RE: [PATCH net-next] net: mana: Force full-page RX buffers for 4K page size on specific systems.

From: Haiyang Zhang

Date: Mon Mar 02 2026 - 11:45:56 EST




> -----Original Message-----
> From: Dipayaan Roy <dipayanroy@xxxxxxxxxxxxxxxxxxx>
> Sent: Friday, February 27, 2026 5:15 AM
> To: KY Srinivasan <kys@xxxxxxxxxxxxx>; Haiyang Zhang
> <haiyangz@xxxxxxxxxxxxx>; wei.liu@xxxxxxxxxx; Dexuan Cui
> <DECUI@xxxxxxxxxxxxx>; andrew+netdev@xxxxxxx; davem@xxxxxxxxxxxxx;
> edumazet@xxxxxxxxxx; kuba@xxxxxxxxxx; pabeni@xxxxxxxxxx; leon@xxxxxxxxxx;
> Long Li <longli@xxxxxxxxxxxxx>; Konstantin Taranov
> <kotaranov@xxxxxxxxxxxxx>; horms@xxxxxxxxxx;
> shradhagupta@xxxxxxxxxxxxxxxxxxx; ssengar@xxxxxxxxxxxxxxxxxxx;
> ernis@xxxxxxxxxxxxxxxxxxx; Shiraz Saleem <shirazsaleem@xxxxxxxxxxxxx>;
> linux-hyperv@xxxxxxxxxxxxxxx; netdev@xxxxxxxxxxxxxxx; linux-
> kernel@xxxxxxxxxxxxxxx; linux-rdma@xxxxxxxxxxxxxxx; Dipayaan Roy
> <dipayanroy@xxxxxxxxxxxxx>
> Subject: [PATCH net-next] net: mana: Force full-page RX buffers for 4K
> page size on specific systems.
>
> On certain systems configured with 4K PAGE_SIZE, utilizing page_pool
> fragments for RX buffers results in a significant throughput regression.
> Profiling reveals that this regression correlates with high overhead in
> the
> fragment allocation and reference counting paths on these specific
> platforms, rendering the multi-buffer-per-page strategy counterproductive.
>
> To mitigate this, bypass the page_pool fragment path and force a single RX
> packet per page allocation when all the following conditions are met:
> 1. The system is configured with a 4K PAGE_SIZE.
> 2. A processor-specific quirk is detected via SMBIOS Type 4 data.
>
> This approach restores expected line-rate performance by ensuring
> predictable RX refill behavior on affected hardware.
>
> There is no behavioral change for systems using larger page sizes
> (16K/64K), or platforms where this processor-specific quirk do not
> apply.
>
> Signed-off-by: Dipayaan Roy <dipayanroy@xxxxxxxxxxxxxxxxxxx>

Reviewed-by: Haiyang Zhang <haiyangz@xxxxxxxxxxxxx>
Thanks.