[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20221219092139.222456-1-zyytlz.wz@163.com>
Date: Mon, 19 Dec 2022 17:21:39 +0800
From: Zheng Wang <zyytlz.wz@....com>
To: zhi.a.wang@...el.com
Cc: 1002992920@...com, airlied@...il.com, airlied@...ux.ie,
alex000young@...il.com, dri-devel@...ts.freedesktop.org,
gregkh@...uxfoundation.org, hackerzheng666@...il.com,
intel-gfx@...ts.freedesktop.org,
intel-gvt-dev@...ts.freedesktop.org,
joonas.lahtinen@...ux.intel.com, linux-kernel@...r.kernel.org,
security@...nel.org, tvrtko.ursulin@...ux.intel.com,
zhenyuw@...ux.intel.com, zyytlz.wz@....com
Subject: Re: [Intel-gfx] [PATCH v3] drm/i915/gvt: fix double free bug in split_2MB_gtt_entry
Wang, Zhi A <zhi.a.wang@...el.com> 于2022年12月19日周一 16:22写道:
>
> I think it is a case-by-case thing. For example:
>
> The current scenario in this function looks like below:
>
> caller pass spt a
> function
> alloc spt b
> something error
> free spt a
> return error
>
> The problem is: the function wrongly frees the spt a instead free what
> it allocates.
Thanks for your clear explaination. It’s really helpfult to me.
I think I might know how to fix now.
> A proper fix should be:
>
> caller pass spt a
> function
> alloc spt b
> something error
> *free spt b*
> return error
>
As it's a case-by-case thing, I'll extract the un-done-mapping-dma part from
ppgtt_invalidate_spt and put it in error path. Then I'll add the code of freeing
new allocated spt. If I misunderstand your meaning, feel free to let me know.
Working on a new fix now.
Best regards,
Zheng Wang
Powered by blists - more mailing lists