NanoPi Neo 4 - USB 3.0 issues under heavy load

[Jim MacKenzie]

I've found a problem with Armbian on the NanoPi Neo 4 with USB 3.0.

 

I have a RAID6 via four USB 3.0 external hard disks (I know, not optimum but it's been working fine for years off an older ARM box using USB 2.0).  While the RAID works fine on the USB 2.0 port (established this was true afterward), when connected to the USB 3.0 port, the USB 3.0 hub I'm using disconnects after only a few seconds, disconnecting all the drives (of course).

 

I initially used an unpowered USB 3.0 hub (I forget the manufacturer; I can check).  I figured at first, this might be a power issue (even though the external drives are all self-powered).  I found a 10-port powered Cable Matters USB 3.0 hub in my basement, and deployed it today, thinking that would solve my issues  But it doesn't.  Not a thing changes; the same problem recurs.

 

I moved the hub's USB cable to the USB 2.0 port on the NanoPi and that cures the issue.  So my RAID is slow, but it works.  (And it's faster than it was on my old box, so it's good enough for now.)

 

I'm not sure if this is a known issue, but thought I would report it.

 

 

[Igor]

6 hours ago, Jim MacKenzie said:

I'm not sure if this is a known issue

I think this is related:

So it looks like one of a known ones. If we had more 

... this could be clear from quick forum scan.

[Jim MacKenzie]

Thanks.  I'll check out that thread.

Not sure if it's significant, but I notice in /boot/armbianEnv.txt:

usbstoragequirks=[snip],0x0bc2:0x331a:u,0x0bc2:0x3322:u

 

And:

 

jim@nobby:~$ lsusb | grep Seagate
Bus 005 Device 009: ID 0bc2:3322 Seagate RSS LLC
Bus 005 Device 008: ID 0bc2:331a Seagate RSS LLC
Bus 005 Device 007: ID 0bc2:3322 Seagate RSS LLC
Bus 005 Device 006: ID 0bc2:3322 Seagate RSS LLC
 

I'm not sure what "usbstoragequirks" does, but perhaps it relates to this issue.

[mhyon]

For USB 2.0, are you using a 4-pin header connector into GPIO2 (pins 1-4)?

[Igor]

On 2/24/2019 at 4:09 PM, Jim MacKenzie said:

I'm not sure what "usbstoragequirks" does, but perhaps it relates to this issue.

http://linux-sunxi.org/USB/UAS#Blacklisted_.2F_UAS_disabled

[Jim MacKenzie]

On 5/3/2019 at 9:59 AM, mhyon said:

For USB 2.0, are you using a 4-pin header connector into GPIO2 (pins 1-4)?

Somehow I missed answering this question.

I'm just using the USB 2.0 port on the device.  Mine has a pair of ports, one black (2.0) and one blue (3.0).

 

Also, my system is still running kernel 4.4.192-rk3399 from October of 2019.  Perhaps an upgrade to a mainline kernel might solve this issue, but I'm not sure which mainline kernel to use.  (Debian Buster version of Armbian.)

[mhyon]

37 minutes ago, Jim MacKenzie said:

Somehow I missed answering this question.

I'm just using the USB 2.0 port on the device.  Mine has a pair of ports, one black (2.0) and one blue (3.0).

 

Also, my system is still running kernel 4.4.192-rk3399 from October of 2019.  Perhaps an upgrade to a mainline kernel might solve this issue, but I'm not sure which mainline kernel to use.  (Debian Buster version of Armbian.)

Ah ok.  Mine has 4 blue (3.0) ports.

[Jim MacKenzie]

57 minutes ago, mhyon said:

Ah ok.  Mine has 4 blue (3.0) ports.

 

I was misremembering - I think mine has four ports, too.  But at least one and possibly two are really USB 2.0.  (This device lives away from here, so I can't look at it.)