linux/drivers/media/v4l2-core/videobuf-dma-contig.c

511 lines
12 KiB
C
Raw Normal View History

/*
* helper functions for physically contiguous capture buffers
*
* The functions support hardware lacking scatter gather support
* (i.e. the buffers must be linear in physical memory)
*
* Copyright (c) 2008 Magnus Damm
*
* Based on videobuf-vmalloc.c,
* (c) 2007 Mauro Carvalho Chehab, <mchehab@infradead.org>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2
*/
#include <linux/init.h>
#include <linux/module.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/dma-mapping.h>
#include <linux/sched.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/slab.h>
#include <media/videobuf-dma-contig.h>
struct videobuf_dma_contig_memory {
u32 magic;
void *vaddr;
dma_addr_t dma_handle;
bool cached;
unsigned long size;
};
#define MAGIC_DC_MEM 0x0733ac61
#define MAGIC_CHECK(is, should) \
if (unlikely((is) != (should))) { \
pr_err("magic mismatch: %x expected %x\n", (is), (should)); \
BUG(); \
}
static int __videobuf_dc_alloc(struct device *dev,
struct videobuf_dma_contig_memory *mem,
unsigned long size, gfp_t flags)
{
mem->size = size;
if (mem->cached) {
mem->vaddr = alloc_pages_exact(mem->size, flags | GFP_DMA);
if (mem->vaddr) {
int err;
mem->dma_handle = dma_map_single(dev, mem->vaddr,
mem->size,
DMA_FROM_DEVICE);
err = dma_mapping_error(dev, mem->dma_handle);
if (err) {
dev_err(dev, "dma_map_single failed\n");
free_pages_exact(mem->vaddr, mem->size);
mem->vaddr = NULL;
return err;
}
}
} else
mem->vaddr = dma_alloc_coherent(dev, mem->size,
&mem->dma_handle, flags);
if (!mem->vaddr) {
dev_err(dev, "memory alloc size %ld failed\n", mem->size);
return -ENOMEM;
}
dev_dbg(dev, "dma mapped data is at %p (%ld)\n", mem->vaddr, mem->size);
return 0;
}
static void __videobuf_dc_free(struct device *dev,
struct videobuf_dma_contig_memory *mem)
{
if (mem->cached) {
if (!mem->vaddr)
return;
dma_unmap_single(dev, mem->dma_handle, mem->size,
DMA_FROM_DEVICE);
free_pages_exact(mem->vaddr, mem->size);
} else
dma_free_coherent(dev, mem->size, mem->vaddr, mem->dma_handle);
mem->vaddr = NULL;
}
static void videobuf_vm_open(struct vm_area_struct *vma)
{
struct videobuf_mapping *map = vma->vm_private_data;
dev_dbg(map->q->dev, "vm_open %p [count=%u,vma=%08lx-%08lx]\n",
map, map->count, vma->vm_start, vma->vm_end);
map->count++;
}
static void videobuf_vm_close(struct vm_area_struct *vma)
{
struct videobuf_mapping *map = vma->vm_private_data;
struct videobuf_queue *q = map->q;
int i;
dev_dbg(q->dev, "vm_close %p [count=%u,vma=%08lx-%08lx]\n",
map, map->count, vma->vm_start, vma->vm_end);
map->count--;
if (0 == map->count) {
struct videobuf_dma_contig_memory *mem;
dev_dbg(q->dev, "munmap %p q=%p\n", map, q);
videobuf_queue_lock(q);
/* We need first to cancel streams, before unmapping */
if (q->streaming)
videobuf_queue_cancel(q);
for (i = 0; i < VIDEO_MAX_FRAME; i++) {
if (NULL == q->bufs[i])
continue;
if (q->bufs[i]->map != map)
continue;
mem = q->bufs[i]->priv;
if (mem) {
/* This callback is called only if kernel has
allocated memory and this memory is mmapped.
In this case, memory should be freed,
in order to do memory unmap.
*/
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
/* vfree is not atomic - can't be
called with IRQ's disabled
*/
dev_dbg(q->dev, "buf[%d] freeing %p\n",
i, mem->vaddr);
__videobuf_dc_free(q->dev, mem);
mem->vaddr = NULL;
}
q->bufs[i]->map = NULL;
q->bufs[i]->baddr = 0;
}
kfree(map);
videobuf_queue_unlock(q);
}
}
static const struct vm_operations_struct videobuf_vm_ops = {
.open = videobuf_vm_open,
.close = videobuf_vm_close,
};
/**
* videobuf_dma_contig_user_put() - reset pointer to user space buffer
* @mem: per-buffer private videobuf-dma-contig data
*
* This function resets the user space pointer
*/
static void videobuf_dma_contig_user_put(struct videobuf_dma_contig_memory *mem)
{
mem->dma_handle = 0;
mem->size = 0;
}
/**
* videobuf_dma_contig_user_get() - setup user space memory pointer
* @mem: per-buffer private videobuf-dma-contig data
* @vb: video buffer to map
*
* This function validates and sets up a pointer to user space memory.
* Only physically contiguous pfn-mapped memory is accepted.
*
* Returns 0 if successful.
*/
static int videobuf_dma_contig_user_get(struct videobuf_dma_contig_memory *mem,
struct videobuf_buffer *vb)
{
struct mm_struct *mm = current->mm;
struct vm_area_struct *vma;
unsigned long prev_pfn, this_pfn;
unsigned long pages_done, user_address;
unsigned int offset;
int ret;
offset = vb->baddr & ~PAGE_MASK;
mem->size = PAGE_ALIGN(vb->size + offset);
ret = -EINVAL;
down_read(&mm->mmap_sem);
vma = find_vma(mm, vb->baddr);
if (!vma)
goto out_up;
if ((vb->baddr + mem->size) > vma->vm_end)
goto out_up;
pages_done = 0;
prev_pfn = 0; /* kill warning */
user_address = vb->baddr;
while (pages_done < (mem->size >> PAGE_SHIFT)) {
ret = follow_pfn(vma, user_address, &this_pfn);
if (ret)
break;
if (pages_done == 0)
mem->dma_handle = (this_pfn << PAGE_SHIFT) + offset;
else if (this_pfn != (prev_pfn + 1))
ret = -EFAULT;
if (ret)
break;
prev_pfn = this_pfn;
user_address += PAGE_SIZE;
pages_done++;
}
out_up:
up_read(&current->mm->mmap_sem);
return ret;
}
static struct videobuf_buffer *__videobuf_alloc_vb(size_t size, bool cached)
{
struct videobuf_dma_contig_memory *mem;
struct videobuf_buffer *vb;
vb = kzalloc(size + sizeof(*mem), GFP_KERNEL);
if (vb) {
vb->priv = ((char *)vb) + size;
mem = vb->priv;
mem->magic = MAGIC_DC_MEM;
mem->cached = cached;
}
return vb;
}
static struct videobuf_buffer *__videobuf_alloc_uncached(size_t size)
{
return __videobuf_alloc_vb(size, false);
}
static struct videobuf_buffer *__videobuf_alloc_cached(size_t size)
{
return __videobuf_alloc_vb(size, true);
}
static void *__videobuf_to_vaddr(struct videobuf_buffer *buf)
{
struct videobuf_dma_contig_memory *mem = buf->priv;
BUG_ON(!mem);
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
return mem->vaddr;
}
static int __videobuf_iolock(struct videobuf_queue *q,
struct videobuf_buffer *vb,
struct v4l2_framebuffer *fbuf)
{
struct videobuf_dma_contig_memory *mem = vb->priv;
BUG_ON(!mem);
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
switch (vb->memory) {
case V4L2_MEMORY_MMAP:
dev_dbg(q->dev, "%s memory method MMAP\n", __func__);
/* All handling should be done by __videobuf_mmap_mapper() */
if (!mem->vaddr) {
dev_err(q->dev, "memory is not alloced/mmapped.\n");
return -EINVAL;
}
break;
case V4L2_MEMORY_USERPTR:
dev_dbg(q->dev, "%s memory method USERPTR\n", __func__);
/* handle pointer from user space */
if (vb->baddr)
return videobuf_dma_contig_user_get(mem, vb);
/* allocate memory for the read() method */
if (__videobuf_dc_alloc(q->dev, mem, PAGE_ALIGN(vb->size),
GFP_KERNEL))
return -ENOMEM;
break;
case V4L2_MEMORY_OVERLAY:
default:
dev_dbg(q->dev, "%s memory method OVERLAY/unknown\n", __func__);
return -EINVAL;
}
return 0;
}
static int __videobuf_sync(struct videobuf_queue *q,
struct videobuf_buffer *buf)
{
struct videobuf_dma_contig_memory *mem = buf->priv;
BUG_ON(!mem);
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
dma_sync_single_for_cpu(q->dev, mem->dma_handle, mem->size,
DMA_FROM_DEVICE);
return 0;
}
static int __videobuf_mmap_mapper(struct videobuf_queue *q,
struct videobuf_buffer *buf,
struct vm_area_struct *vma)
{
struct videobuf_dma_contig_memory *mem;
struct videobuf_mapping *map;
int retval;
unsigned long size;
unsigned long pos, start = vma->vm_start;
struct page *page;
dev_dbg(q->dev, "%s\n", __func__);
/* create mapping + update buffer list */
map = kzalloc(sizeof(struct videobuf_mapping), GFP_KERNEL);
if (!map)
return -ENOMEM;
buf->map = map;
map->q = q;
buf->baddr = vma->vm_start;
mem = buf->priv;
BUG_ON(!mem);
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
if (__videobuf_dc_alloc(q->dev, mem, PAGE_ALIGN(buf->bsize),
GFP_KERNEL | __GFP_COMP))
goto error;
/* Try to remap memory */
size = vma->vm_end - vma->vm_start;
size = (size < mem->size) ? size : mem->size;
if (!mem->cached) {
vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);
retval = remap_pfn_range(vma, vma->vm_start,
mem->dma_handle >> PAGE_SHIFT,
size, vma->vm_page_prot);
if (retval) {
dev_err(q->dev, "mmap: remap failed with error %d. ",
retval);
dma_free_coherent(q->dev, mem->size,
mem->vaddr, mem->dma_handle);
goto error;
}
} else {
pos = (unsigned long)mem->vaddr;
while (size > 0) {
page = virt_to_page((void *)pos);
if (NULL == page) {
dev_err(q->dev, "mmap: virt_to_page failed\n");
__videobuf_dc_free(q->dev, mem);
goto error;
}
retval = vm_insert_page(vma, start, page);
if (retval) {
dev_err(q->dev, "mmap: insert failed with error %d\n",
retval);
__videobuf_dc_free(q->dev, mem);
goto error;
}
start += PAGE_SIZE;
pos += PAGE_SIZE;
if (size > PAGE_SIZE)
size -= PAGE_SIZE;
else
size = 0;
}
}
vma->vm_ops = &videobuf_vm_ops;
vma->vm_flags |= VM_DONTEXPAND;
vma->vm_private_data = map;
dev_dbg(q->dev, "mmap %p: q=%p %08lx-%08lx (%lx) pgoff %08lx buf %d\n",
map, q, vma->vm_start, vma->vm_end,
(long int)buf->bsize, vma->vm_pgoff, buf->i);
videobuf_vm_open(vma);
return 0;
error:
kfree(map);
return -ENOMEM;
}
static struct videobuf_qtype_ops qops = {
.magic = MAGIC_QTYPE_OPS,
.alloc_vb = __videobuf_alloc_uncached,
.iolock = __videobuf_iolock,
.mmap_mapper = __videobuf_mmap_mapper,
.vaddr = __videobuf_to_vaddr,
};
static struct videobuf_qtype_ops qops_cached = {
.magic = MAGIC_QTYPE_OPS,
.alloc_vb = __videobuf_alloc_cached,
.iolock = __videobuf_iolock,
.sync = __videobuf_sync,
.mmap_mapper = __videobuf_mmap_mapper,
.vaddr = __videobuf_to_vaddr,
};
void videobuf_queue_dma_contig_init(struct videobuf_queue *q,
const struct videobuf_queue_ops *ops,
struct device *dev,
spinlock_t *irqlock,
enum v4l2_buf_type type,
enum v4l2_field field,
unsigned int msize,
void *priv,
struct mutex *ext_lock)
{
videobuf_queue_core_init(q, ops, dev, irqlock, type, field, msize,
priv, &qops, ext_lock);
}
EXPORT_SYMBOL_GPL(videobuf_queue_dma_contig_init);
void videobuf_queue_dma_contig_init_cached(struct videobuf_queue *q,
const struct videobuf_queue_ops *ops,
struct device *dev,
spinlock_t *irqlock,
enum v4l2_buf_type type,
enum v4l2_field field,
unsigned int msize,
void *priv, struct mutex *ext_lock)
{
videobuf_queue_core_init(q, ops, dev, irqlock, type, field, msize,
priv, &qops_cached, ext_lock);
}
EXPORT_SYMBOL_GPL(videobuf_queue_dma_contig_init_cached);
dma_addr_t videobuf_to_dma_contig(struct videobuf_buffer *buf)
{
struct videobuf_dma_contig_memory *mem = buf->priv;
BUG_ON(!mem);
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
return mem->dma_handle;
}
EXPORT_SYMBOL_GPL(videobuf_to_dma_contig);
void videobuf_dma_contig_free(struct videobuf_queue *q,
struct videobuf_buffer *buf)
{
struct videobuf_dma_contig_memory *mem = buf->priv;
/* mmapped memory can't be freed here, otherwise mmapped region
would be released, while still needed. In this case, the memory
release should happen inside videobuf_vm_close().
So, it should free memory only if the memory were allocated for
read() operation.
*/
if (buf->memory != V4L2_MEMORY_USERPTR)
return;
if (!mem)
return;
MAGIC_CHECK(mem->magic, MAGIC_DC_MEM);
/* handle user space pointer case */
if (buf->baddr) {
videobuf_dma_contig_user_put(mem);
return;
}
/* read() method */
if (mem->vaddr) {
__videobuf_dc_free(q->dev, mem);
mem->vaddr = NULL;
}
}
EXPORT_SYMBOL_GPL(videobuf_dma_contig_free);
MODULE_DESCRIPTION("helper module to manage video4linux dma contig buffers");
MODULE_AUTHOR("Magnus Damm");
MODULE_LICENSE("GPL");