Mity CPU Platforms » MityDSP-L138 (ARM9 Based Platforms)

Hannes,

Can you upload your kernel .config file? We have seen similar problems on our 335x product line that were traced to an improper kernel configuration.

Thanks,

-Bob

Hi Bob,

Here's the file.

Thanks
Hannes

Could you post a full boot log? I'd like to compare it with what I'm getting. Also would be nice to see the connection log for the working port for comparison.

OT: Hannes, can you please double check your email address in your profile, I (admin) am getting bounced email warnings for every update to this thread.
thank you
/Tim Iskander

There you are.

That's the dmesg right after startup, with the card reader plugged in during startup. I enabled USB-MSD debug messages, so there's a bit more noise from that.

I'm also getting

udc_core: exports duplicate symbol usb_add_gadget_udc (owned by kernel)

Heres my bootlog. I'm having a hard time getting the mini usb working at all, in either host or peripheral mode. I haven't tried switching from PIO only to the DMA driver you enabled. Did you find that fixed something for you?

I tried changing between PIO and DMA before. That changed nothing about the problem - It enumerates the device and then stops.

Another thing:
If I plug in an external USB hub, the Linux detects that. Afterwards if i plug anything in, nothing happens.

But if i attach the hub with the device already plugged it, the Linux enumerates both devices. If i detach the device in the hub, the unplug doesn't get detected again.

I tried with a bunch of other devices (mostly USBHID). They also get detected and enumerated, the driver gets loaded, then nothing happens at all.

Ok I was able to get things to the point where it will atleast detect the insertion of the flash drive by changing the usb register from MUSB_OTG to MUSB_HOST. I suspect that this is where you are at.

Exactly. But the mass storage driver itself gets loaded and then stalls (no /dev/sd* gets created).

Wondering if this is perhaps a udev issue?

I don't think so. Udev primarily loads the driver and gets called once the driver does something. In the boot log you can see that the usb enumeration hangs during kernel initialization (usb is compiled in) and before init (and therefor udev) is running.

If I do the same on the ohci port, it works - including initilization of the usb-msd before init is running.

Got it working.

-# CONFIG_USB_SUSPEND is not set
+CONFIG_USB_SUSPEND=y
+CONFIG_USB_OTG=y
+# CONFIG_USB_OTG_WHITELIST is not set
+# CONFIG_USB_OTG_BLACKLIST_HUB is not set

I'll push a patch in a couple minutes.

Hannes, please let us know if this resolves your issue. We actually fixed this same problem on our MitySOM-335x module a few days before you posted your question.

For the sake of those searching the forum in the future, the MitySOM-335x fix is in this commit.

Thanks for the detailed troubleshooting information.

-Bob

That changed stuff a bit. I now get more messages and the usb-msd driver seems to wake up.

But during initialization, i get error messages of disconnect and reconnect - see the attached log. I'm not sure whether the problem's in the cable or in the software. I'll only be able to check next week, hopefully with more information then.

Additionally, that patch seems to have made the ethernet of the device unusable. But that's not important right now.

How were you experiencing the unstable ethernet? Just transferring a file?

Also you can buy the cables prebuilt.

USB mini 5 pin to USB A Female

http://www.amazon.com/USB-Female-Mini-Male-Adapter/sim/B0016RNX2I/2

I was able to scp a 12 MB file without any errors

When I enabled the dma I got the following errors when i plugged in the flash drive.

root@mityomapl138 ~ $ usb 1-1: new high-speed USB device number 2 using musb-hdrc
scsi1 : usb-storage 1-1:1.0

root@mityomapl138 ~ $ usb_rx_ch_teardown 1094: teardown Desc not found
usb 1-1: reset high-speed USB device number 2 using musb-hdrc
usb 1-1: reset high-speed USB device number 2 using musb-hdrc
usb 1-1: reset high-speed USB device number 2 using musb-hdrc
usb 1-1: reset high-speed USB device number 2 using musb-hdrc
usb 1-1: reset high-speed USB device number 2 using musb-hdrc
scsi 1:0:0:0: Device offlined - not ready after error recovery

So perhaps there is a problem with the dma driver as well. Can you try your device with PIO only set?

Hey again,

i tried changing it from DMA to PIO. Now, everything seems to work. Device is readable, writable and mountable. Performonce is ok - about 8MB/s either way.

The ethernet problem is gone. I'm not sure where the problem was. eth0 was up and had an IP, but it didn't send/receive anything on the network. Didn't try if ARP worked or not. But now it works again, so we're fine here.

If you don't mind, could you post your config.gz file?

There you are. Note that i had to change the initialization in the code as noted above.