Re: [PATCH] RDMA/core: Trigger a WARN_ON if the driver causes uobjects to become leaked

From: Gal Pressman <galpress@amazon.com>
To: Jason Gunthorpe <jgg@nvidia.com>, <linux-rdma@vger.kernel.org>
Cc: Leon Romanovsky <leonro@mellanox.com>
Subject: Re: [PATCH] RDMA/core: Trigger a WARN_ON if the driver causes uobjects to become leaked
Date: Wed, 26 Aug 2020 11:10:18 +0300	[thread overview]
Message-ID: <6cf9106a-7895-4e88-6ea4-b0292ac2c2f8@amazon.com> (raw)
In-Reply-To: <0-v1-b1e0ed400ba9+f7-warn_destroy_ufile_hw_jgg@nvidia.com>

On 25/08/2020 19:35, Jason Gunthorpe wrote:
> Drivers that fail destroy can cause uverbs to leak uobjects. Drivers are
> required to always eventually destroy their ubojects, so trigger a WARN_ON
> to detect this driver bug.
> 
> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
> ---
>  drivers/infiniband/core/rdma_core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/infiniband/core/rdma_core.c b/drivers/infiniband/core/rdma_core.c
> index 3962da54ffbf47..d2b5417a4d5170 100644
> --- a/drivers/infiniband/core/rdma_core.c
> +++ b/drivers/infiniband/core/rdma_core.c
> @@ -895,8 +895,9 @@ void uverbs_destroy_ufile_hw(struct ib_uverbs_file *ufile,
>  		if (__uverbs_cleanup_ufile(ufile, reason)) {
>  			/*
>  			 * No entry was cleaned-up successfully during this
> -			 * iteration
> +			 * iteration. It is a driver bug to fail destruction.
>  			 */
> +			WARN_ON(!list_empty(&ufile->uobjects));
>  			break;
>  		}
>  
> 

Looks good,
Reviewed-by: Gal Pressman <galpress@amazon.com>