Re: [RFC] make kmemleak scan __ro_after_init section (was: Re: [PATCH 0/5] genetlink improvements)

From: Cong Wang
Date: Thu Nov 03 2016 - 01:40:42 EST

On Wed, Nov 2, 2016 at 4:47 PM, Jakub Kicinski <kubakici@xxxxx> wrote:
> Thanks for looking into this! Bisect led me to the following commit:
> commit 56989f6d8568c21257dcec0f5e644d5570ba3281
> Author: Johannes Berg <johannes.berg@xxxxxxxxx>
> Date: Mon Oct 24 14:40:05 2016 +0200
> genetlink: mark families as __ro_after_init
> Now genl_register_family() is the only thing (other than the
> users themselves, perhaps, but I didn't find any doing that)
> writing to the family struct.
> In all families that I found, genl_register_family() is only
> called from __init functions (some indirectly, in which case
> I've add __init annotations to clarifly things), so all can
> actually be marked __ro_after_init.
> This protects the data structure from accidental corruption.
> Signed-off-by: Johannes Berg <johannes.berg@xxxxxxxxx>
> Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
> I realized that kmemleak is not scanning the __ro_after_init section...
> Following patch solves the false positives but I wonder if it's the
> right/acceptable solution.

Nice work! Looks reasonable to me, but I am definitely not familiar
with kmemleak. ;)