When reading the page_pool code the first impression is that keeping
two separate counters, one being the page refcnt and the other being
fragment pp_frag_count, is counter-intuitive.
However without that fragment counter we don't know when to reliably
destroy or sync the outstanding DMA mappings. So let's add a comment
explaining this part.
Reviewed-by: Alexander Duyck <[email protected]>
Signed-off-by: Ilias Apalodimas <[email protected]>
---
Changes since v2:
- Removed a uneeded commas on the comment
Changes since v1:
- Update the comment withe the correct description for pp_frag_count
include/net/page_pool.h | 10 ++++++++++
1 file changed, 10 insertions(+)
diff --git a/include/net/page_pool.h b/include/net/page_pool.h
index 34bf531ffc8d..ddfa0b328677 100644
--- a/include/net/page_pool.h
+++ b/include/net/page_pool.h
@@ -277,6 +277,16 @@ void page_pool_put_defragged_page(struct page_pool *pool, struct page *page,
unsigned int dma_sync_size,
bool allow_direct);
+/* pp_frag_count represents the number of writers who can update the page
+ * either by updating skb->data or via DMA mappings for the device.
+ * We can't rely on the page refcnt for that as we don't know who might be
+ * holding page references and we can't reliably destroy or sync DMA mappings
+ * of the fragments.
+ *
+ * When pp_frag_count reaches 0 we can either recycle the page if the page
+ * refcnt is 1 or return it back to the memory allocator and destroy any
+ * mappings we have.
+ */
static inline void page_pool_fragment_page(struct page *page, long nr)
{
atomic_long_set(&page->pp_frag_count, nr);
--
2.38.1
On 17/02/2023 23.21, Ilias Apalodimas wrote:
> When reading the page_pool code the first impression is that keeping
> two separate counters, one being the page refcnt and the other being
> fragment pp_frag_count, is counter-intuitive.
>
> However without that fragment counter we don't know when to reliably
> destroy or sync the outstanding DMA mappings. So let's add a comment
> explaining this part.
>
> Reviewed-by: Alexander Duyck <[email protected]>
> Signed-off-by: Ilias Apalodimas <[email protected]>
Acked-by: Jesper Dangaard Brouer <[email protected]>
> ---
> Changes since v2:
> - Removed a uneeded commas on the comment
> Changes since v1:
> - Update the comment withe the correct description for pp_frag_count
> include/net/page_pool.h | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/include/net/page_pool.h b/include/net/page_pool.h
> index 34bf531ffc8d..ddfa0b328677 100644
> --- a/include/net/page_pool.h
> +++ b/include/net/page_pool.h
> @@ -277,6 +277,16 @@ void page_pool_put_defragged_page(struct page_pool *pool, struct page *page,
> unsigned int dma_sync_size,
> bool allow_direct);
>
> +/* pp_frag_count represents the number of writers who can update the page
> + * either by updating skb->data or via DMA mappings for the device.
> + * We can't rely on the page refcnt for that as we don't know who might be
> + * holding page references and we can't reliably destroy or sync DMA mappings
> + * of the fragments.
> + *
> + * When pp_frag_count reaches 0 we can either recycle the page if the page
> + * refcnt is 1 or return it back to the memory allocator and destroy any
> + * mappings we have.
> + */
> static inline void page_pool_fragment_page(struct page *page, long nr)
> {
> atomic_long_set(&page->pp_frag_count, nr);
> --
> 2.38.1
>
On Sat, 2023-02-18 at 00:21 +0200, Ilias Apalodimas wrote:
> When reading the page_pool code the first impression is that keeping
> two separate counters, one being the page refcnt and the other being
> fragment pp_frag_count, is counter-intuitive.
>
> However without that fragment counter we don't know when to reliably
> destroy or sync the outstanding DMA mappings. So let's add a comment
> explaining this part.
>
> Reviewed-by: Alexander Duyck <[email protected]>
> Signed-off-by: Ilias Apalodimas <[email protected]>
# Form letter - net-next is closed
The merge window for v6.3 has begun and therefore net-next is closed
for new drivers, features, code refactoring and optimizations.
We are currently accepting bug fixes only.
Please repost when net-next reopens after Mar 6th.
RFC patches sent for review only are obviously welcome at any time.
On Sat, 18 Feb 2023 00:21:30 +0200 Ilias Apalodimas wrote:
> When reading the page_pool code the first impression is that keeping
> two separate counters, one being the page refcnt and the other being
> fragment pp_frag_count, is counter-intuitive.
>
> However without that fragment counter we don't know when to reliably
> destroy or sync the outstanding DMA mappings. So let's add a comment
> explaining this part.
I discussed with Paolo off-list, since it's just a comment change
I'll push it in.
On Tue, 21 Feb 2023 at 19:15, Jakub Kicinski <[email protected]> wrote:
>
> On Sat, 18 Feb 2023 00:21:30 +0200 Ilias Apalodimas wrote:
> > When reading the page_pool code the first impression is that keeping
> > two separate counters, one being the page refcnt and the other being
> > fragment pp_frag_count, is counter-intuitive.
> >
> > However without that fragment counter we don't know when to reliably
> > destroy or sync the outstanding DMA mappings. So let's add a comment
> > explaining this part.
>
> I discussed with Paolo off-list, since it's just a comment change
> I'll push it in.
Fair enough. Thanks Jakub.
Regards
/Ilias
Hello:
This patch was applied to netdev/net-next.git (master)
by Jakub Kicinski <[email protected]>:
On Sat, 18 Feb 2023 00:21:30 +0200 you wrote:
> When reading the page_pool code the first impression is that keeping
> two separate counters, one being the page refcnt and the other being
> fragment pp_frag_count, is counter-intuitive.
>
> However without that fragment counter we don't know when to reliably
> destroy or sync the outstanding DMA mappings. So let's add a comment
> explaining this part.
>
> [...]
Here is the summary with links:
- [v3] page_pool: add a comment explaining the fragment counter usage
https://git.kernel.org/netdev/net-next/c/4d4266e3fd32
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html