On 13/11/2024 11:13, Shinas Rasheed wrote:
> During unload, at times the OQ parsed in the napi callbacks
> have been observed to be null, causing system crash.
> Add protective checks to avoid the same, for cn9k cards.
>
> Fixes: 1f2c2d0cee02 ("octeon_ep: add hardware configuration APIs")
> Signed-off-by: Shinas Rasheed <srasheed@marvell.com>
> ---
> V4:
> - No changes
>
> V3: https://lore.kernel.org/all/20241108074543.1123036-4-srasheed@marvell.com/
> - Added back "Fixes" to the changelist
>
> V2: https://lore.kernel.org/all/20241107132846.1118835-4-srasheed@marvell.com/
> - Split into a separate patch
> - Added more context
>
> V1: https://lore.kernel.org/all/20241101103416.1064930-3-srasheed@marvell.com/
>
> drivers/net/ethernet/marvell/octeon_ep/octep_cn9k_pf.c | 9 ++++++++-
> 1 file changed, 8 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/ethernet/marvell/octeon_ep/octep_cn9k_pf.c b/drivers/net/ethernet/marvell/octeon_ep/octep_cn9k_pf.c
> index b5805969404f..b87336b2e4b9 100644
> --- a/drivers/net/ethernet/marvell/octeon_ep/octep_cn9k_pf.c
> +++ b/drivers/net/ethernet/marvell/octeon_ep/octep_cn9k_pf.c
> @@ -617,7 +617,14 @@ static irqreturn_t octep_rsvd_intr_handler_cn93_pf(void *dev)
> static irqreturn_t octep_ioq_intr_handler_cn93_pf(void *data)
> {
> struct octep_ioq_vector *vector = (struct octep_ioq_vector *)data;
> - struct octep_oq *oq = vector->oq;
> + struct octep_oq *oq;
> +
> + if (!vector)
> + return IRQ_HANDLED;
Sorry for not flagging in previous review, but the question is again the
same - how that can happen? This function is irq handler, which is
called from octep_ioq_intr_handler() only if ioq_vector was properly
resolved. This check makes no sense here.
> + oq = vector->oq;
> +
> + if (!oq || !(oq->napi))
> + return IRQ_HANDLED;
>
> napi_schedule_irqoff(oq->napi);
> return IRQ_HANDLED;