Re: [RFC][PATCH 2/2] dma-heap: Add a system-uncached heap

On 2020-07-29 06:16, John Stultz wrote:

This adds a heap that allocates non-contiguous buffers that are marked as writecombined, so they are not cached by the CPU.

This is useful, as most graphics buffers are usually not touched by the CPU or only written into once by the CPU. So when mapping the buffer over and over between devices, we can skip the CPU syncing, which saves a lot of cache management overhead, greatly improving performance.

For folk using ION, there was a ION_FLAG_CACHED flag, which signaled if the returned buffer should be CPU cacheable or not. With DMA-BUF heaps, we have no such flag, and by default the current heaps (system and cma) produce CPU cachable buffers. So for folks transitioning from ION to DMA-BUF Heaps, this fills in some of that missing functionality.

This does have a few "ugly" bits that were required to get the buffer properly flushed out initially which I'd like to improve. So feedback would be very welcome!

Many thanks to Liam Mark for his help to get this working.

Cc: Sumit Semwal sumit.semwal@linaro.org Cc: Andrew F. Davis afd@ti.com Cc: Benjamin Gaignard benjamin.gaignard@linaro.org Cc: Liam Mark lmark@codeaurora.org Cc: Laura Abbott labbott@kernel.org Cc: Brian Starkey Brian.Starkey@arm.com Cc: Hridya Valsaraju hridya@google.com Cc: linux-media@vger.kernel.org Cc: dri-devel@lists.freedesktop.org Signed-off-by: John Stultz john.stultz@linaro.org

---

drivers/dma-buf/heaps/Kconfig | 10 + drivers/dma-buf/heaps/Makefile | 1 + drivers/dma-buf/heaps/system_uncached_heap.c | 392 +++++++++++++++++++ 3 files changed, 403 insertions(+) create mode 100644 drivers/dma-buf/heaps/system_uncached_heap.c

diff --git a/drivers/dma-buf/heaps/Kconfig b/drivers/dma-buf/heaps/Kconfig index a5eef06c4226..420b0ed0a512 100644 --- a/drivers/dma-buf/heaps/Kconfig +++ b/drivers/dma-buf/heaps/Kconfig @@ -5,6 +5,16 @@ config DMABUF_HEAPS_SYSTEM Choose this option to enable the system dmabuf heap. The system heap is backed by pages from the buddy allocator. If in doubt, say Y.

+config DMABUF_HEAPS_SYSTEM_UNCACHED

• bool "DMA-BUF Uncached System Heap"
• depends on DMABUF_HEAPS
• help

•  Choose this option to enable the uncached system dmabuf heap. This
  

•  heap is backed by pages from the buddy allocator, but pages are setup
  

•  for write combining. This avoids cache management overhead, and can
  

•  be faster if pages are mostly untouched by the cpu.  If in doubt,
  

•  say Y.
  

• config DMABUF_HEAPS_CMA bool "DMA-BUF CMA Heap" depends on DMABUF_HEAPS && DMA_CMA

diff --git a/drivers/dma-buf/heaps/Makefile b/drivers/dma-buf/heaps/Makefile index 6e54cdec3da0..085685ec478f 100644 --- a/drivers/dma-buf/heaps/Makefile +++ b/drivers/dma-buf/heaps/Makefile @@ -1,4 +1,5 @@ # SPDX-License-Identifier: GPL-2.0 obj-y += heap-helpers.o obj-$(CONFIG_DMABUF_HEAPS_SYSTEM) += system_heap.o +obj-$(CONFIG_DMABUF_HEAPS_SYSTEM_UNCACHED) += system_uncached_heap.o obj-$(CONFIG_DMABUF_HEAPS_CMA) += cma_heap.o diff --git a/drivers/dma-buf/heaps/system_uncached_heap.c b/drivers/dma-buf/heaps/system_uncached_heap.c new file mode 100644 index 000000000000..d23908038376 --- /dev/null +++ b/drivers/dma-buf/heaps/system_uncached_heap.c @@ -0,0 +1,392 @@ +// SPDX-License-Identifier: GPL-2.0 +/*

• • Uncached System DMA-Heap exporter
• • Copyright (C) 2020 Linaro Ltd.
• • Based off of Andrew Davis' SRAM heap:
• • Copyright (C) 2019 Texas Instruments Incorporated - http://www.ti.com/
• • Andrew F. Davis afd@ti.com
• */

+#include <linux/dma-mapping.h> +#include <linux/err.h> +#include <linux/highmem.h> +#include <linux/io.h> +#include <linux/mm.h> +#include <linux/scatterlist.h> +#include <linux/slab.h> +#include <linux/dma-buf.h> +#include <linux/dma-heap.h>

+struct uncached_heap {

• struct dma_heap *heap;

+};

+struct uncached_heap_buffer {

• struct dma_heap *heap;
• struct list_head attachments;
• struct mutex lock;
• unsigned long len;
• struct sg_table sg_table;
• int vmap_cnt;
• void *vaddr;

+};

+struct dma_heap_attachment {

• struct device *dev;
• struct sg_table *table;
• struct list_head list;

+};

+static struct sg_table *dup_sg_table(struct sg_table *table) +{

• struct sg_table *new_table;
• int ret, i;
• struct scatterlist *sg, *new_sg;
• new_table = kzalloc(sizeof(*new_table), GFP_KERNEL);
• if (!new_table)

• return ERR_PTR(-ENOMEM);
  

• ret = sg_alloc_table(new_table, table->nents, GFP_KERNEL);
• if (ret) {

• kfree(new_table);
  

• return ERR_PTR(-ENOMEM);
  

• }
• new_sg = new_table->sgl;
• for_each_sg(table->sgl, sg, table->nents, i) {

Consider using the new sgtable helpers that are just about to land - in this case, for_each_sgtable_sg().

• memcpy(new_sg, sg, sizeof(*sg));
  

• new_sg->dma_address = 0;
  

This seems a little bit hairy, as in theory a consumer could still treat a nonzero DMA length as the address being valid. Rather than copying the whole entry then trying to undo parts of that, maybe just:

sg_set_page(new_sg, sg_page(sg), sg->len, sg->offset);

?

• new_sg = sg_next(new_sg);
  

• }
• return new_table;

+}

+static int dma_heap_attach(struct dma_buf *dmabuf,

• 	   struct dma_buf_attachment *attachment)
  

+{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• struct dma_heap_attachment *a;
• struct sg_table *table;
• a = kzalloc(sizeof(*a), GFP_KERNEL);
• if (!a)

• return -ENOMEM;
  

• table = dup_sg_table(&buffer->sg_table);
• if (IS_ERR(table)) {

• kfree(a);
  

• return -ENOMEM;
  

• }
• a->table = table;
• a->dev = attachment->dev;
• INIT_LIST_HEAD(&a->list);
• attachment->priv = a;
• mutex_lock(&buffer->lock);
• list_add(&a->list, &buffer->attachments);
• mutex_unlock(&buffer->lock);
• return 0;

+}

+static void dma_heap_detatch(struct dma_buf *dmabuf,

• 	     struct dma_buf_attachment *attachment)
  

+{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• struct dma_heap_attachment *a = attachment->priv;
• mutex_lock(&buffer->lock);
• list_del(&a->list);
• mutex_unlock(&buffer->lock);
• sg_free_table(a->table);
• kfree(a->table);
• kfree(a);

+}

+static struct sg_table *dma_heap_map_dma_buf(struct dma_buf_attachment *attachment,

• 			     enum dma_data_direction direction)
  

+{

• struct dma_heap_attachment *a = attachment->priv;
• struct sg_table *table = a->table;
• if (!dma_map_sg_attrs(attachment->dev, table->sgl, table->nents, direction,

• 	      DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_WRITE_COMBINE))
  

dma_map_sgtable()

Also, DMA_ATTR_WRITE_COMBINE is meaningless for streaming DMA.

• return ERR_PTR(-ENOMEM);
  

• return table;

+}

+static void dma_heap_unmap_dma_buf(struct dma_buf_attachment *attachment,

• 		   struct sg_table *table,
  

• 		   enum dma_data_direction direction)
  

+{

• dma_unmap_sg_attrs(attachment->dev, table->sgl, table->nents, direction,

• 	   DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_WRITE_COMBINE);
  

Similarly.

+}

+static int dma_heap_mmap(struct dma_buf *dmabuf, struct vm_area_struct *vma) +{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• struct sg_table *table = &buffer->sg_table;
• unsigned long addr = vma->vm_start;
• unsigned long offset = vma->vm_pgoff * PAGE_SIZE;
• struct scatterlist *sg;
• int i;
• int ret;
• vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
• for_each_sg(table->sgl, sg, table->nents, i) {

• struct page *page = sg_page(sg);
  

• unsigned long remainder = vma->vm_end - addr;
  

• unsigned long len = sg->length;
  

• if (offset >= sg->length) {
  

• 	offset -= sg->length;
  

• 	continue;
  

• } else if (offset) {
  

• 	page += offset / PAGE_SIZE;
  

• 	len = sg->length - offset;
  

• 	offset = 0;
  

• }
  

• len = min(len, remainder);
  

• ret = remap_pfn_range(vma, addr, page_to_pfn(page), len,
  

• 		      vma->vm_page_prot);
  

• if (ret)
  

• 	return ret;
  

• addr += len;
  

• if (addr >= vma->vm_end)
  

• 	return 0;
  

• }
• return 0;

+}

+static void *dma_heap_do_vmap(struct uncached_heap_buffer *buffer) +{

• struct sg_table *table = &buffer->sg_table;
• int npages = PAGE_ALIGN(buffer->len) / PAGE_SIZE;
• struct page **pages = vmalloc(sizeof(struct page *) * npages);
• struct page **tmp = pages;
• struct scatterlist *sg;
• pgprot_t pgprot;
• void *vaddr;
• int i, j;
• if (!pages)

• return ERR_PTR(-ENOMEM);
  

• pgprot = pgprot_writecombine(PAGE_KERNEL);
• for_each_sg(table->sgl, sg, table->nents, i) {

for_each_sg_page()

• int npages_this_entry = PAGE_ALIGN(sg->length) / PAGE_SIZE;
  

• struct page *page = sg_page(sg);
  

• WARN_ON(i >= npages);
  

• for (j = 0; j < npages_this_entry; j++)
  

• 	*(tmp++) = page++;
  

• }
• vaddr = vmap(pages, npages, VM_MAP, pgprot);
• vfree(pages);
• if (!vaddr)

• return ERR_PTR(-ENOMEM);
  

• return vaddr;

+}

+static void *dma_heap_buffer_vmap_get(struct uncached_heap_buffer *buffer) +{

• void *vaddr;
• if (buffer->vmap_cnt) {

• buffer->vmap_cnt++;
  

• return buffer->vaddr;
  

• }
• vaddr = dma_heap_do_vmap(buffer);
• if (IS_ERR(vaddr))

• return vaddr;
  

• buffer->vaddr = vaddr;
• buffer->vmap_cnt++;
• return vaddr;

+}

+static void dma_heap_buffer_vmap_put(struct uncached_heap_buffer *buffer) +{

• if (!--buffer->vmap_cnt) {

• vunmap(buffer->vaddr);
  

• buffer->vaddr = NULL;
  

• }

+}

+static void *dma_heap_vmap(struct dma_buf *dmabuf) +{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• void *vaddr;
• mutex_lock(&buffer->lock);
• vaddr = dma_heap_buffer_vmap_get(buffer);
• mutex_unlock(&buffer->lock);
• return vaddr;

+}

+static void dma_heap_vunmap(struct dma_buf *dmabuf, void *vaddr) +{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• mutex_lock(&buffer->lock);
• dma_heap_buffer_vmap_put(buffer);
• mutex_unlock(&buffer->lock);

+}

+static void dma_heap_dma_buf_release(struct dma_buf *dmabuf) +{

• struct uncached_heap_buffer *buffer = dmabuf->priv;
• struct sg_table *table;
• struct scatterlist *sg;
• int i;
• table = &buffer->sg_table;
• for_each_sg(table->sgl, sg, table->nents, i)

• __free_page(sg_page(sg));
  

• sg_free_table(table);
• kfree(buffer);

+}

+const struct dma_buf_ops uncached_heap_buf_ops = {

• .attach = dma_heap_attach,
• .detach = dma_heap_detatch,
• .map_dma_buf = dma_heap_map_dma_buf,
• .unmap_dma_buf = dma_heap_unmap_dma_buf,
• .mmap = dma_heap_mmap,
• .vmap = dma_heap_vmap,
• .vunmap = dma_heap_vunmap,
• .release = dma_heap_dma_buf_release,

+};

+static int uncached_heap_allocate(struct dma_heap *heap,

• 		  unsigned long len,
  

• 		  unsigned long fd_flags,
  

• 		  unsigned long heap_flags)
  

+{

• struct uncached_heap_buffer *buffer;
• DEFINE_DMA_BUF_EXPORT_INFO(exp_info);
• struct dma_buf *dmabuf;
• struct sg_table *table;
• struct scatterlist *sg;
• pgoff_t pagecount;
• pgoff_t pg;
• int i, ret = -ENOMEM;
• buffer = kzalloc(sizeof(*buffer), GFP_KERNEL);
• if (!buffer)

• return -ENOMEM;
  

• INIT_LIST_HEAD(&buffer->attachments);
• mutex_init(&buffer->lock);
• buffer->heap = heap;
• buffer->len = len;
• table = &buffer->sg_table;
• pagecount = len / PAGE_SIZE;
• if (sg_alloc_table(table, pagecount, GFP_KERNEL))

• goto free_buffer;
  

• sg = table->sgl;
• for (pg = 0; pg < pagecount; pg++) {

• struct page *page;
  

• /*
  

•  * Avoid trying to allocate memory if the process
  

•  * has been killed by by SIGKILL
  

•  */
  

• if (fatal_signal_pending(current))
  

• 	goto free_pages;
  

• page = alloc_page(GFP_KERNEL | __GFP_ZERO);
  

• if (!page)
  

• 	goto free_pages;
  

• sg_set_page(sg, page, page_size(page), 0);
  

• sg = sg_next(sg);
  

• }
• /* create the dmabuf */
• exp_info.ops = &uncached_heap_buf_ops;
• exp_info.size = buffer->len;
• exp_info.flags = fd_flags;
• exp_info.priv = buffer;
• dmabuf = dma_buf_export(&exp_info);
• if (IS_ERR(dmabuf)) {

• ret = PTR_ERR(dmabuf);
  

• goto free_pages;
  

• }
• ret = dma_buf_fd(dmabuf, fd_flags);
• if (ret < 0) {

• dma_buf_put(dmabuf);
  

• /* just return, as put will call release and that will free */
  

• return ret;
  

• }
• /*

• * XXX This is hackish. While the buffer will be uncached, we need
  

• * to initially flush cpu cache, since the the __GFP_ZERO on the
  

• * allocation means the zeroing was done by the cpu and thus it is likely
  

• * cached. Map & flush it out now so we don't get corruption later on.
  

• *
  

• * Ideally we could do this without using the heap device as a dummy dev.
  

• */
  

• dma_map_sg_attrs(dma_heap_get_dev(heap), table->sgl, table->nents,

• 	 DMA_BIDIRECTIONAL, DMA_ATTR_WRITE_COMBINE);
  

Again, DMA_ATTR_WRITE_COMBINE is meaningless here.

• dma_sync_sg_for_device(dma_heap_get_dev(heap), table->sgl, table->nents,

• 	       DMA_BIDIRECTIONAL);
  

This doesn't do anything that the map hasn't already just done.

• return ret;

+free_pages:

• for_each_sg(table->sgl, sg, table->nents, i)

• __free_page(sg_page(sg));
  

• sg_free_table(table);

+free_buffer:

• kfree(buffer);
• return ret;

+}

+static struct dma_heap_ops uncached_heap_ops = {

• .allocate = uncached_heap_allocate,

+};

+static int uncached_heap_create(void) +{

• struct uncached_heap *heap;
• struct dma_heap_export_info exp_info;
• static u64 dummy_mask;
• heap = kzalloc(sizeof(*heap), GFP_KERNEL);
• if (!heap)

• return -ENOMEM;
  

• exp_info.name = "system-uncached";
• exp_info.ops = &uncached_heap_ops;
• exp_info.priv = heap;
• heap->heap = dma_heap_add(&exp_info);
• if (IS_ERR(heap->heap)) {

• int ret = PTR_ERR(heap->heap);
  

• kfree(heap);
  

• return ret;
  

• }
• dma_heap_get_dev(heap->heap)->dma_mask = &dummy_mask;
• dma_set_mask(dma_heap_get_dev(heap->heap), DMA_BIT_MASK(64));

Much as I'd hate to encourage using dma_coerce_mask_and_coherent(), I'm not sure this is really any better :/

Robin.

• return 0;

+} +device_initcall(uncached_heap_create);