Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Sam Bobroff <sbobroff@linux.ibm.com>
To:     airlied@redhat.com, airlied@linux.ie,
        dri-devel@lists.freedesktop.org, linux-kernel@vger.kernel.org
Subject: [PATCH 0/2] Two AST driver fixes
Date:   Mon,  5 Nov 2018 16:57:46 +1100
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Message-Id: <cover.1541397462.git.sbobroff@linux.ibm.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

Hello,

Here are two (attempted) fixes for the AST DRM driver. The issues they fix are
both seen when the ast driver is unloaded (tested on Power9, although it looks
like the second one is architecture independent).

I'm fairly confident about the first fix, as it looks pretty straight forward.

I'm not sure if the second patch uses the right approach, so I'd appreciate some
feedback. I've had a good look around, and it does seem to fix the problem but
I don't know this area at all.

Note that while it does seem odd to take a reference just before calling
drm_framebuffer_remove(), that seems to be be necessary because
drm_framebuffer_remove() is expecting the framebuffer to be dynamically
allocated (but the AST driver directly contains it) (I saw some references to
this being legacy behaviour).

Cheers,
Sam.

Sam Bobroff (2):
  drm/ast: Fix incorrect free on ioregs
  drm/ast: Fix connector leak during driver unload

 drivers/gpu/drm/ast/ast_fb.c   | 4 ++++
 drivers/gpu/drm/ast/ast_main.c | 3 ++-
 2 files changed, 6 insertions(+), 1 deletion(-)

-- 
2.19.0.2.gcad72f5712