linux-kernel - Re: [GIT PULL] Networking for v6.17

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: <CAHk-=wiQ0p09UvRVZ3tGqmRgstgZ75o7ppcaPfCa6oVJOEEzeQ@mail.gmail.com>
Date: Tue, 5 Aug 2025 21:22:18 +0300
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Jakub Kicinski <kuba@...nel.org>, John Ernberg <john.ernberg@...ia.se>
Cc: davem@...emloft.net, netdev@...r.kernel.org, linux-kernel@...r.kernel.org, 
	pabeni@...hat.com
Subject: Re: [GIT PULL] Networking for v6.17

On Tue, 5 Aug 2025 at 19:22, Linus Torvalds
<torvalds@...ux-foundation.org> wrote:
>
> On Sun, 27 Jul 2025 at 04:35, Jakub Kicinski <kuba@...nel.org> wrote:
> >
> > Networking changes for 6.17.
>
> So I found out the hard way while traveling that this networking pull
> seems to have broken USB tethering for me. Which I only use when
> traveling, but then I do often end up relying on my phone as the
> source of internet (the phone being on the single-device flight wifi,
> and tethering to the laptop which is why hotspot doesn't necessarily
> work).
>
> It *might* be something else, and I'm bisecting it right now, but the
> networking pull is the obvious first suspect, and my first three
> bisection steps have taken me into that pull.

To absolutely zero surprise, it continued to bisect into the
networking pull, and this is the end result:

  0d9cfc9b8cb17dbc29a98792d36ec39a1cf1395f is the first bad commit
  commit 0d9cfc9b8cb17dbc29a98792d36ec39a1cf1395f
  Author: John Ernberg <john.ernberg@...ia.se>
  Date:   Wed Jul 23 10:25

      net: usbnet: Avoid potential RCU stall on LINK_CHANGE event

and I'll test with that just reverted on top of current -tip. But it
bisected right to that commit, and the commit certainly makes sense as
a "that could break usbnet" commit, so I expect that the revert will
indeed fix it.

Considering that I will need usb tethering while traveling during the
rest of the merge window, I almost certainly will just revert it for
good tomorrow, but if somebody comes up with a fix for this that
doesn't involve a revert, I'm all ears.

             Linus