Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: Issue with RTL8111 NIC after upgrade to kernel 4.19
To:     Norbert Jurkeit <norbert.jurkeit@web.de>, nic_swsd@realtek.com,
        Andrew Lunn <andrew@lunn.ch>,
        Florian Fainelli <f.fainelli@gmail.com>
Cc:     davem@davemloft.net, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org, michael.wiktowy@gmail.com,
        jcline@redhat.com, marc.c.dionne@gmail.com
References: <d0a53002-bf14-6cd9-ab0f-5d3f9aa395ce@web.de>
From:   Heiner Kallweit <hkallweit1@gmail.com>
Message-ID: <38dad61b-bc7f-7038-6d1b-f5c4afe3841c@gmail.com>
Date:   Wed, 21 Nov 2018 20:57:00 +0100
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.3.1
MIME-Version: 1.0
In-Reply-To: <d0a53002-bf14-6cd9-ab0f-5d3f9aa395ce@web.de>
Content-Type: text/plain; charset=windows-1252
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 21.11.2018 11:51, Norbert Jurkeit wrote:
> Dear Linux developers,
> 
> after upgrade from distro kernel 4.18.18 to 4.19.2 some Fedora users (including me) encounter issues with RTL8111 based Ethernet interfaces. The interface is not started correctly after boot with kernel 4.19.2 (i.e. is reported by ifconfig as up, but not running, the LEDs besides the connector stay off and no data is transferred).
> 
> At least for me the issue seems to occur only after reboot from a previous session with kernel 4.18.18, but not after power-on, which means the hardware is not fully initialized by the new r8169 driver. Marc Dionne, another affected Fedora user, thinks that module "realtek" needs to be loaded, which does not happen automatically. Please see https://bugzilla.redhat.com/show_bug.cgi?id=1650984 for more details.
> 

Thanks for reporting the issue. I can't reproduce the error on my
system and obviously it affects only some users, so it seems to be
chip-version- or system-dependent. Having to load the realtek PHY
driver module manually is weird as this should be done by
request_module() in phy_device_create().

request_module() is supposed to be synchronous, however after some
reading this may not be 100% guaranteed. Maybe the module init
function on some systems isn't finished yet when request_module()
returns. As a result the genphy driver may be used instead of
the PHY version-specific driver.

That the issue occurs after the switch to 4.19 isn't a big surprise
because with this version the r8169 driver switched to phylib.

First I'd like to get an overview of the affected systems. I'll
comment in the linked bug ticket to address the affected users.


> As Fedora kernel maintainers usually don't apply many changes to the upstream kernel, users of other distros will likely be affected as well. We therefore kindly ask you to investigate the issue and fix it in future upstream kernels.
> 
> Many thanks and best regards,
> 
> Norbert Jurkeit
> 
> 
Heiner