Re: [Xen-devel] Xen 3.4 & tg3 gigabit ethernet stalls

[Pasi Kärkkäinen <pasik@xxxxxx>]

On Thu, Mar 11, 2010 at 07:42:57PM +0100, René Bühlmann wrote:
> Pasi Kärkkäinen schrieb:
>> On Wed, Mar 10, 2010 at 07:19:50PM +0100, René Bühlmann wrote:
>>   
>>> Hi all,
>>>
>>> Some Weeks ago, i've upgraded my Xen 3.3 installation to a 3.4 Xen. 
>>> The  same time, i upgraded Dom0 Kernel to 2.6.31.2 which also made 
>>> the  interface to switch from 100Mbit to 1Gbit. Since then, the tg3 
>>> interface  stalls about once a week and i need to reboot Xen. The 
>>> problem looks  similar to   
>>> http://lists.xensource.com/archives/html/xen-devel/2009-07/msg00139.html 
>>>  but the solution there (adding cpuidle=0 cpufreq=none to kernel 
>>> param.)  did not help for me.
>>>
>>> Does anyone has a solution or workaround for this or how could i 
>>> debug  the problem?
>>>
>>>     
>>
>> Have you monitored the recent tg3 driver changes in 2.6.32.x and/or 
>> 2.6.33? I remember seeing some patches and discussion about tg3..  
>> maybe it's a tg3 driver bug?
> I went through the tg3 parts of the kernel changelog. I'm not sure if  
> these patches are related to my problem. I will try 2.6.33 as soon as  
> pv_ops gets ported to it.
>>  
>>
>> Are you using pv_ops dom0 kernel, or a kernel with forwardported 
>> patches?   
> I'm using pv_ops dom0 from Jeremy's stable branch.
>

You could just copy the newer tg3 driver from newer kernel to
your kernel tree and rebuild.

-- Pasi

>> -- Pasi
>>
>>   
>>> Thanks
>>> René
>>>
>>> Here is the dmesg output:
>>>
>>> WARNING: at net/sched/sch_generic.c:246 dev_watchdog+0x23e/0x250()
>>> Hardware name: ProLiant ML110 G4
>>> NETDEV WATCHDOG: peth1 (tg3): transmit queue 0 timed out
>>> Modules linked in:
>>> Pid: 0, comm: swapper Not tainted 2.6.31.6 #2
>>> Call Trace:
>>> <IRQ>  [<ffffffff8143bbbe>] ? dev_watchdog+0x23e/0x250
>>> [<ffffffff8143bbbe>] ? dev_watchdog+0x23e/0x250
>>> [<ffffffff81045054>] ? warn_slowpath_common+0x74/0xd0
>>> [<ffffffff814c1fdf>] ? br_handle_frame_finish+0x14f/0x190
>>> [<ffffffff81045111>] ? warn_slowpath_fmt+0x51/0x60
>>> [<ffffffff8100f10f>] ? xen_restore_fl_direct_end+0x0/0x1
>>> [<ffffffff81503f4c>] ? _spin_unlock_irqrestore+0xc/0x10
>>> [<ffffffff8104150f>] ? try_to_wake_up+0xbf/0x1e0
>>> [<ffffffff8100f009>] ? xen_clocksource_get_cycles+0x9/0x20
>>> [<ffffffff81218111>] ? strlcpy+0x41/0x50
>>> [<ffffffff81426e6b>] ? netdev_drivername+0x3b/0x40
>>> [<ffffffff8143bbbe>] ? dev_watchdog+0x23e/0x250
>>> [<ffffffff8100e9a9>] ? xen_force_evtchn_callback+0x9/0x10
>>> [<ffffffff8100f122>] ? check_events+0x12/0x20
>>> [<ffffffff8143b980>] ? dev_watchdog+0x0/0x250
>>> [<ffffffff8104f1ac>] ? run_timer_softirq+0x13c/0x210
>>> [<ffffffff8104a935>] ? __do_softirq+0xa5/0x140
>>> [<ffffffff810141ac>] ? call_softirq+0x1c/0x30
>>> [<ffffffff8101613d>] ? do_softirq+0x4d/0x90
>>> [<ffffffff812740be>] ? xen_evtchn_do_upcall+0x14e/0x1d0
>>> [<ffffffff810141fe>] ? xen_do_hypervisor_callback+0x1e/0x30
>>> <EOI>  [<ffffffff810093aa>] ? hypercall_page+0x3aa/0x1010
>>> [<ffffffff810093aa>] ? hypercall_page+0x3aa/0x1010
>>> [<ffffffff8100ea3c>] ? xen_safe_halt+0xc/0x20
>>> [<ffffffff8100b825>] ? xen_idle+0x25/0x50
>>> [<ffffffff810122a6>] ? cpu_idle+0x66/0xa0
>>> [<ffffffff818208df>] ? start_kernel+0x2e6/0x328
>>> [<ffffffff818229d5>] ? xen_start_kernel+0x65c/0x6dc
>>> ---[ end trace 8108a21093ed2967 ]---
>>> tg3: peth1: transmit timed out, resetting
>>> tg3: DEBUG: MAC_TX_STATUS[ffffffff] MAC_RX_STATUS[ffffffff]
>>> tg3: DEBUG: RDMAC_STATUS[ffffffff] WDMAC_STATUS[ffffffff]
>>> tg3: tg3_stop_block timed out, ofs=2c00 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=2000 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=2400 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=2800 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=3000 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=1400 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=1800 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=c00 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=4800 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=1000 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=1c00 enable_bit=2
>>> tg3: tg3_abort_hw timed out for peth1, TX_MODE_ENABLE will not clear  
>>> MAC_TX_MODE=ffffffff
>>> tg3: tg3_stop_block timed out, ofs=3c00 enable_bit=2
>>> tg3: tg3_stop_block timed out, ofs=4c00 enable_bit=2
>>> tg3: peth1: No firmware running.
>>> tg3: tg3_abort_hw timed out for peth1, TX_MODE_ENABLE will not clear  
>>> MAC_TX_MODE=ffffffff
>>> tg3: peth1: Link is down.
>>> xenbr1: port 1(peth1) entering disabled state
>>>
>>>     
>>
>>
>>
>>   
>>> _______________________________________________
>>> Xen-devel mailing list
>>> Xen-devel@xxxxxxxxxxxxxxxxxxx
>>> http://lists.xensource.com/xen-devel
>>>     
>>
>>
>>   
>
>



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel