netdev - Re: Regression in 4.19 net/phy/realtek: garbled sysfs output

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: <7d9dd587-3ccf-b496-6af9-e106f2f9d126@applied-asynchrony.com>
Date:   Thu, 25 Oct 2018 11:29:34 +0200
From:   Holger Hoffstätte <holger@...lied-asynchrony.com>
To:     Andrew Lunn <andrew@...n.ch>
Cc:     Netdev <netdev@...r.kernel.org>,
        Jassi Brar <jaswinder.singh@...aro.org>,
        "David S. Miller" <davem@...emloft.net>,
        Heiner Kallweit <hkallweit1@...il.com>
Subject: Re: Regression in 4.19 net/phy/realtek: garbled sysfs output

On 10/24/18 22:12, Andrew Lunn wrote:
> On Wed, Oct 24, 2018 at 09:36:02PM +0200, Holger Hoffstätte wrote:
>> Hi,
>>
>> Since 4.19 r8169 depends on phylib:
>>
>> $lsmod | grep r8169
>> r8169                  81920  0
>> libphy                 57344  2 r8169,realtek
>>
>> Unfortunately this now gives me the following sysfs error:
>>
>> $cd /sys/module/realtek/drivers
>> $ls -l
>> ls: cannot access 'mdio_bus:RTL8201F 10/100Mbps Ethernet': No such file or directory
>> total 0
>> lrwxrwxrwx 1 root root 0 Oct 24 21:09 'mdio_bus:RTL8201CP Ethernet' -> '../../../bus/mdio_bus/drivers/RTL8201CP Ethernet'
>> l????????? ? ?    ?    ?            ? 'mdio_bus:RTL8201F 10/100Mbps Ethernet'
>> lrwxrwxrwx 1 root root 0 Oct 24 21:09 'mdio_bus:RTL8211 Gigabit Ethernet' -> '../../../bus/mdio_bus/drivers/RTL8211 Gigabit Ethernet'
>> [..]
>>
>> Apparently the forward slash in "10/100Mbps Ethernet" is interpreted as
>> directory separator that leads nowhere, and was introduced in commit
>> 513588dd44b ("net: phy: realtek: add RTL8201F phy-id and functions").
>>
>> Would it be acceptable to change the name simply to "RTL8201F Ethernet"?
> 
> Hi Holger
> 
> Or use "RTL8201F Fast Ethernet"

Yes, even better since it's correct. :)
As expected changing the name .name entry fixes the sysfs behaviour.

diff --git a/drivers/net/phy/realtek.c b/drivers/net/phy/realtek.c
index 7fc8508b5..271e8adc3 100644
--- a/drivers/net/phy/realtek.c
+++ b/drivers/net/phy/realtek.c
@@ -220,7 +220,7 @@ static struct phy_driver realtek_drvs[] = {
  		.flags          = PHY_HAS_INTERRUPT,
  	}, {
  		.phy_id		= 0x001cc816,
-		.name		= "RTL8201F 10/100Mbps Ethernet",
+		.name		= "RTL8201F Fast Ethernet",
  		.phy_id_mask	= 0x001fffff,
  		.features	= PHY_BASIC_FEATURES,
  		.flags		= PHY_HAS_INTERRUPT,

> I wonder if other drivers have similar problems?
> 
> davicom.c:      .name           = "Davicom DM9161B/C",
> intel-xway.c:           .name           = "Intel XWAY PHY11G (PEF 7071/PEF 7072) v1.3",
> intel-xway.c:           .name           = "Intel XWAY PHY11G (PEF 7071/PEF 7072) v1.4",
> intel-xway.c:           .name           = "Intel XWAY PHY11G (PEF 7071/PEF 7072) v1.5 / v1.6",
> intel-xway.c:           .name           = "Intel XWAY PHY22F (PEF 7061) v1.5 / v1.6",
> smsc.c:	 .name	       = "SMSC LAN8710/LAN8720",

I'm open to suggestions about how to rename those identifiers.
"|" seems to work but IMHO looks a bit weird:

"Davicom DM9161B/C" -> "Davicom DM9161B|C"
"(PEF 7071/PEF 7072) v1.5 / v1.6" -> "(PEF 7071|7072) v1.5|6"

We can go full regex, which will probably get me voted off the island:
"(PEF 7071/PEF 7072) v1.5 / v1.6" -> "(PEF {7071,7072}) v1.{5,6}"

Cast your votes now!

cheers,
Holger