Mity CPU Platforms » MitySOM-5CSX Altera Cyclone V

The kernel used on MitySOM is 3.16.0.

Hello Dominik,

Have you tried a quick test of using a smaller flash drive with the MitySOM? Just to see if we can get a baseline, I've used 8GB/16GB flash drives on the MitySOM without issue.

Thanks!

Dan

Hello Dan,

Yes. I tested with 8 and 16GB USB sticks as well and it works just fine, so the culprit might be in the unusual size.

So this might be a red herring but in the pi's dmesg it has the following line, which the MitySOM doesn't:

sd 0:0:0:0: Attached scsi generic sg0 type 0

It looks like it's using the SCSI generic driver, which isn't part of the MitySOM defconfig we maintain. Looks like the kconfig for it is: CONFIG_CHR_DEV_SG.

You might also might want to do a quick diff of the mitysom5csx_devkit_defconfig and the RPI's 3.16 defconfig to see if there might be some drivers for USB/SCSI that our defconfig doesn't have.

Dan

Hey Dan,

Thanks for your reply.
Building CHR_DEV_SG into mitysom kernel doesn't seem to fix anything. The SG initialization message appears now though. I've also tried to use /dev/sg0 device as the dd input with and without DEV_SD driver included into kernel, without success.

I also find it odd that sd_read_capacity is being called twice (printing twice the total device size) but I cannot identify a reason of that.

[    2.434541] cdc_acm 1-1:1.0: ttyACM0: USB ACM device
[    2.438834] usb-storage 1-1:1.2: USB Mass Storage device detected
[    2.443855] scsi0 : usb-storage 1-1:1.2
[    3.445600] scsi 0:0:0:0: Direct-Access     ATMEL    Mass Storage MSD 0.01 PQ: 0 ANSI: 6
[    3.452968] sd 0:0:0:0: Attached scsi generic sg0 type 0
[    3.453085] sd 0:0:0:0: [sda] 2147483648 4096-byte logical blocks: (8.79 TB/8.00 TiB)
[    3.468013] sd 0:0:0:0: [sda] Write Protect is off
[    3.471492] sd 0:0:0:0: [sda] Mode Sense: 00 00 00 00
[    3.487538] sd 0:0:0:0: [sda] Asking for cache data failed
[    3.491704] sd 0:0:0:0: [sda] Assuming drive cache: write through
[    3.497268] sd 0:0:0:0: [sda] 2147483648 4096-byte logical blocks: (8.79 TB/8.00 TiB)
[    3.536372] sd 0:0:0:0: [sda] Attached SCSI removable disk

Do you have any other ideas?

Kind regards,
Dominik

Hi Dominik,

The only thing I could really think of was try the very verbose USB output like you initially suggested or add printk's to the USB driver. Try it with a flash drive and then the 8TB devices and compare the outputs.

I'm sorry, if I can think of a better way to test this I'll be sure to let you know.

Dan

Hey Dan!

I found the culprit! :)
As it often happens in linux kernel, the solution turns out to be very trivial.
It requires enabling CONFIG_LBDAF, which extends kernel partition device sector count width from 32 to 64 bits.
http://cateee.net/lkddb/web-lkddb/LBDAF.html

Kind regards,
Dominik

Hi Dominik,

Glad you found the solution! I know I've spent many hours working on kernel issues to have them solved with one line changes.

Thank you for the update!

Dan