Re: [PATCH net-next,v9 2/2] net: mana: force full-page RX buffers via ethtool private flag

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Forcing `apc->port_is_up = true` in the `mana_attach` error path causes a guaranteed NULL pointer dereference in the reset worker.
--

commit d36beb9b51f56b9797284c9fa1dd515410680980
Author: Dipayaan Roy <dipayanroy@linux.microsoft.com>
Subject: net: mana: force full-page RX buffers via ethtool private flag

This patch adds an ethtool private flag full-page-rx that allows the
user to force one RX buffer per page, bypassing the page_pool fragment path
on ARM64 platforms with 4K PAGE_SIZE.

> diff --git a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> index 7e79681634db3..f22bbb3259489 100644
> --- a/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
> +++ b/drivers/net/ethernet/microsoft/mana/mana_ethtool.c
[ ... ]
> @@ -590,6 +609,88 @@ static int mana_get_link_ksettings(struct net_device *ndev,
[ ... ]
> +static int mana_set_priv_flags(struct net_device *ndev, u32 priv_flags)
> +{
[ ... ]
> +		err = mana_attach(ndev);
> +		if (err) {
> +			netdev_err(ndev, "mana_attach failed: %d\n", err);
> +			apc->priv_flags = old_priv_flags;
> +
> +			/* Restore 'port_is_up' so the reset work handler
> +			 * can properly detach/attach. Without this,
> +			 * the handler sees port_is_up=false and skips
> +			 * queue allocation, leaving the port dead.
> +			 */
> +			apc->port_is_up = true;
> +			schedule_port_reset = true;
> +		}

Can this cause a NULL pointer dereference in the reset worker?

If mana_attach() fails, its error path cleans up the partially allocated state.
Specifically, if mana_alloc_queues() fails, it calls mana_cleanup_port_context(),
which frees the tx_qp array and sets apc->tx_qp = NULL.

When apc->port_is_up is forced to true and the reset work is scheduled here:

mana_set_priv_flags()
  err = mana_attach(ndev);
  if (err) {
      ...
      apc->port_is_up = true;
      schedule_port_reset = true;
  }

The reset worker will eventually call mana_detach(). Because apc->port_is_up
is true, mana_detach() will attempt to deallocate the queues:

mana_detach()
  apc->port_st_save = apc->port_is_up;
  ...
  if (apc->port_st_save) {
      err = mana_dealloc_queues(ndev);

If mana_dealloc_queues() unconditionally iterates over apc->num_queues and
accesses apc->tx_qp[i], does this lead to an immediate kernel panic since
apc->tx_qp was already set to NULL during the earlier cleanup?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260508142921.497921-1-dipayanroy@linux.microsoft.com?part=2