Bug 32672 - Sis 191
Summary: Sis 191
Status: RESOLVED OBSOLETE
Alias: None
Product: Drivers
Classification: Unclassified
Component: Network (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: Francois Romieu
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-04-05 05:02 UTC by Ben Wang
Modified: 2016-12-15 20:20 UTC (History)
4 users (show)

See Also:
Kernel Version: 2.6.38
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Ben Wang 2011-04-05 05:02:08 UTC
I have the SIS191 gigabit ethernet card on HP EX470. Had the MTU bug and changed it from 1500 to 1024. All seems to work fine until I started transfer large files. At some random point, sometimes after transferring a few hundred MB, sometimes even a few GB of data ca be transferred, the driver will stop working. Both TX and RX large files will fail.

Logging in at the console, the interface is still up according to ifconfig. But no traffic can go out. Bringing the interface down and restarting it brings everything back to normal. Nothing special on dmesg when the interface goes down. Kernel doesn't hang or panic.

Tried Ubuntu and Gentoo, many different versions of the kernel. All failed the same way. Currently running 2.6.38.

This output of lspci:

00:04.0 Ethernet controller [0200]: Silicon Integrated Systems [SiS] 191 Gigabit Ethernet Adapter [1039:0191] (rev 01)

Related dmesg output:

sis190: sis190 Gigabit Ethernet driver 1.4 loaded
sis190 0000:00:04.0: PCI INT A -> GSI 19 (level, low) -> IRQ 19
sis190 0000:00:04.0: setting latency timer to 64
sis190: 0000:00:04.0: Read MAC address from EEPROM
sis190: 0000:00:04.0: Atheros PHY transceiver at address 0
sis190 0000:00:04.0: eth0: 0000:00:04.0: SiS 191 PCI Gigabit Ethernet adapter at ffffc90000336000 (IRQ: 19), 00:0a:e4:88:cb:42
sis190 0000:00:04.0: eth0: RGMII mode.
sis190 0000:00:04.0: eth0: Enabling Auto-negotiation

...

ADDRCONF(NETDEV_UP): eth0: link is not ready
sis190 0000:00:04.0: eth0: mii ext = 0000
sis190 0000:00:04.0: eth0: mii lpa=cde1 adv=01e1 exp=000f
sis190 0000:00:04.0: eth0: link on 1000 Mbps Full Duplex mode

ifconfig output:

eth0      Link encap:Ethernet  HWaddr 00:0a:e4:88:cb:42  
          inet addr:192.168.1.9  Bcast:192.168.1.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1024  Metric:1
          RX packets:21543 errors:0 dropped:1 overruns:0 frame:0
          TX packets:15848 errors:0 dropped:0 overruns:0 carrier:0
          collisions:2 txqueuelen:1000 
          RX bytes:14964283 (14.2 MiB)  TX bytes:2665690 (2.5 MiB)
          Interrupt:19 Base address:0xdead 

uname -a:

Linux hpserver 2.6.38-gentoo-r1 #2 SMP Sun Apr 3 13:18:53 PDT 2011 x86_64 AMD Athlon(tm) X2 Dual Core Processor BE-2300 AuthenticAMD GNU/Linux

I heard that 32 bit kernel doesn't have any problem but haven't tried it. I'd like to fix this issue myself if someone can point me to the right direction.

Let me if any more info is needed. Thanks!
Comment 1 anonymous 2011-06-15 07:19:16 UTC
I had the same bug on my notebook Asus X59SL.
Comment 2 Alan 2013-12-23 11:55:02 UTC
If this is still seen on a modern kernel please update
Comment 3 Ivan 2014-05-26 21:38:41 UTC
I can confirm that this is still an issue under Kernel 3.2.0-4 on Debian Wheezy, as well as under the backported Kernel 3.14.

It seems to be random how long it takes for the interface to go down, but it will go down quicker if you are doing large transfers. There are no errors relating to this anywhere that I can find, and the system believes the connection to still be up, despite all attempts to use it failing.

My lspci output:
00:04.0 Ethernet controller: Silicon Integrated Systems [SiS] 191 Gigabit Ethernet Adapter (rev 01)

I am not sure what output would be useful in this case, as there don't appear to be any errors anywhere.

Note You need to log in before you can comment on or make changes to this bug.