[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Fri, 31 Dec 2021 21:39:49 -0600
From: Frank Rowand <frowand.list@...il.com>
To: Rob Herring <robh+dt@...nel.org>, erhard_f@...lbox.org,
yinxiujiang@...inos.cn
Cc: devicetree@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 0/2] of: unittest: re-implement overlay tracking
Hi Erhard,
Can you please check whether this patch series fixes the shift-out-of-bounds
problem you reported?
-Frank
On 12/31/21 9:33 PM, frowand.list@...il.com wrote:
> From: Frank Rowand <frank.rowand@...y.com>
>
> Some overlays are tracked when they are applied. The tracked overlays
> are later removed after the overlay tests are completed. The old
> implementation makes assumptions about the expected values for
> overlay changeset id created by the overlay apply which result
> in fragile code. The new code removes the assumptions.
>
> A symptom that exposes a problem with the tracking code is a
> warning "UBSAN: shift-out-of-bounds in drivers/of/unittest.c:1933:36",
> Kernel Version: 5.15-rc7, PPC-64, Talos II. This results from variable
> "id" value of -1 in the final line of of_unittest_untrack_overlay().
>
> The first patch in the series cleans up the inconsistent use of overlay
> changeset id and the obsolete overlay id. The id is a core concept in
> the overlay tracking that is re-implemented in the second patch in
> the series.
>
> Frank Rowand (2):
> of: unittest: change references to obsolete overlay id
> of: unittest: re-implement overlay tracking
>
> drivers/of/unittest.c | 154 +++++++++++++++++++-----------------------
> 1 file changed, 71 insertions(+), 83 deletions(-)
>
Powered by blists - more mailing lists