[LEDE-DEV] SUNXI: Network problem with kernel 4.9.51

Gerhard Bertelsmann info at gerhard-bertelsmann.de
Tue Sep 26 02:25:57 PDT 2017


Hi Hauke,

Am 2017-09-25 22:10, schrieb Hauke Mehrtens:
> On 09/25/2017 08:47 AM, Gerhard Bertelsmann wrote:
>> Hi Hauke,
>> 
>> Am 2017-09-25 00:14, schrieb Hauke Mehrtens:
>>> On 09/24/2017 04:39 PM, Hauke Mehrtens wrote:
>>>> On 09/24/2017 12:27 AM, Lucian Cristian wrote:
>>>>> On 23.09.2017 19:03, Gerhard Bertelsmann wrote:
>>>>>> Hi Zoltan,
>>>>>> 
>>>>>> Am 2017-09-23 17:36, schrieb Zoltan HERPAI:
>>>>>>> Hi Gerhard,
>>>>>>> 
>>>>>>> On Sat, 23 Sep 2017, Gerhard Bertelsmann wrote:
>>>>>>> 
>>>>>>>> Hi,
>>>>>>>> 
>>>>>>>> I'm trying Lede HEAD with a BananaPI (SUNXI7 A20) and have an 
>>>>>>>> issue
>>>>>>>> with the network connection:
>>>>>>>> 
>>>>>>>> root at Modellbahn-BPi:/# ping -qc 100 192.168.0.9
>>>>>>>> PING 192.168.0.9 (192.168.0.9): 56 data bytes
>>>>>>> [snip]
>>>>>>> 
>>>>>>>> Does anybody have a clue why this happens ?
>>>>>>> 
>>>>>>> Can you please post a full dmesg.
>>>>>> 
>>>>>> yes, of course:
>>>>>> 
>>>>>> root at Modellbahn-BPi:~# dmesg
>>>>>> [    0.000000] Booting Linux on physical CPU 0x0
>>>> .......
>>>>>> 
>>>>>> root at Modellbahn-BPi:~# ping -qc 100 192.168.0.9
>>>>>> PING 192.168.0.9 (192.168.0.9): 56 data bytes
>>>>>> 
>>>>>> --- 192.168.0.9 ping statistics ---
>>>>>> 100 packets transmitted, 94 packets received, 6% packet loss
>>>>>> round-trip min/avg/max = 0.354/0.518/1.182 ms
>>>>>> 
>>>>>> root at Modellbahn-BPi:~# ip -s -d link show eth0
>>>>>> 2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc fq_codel
>>>>>> master br-lan state UP mode DEFAULT group default qlen 1000
>>>>>>     link/ether 02:03:0a:43:06:44 brd ff:ff:ff:ff:ff:ff promiscuity 
>>>>>> 1
>>>>>>     bridge_slave state forwarding priority 32 cost 4 hairpin off 
>>>>>> guard
>>>>>> off root_block off fastleave off learning on flood on port_id 
>>>>>> 0x8001
>>>>>> port_no 0x1 designated_port 32769
>>>>>>     RX: bytes  packets  errors  dropped overrun mcast
>>>>>>     88003      560      0       9       0       0
>>>>>>     TX: bytes  packets  errors  dropped carrier collsns
>>>>>>     20196      210      0       0       0       0
>>>>>> 
>>>>>> Regards,
>>>>>> 
>>>>>> Gerd
>>>>>> 
>>>>>> _______________________________________________
>>>>>> Lede-dev mailing list
>>>>>> Lede-dev at lists.infradead.org
>>>>>> http://lists.infradead.org/mailman/listinfo/lede-dev
>>>>> 
>>>>> I don't have packet loss but I have slow upload (A20 Olimex Lime2) 
>>>>> but
>>>>> maybe is the uboot bug found on the internet if is not already in 
>>>>> the
>>>>> uboot mainline, if I increase the number of parallel connection the
>>>>> speed goes up to 200Mb/s
>>>>> 
>>>>> UP:
>>>>> 
>>>>> [ ID] Interval           Transfer     Bandwidth       Retr  Cwnd
>>>>> [  4]   0.00-1.00   sec  2.18 MBytes  18.2 Mbits/sec  148   7.03 
>>>>> KBytes
>>>>> [  4]   1.00-2.00   sec  4.73 MBytes  39.7 Mbits/sec  317   4.22 
>>>>> KBytes
>>>>> [  4]   2.00-3.00   sec  4.38 MBytes  36.7 Mbits/sec  294   4.22 
>>>>> KBytes
>>>>> [  4]   3.00-4.00   sec  4.92 MBytes  41.2 Mbits/sec  344   2.81 
>>>>> KBytes
>>>>> [  4]   4.00-5.00   sec  3.01 MBytes  25.3 Mbits/sec  210   12.7 
>>>>> KBytes
>>>>> [  4]   5.00-6.00   sec  7.56 MBytes  63.4 Mbits/sec  503   2.81 
>>>>> KBytes
>>>>> [  4]   6.00-7.00   sec  3.13 MBytes  26.2 Mbits/sec  216   1.41 
>>>>> KBytes
>>>>> [  4]   7.00-8.00   sec   665 KBytes  5.45 Mbits/sec   58   2.81 
>>>>> KBytes
>>>>> [  4]   8.00-9.00   sec  7.97 MBytes  66.9 Mbits/sec  554   7.03 
>>>>> KBytes
>>>>> [  4]   9.00-10.00  sec  1.42 MBytes  11.9 Mbits/sec  137   2.81 
>>>>> KBytes
>>>>> - - - - - - - - - - - - - - - - - - - - - - - - -
>>>>> [ ID] Interval           Transfer     Bandwidth       Retr
>>>>> [  4]   0.00-10.00  sec  39.9 MBytes  33.5 Mbits/sec 
>>>>> 2781            
>>>>> sender
>>>>> [  4]   0.00-10.00  sec  39.6 MBytes  33.2 
>>>>> Mbits/sec                 
>>>>> receiver
>>>>> 
>>>>> DOWN:
>>>>> 
>>>>> [ ID] Interval           Transfer     Bandwidth
>>>>> [  4]   0.00-1.00   sec  55.2 MBytes   463 Mbits/sec
>>>>> [  4]   1.00-2.00   sec  58.4 MBytes   490 Mbits/sec
>>>>> [  4]   2.00-3.00   sec  55.3 MBytes   464 Mbits/sec
>>>>> [  4]   3.00-4.00   sec  55.4 MBytes   465 Mbits/sec
>>>>> [  4]   4.00-5.00   sec  55.8 MBytes   468 Mbits/sec
>>>>> [  4]   5.00-6.00   sec  56.1 MBytes   471 Mbits/sec
>>>>> [  4]   6.00-7.00   sec  55.1 MBytes   463 Mbits/sec
>>>>> [  4]   7.00-8.00   sec  55.6 MBytes   466 Mbits/sec
>>>>> [  4]   8.00-9.00   sec  55.3 MBytes   464 Mbits/sec
>>>>> [  4]   9.00-10.00  sec  55.4 MBytes   465 Mbits/sec
>>>>> - - - - - - - - - - - - - - - - - - - - - - - - -
>>>>> [ ID] Interval           Transfer     Bandwidth       Retr
>>>>> [  4]   0.00-10.00  sec   559 MBytes   469 Mbits/sec 0            
>>>>> sender
>>>>> [  4]   0.00-10.00  sec   559 MBytes   469 
>>>>> Mbits/sec                 
>>>>> receiver
>>>> 
>>>> Hi,
>>>> 
>>>> I think the U-Boot should not influence this.
>>>> 
>>>> @Gred could you please remove all the patches from
>>>> target/linux/sunxi/patches-4.9/ and test with such an image and then
>>>> report back your results please? It could be that something when 
>>>> wrong
>>>> when I backported some clock patches.
>>>> 
>>>> Hauke
>>> 
>>> Hi Gerd,
>>> 
>>> Could you please also test in addition my sunxi feature branch here:
>>> https://git.lede-project.org/?p=lede/hauke/staging.git;a=shortlog;h=refs/heads/sunxi
>>> 
>>> 
>>> It contains an updated version of the stmmac driver, I did this 
>>> mostly
>>> to add support for the A64, but it could also fix your problems.
>>> 
>>> Hauke
>> 
>> thanks for your assistance.
>> 
>> I've tested both versions, stock Lede HEAD without patches and your 
>> git.
>> Both showing the same problem with packet lost and poor upload 
>> bandwidth.
>> 
>> 
>> Regards
>> 
>> Gerd
> 
> Ok they I can probably not help any more, this looks more like a bug in
> the upstream kernel introduced between kernel 4.4 and 4.9. It probably
> makes sense to talk directly to the Linux stmmac maintainers about this
> problem.
> 
> Hauke

IMHO Cristian mentioned it already, there seems to some magic in the old
Uboot versions aka CONFIG_GMAC_TX_DELAY.

I adapted 
https://groups.google.com/forum/#!msg/linux-sunxi/Y_Zh5juEJG4/egjJojeTgS8J
and received much better results:

U-Boot 2017.07 (Sep 24 2017 - 14:28:28 +0000) Allwinner Technology

CPU:   Allwinner A20 (SUN7I)
Model: LeMaker Banana Pi
I2C:   ready
DRAM:  1 GiB
MMC:   SUNXI SD/MMC: 0
*** Warning - bad CRC, using default environment


=> md.l 0x1c20164 1
01c20164: 00000000                               ....
=> mw.l 0x1c20164 0xc06 1
=> boot

...

root at LEDE:~# uname -a
Linux LEDE 4.9.51 #0 SMP PREEMPT Sun Sep 24 14:28:28 2017 armv7l
GNU/Linux

root at LEDE:~# ping -qc 100 192.168.0.9
PING 192.168.0.9 (192.168.0.9): 56 data bytes

--- 192.168.0.9 ping statistics ---
100 packets transmitted, 100 packets received, 0% packet loss
round-trip min/avg/max = 0.363/0.514/0.813 ms


root at LEDE:~# iperf -c 192.168.0.9 -d
------------------------------------------------------------
Server listening on TCP port 5001
TCP window size: 85.3 KByte (default)
------------------------------------------------------------
------------------------------------------------------------
Client connecting to 192.168.0.9, TCP port 5001
TCP window size: 87.5 KByte (default)
------------------------------------------------------------
[  5] local 192.168.0.159 port 54946 connected with 192.168.0.9 port 
5001
[  4] local 192.168.0.159 port 5001 connected with 192.168.0.9 port 
52498
[ ID] Interval       Transfer     Bandwidth
[  5]  0.0-10.0 sec   695 MBytes   583 Mbits/sec
[  4]  0.0-10.0 sec   330 MBytes   276 Mbits/sec

I'm not an Uboot expert, maybe somebody else could migrate
this parameter from the old Uboot to the new one.

Regards

Gerd




More information about the Lede-dev mailing list