Re: [PATCH net-next 1/2] net: ethernet: mtk_eth_soc: add 9K jumbo frame

Re: [PATCH net-next 1/2] net: ethernet: mtk_eth_soc: add 9K jumbo frame

[Simon Horman]

On Mon, Nov 02, 2026 at 12:48:45AM +0200, Sandeep Haemoon wrote:
> The driver caps max_mtu at 2030 bytes and programs the MAC max-rx field
> for at most 2048-byte frames, so MTU > 2K is rejected and 9K jumbo
> frames are not usable on the 10G XGMII MACs.
> 
> Add a MTK_NETSYS_RX_9K capability for MT7988 and:
> - advertise 9K max_mtu on XGMII MACs, derived from the DT phy_mode at
>   probe time so DSA can raise the conduit MTU before the netdev is
>   opened;
> - program XMAC_RX_CFG2 on the XGMII MACs;
> - derive the shared RX buffer length from the largest MTU and size the
>   RX rings from it;
> - program the PDMA RX SDL max frame so the DMA engine accepts jumbo
>   frames;
> - reject a runtime MTU change that would need a different ring buffer
>   size while DMA is running; the rings are sized at the next open.
> 
> This is a forward-port of the MediaTek downstream jumbo series:
> Mason Chang's "change default rx buffer length", "add 9k jumbo frame
> support" and "add dynamic rx buffer adjustment support" (MediaTek
> OpenWrt feed 999-eth-30/31/33, carried in
> frank-w/BPI-Router-Linux 92bfdbdd4de1/b4c6cb1d5532/ca4c976dce22),
> plus Frank Wunderlich's "try to fix crash the right way"
> (a3207a383794).
> 
> Link: https://github.com/frank-w/BPI-Router-Linux/commit/b4c6cb1d5532
> Link: https://github.com/mediatek/mtk-openwrt-feeds
> Signed-off-by: Sandeep Haemoon <sandeeph@loljews.com>

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net-next 2/2] net: ethernet: mtk_eth_soc: skip page_pool for

[Simon Horman]

On Mon, Nov 02, 2026 at 12:48:46AM +0200, Sandeep Haemoon wrote:
> page_pool allocates order-0 pages and the pool is created with
> MTK_PP_MAX_BUF_SIZE.  Rings that need larger buffers (9K jumbo or the
> HW-LRO rings) cannot use it.  Only create/use a page_pool when
> ring->frag_size fits in a page; the large-buffer path already uses
> mtk_max_buf_alloc().
> 
> This is required by the 9K support in the previous patch; the
> large-buffer path it relies on comes from the MediaTek downstream
> series.
> 
> Link: https://github.com/mediatek/mtk-openwrt-feeds
> Signed-off-by: Sandeep Haemoon <sandeeph@loljews.com>

It may not be particularly important, but I do wonder
if the order of the patches in this patch-set should
be reversed: as it is 9k support seems to be added by 1/2
and fixed in 2/2.

That notwithstanding:

Reviewed-by: Simon Horman <horms@kernel.org>

Re: [PATCH net-next 1/2] net: ethernet: mtk_eth_soc: add 9K jumbo frame

[Jakub Kicinski]

On Mon, 2 Nov 2026 00:48:45 +0200 Sandeep Haemoon wrote:
> +static void mtk_set_mcr_max_rx(struct mtk_mac *mac, u32 val);

Avoid the forward declarations, please.

When you repost make sure that patch 2 is in reply to the cover letter,
not chained off patch 1.
-- 
pw-bot: cr