Re: [PATCH v2 16/16] mt76: use netmem descriptor and APIs for page pool

On Wed, May 28, 2025 at 12:34 AM Toke Høiland-Jørgensen [off-list ref] wrote:

Byungchul Park [off-list ref] writes:

quoted

On Wed, May 28, 2025 at 11:29:11AM +0900, Byungchul Park wrote:

quoted

To simplify struct page, the effort to separate its own descriptor from
struct page is required and the work for page pool is on going.

Use netmem descriptor and APIs for page pool in mt76 code.

Signed-off-by: Byungchul Park <byungchul@sk.com>
---
 drivers/net/wireless/mediatek/mt76/dma.c      |  6 ++---
 drivers/net/wireless/mediatek/mt76/mt76.h     | 12 +++++-----
 .../net/wireless/mediatek/mt76/sdio_txrx.c    | 24 +++++++++----------
 drivers/net/wireless/mediatek/mt76/usb.c      | 10 ++++----
 4 files changed, 26 insertions(+), 26 deletions(-)

diff --git a/drivers/net/wireless/mediatek/mt76/dma.c b/drivers/net/wireless/mediatek/mt76/dma.c
index 35b4ec91979e..cceff435ec4a 100644
--- a/drivers/net/wireless/mediatek/mt76/dma.c
+++ b/drivers/net/wireless/mediatek/mt76/dma.c

@@ -820,10 +820,10 @@ mt76_add_fragment(struct mt76_dev *dev, struct mt76_queue *q, void *data,
     int nr_frags = shinfo->nr_frags;

     if (nr_frags < ARRAY_SIZE(shinfo->frags)) {
-            struct page *page = virt_to_head_page(data);
-            int offset = data - page_address(page) + q->buf_offset;
+            netmem_ref netmem = netmem_compound_head(virt_to_netmem(data));
+            int offset = data - netmem_address(netmem) + q->buf_offset;

-            skb_add_rx_frag(skb, nr_frags, page, offset, len, q->buf_size);
+            skb_add_rx_frag_netmem(skb, nr_frags, netmem, offset, len, q->buf_size);
     } else {
             mt76_put_page_pool_buf(data, allow_direct);
     }

diff --git a/drivers/net/wireless/mediatek/mt76/mt76.h b/drivers/net/wireless/mediatek/mt76/mt76.h
index 5f8d81cda6cd..f075c1816554 100644
--- a/drivers/net/wireless/mediatek/mt76/mt76.h
+++ b/drivers/net/wireless/mediatek/mt76/mt76.h

@@ -1795,21 +1795,21 @@ int mt76_rx_token_consume(struct mt76_dev *dev, void *ptr,
 int mt76_create_page_pool(struct mt76_dev *dev, struct mt76_queue *q);
 static inline void mt76_put_page_pool_buf(void *buf, bool allow_direct)
 {
-    struct page *page = virt_to_head_page(buf);
+    netmem_ref netmem = netmem_compound_head(virt_to_netmem(buf));

-    page_pool_put_full_page(page->pp, page, allow_direct);

To Mina,

They touch ->pp field.  That's why I thought they use page pool.  Am I
missing something?

It does, since commit: 2f5c3c77fc9b ("wifi: mt76: switch to page_pool allocator")

I am very sorry, I was clearly hallucinating when I first looked at
this. In my defence it was kinda late at night in my time zone :-D

Since this driver supports page_pool, I think it's reasonable to
convert it to netmem. It may not be intreseted in supporting devmem or
io_uring zcrx but there may be more netmem features in the future and
future proofing it sounds good to me.

Reviewed-by: Mina Almasry <redacted>

(although I would like to see a virt_to_head_netmem helper FWIW).

-- 
Thanks,
Mina