Re: [PATCH for-next 2/2] RDMA/bnxt_re: Allocate dev_attr information dynamically

[Leon Romanovsky <leon@kernel.org>]

On Thu, Jan 09, 2025 at 10:18:13AM -0800, Selvin Xavier wrote:
> From: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
> 
> In order to optimize the size of driver private structure,
> the memory for dev_attr is allocated dynamically during the
> chip context initialization. In order to make certain runtime
> decisions, store dev_attr in the qplib_res structure.
> 
> Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
> Signed-off-by: Selvin Xavier <selvin.xavier@broadcom.com>
> ---
>  drivers/infiniband/hw/bnxt_re/bnxt_re.h     |  2 +-
>  drivers/infiniband/hw/bnxt_re/hw_counters.c |  2 +-
>  drivers/infiniband/hw/bnxt_re/ib_verbs.c    | 38 ++++++++++++++---------------
>  drivers/infiniband/hw/bnxt_re/main.c        | 36 +++++++++++++++++----------
>  drivers/infiniband/hw/bnxt_re/qplib_res.c   |  7 +++---
>  drivers/infiniband/hw/bnxt_re/qplib_res.h   |  4 +--
>  drivers/infiniband/hw/bnxt_re/qplib_sp.c    |  4 +--
>  drivers/infiniband/hw/bnxt_re/qplib_sp.h    |  3 +--
>  8 files changed, 51 insertions(+), 45 deletions(-)

<...>

> index 33956fc..7c7057b 100644
> --- a/drivers/infiniband/hw/bnxt_re/main.c
> +++ b/drivers/infiniband/hw/bnxt_re/main.c
> @@ -148,6 +148,10 @@ static void bnxt_re_destroy_chip_ctx(struct bnxt_re_dev *rdev)
>  
>  	if (!rdev->chip_ctx)
>  		return;
> +
> +	kfree(rdev->dev_attr);
> +	rdev->dev_attr = NULL;
> +

I'm taking this patch, but please let's stop this practice of setting NULL
for the pointers which are not going to be reused. Such assignment hides bugs.

Thanks