index << IO_TLB_SHIFT);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2008-04-29 15:59:36 +08:00
|
|
|
/*
|
|
|
|
* Update the indices to avoid searching in the next
|
|
|
|
* round.
|
|
|
|
*/
|
|
|
|
io_tlb_index = ((index + nslots) < io_tlb_nslabs
|
|
|
|
? (index + nslots) : 0);
|
|
|
|
|
|
|
|
goto found;
|
|
|
|
}
|
|
|
|
index += stride;
|
|
|
|
if (index >= io_tlb_nslabs)
|
|
|
|
index = 0;
|
|
|
|
} while (index != wrap);
|
|
|
|
|
|
|
|
not_found:
|
|
|
|
spin_unlock_irqrestore(&io_tlb_lock, flags);
|
|
|
|
return NULL;
|
|
|
|
found:
|
2005-04-17 06:20:36 +08:00
|
|
|
spin_unlock_irqrestore(&io_tlb_lock, flags);
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Save away the mapping from the original address to the DMA address.
|
|
|
|
* This is needed when we sync the memory. Then we sync the buffer if
|
|
|
|
* needed.
|
|
|
|
*/
|
2008-12-23 02:26:08 +08:00
|
|
|
for (i = 0; i < nslots; i++)
|
|
|
|
io_tlb_orig_addr[index+i] = phys + (i << IO_TLB_SHIFT);
|
2005-04-17 06:20:36 +08:00
|
|
|
if (dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL)
|
2008-12-23 02:26:09 +08:00
|
|
|
swiotlb_bounce(phys, dma_addr, size, DMA_TO_DEVICE);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
return dma_addr;
|
|
|
|
}
|
|
|
|
|
swiotlb: add swiotlb_tbl_map_single library function
swiotlb_tbl_map_single() takes the dma address of iotlb instead of
using swiotlb_virt_to_bus().
[v2: changed swiotlb_tlb to swiotlb_tbl]
[v3: changed u64 to dma_addr_t]
This patch:
This is a set of patches that separate the address translation
(virt_to_phys, virt_to_bus, etc) and allocation of the SWIOTLB buffer
from the SWIOTLB library.
The idea behind this set of patches is to make it possible to have separate
mechanisms for translating virtual to physical or virtual to DMA addresses
on platforms which need an SWIOTLB, and where physical != PCI bus address
and also to allocate the core IOTLB memory outside SWIOTLB.
One customers of this is the pv-ops project, which can switch between
different modes of operation depending on the environment it is running in:
bare-metal or virtualized (Xen for now). Another is the Wii DMA - used to
implement the MEM2 DMA facility needed by its EHCI controller (for details:
http://lkml.org/lkml/2010/5/18/303)
On bare-metal SWIOTLB is used when there are no hardware IOMMU. In virtualized
environment it used when PCI pass-through is enabled for the guest. The problems
with PCI pass-through is that the guest's idea of PFN's is not the real thing.
To fix that, there is translation layer for PFN->machine frame number and vice-versa.
To bubble that up to the SWIOTLB layer there are two possible solutions.
One solution has been to wholesale copy the SWIOTLB, stick it in
arch/x86/xen/swiotlb.c and modify the virt_to_phys, phys_to_virt and others
to use the Xen address translation functions. Unfortunately, since the kernel can
run on bare-metal, there would be big code overlap with the real SWIOTLB.
(git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git xen/dom0/swiotlb-new)
Another approach, which this set of patches explores, is to abstract the
address translation and address determination functions away from the
SWIOTLB book-keeping functions. This way the core SWIOTLB library functions
are present in one place, while the address related functions are in
a separate library that can be loaded when running under non-bare-metal platform.
Changelog:
Since the last posting [v8.2] Konrad has done:
- Added this changelog in the patch and referenced in the other patches
this description.
- 'enum dma_data_direction direction' to 'enum dma.. dir' so to be
unified.
[v8-v8.2 changes:]
- Rolled-up the last two patches in one.
- Rebased against linus latest. That meant dealing with swiotlb_sync_single_range_* changes.
- added Acked-by: Fujita Tomonori and Tested-by: Albert Herranz
[v7-v8 changes:]
- Minimized the list of exported functions.
- Integrated Fujita's patches and changed "swiotlb_tlb" to "swiotlb_tbl" in them.
[v6-v7 changes:]
- Minimized the amount of exported functions/variable with a prefix of: "swiotbl_tbl".
- Made the usage of 'int dir' to be 'enum dma_data_direction'.
[v5-v6 changes:]
- Made the exported functions/variables have the 'swiotlb_bk' prefix.
- dropped the checkpatches/other reworks
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: Albert Herranz <albert_herranz@yahoo.es>
2010-05-11 03:14:54 +08:00
|
|
|
/*
|
|
|
|
* Allocates bounce buffer and returns its kernel virtual address.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static void *
|
|
|
|
map_single(struct device *hwdev, phys_addr_t phys, size_t size, int dir)
|
|
|
|
{
|
|
|
|
dma_addr_t start_dma_addr = swiotlb_virt_to_bus(hwdev, io_tlb_start);
|
|
|
|
|
|
|
|
return swiotlb_tbl_map_single(hwdev, start_dma_addr, phys, size, dir);
|
|
|
|
}
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* dma_addr is the kernel virtual address of the bounce buffer to unmap.
|
|
|
|
*/
|
|
|
|
static void
|
2009-04-08 22:09:19 +08:00
|
|
|
do_unmap_single(struct device *hwdev, char *dma_addr, size_t size, int dir)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
|
|
|
unsigned long flags;
|
|
|
|
int i, count, nslots = ALIGN(size, 1 << IO_TLB_SHIFT) >> IO_TLB_SHIFT;
|
|
|
|
int index = (dma_addr - io_tlb_start) >> IO_TLB_SHIFT;
|
2008-12-23 02:26:08 +08:00
|
|
|
phys_addr_t phys = io_tlb_orig_addr[index];
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* First, sync the memory before unmapping the entry
|
|
|
|
*/
|
2008-12-23 02:26:08 +08:00
|
|
|
if (phys && ((dir == DMA_FROM_DEVICE) || (dir == DMA_BIDIRECTIONAL)))
|
2008-12-23 02:26:09 +08:00
|
|
|
swiotlb_bounce(phys, dma_addr, size, DMA_FROM_DEVICE);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Return the buffer to the free list by setting the corresponding
|
tree-wide: fix assorted typos all over the place
That is "success", "unknown", "through", "performance", "[re|un]mapping"
, "access", "default", "reasonable", "[con]currently", "temperature"
, "channel", "[un]used", "application", "example","hierarchy", "therefore"
, "[over|under]flow", "contiguous", "threshold", "enough" and others.
Signed-off-by: André Goddard Rosa <andre.goddard@gmail.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
2009-11-14 23:09:05 +08:00
|
|
|
* entries to indicate the number of contiguous entries available.
|
2005-04-17 06:20:36 +08:00
|
|
|
* While returning the entries to the free list, we merge the entries
|
|
|
|
* with slots below and above the pool being returned.
|
|
|
|
*/
|
|
|
|
spin_lock_irqsave(&io_tlb_lock, flags);
|
|
|
|
{
|
|
|
|
count = ((index + nslots) < ALIGN(index + 1, IO_TLB_SEGSIZE) ?
|
|
|
|
io_tlb_list[index + nslots] : 0);
|
|
|
|
/*
|
|
|
|
* Step 1: return the slots to the free list, merging the
|
|
|
|
* slots with superceeding slots
|
|
|
|
*/
|
|
|
|
for (i = index + nslots - 1; i >= index; i--)
|
|
|
|
io_tlb_list[i] = ++count;
|
|
|
|
/*
|
|
|
|
* Step 2: merge the returned slots with the preceding slots,
|
|
|
|
* if available (non zero)
|
|
|
|
*/
|
|
|
|
for (i = index - 1; (OFFSET(i, IO_TLB_SEGSIZE) != IO_TLB_SEGSIZE -1) && io_tlb_list[i]; i--)
|
|
|
|
io_tlb_list[i] = ++count;
|
|
|
|
}
|
|
|
|
spin_unlock_irqrestore(&io_tlb_lock, flags);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2005-09-30 05:44:57 +08:00
|
|
|
sync_single(struct device *hwdev, char *dma_addr, size_t size,
|
|
|
|
int dir, int target)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2008-12-23 02:26:08 +08:00
|
|
|
int index = (dma_addr - io_tlb_start) >> IO_TLB_SHIFT;
|
|
|
|
phys_addr_t phys = io_tlb_orig_addr[index];
|
|
|
|
|
|
|
|
phys += ((unsigned long)dma_addr & ((1 << IO_TLB_SHIFT) - 1));
|
2007-07-21 19:37:24 +08:00
|
|
|
|
2005-09-30 05:44:57 +08:00
|
|
|
switch (target) {
|
|
|
|
case SYNC_FOR_CPU:
|
|
|
|
if (likely(dir == DMA_FROM_DEVICE || dir == DMA_BIDIRECTIONAL))
|
2008-12-23 02:26:09 +08:00
|
|
|
swiotlb_bounce(phys, dma_addr, size, DMA_FROM_DEVICE);
|
2006-03-25 01:47:11 +08:00
|
|
|
else
|
|
|
|
BUG_ON(dir != DMA_TO_DEVICE);
|
2005-09-30 05:44:57 +08:00
|
|
|
break;
|
|
|
|
case SYNC_FOR_DEVICE:
|
|
|
|
if (likely(dir == DMA_TO_DEVICE || dir == DMA_BIDIRECTIONAL))
|
2008-12-23 02:26:09 +08:00
|
|
|
swiotlb_bounce(phys, dma_addr, size, DMA_TO_DEVICE);
|
2006-03-25 01:47:11 +08:00
|
|
|
else
|
|
|
|
BUG_ON(dir != DMA_FROM_DEVICE);
|
2005-09-30 05:44:57 +08:00
|
|
|
break;
|
|
|
|
default:
|
2005-04-17 06:20:36 +08:00
|
|
|
BUG();
|
2005-09-30 05:44:57 +08:00
|
|
|
}
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void *
|
|
|
|
swiotlb_alloc_coherent(struct device *hwdev, size_t size,
|
2005-10-21 15:21:03 +08:00
|
|
|
dma_addr_t *dma_handle, gfp_t flags)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-02-06 10:51:25 +08:00
|
|
|
dma_addr_t dev_addr;
|
2005-04-17 06:20:36 +08:00
|
|
|
void *ret;
|
|
|
|
int order = get_order(size);
|
2009-04-07 10:01:15 +08:00
|
|
|
u64 dma_mask = DMA_BIT_MASK(32);
|
2008-11-17 15:24:34 +08:00
|
|
|
|
|
|
|
if (hwdev && hwdev->coherent_dma_mask)
|
|
|
|
dma_mask = hwdev->coherent_dma_mask;
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-03-07 05:31:45 +08:00
|
|
|
ret = (void *)__get_free_pages(flags, order);
|
2009-12-16 08:47:43 +08:00
|
|
|
if (ret && swiotlb_virt_to_bus(hwdev, ret) + size - 1 > dma_mask) {
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* The allocated memory isn't reachable by the device.
|
|
|
|
*/
|
|
|
|
free_pages((unsigned long) ret, order);
|
|
|
|
ret = NULL;
|
|
|
|
}
|
|
|
|
if (!ret) {
|
|
|
|
/*
|
|
|
|
* We are either out of memory or the device can't DMA
|
2009-04-08 22:09:15 +08:00
|
|
|
* to GFP_DMA memory; fall back on map_single(), which
|
|
|
|
* will grab memory from the lowest available address range.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2008-12-23 02:26:08 +08:00
|
|
|
ret = map_single(hwdev, 0, size, DMA_FROM_DEVICE);
|
2008-09-08 17:53:48 +08:00
|
|
|
if (!ret)
|
2005-04-17 06:20:36 +08:00
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
memset(ret, 0, size);
|
2008-12-23 02:26:05 +08:00
|
|
|
dev_addr = swiotlb_virt_to_bus(hwdev, ret);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/* Confirm address can be DMA'd by device */
|
2009-12-16 08:47:43 +08:00
|
|
|
if (dev_addr + size - 1 > dma_mask) {
|
2007-02-06 10:51:25 +08:00
|
|
|
printk("hwdev DMA mask = 0x%016Lx, dev_addr = 0x%016Lx\n",
|
2008-11-17 15:24:34 +08:00
|
|
|
(unsigned long long)dma_mask,
|
2007-02-06 10:51:25 +08:00
|
|
|
(unsigned long long)dev_addr);
|
2008-10-23 17:42:03 +08:00
|
|
|
|
|
|
|
/* DMA_TO_DEVICE to avoid memcpy in unmap_single */
|
2009-04-08 22:09:19 +08:00
|
|
|
do_unmap_single(hwdev, ret, size, DMA_TO_DEVICE);
|
2008-10-23 17:42:03 +08:00
|
|
|
return NULL;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
*dma_handle = dev_addr;
|
|
|
|
return ret;
|
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_alloc_coherent);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
void
|
|
|
|
swiotlb_free_coherent(struct device *hwdev, size_t size, void *vaddr,
|
2009-07-23 10:18:49 +08:00
|
|
|
dma_addr_t dev_addr)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-07-10 09:05:02 +08:00
|
|
|
phys_addr_t paddr = dma_to_phys(hwdev, dev_addr);
|
2009-07-23 10:18:49 +08:00
|
|
|
|
2007-08-11 04:10:27 +08:00
|
|
|
WARN_ON(irqs_disabled());
|
2009-07-23 10:18:49 +08:00
|
|
|
if (!is_swiotlb_buffer(paddr))
|
|
|
|
free_pages((unsigned long)vaddr, get_order(size));
|
2005-04-17 06:20:36 +08:00
|
|
|
else
|
|
|
|
/* DMA_TO_DEVICE to avoid memcpy in unmap_single */
|
2009-04-08 22:09:19 +08:00
|
|
|
do_unmap_single(hwdev, vaddr, size, DMA_TO_DEVICE);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_free_coherent);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
static void
|
|
|
|
swiotlb_full(struct device *dev, size_t size, int dir, int do_panic)
|
|
|
|
{
|
|
|
|
/*
|
|
|
|
* Ran out of IOMMU space for this operation. This is very bad.
|
|
|
|
* Unfortunately the drivers cannot handle this operation properly.
|
2005-09-30 06:52:13 +08:00
|
|
|
* unless they check for dma_mapping_error (most don't)
|
2005-04-17 06:20:36 +08:00
|
|
|
* When the mapping is small enough return a static buffer to limit
|
|
|
|
* the damage, or panic when the transfer is too big.
|
|
|
|
*/
|
2007-02-06 10:51:25 +08:00
|
|
|
printk(KERN_ERR "DMA: Out of SW-IOMMU space for %zu bytes at "
|
2009-01-07 02:44:37 +08:00
|
|
|
"device %s\n", size, dev ? dev_name(dev) : "?");
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-08-21 07:27:56 +08:00
|
|
|
if (size <= io_tlb_overflow || !do_panic)
|
|
|
|
return;
|
|
|
|
|
|
|
|
if (dir == DMA_BIDIRECTIONAL)
|
|
|
|
panic("DMA: Random memory could be DMA accessed\n");
|
|
|
|
if (dir == DMA_FROM_DEVICE)
|
|
|
|
panic("DMA: Random memory could be DMA written\n");
|
|
|
|
if (dir == DMA_TO_DEVICE)
|
|
|
|
panic("DMA: Random memory could be DMA read\n");
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Map a single buffer of the indicated size for DMA in streaming mode. The
|
2005-09-30 06:52:13 +08:00
|
|
|
* physical address to use is returned.
|
2005-04-17 06:20:36 +08:00
|
|
|
*
|
|
|
|
* Once the device is given the dma address, the device owns this memory until
|
2009-04-08 22:09:15 +08:00
|
|
|
* either swiotlb_unmap_page or swiotlb_dma_sync_single is performed.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
2009-01-05 22:59:03 +08:00
|
|
|
dma_addr_t swiotlb_map_page(struct device *dev, struct page *page,
|
|
|
|
unsigned long offset, size_t size,
|
|
|
|
enum dma_data_direction dir,
|
|
|
|
struct dma_attrs *attrs)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-01-05 22:59:03 +08:00
|
|
|
phys_addr_t phys = page_to_phys(page) + offset;
|
2009-07-10 09:05:02 +08:00
|
|
|
dma_addr_t dev_addr = phys_to_dma(dev, phys);
|
2005-04-17 06:20:36 +08:00
|
|
|
void *map;
|
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(dir == DMA_NONE);
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
2009-04-08 22:09:15 +08:00
|
|
|
* If the address happens to be in the device's DMA window,
|
2005-04-17 06:20:36 +08:00
|
|
|
* we can safely return the device addr and not worry about bounce
|
|
|
|
* buffering it.
|
|
|
|
*/
|
2009-07-10 09:04:58 +08:00
|
|
|
if (dma_capable(dev, dev_addr, size) && !swiotlb_force)
|
2005-04-17 06:20:36 +08:00
|
|
|
return dev_addr;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Oh well, have to allocate and map a bounce buffer.
|
|
|
|
*/
|
2009-01-05 22:59:03 +08:00
|
|
|
map = map_single(dev, phys, size, dir);
|
2005-04-17 06:20:36 +08:00
|
|
|
if (!map) {
|
2009-01-05 22:59:03 +08:00
|
|
|
swiotlb_full(dev, size, dir, 1);
|
2005-04-17 06:20:36 +08:00
|
|
|
map = io_tlb_overflow_buffer;
|
|
|
|
}
|
|
|
|
|
2009-01-05 22:59:03 +08:00
|
|
|
dev_addr = swiotlb_virt_to_bus(dev, map);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Ensure that the address returned is DMA'ble
|
|
|
|
*/
|
2009-07-10 09:04:58 +08:00
|
|
|
if (!dma_capable(dev, dev_addr, size))
|
2005-04-17 06:20:36 +08:00
|
|
|
panic("map_single: bounce buffer is not DMA'ble");
|
|
|
|
|
|
|
|
return dev_addr;
|
|
|
|
}
|
2009-01-05 22:59:03 +08:00
|
|
|
EXPORT_SYMBOL_GPL(swiotlb_map_page);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Unmap a single streaming mode DMA translation. The dma_addr and size must
|
2009-04-08 22:09:15 +08:00
|
|
|
* match what was provided for in a previous swiotlb_map_page call. All
|
2005-04-17 06:20:36 +08:00
|
|
|
* other usages are undefined.
|
|
|
|
*
|
|
|
|
* After this call, reads by the cpu to the buffer are guaranteed to see
|
|
|
|
* whatever the device wrote there.
|
|
|
|
*/
|
2009-04-08 22:09:19 +08:00
|
|
|
static void unmap_single(struct device *hwdev, dma_addr_t dev_addr,
|
|
|
|
size_t size, int dir)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-07-10 09:05:02 +08:00
|
|
|
phys_addr_t paddr = dma_to_phys(hwdev, dev_addr);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(dir == DMA_NONE);
|
2009-04-08 22:09:19 +08:00
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
if (is_swiotlb_buffer(paddr)) {
|
|
|
|
do_unmap_single(hwdev, phys_to_virt(paddr), size, dir);
|
2009-04-08 22:09:19 +08:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (dir != DMA_FROM_DEVICE)
|
|
|
|
return;
|
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
/*
|
|
|
|
* phys_to_virt doesn't work with hihgmem page but we could
|
|
|
|
* call dma_mark_clean() with hihgmem page here. However, we
|
|
|
|
* are fine since dma_mark_clean() is null on POWERPC. We can
|
|
|
|
* make dma_mark_clean() take a physical address if necessary.
|
|
|
|
*/
|
|
|
|
dma_mark_clean(phys_to_virt(paddr), size);
|
2009-04-08 22:09:19 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
void swiotlb_unmap_page(struct device *hwdev, dma_addr_t dev_addr,
|
|
|
|
size_t size, enum dma_data_direction dir,
|
|
|
|
struct dma_attrs *attrs)
|
|
|
|
{
|
|
|
|
unmap_single(hwdev, dev_addr, size, dir);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2009-01-05 22:59:03 +08:00
|
|
|
EXPORT_SYMBOL_GPL(swiotlb_unmap_page);
|
2008-12-28 14:02:07 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
/*
|
|
|
|
* Make physical memory consistent for a single streaming mode DMA translation
|
|
|
|
* after a transfer.
|
|
|
|
*
|
2009-04-08 22:09:15 +08:00
|
|
|
* If you perform a swiotlb_map_page() but wish to interrogate the buffer
|
2005-09-30 06:52:13 +08:00
|
|
|
* using the cpu, yet do not wish to teardown the dma mapping, you must
|
|
|
|
* call this function before doing so. At the next point you give the dma
|
2005-04-17 06:20:36 +08:00
|
|
|
* address back to the card, you must first perform a
|
|
|
|
* swiotlb_dma_sync_for_device, and then the device again owns the buffer
|
|
|
|
*/
|
[PATCH] swiotlb uninlinings
Optimise swiotlb.c for size.
text data bss dec hex filename
5009 89 64 5162 142a lib/swiotlb.o-before
4666 89 64 4819 12d3 lib/swiotlb.o-after
For some reason my gcc (4.0.2) doesn't want to tailcall these things.
swiotlb_sync_sg_for_device:
pushq %rbp #
movl $1, %r8d #,
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret
.size swiotlb_sync_sg_for_device, .-swiotlb_sync_sg_for_device
.section .text.swiotlb_sync_sg_for_cpu,"ax",@progbits
.globl swiotlb_sync_sg_for_cpu
.type swiotlb_sync_sg_for_cpu, @function
swiotlb_sync_sg_for_cpu:
pushq %rbp #
xorl %r8d, %r8d #
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 16:52:17 +08:00
|
|
|
static void
|
2005-09-30 05:43:32 +08:00
|
|
|
swiotlb_sync_single(struct device *hwdev, dma_addr_t dev_addr,
|
2005-09-30 05:44:57 +08:00
|
|
|
size_t size, int dir, int target)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2009-07-10 09:05:02 +08:00
|
|
|
phys_addr_t paddr = dma_to_phys(hwdev, dev_addr);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(dir == DMA_NONE);
|
2009-04-08 22:09:20 +08:00
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
if (is_swiotlb_buffer(paddr)) {
|
|
|
|
sync_single(hwdev, phys_to_virt(paddr), size, dir, target);
|
2009-04-08 22:09:20 +08:00
|
|
|
return;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (dir != DMA_FROM_DEVICE)
|
|
|
|
return;
|
|
|
|
|
2009-07-23 10:18:49 +08:00
|
|
|
dma_mark_clean(phys_to_virt(paddr), size);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-09-30 05:43:32 +08:00
|
|
|
void
|
|
|
|
swiotlb_sync_single_for_cpu(struct device *hwdev, dma_addr_t dev_addr,
|
2009-01-05 22:59:02 +08:00
|
|
|
size_t size, enum dma_data_direction dir)
|
2005-09-30 05:43:32 +08:00
|
|
|
{
|
2005-09-30 05:44:57 +08:00
|
|
|
swiotlb_sync_single(hwdev, dev_addr, size, dir, SYNC_FOR_CPU);
|
2005-09-30 05:43:32 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_sync_single_for_cpu);
|
2005-09-30 05:43:32 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
void
|
|
|
|
swiotlb_sync_single_for_device(struct device *hwdev, dma_addr_t dev_addr,
|
2009-01-05 22:59:02 +08:00
|
|
|
size_t size, enum dma_data_direction dir)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2005-09-30 05:44:57 +08:00
|
|
|
swiotlb_sync_single(hwdev, dev_addr, size, dir, SYNC_FOR_DEVICE);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_sync_single_for_device);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Map a set of buffers described by scatterlist in streaming mode for DMA.
|
2009-04-08 22:09:15 +08:00
|
|
|
* This is the scatter-gather version of the above swiotlb_map_page
|
2005-04-17 06:20:36 +08:00
|
|
|
* interface. Here the scatter gather list elements are each tagged with the
|
|
|
|
* appropriate dma address and length. They are obtained via
|
|
|
|
* sg_dma_{address,length}(SG).
|
|
|
|
*
|
|
|
|
* NOTE: An implementation may be able to use a smaller number of
|
|
|
|
* DMA address/length pairs than there are SG table elements.
|
|
|
|
* (for example via virtual mapping capabilities)
|
|
|
|
* The routine returns the number of addr/length pairs actually
|
|
|
|
* used, at most nents.
|
|
|
|
*
|
2009-04-08 22:09:15 +08:00
|
|
|
* Device ownership issues as mentioned above for swiotlb_map_page are the
|
2005-04-17 06:20:36 +08:00
|
|
|
* same here.
|
|
|
|
*/
|
|
|
|
int
|
2008-04-29 16:00:32 +08:00
|
|
|
swiotlb_map_sg_attrs(struct device *hwdev, struct scatterlist *sgl, int nelems,
|
2009-01-05 22:59:02 +08:00
|
|
|
enum dma_data_direction dir, struct dma_attrs *attrs)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-05-11 20:56:18 +08:00
|
|
|
struct scatterlist *sg;
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(dir == DMA_NONE);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2007-05-11 20:56:18 +08:00
|
|
|
for_each_sg(sgl, sg, nelems, i) {
|
2009-01-10 02:32:10 +08:00
|
|
|
phys_addr_t paddr = sg_phys(sg);
|
2009-07-10 09:05:02 +08:00
|
|
|
dma_addr_t dev_addr = phys_to_dma(hwdev, paddr);
|
2008-12-23 02:26:08 +08:00
|
|
|
|
2009-07-10 09:04:52 +08:00
|
|
|
if (swiotlb_force ||
|
2009-07-10 09:04:58 +08:00
|
|
|
!dma_capable(hwdev, dev_addr, sg->length)) {
|
2008-12-23 02:26:08 +08:00
|
|
|
void *map = map_single(hwdev, sg_phys(sg),
|
|
|
|
sg->length, dir);
|
2005-12-20 21:45:19 +08:00
|
|
|
if (!map) {
|
2005-04-17 06:20:36 +08:00
|
|
|
/* Don't panic here, we expect map_sg users
|
|
|
|
to do proper error handling. */
|
|
|
|
swiotlb_full(hwdev, sg->length, dir, 0);
|
2008-04-29 16:00:32 +08:00
|
|
|
swiotlb_unmap_sg_attrs(hwdev, sgl, i, dir,
|
|
|
|
attrs);
|
2007-05-11 20:56:18 +08:00
|
|
|
sgl[0].dma_length = 0;
|
2005-04-17 06:20:36 +08:00
|
|
|
return 0;
|
|
|
|
}
|
2008-12-23 02:26:05 +08:00
|
|
|
sg->dma_address = swiotlb_virt_to_bus(hwdev, map);
|
2005-04-17 06:20:36 +08:00
|
|
|
} else
|
|
|
|
sg->dma_address = dev_addr;
|
|
|
|
sg->dma_length = sg->length;
|
|
|
|
}
|
|
|
|
return nelems;
|
|
|
|
}
|
2008-04-29 16:00:32 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_map_sg_attrs);
|
|
|
|
|
|
|
|
int
|
|
|
|
swiotlb_map_sg(struct device *hwdev, struct scatterlist *sgl, int nelems,
|
|
|
|
int dir)
|
|
|
|
{
|
|
|
|
return swiotlb_map_sg_attrs(hwdev, sgl, nelems, dir, NULL);
|
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_map_sg);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Unmap a set of streaming mode DMA translations. Again, cpu read rules
|
2009-04-08 22:09:15 +08:00
|
|
|
* concerning calls here are the same as for swiotlb_unmap_page() above.
|
2005-04-17 06:20:36 +08:00
|
|
|
*/
|
|
|
|
void
|
2008-04-29 16:00:32 +08:00
|
|
|
swiotlb_unmap_sg_attrs(struct device *hwdev, struct scatterlist *sgl,
|
2009-01-05 22:59:02 +08:00
|
|
|
int nelems, enum dma_data_direction dir, struct dma_attrs *attrs)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-05-11 20:56:18 +08:00
|
|
|
struct scatterlist *sg;
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
|
|
|
|
2006-03-25 01:47:11 +08:00
|
|
|
BUG_ON(dir == DMA_NONE);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
2009-04-08 22:09:19 +08:00
|
|
|
for_each_sg(sgl, sg, nelems, i)
|
|
|
|
unmap_single(hwdev, sg->dma_address, sg->dma_length, dir);
|
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-04-29 16:00:32 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_unmap_sg_attrs);
|
|
|
|
|
|
|
|
void
|
|
|
|
swiotlb_unmap_sg(struct device *hwdev, struct scatterlist *sgl, int nelems,
|
|
|
|
int dir)
|
|
|
|
{
|
|
|
|
return swiotlb_unmap_sg_attrs(hwdev, sgl, nelems, dir, NULL);
|
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_unmap_sg);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
|
|
|
* Make physical memory consistent for a set of streaming mode DMA translations
|
|
|
|
* after a transfer.
|
|
|
|
*
|
|
|
|
* The same as swiotlb_sync_single_* but for a scatter-gather list, same rules
|
|
|
|
* and usage.
|
|
|
|
*/
|
[PATCH] swiotlb uninlinings
Optimise swiotlb.c for size.
text data bss dec hex filename
5009 89 64 5162 142a lib/swiotlb.o-before
4666 89 64 4819 12d3 lib/swiotlb.o-after
For some reason my gcc (4.0.2) doesn't want to tailcall these things.
swiotlb_sync_sg_for_device:
pushq %rbp #
movl $1, %r8d #,
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret
.size swiotlb_sync_sg_for_device, .-swiotlb_sync_sg_for_device
.section .text.swiotlb_sync_sg_for_cpu,"ax",@progbits
.globl swiotlb_sync_sg_for_cpu
.type swiotlb_sync_sg_for_cpu, @function
swiotlb_sync_sg_for_cpu:
pushq %rbp #
xorl %r8d, %r8d #
movq %rsp, %rbp #,
call swiotlb_sync_sg #
leave
ret
Cc: Jan Beulich <jbeulich@novell.com>
Cc: Andi Kleen <ak@suse.de>
Cc: "Luck, Tony" <tony.luck@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-02-12 16:52:17 +08:00
|
|
|
static void
|
2007-05-11 20:56:18 +08:00
|
|
|
swiotlb_sync_sg(struct device *hwdev, struct scatterlist *sgl,
|
2005-09-30 05:44:57 +08:00
|
|
|
int nelems, int dir, int target)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2007-05-11 20:56:18 +08:00
|
|
|
struct scatterlist *sg;
|
2005-04-17 06:20:36 +08:00
|
|
|
int i;
|
|
|
|
|
2009-04-08 22:09:20 +08:00
|
|
|
for_each_sg(sgl, sg, nelems, i)
|
|
|
|
swiotlb_sync_single(hwdev, sg->dma_address,
|
2005-09-30 05:44:57 +08:00
|
|
|
sg->dma_length, dir, target);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
|
2005-09-30 05:43:32 +08:00
|
|
|
void
|
|
|
|
swiotlb_sync_sg_for_cpu(struct device *hwdev, struct scatterlist *sg,
|
2009-01-05 22:59:02 +08:00
|
|
|
int nelems, enum dma_data_direction dir)
|
2005-09-30 05:43:32 +08:00
|
|
|
{
|
2005-09-30 05:44:57 +08:00
|
|
|
swiotlb_sync_sg(hwdev, sg, nelems, dir, SYNC_FOR_CPU);
|
2005-09-30 05:43:32 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_sync_sg_for_cpu);
|
2005-09-30 05:43:32 +08:00
|
|
|
|
2005-04-17 06:20:36 +08:00
|
|
|
void
|
|
|
|
swiotlb_sync_sg_for_device(struct device *hwdev, struct scatterlist *sg,
|
2009-01-05 22:59:02 +08:00
|
|
|
int nelems, enum dma_data_direction dir)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2005-09-30 05:44:57 +08:00
|
|
|
swiotlb_sync_sg(hwdev, sg, nelems, dir, SYNC_FOR_DEVICE);
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_sync_sg_for_device);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
int
|
2008-07-26 10:44:49 +08:00
|
|
|
swiotlb_dma_mapping_error(struct device *hwdev, dma_addr_t dma_addr)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2008-12-23 02:26:05 +08:00
|
|
|
return (dma_addr == swiotlb_virt_to_bus(hwdev, io_tlb_overflow_buffer));
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
2008-12-28 14:02:07 +08:00
|
|
|
EXPORT_SYMBOL(swiotlb_dma_mapping_error);
|
2005-04-17 06:20:36 +08:00
|
|
|
|
|
|
|
/*
|
2005-09-30 06:52:13 +08:00
|
|
|
* Return whether the given device DMA address mask can be supported
|
2005-04-17 06:20:36 +08:00
|
|
|
* properly. For example, if your device can only drive the low 24-bits
|
2005-09-30 06:52:13 +08:00
|
|
|
* during bus mastering, then you would pass 0x00ffffff as the mask to
|
2005-04-17 06:20:36 +08:00
|
|
|
* this function.
|
|
|
|
*/
|
|
|
|
int
|
2007-02-06 10:51:25 +08:00
|
|
|
swiotlb_dma_supported(struct device *hwdev, u64 mask)
|
2005-04-17 06:20:36 +08:00
|
|
|
{
|
2008-12-23 02:26:05 +08:00
|
|
|
return swiotlb_virt_to_bus(hwdev, io_tlb_end - 1) <= mask;
|
2005-04-17 06:20:36 +08:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(swiotlb_dma_supported);
|