[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20211005090856.ccuccgbjejs5g7gd@wrt>
Date: Tue, 5 Oct 2021 11:08:56 +0200
From: andreas@...mhold.de
To: Punit Agrawal <punitagrawal@...il.com>
Cc: Andreas Rammhold <andreas@...mhold.de>,
Jakub Kicinski <kuba@...nel.org>,
Heiko Stübner <heiko@...ech.de>,
netdev@...r.kernel.org, linux-rockchip@...ts.infradead.org,
linux-arm-kernel@...ts.infradead.org, davem@...emloft.net,
peppe.cavallaro@...com, alexandre.torgue@...s.st.com,
Michael Riesch <michael.riesch@...fvision.net>
Subject: Re: [PATCH] net: stmmac: dwmac-rk: Fix ethernet on rk3399 based
devices
On 21:06 04.10.21, Punit Agrawal wrote:
> Andreas Rammhold <andreas@...mhold.de> writes:
>
> > On 17:20 02.10.21, Jakub Kicinski wrote:
> >> On Sat, 2 Oct 2021 23:33:03 +0200 Andreas Rammhold wrote:
> >> > On 16:02 01.10.21, Jakub Kicinski wrote:
> >> > > On Wed, 29 Sep 2021 23:02:35 +0200 Heiko Stübner wrote:
> >> > > > On a rk3399-puma which has the described issue,
> >> > > > Tested-by: Heiko Stuebner <heiko@...ech.de>
> >> > >
> >> > > Applied, thanks!
> >> >
> >> > This also fixed the issue on a RockPi4.
> >> >
> >> > Will any of you submit this to the stable kernels (as this broke within
> >> > 3.13 for me) or shall I do that?
> >>
> >> I won't. The patch should be in Linus's tree in around 1 week - at which
> >> point anyone can request the backport.
> >>
> >> That said, as you probably know, 4.4 is the oldest active stable branch,
> >> the ship has sailed for anything 3.x.
> >
> > I am sorry. I meant 5.13.
>
> AFAICT, 2d26f6e39afb ("net: stmmac: dwmac-rk: fix unbalanced
> pm_runtime_enable warnings") is not in 5.13 stable.
>
> Either you're not using the stable kernel or there's another issue
> breaking things on the RockPi4.
I was using the 5.13 branch, somewhere after 5.13.12 the network started
to fail on bootup. Due to the nature of the system I don't have
persistent logs on it. When I next looked at the system (and updated to
5.14.x) the issue still occured until I applied this patch on to 5.14.8.
Might have been the same use-facing issue but different bugs?
I can try to narrow the issue down further. It'll probably take a few
evenings to test this out.
Powered by blists - more mailing lists