netdev - Re: [PATCH 0/2] vmxnet3: Fix inconsistent DMA accesses

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: <CAOZ5it1S+fiV5Nz8Fivq0MM3dLFS+Rv90v9izgGkZMJhz69fXQ@mail.gmail.com>
Date: Tue, 19 Nov 2024 10:10:54 -0700
From: Brian Johannesmeyer <bjohannesmeyer@...il.com>
To: Jakub Kicinski <kuba@...nel.org>
Cc: Ronak Doshi <ronak.doshi@...adcom.com>, 
	Broadcom internal kernel review list <bcm-kernel-feedback-list@...adcom.com>, Andrew Lunn <andrew+netdev@...n.ch>, 
	"David S . Miller" <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>, 
	Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org, linux-kernel@...r.kernel.org, 
	Raphael Isemann <teemperor@...il.com>
Subject: Re: [PATCH 0/2] vmxnet3: Fix inconsistent DMA accesses

> What is the purpose of the first patch? Is it sufficient to make
> the device work correctly?
The purpose of the first patch is to fix the inconsistent accesses in
`vmxnet3_probe_device()`. This only partially fixes the issue,
however, because there are inconsistent accesses elsewhere in the
driver. So, no, it does not make the device work *entirely* correctly.

> If yes, why do we need patch 2.
> If no, why do we have patch 1, instead of a revert / patch 2...
The answer is that the way I submitted this patch series was a
mistake. Specifically, I submitted it as: (i) patch 1 is applied on
master, *and* (ii) patch 2 is applied on patch 1.

Instead, the way I should have submitted it was: (i) patch 1 is
applied on master, *or* (ii) a corrected version of patch 2 is applied
on master. (By "a corrected version of patch 2", I mean not
pointlessly reverting patch 1.)

The difference being:
- If `adapter` *should* be mapped to DMA, then patch 1 is the way to go. Or,
- If `adapter` *should not* be mapped to DMA, then a corrected version
of patch 2 is the way to go.

I hope this clears things up. I'm sorry for the confusion I caused. I
will submit a V2 patch series that makes this clear.

Thanks,

Brian