linux-kernel - Re: [Regression] "iommu/amd: Relax locking in dma_ops path" makes tg3 ethernet transmit queue timeout

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

Message-Id: <5D7A63DF-D360-4453-845C-924F4C13E3DB@canonical.com>
Date:   Fri, 21 Aug 2020 21:50:33 +0800
From:   Kai-Heng Feng <kai.heng.feng@...onical.com>
To:     Joerg Roedel <jroedel@...e.de>
Cc:     iommu@...ts.linux-foundation.org,
        open list <linux-kernel@...r.kernel.org>
Subject: Re: [Regression] "iommu/amd: Relax locking in dma_ops path" makes tg3
 ethernet transmit queue timeout

Hi Joerg,

> On Aug 21, 2020, at 21:43, Joerg Roedel <jroedel@...e.de> wrote:
> 
> Hi Kai,
> 
> On Mon, Jun 29, 2020 at 08:33:22PM +0800, Kai-Heng Feng wrote:
>> I am still seeing the issue on v5.8-rc3. The issue goes away as soon
>> as "iommu=off" is added.
> 
> Can you probably help with debugging this issue? I've had no luck so far
> getting my hands on a machine with tg3 hardware and an AMD IOMMU.

Of course, I still have the system at my side.

The offending commit is 92d420ec028d ("iommu/amd: Relax locking in dma_ops path"), however be62dbf554c5 ("iommu/amd: Convert AMD iommu driver to the dma-iommu api") removed .map_page entirely so I don't know where to start.

Kai-Heng

> 
> Regards,
> 
> 	Joerg