netdev - Re: [PATCH] net: ethernet: sun4i-emac: free dma descriptor

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <tencent_7731B6630F8BF7EC0292D2F3ADAEA6F70207@qq.com>
Date: Wed, 10 Sep 2025 18:42:18 +0800
From: Conley Lee <conleylee@...mail.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: vkoul@...nel.org, davem@...emloft.net, wens@...e.org,
	mripard@...nel.org, netdev@...r.kernel.org,
	dmaengine@...r.kernel.org, linux-arm-kernel@...ts.infradead.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] net: ethernet: sun4i-emac: free dma descriptor

On 09/09/25 at 06:15下午, Jakub Kicinski wrote:
> Date: Tue, 9 Sep 2025 18:15:47 -0700
> From: Jakub Kicinski <kuba@...nel.org>
> To: 李克斯 <conleylee@...mail.com>
> Cc: vkoul@...nel.org, davem@...emloft.net, wens@...e.org,
>  mripard@...nel.org, netdev@...r.kernel.org, dmaengine@...r.kernel.org,
>  linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
> Subject: Re: [PATCH] net: ethernet: sun4i-emac: free dma descriptor
> 
> On Tue, 9 Sep 2025 14:36:42 +0800 李克斯 wrote:
> > Thank you for the suggestion. I've reviewed the documentation, and
> > setting the reuse flag while reusing descriptors might be a good
> > optimization. I'll make the changes and run some tests. If everything
> > works well, I'll submit a new patch.
> 
> To be clear if you're saying the driver is buggy and can crash right
> now we need to fix it first and then optimize it later, as separate
> commits. So that LTS kernels can backport the fix.
> 
> The questions I'm asking are because I don't understand whether the
> upstream kernel is buggy and how exactly..

After carefully reviewing the documentation and how other drivers use
dmaengine, I think that if the reuse flag is not set for the descriptor,
manual release is unnecessary. Therefore, the current driver implementation
does not contain a bug. This fix patch was a mistake.

Thank you for your thorough review.