Re: [PATCH] IB/rdmavt: add lock to call to rvt_error_qp to prevent a race condition

From: Jason Gunthorpe <jgg@nvidia.com>
To: Niels Dossche <dossche.niels@gmail.com>
Cc: linux-rdma@vger.kernel.org, linux-kernel@vger.kernel.org,
	Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com>,
	Mike Marciniszyn <mike.marciniszyn@cornelisnetworks.com>
Subject: Re: [PATCH] IB/rdmavt: add lock to call to rvt_error_qp to prevent a race condition
Date: Fri, 8 Apr 2022 13:30:46 -0300	[thread overview]
Message-ID: <20220408163046.GA3612436@nvidia.com> (raw)
In-Reply-To: <20220228165330.41546-1-dossche.niels@gmail.com>

On Mon, Feb 28, 2022 at 05:53:30PM +0100, Niels Dossche wrote:
> The documentation of the function rvt_error_qp says both r_lock and
> s_lock need to be held when calling that function.
> It also asserts using lockdep that both of those locks are held.
> However, the commit I referenced in Fixes accidentally makes the call
> to rvt_error_qp in rvt_ruc_loopback no longer covered by r_lock.
> This results in the lockdep assertion failing and also possibly in a
> race condition.
> 
> Fixes: d757c60eca9b ("IB/rdmavt: Fix concurrency panics in QP post_send and modify to error")
> Signed-off-by: Niels Dossche <dossche.niels@gmail.com>
> ---
>  drivers/infiniband/sw/rdmavt/qp.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)

Applied to for-rc, thanks

Jason