[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <YL8BDyMQKym5v7ob@unreal>
Date: Tue, 8 Jun 2021 08:33:03 +0300
From: Leon Romanovsky <leon@...nel.org>
To: Jason Gunthorpe <jgg@...pe.ca>
Cc: Pavel Skripkin <paskripkin@...il.com>, dledford@...hat.com,
shayd@...dia.com, linux-rdma@...r.kernel.org,
linux-kernel@...r.kernel.org, stable@...r.kernel.org
Subject: Re: [PATCH] infiniband: core: fix memory leak
On Mon, Jun 07, 2021 at 08:44:49PM -0300, Jason Gunthorpe wrote:
> On Sun, Jun 06, 2021 at 08:21:51AM +0300, Leon Romanovsky wrote:
>
> > I think that better fix is:
> > https://lore.kernel.org/linux-rdma/f72e27d5c82cd9beec7670141afa62786836c569.1622956637.git.leonro@nvidia.com/T/#u
>
> Can you resend that with the bug report in the comment?
Done,
https://lore.kernel.org/linux-rdma/e2eed941f912b2068e371fd37f43b8cf5082a0e6.1623129597.git.leonro@nvidia.com/T/#u
>
> What is the issue? Some error path missed a restrack_del? which one?
All paths that called to rdma_restrack_new(), but didn't set "id_priv->cma_dev" yet.
For example, cma_ib_new_conn_id() -> __rdma_create_id() -> do something -> exit on
error -> miss call to rdma_restrack_del()
>
> And I still dislike that patch for adding restrack_del and undoing the
> previous patch which was trying to get rid of it, but if it actually
> fixes a bug...
And I didn't like the previous patch :).
Thanks
>
> Jason
Powered by blists - more mailing lists