* [RFC PATCH 1/2] dmapool: Improve dma api kernel-doc comments
2024-11-13 6:34 [RFC PATCH 0/2] modernize DMA api documentation anish kumar
@ 2024-11-13 6:34 ` anish kumar
2024-11-13 6:34 ` [RFC PATCH 2/2] dmapool: Documentation: use the kernel-doc comment anish kumar
2024-11-14 6:12 ` [RFC PATCH 0/2] modernize DMA api documentation Randy Dunlap
2 siblings, 0 replies; 4+ messages in thread
From: anish kumar @ 2024-11-13 6:34 UTC (permalink / raw)
To: corbet, tj, akpm, ptesarik, xavier_qy, vbabka, tintinm2017
Cc: linux-mm, linux-doc, linux-kernel, anish kumar
Move explanatory content from the documentation
to source code comment written in kernel-doc format.
This allows kernel-doc to generate more comprehensive
and accurate documentation. Additionally, improving
the kernel-doc comment describes the api behaviour
and usage better.
Signed-off-by: anish kumar <yesanishhere@gmail.com>
---
mm/dmapool.c | 40 ++++++++++++++++++++++++++++------------
1 file changed, 28 insertions(+), 12 deletions(-)
diff --git a/mm/dmapool.c b/mm/dmapool.c
index f0bfc6c490f4..8b32a52becd6 100644
--- a/mm/dmapool.c
+++ b/mm/dmapool.c
@@ -207,11 +207,16 @@ static void pool_block_push(struct dma_pool *pool, struct dma_block *block,
* @boundary: returned blocks won't cross this power of two boundary
* Context: not in_interrupt()
*
- * Given one of these pools, dma_pool_alloc()
- * may be used to allocate memory. Such memory will all have "consistent"
- * DMA mappings, accessible by the device and its driver without using
- * cache flushing primitives. The actual size of blocks allocated may be
- * larger than requested because of alignment.
+ * This api initializes a pool of DMA-coherent buffers for use with a given
+ * device. It must be called in a context which can sleep. The device's
+ * hardware alignment requirement for this type of data is "align". If your
+ * device has no boundary crossing restrictions, pass 0 for alloc; passing
+ * 4096 says memory allocated from this pool must not cross 4KByte boundaries.
+ *
+ * Given one of these pools, dma_pool_alloc() may be used to allocate memory.
+ * Such memory will all have "consistent" DMA mappings, accessible by the
+ * device and its driver without using cache flushing primitives. The actual
+ * size of blocks allocated may be larger than requested because of alignment.
*
* If @boundary is nonzero, objects returned from dma_pool_alloc() won't
* cross that size boundary. This is useful for devices which have
@@ -356,6 +361,7 @@ static struct dma_page *pool_alloc_page(struct dma_pool *pool, gfp_t mem_flags)
*
* Caller guarantees that no more memory from the pool is in use,
* and that nothing will try to use the pool after this call.
+ * It must be called in a context which can sleep.
*/
void dma_pool_destroy(struct dma_pool *pool)
{
@@ -392,14 +398,24 @@ void dma_pool_destroy(struct dma_pool *pool)
EXPORT_SYMBOL(dma_pool_destroy);
/**
- * dma_pool_alloc - get a block of consistent memory
- * @pool: dma pool that will produce the block
- * @mem_flags: GFP_* bitmask
- * @handle: pointer to dma address of block
+ * dma_pool_alloc - Get a block of consistent memory from a DMA pool
+ * @pool: DMA pool that will produce the block
+ * @mem_flags: GFP_* bitmask specifying memory allocation flags
+ * @handle: Pointer to a DMA address that will hold the address of the block
+ *
+ * Return: The kernel virtual address of a currently unused block of memory,
+ * and reports its DMA address through the handle. If such a memory
+ * block can't be allocated, %NULL is returned.
+ *
+ * This function allocates memory from the specified DMA pool. The returned
+ * memory will meet the size and alignment requirements specified when the
+ * pool was created. Pass GFP_ATOMIC to prevent blocking, or if permitted
+ * (not in interrupt context, and not holding SMP locks), pass GFP_KERNEL
+ * to allow blocking.
*
- * Return: the kernel virtual address of a currently unused block,
- * and reports its dma address through the handle.
- * If such a memory block can't be allocated, %NULL is returned.
+ * Similar to dma_alloc_coherent(), this function returns two addresses:
+ * - A CPU-accessible virtual address
+ * - A DMA address usable by the pool's associated device.
*/
void *dma_pool_alloc(struct dma_pool *pool, gfp_t mem_flags,
dma_addr_t *handle)
--
2.39.5 (Apple Git-154)
^ permalink raw reply [flat|nested] 4+ messages in thread* [RFC PATCH 2/2] dmapool: Documentation: use the kernel-doc comment
2024-11-13 6:34 [RFC PATCH 0/2] modernize DMA api documentation anish kumar
2024-11-13 6:34 ` [RFC PATCH 1/2] dmapool: Improve dma api kernel-doc comments anish kumar
@ 2024-11-13 6:34 ` anish kumar
2024-11-14 6:12 ` [RFC PATCH 0/2] modernize DMA api documentation Randy Dunlap
2 siblings, 0 replies; 4+ messages in thread
From: anish kumar @ 2024-11-13 6:34 UTC (permalink / raw)
To: corbet, tj, akpm, ptesarik, xavier_qy, vbabka, tintinm2017
Cc: linux-mm, linux-doc, linux-kernel, anish kumar
In an effort to modernize the documentation for
dma api, move the api explanation to kernel-doc comment
in the source code and use the kernel-doc here in
the documentation.
Signed-off-by: anish kumar <yesanishhere@gmail.com>
---
Documentation/core-api/dma-api.rst | 66 ++++++------------------------
1 file changed, 12 insertions(+), 54 deletions(-)
diff --git a/Documentation/core-api/dma-api.rst b/Documentation/core-api/dma-api.rst
index 8e3cce3d0a23..8e89f328ba54 100644
--- a/Documentation/core-api/dma-api.rst
+++ b/Documentation/core-api/dma-api.rst
@@ -86,65 +86,23 @@ not __get_free_pages(). Also, they understand common hardware constraints
for alignment, like queue heads needing to be aligned on N-byte boundaries.
-::
-
- struct dma_pool *
- dma_pool_create(const char *name, struct device *dev,
- size_t size, size_t align, size_t alloc);
-
-dma_pool_create() initializes a pool of DMA-coherent buffers
-for use with a given device. It must be called in a context which
-can sleep.
-
-The "name" is for diagnostics (like a struct kmem_cache name); dev and size
-are like what you'd pass to dma_alloc_coherent(). The device's hardware
-alignment requirement for this type of data is "align" (which is expressed
-in bytes, and must be a power of two). If your device has no boundary
-crossing restrictions, pass 0 for alloc; passing 4096 says memory allocated
-from this pool must not cross 4KByte boundaries.
-
-::
-
- void *
- dma_pool_zalloc(struct dma_pool *pool, gfp_t mem_flags,
- dma_addr_t *handle)
-
-Wraps dma_pool_alloc() and also zeroes the returned memory if the
-allocation attempt succeeded.
-
+.. kernel-doc:: mm/dmapool.c
+ :identifiers: dma_pool_create
-::
-
- void *
- dma_pool_alloc(struct dma_pool *pool, gfp_t gfp_flags,
- dma_addr_t *dma_handle);
-
-This allocates memory from the pool; the returned memory will meet the
-size and alignment requirements specified at creation time. Pass
-GFP_ATOMIC to prevent blocking, or if it's permitted (not
-in_interrupt, not holding SMP locks), pass GFP_KERNEL to allow
-blocking. Like dma_alloc_coherent(), this returns two values: an
-address usable by the CPU, and the DMA address usable by the pool's
-device.
-
-::
+.. kernel-doc:: mm/dmapool.c
+ :identifiers: dma_pool_alloc
- void
- dma_pool_free(struct dma_pool *pool, void *vaddr,
- dma_addr_t addr);
+dma_pool_zalloc wraps dma_pool_alloc() and also zeroes the returned memory
+if the allocation attempt succeeded.
-This puts memory back into the pool. The pool is what was passed to
-dma_pool_alloc(); the CPU (vaddr) and DMA addresses are what
-were returned when that routine allocated the memory being freed.
+.. kernel-doc:: mm/dmapool.c
+ :identifiers: dma_pool_create
-::
-
- void
- dma_pool_destroy(struct dma_pool *pool);
+.. kernel-doc:: mm/dmapool.c
+ :identifiers: dma_pool_free
-dma_pool_destroy() frees the resources of the pool. It must be
-called in a context which can sleep. Make sure you've freed all allocated
-memory back to the pool before you destroy it.
+.. kernel-doc:: mm/dmapool.c
+ :identifiers: dma_pool_destroy
Part Ic - DMA addressing limitations
--
2.39.5 (Apple Git-154)
^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [RFC PATCH 0/2] modernize DMA api documentation
2024-11-13 6:34 [RFC PATCH 0/2] modernize DMA api documentation anish kumar
2024-11-13 6:34 ` [RFC PATCH 1/2] dmapool: Improve dma api kernel-doc comments anish kumar
2024-11-13 6:34 ` [RFC PATCH 2/2] dmapool: Documentation: use the kernel-doc comment anish kumar
@ 2024-11-14 6:12 ` Randy Dunlap
2 siblings, 0 replies; 4+ messages in thread
From: Randy Dunlap @ 2024-11-14 6:12 UTC (permalink / raw)
To: anish kumar, corbet, tj, akpm, ptesarik, xavier_qy, vbabka, tintinm2017
Cc: linux-mm, linux-doc, linux-kernel
Hi--
On 11/12/24 10:34 PM, anish kumar wrote:
> Dear maintainers,
>
> This patch series is inspired by the ongoing discussion in the following thread:
> https://lore.kernel.org/lkml/20241111063847.GB23992@lst.de/
>
> As part of an effort to modernize the documentation, the first patch in
> this series converts the relevant document to use kernel-doc comments directly
> within the source code, replacing the previous method of manually adding
> documentation. This change will help streamline the documentation workflow and
> ensure it remains synchronized with the code.
>
> I am still getting below errors and wondering if there is a better
> way to write this patch.
>
> /home/ANT.AMAZON.COM/anishkmr/Downloads/linux/Documentation/core-api/dma-api:89:
> ./mm/dmapool.c:229: WARNING: Duplicate C declaration, also defined at
> core-api/mm-api:224. Declaration is:
> '.. c:function:: struct dma_pool *dma_pool_create (const char *name,
> struct device *dev, size_t size, size_t align, size_t boundary)'.
>
> /home/ANT.AMAZON.COM/anishkmr/Downloads/linux/Documentation/core-api/dma-api:92:
> ./mm/dmapool.c:420: WARNING: Duplicate C declaration, also defined at
> core-api/mm-api:404. Declaration is:
> '.. c:function:: void *dma_pool_alloc (struct dma_pool *pool, gfp_t mem_flags,
> dma_addr_t *handle)'.
>
> /home/ANT.AMAZON.COM/anishkmr/Downloads/linux/Documentation/core-api/dma-api:98:
> ./mm/dmapool.c:229: WARNING: Duplicate C declaration, also defined at
> core-api/mm-api:224. Declaration is:
> '.. c:function:: struct dma_pool *dma_pool_create (const char *name,
> struct device *dev, size_t size, size_t align, size_t boundary)'.
>
> /home/ANT.AMAZON.COM/anishkmr/Downloads/linux/Documentation/core-api/dma-api:101:
> ./mm/dmapool.c:466: WARNING: Duplicate C declaration, also defined at
> core-api/mm-api:450. Declaration is:
> '.. c:function:: void dma_pool_free (struct dma_pool *pool, void *vaddr,
> dma_addr_t dma)'.
>
> /home/ANT.AMAZON.COM/anishkmr/Downloads/linux/Documentation/core-api/dma-api:104:
> ./mm/dmapool.c:366: WARNING: Duplicate C declaration, also defined at
> core-api/mm-api:360.
>
If you are going to add these APIs in dma-api.rst, you need to remove them from
mm-api.rst (IMO). That should remove these Duplicate warnings.
> Thank you for considering this patch. I look forward to your feedback.
>
> anish kumar (2):
> dmapool: Improve dma api kernel-doc comments
> dmapool: Documentation: use the kernel-doc comment
>
> Documentation/core-api/dma-api.rst | 66 ++++++------------------------
> mm/dmapool.c | 40 ++++++++++++------
> 2 files changed, 40 insertions(+), 66 deletions(-)
>
--
~Randy
^ permalink raw reply [flat|nested] 4+ messages in thread