g_cdc USB modules cause kernel (3.12.10) crash on AM3352

Hi Marco,

Are you seeing the same behavior any other old TI SDKs or only SDK 7 ?

Have you tried to sent/receive the files through "tftp" command or any other n/w file transfer commands ?

Like the BB black , i have two usb slots. One is the host , the second one is  the device (not OTG because is forced to be device. The ID pin is floating. )

Sorry I'm not aware about AM335x, some how I give you my suggestions and ideas to narrow down the problem.

Both USB device ports are connected to the same USB controller ( or HUB ) or different USB controller ?

Hi Titus :) :)

nice to hear news from you, thanks a lot for the fast answer .

I've seen that behavior only with the sdk7 , the one that i've used to develop my distribution. I've seen something similar of my problem (in other linux forum discussions) but with older version of kernel like 2.6.x .... I don't think it could be useful,but i could be wrong.

Two usb ports are linked directly with a cable. Distribution are identical in kernel and drivers.

platform 4830e000.lcdc: Driver da8xx_lcdc

at the host side , when the other board wake up i get

 [   10.774225] usb 2-1: new high-speed USB device number 2 using musb-hdrc
[   10.915275] usb 2-1: New USB device found, idVendor=0525, idProduct=a4aa
[   10.922454] usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0
[   10.930107] usb 2-1: Product: CDC Composite Gadget
[   10.935237] usb 2-1: Manufacturer: Linux 3.12.10 with musb-hdrc
[   10.959861] cdc_ether 2-1:1.0 usb0: register 'cdc_ether' at usb-musb-hdrc.1.auto-1, CDC Ethernet Device, c6:f8:24:22:b3:f0
[   11.002504] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
[   11.066951] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
 new cdc_link found ...
... link established
root
root@localhost:~ > lsusb
Bus 002 Device 002: ID 0525:a4aa Netchip Technology, Inc. Linux-USB CDC Composite Gadge (Ethernet and ACM)
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub


From device side

Jul 18 15:56:02  kernel: [    9.036391] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
Jul 18 15:56:02  kernel: [    9.047691] of_get_named_gpio_flags: can't parse gpios property of node '/ocp/usb@47400000/usb-phy@47401300[0]'
Jul 18 15:56:02  kernel: [    9.051936] of_get_named_gpio_flags: can't parse gpios property of node '/ocp/usb@47400000/usb-phy@47401b00[0]'
Jul 18 15:56:02  kernel: [    9.096645] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
Jul 18 15:56:02 kernel: [    9.114959] musb-hdrc musb-hdrc.0.auto: Enabled SW babble control
Jul 18 15:56:02  kernel: [    9.145196] musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, bulk combine, bulk split, HB-ISO Rx, HB-ISO Tx, SoftConn)
Jul 18 15:56:02 kernel: [    9.145220] musb-hdrc: MHDRC RTL version 2.0
Jul 18 15:56:02  kernel: [    9.145228] musb-hdrc: setup fifo_mode 4
Jul 18 15:56:02  kernel: [    9.145243] musb-hdrc: 28/31 max ep, 16384/16384 memory
Jul 18 15:56:02  kernel: [    9.145347] musb-hdrc musb-hdrc.0.auto: MUSB HDRC host driver
Jul 18 15:56:02  kernel: [    9.173251] musb-hdrc musb-hdrc.0.auto: new USB bus registered, assigned bus number 1
Jul 18 15:56:02 kernel: [    9.188813] usb usb1: New USB device found, idVendor=1d6b, idProduct=0002
Jul 18 15:56:02 kernel: [    9.196016] usb usb1: New USB device strings: Mfr=3, Product=2, SerialNumber=1
Jul 18 15:56:02 kernel: [    9.203604] usb usb1: Product: MUSB HDRC host driver
Jul 18 15:56:02 kernel: [    9.208811] usb usb1: Manufacturer: Linux 3.12.10 musb-hcd
Jul 18 15:56:02 kernel: [    9.214576] usb usb1: SerialNumber: musb-hdrc.0.auto
Jul 18 15:56:02 kernel: [    9.221748] hub 1-0:1.0: USB hub found
Jul 18 15:56:02 kernel: [    9.225725] hub 1-0:1.0: 1 port detected
Jul 18 15:56:02 kernel: [    9.230322] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
Jul 18 15:56:02 kernel: [    9.239306] musb-hdrc musb-hdrc.1.auto: Enabled SW babble control
Jul 18 15:56:02 kernel: [    9.247018] musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, bulk combine, bulk split, HB-ISO Rx, HB-ISO Tx, SoftConn)
Jul 18 15:56:02 kernel: [    9.247029] musb-hdrc: MHDRC RTL version 2.0
Jul 18 15:56:02 kernel: [    9.247036] musb-hdrc: setup fifo_mode 4
Jul 18 15:56:02 kernel: [    9.247051] musb-hdrc: 28/31 max ep, 16384/16384 memory
Jul 18 15:56:02 kernel: [    9.247144] musb-hdrc musb-hdrc.1.auto: MUSB HDRC host driver
Jul 18 15:56:02 kernel: [    9.255819] musb-hdrc musb-hdrc.1.auto: new USB bus registered, assigned bus number 2
Jul 18 15:56:02 kernel: [    9.264852] usb usb2: New USB device found, idVendor=1d6b, idProduct=0002
Jul 18 15:56:02 kernel: [    9.272020] usb usb2: New USB device strings: Mfr=3, Product=2, SerialNumber=1
Jul 18 15:56:02 kernel: [    9.279595] usb usb2: Product: MUSB HDRC host driver
Jul 18 15:56:02 kernel: [    9.284821] usb usb2: Manufacturer: Linux 3.12.10 musb-hcd
Jul 18 15:56:02 kernel: [    9.290574] usb usb2: SerialNumber: musb-hdrc.1.auto
Jul 18 15:56:02 kernel: [    9.296973] hub 2-0:1.0: USB hub found
Jul 18 15:56:02 kernel: [    9.300943] hub 2-0:1.0: 1 port detected
Jul 18 15:56:02 kernel: [    9.305617] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
Jul 18 15:56:02 kernel: [    9.330692] platform 4830e000.lcdc: Driver da8xx_lcdc requests probe deferral
Jul 18 15:56:03 connmand[1241]: eth0 {update} flags 102467 <UP,RUNNING,LOWER_UP>
Jul 18 15:56:03 connmand[1241]: eth0 {newlink} index 2 address 1C:BA:8C:B7:55:1D mtu 1500
Jul 18 15:56:03  connmand[1241]: eth0 {newlink} index 2 operstate 6 <UP>
Jul 18 15:56:03  kernel: [   10.221754] libphy: 4a101000.mdio:00 - Link is Up - 100/Full
Jul 18 15:56:03  connmand[1241]: Skipping disconnect of carrier, network is connecting.
Jul 18 15:56:03  connmand[1241]: Setting hostname to elenos
Jul 18 15:56:03  avahi-daemon[1364]: Registering new address record for 10.0.100.125 on eth0.IPv4.
Jul 18 15:56:03  connmand[1241]: eth0 {add} address 10.0.100.125/24 label eth0 family 2
Jul 18 15:56:03  connmand[1241]: eth0 {add} route 10.0.100.254 gw 0.0.0.0 scope 253 <LINK>
Jul 18 15:56:03  connmand[1241]: eth0 {add} route 208.67.220.220 gw 10.0.100.254 scope 0 <UNIVERSE>
Jul 18 15:56:04  connmand[1241]: eth0 {add} route 87.106.208.187 gw 10.0.100.254 scope 0 <UNIVERSE>
Jul 18 15:56:04  connmand[1241]: eth0 {del} route 87.106.208.187 gw 10.0.100.254 scope 0 <UNIVERSE>
Jul 18 15:56:04  kernel: [   11.187101] udc musb-hdrc.0.auto: registering UDC driver [g_cdc]
Jul 18 15:56:04  kernel: [   11.204226] using random self ethernet address
Jul 18 15:56:04  kernel: [   11.209006] using random host ethernet address
Jul 18 15:56:04  kernel: [   11.230207] userial_init: registered 4 ttyGS* devices
Jul 18 15:56:04  kernel: [   11.239322] g_cdc gadget: adding config #1 'CDC Composite (ECM + ACM)'/bf0872cc
Jul 18 15:56:04  kernel: [   11.239390] g_cdc gadget: adding 'cdc_ethernet'/dd316200 to config 'CDC Composite (ECM + ACM)'/bf0872cc
Jul 18 15:56:04  connmand[1241]: usb0 {create} index 3 type 1 <ETHER>
Jul 18 15:56:04  connmand[1241]: usb0 {update} flags 4098 <DOWN>
Jul 18 15:56:04  connmand[1241]: usb0 {newlink} index 3 operstate 2 <DOWN>
Jul 18 15:56:04  kernel: [   11.246901] usb0: HOST MAC c6:f8:24:22:b3:f0
Jul 18 15:56:04  kernel: [   11.258493] usb0: MAC 1a:c6:dd:11:c9:34
Jul 18 15:56:04  connmand[1241]: Adding interface usb0 [ gadget ]
Jul 18 15:56:04  connmand[1241]: usb0 {newlink} index 3 address 1A:C6:DD:11:C9:34 mtu 1500
Jul 18 15:56:04  connmand[1241]: usb0 {newlink} index 3 operstate 2 <DOWN>
Jul 18 15:56:04 kernel: [   11.275958] g_cdc gadget: CDC Ethernet: dual speed IN/ep1in OUT/ep1out NOTIFY/ep2in
Jul 18 15:56:04  kernel: [   11.276033] g_cdc gadget: adding 'acm'/dd663f00 to config 'CDC Composite (ECM + ACM)'/bf0872cc
Jul 18 15:56:04  kernel: [   11.276088] g_cdc gadget: acm ttyGS0: dual speed IN/ep3in OUT/ep2out NOTIFY/ep4in
Jul 18 15:56:04  kernel: [   11.276118] g_cdc gadget: cfg 1/bf0872cc speeds: high full
Jul 18 15:56:04  kernel: [   11.276142] g_cdc gadget:   interface 0 = cdc_ethernet/dd316200
Jul 18 15:56:04  kernel: [   11.276164] g_cdc gadget:   interface 1 = cdc_ethernet/dd316200
Jul 18 15:56:04  kernel: [   11.276186] g_cdc gadget:   interface 2 = acm/dd663f00
Jul 18 15:56:04  kernel: [   11.276207] g_cdc gadget:   interface 3 = acm/dd663f00
Jul 18 15:56:04  kernel: [   11.276248] g_cdc gadget: CDC Composite Gadget, version: King Kamehameha Day 2008
Jul 18 15:56:04  kernel: [   11.284347] g_cdc gadget: g_cdc ready
Jul 18 15:56:05  kernel: [   11.876808] g_cdc gadget: high-speed config #1: CDC Composite (ECM + ACM)
Jul 18 15:56:05  kernel: [   11.884088] g_cdc gadget: init ecm
Jul 18 15:56:05  kernel: [   11.884116] g_cdc gadget: notify connect false
Jul 18 15:56:05  kernel: [   11.884155] g_cdc gadget: activate acm ttyGS0
Jul 18 15:56:05  kernel: [   11.884190] g_cdc gadget: acm ttyGS0 serial state 0000
Jul 18 15:56:05 kernel: [   11.884940] g_cdc gadget: activate ecm
Jul 18 15:56:05  kernel: [   11.885024] usb0: qlen 10
Jul 18 15:56:05  kernel: [   11.885045] g_cdc gadget: ecm_close
Jul 18 15:56:05  kernel: [   11.903547] g_cdc gadget: notify connect false
Jul 18 15:56:05  kernel: [   11.935675] g_cdc gadget: notify speed 425984000
Jul 18 15:56:05  connmand[1241]: usb0 {update} flags 69699 <UP,RUNNING,LOWER_UP>
Jul 18 15:56:05  connmand[1241]: usb0 {newlink} index 3 address 1A:C6:DD:11:C9:34 mtu 1500
Jul 18 15:56:05  connmand[1241]: usb0 {newlink} index 3 operstate 6 <UP>
Jul 18 15:56:05  kernel: [   12.458732] usb0: eth_open
Jul 18 15:56:05  kernel: [   12.458778] usb0: eth_start
Jul 18 15:56:05  kernel: [   12.458867] g_cdc gadget: ecm_open
Jul 18 15:56:05  kernel: [   12.458889] g_cdc gadget: notify connect true

i've tried with netcat command but it seems that it doedn't works because it refuse connection while ssh connection is active

i've tried tftp command but it goes on timeout both send and recive file

Now i'm trying to configure proftpd on the host side. the goal is to obtain a transfer with the udp packets. 

OTHER problem : i've already set up a fuse server with sshfs and transfer works fine from device to host but , when i copy or move a file from the folder , the kernl crash and system panic take place.

I think to call ghostbuster in certain moments :)

Best regards

Marco

Marco,

Please disable CPPI DMA and rebuild the kernel for both boards, if the issue goes away, please try the kernel git://git.ti.com/ti-linux-kernel/ti-linux-kernel.git, branch origin/ti-linux-3.12.y. This is the kernel repo for SDK releases, it is currently on 3.12.25 and have some CPPI DMA bug fixes.

Marco,

I have some questions for your setup.

Marco Destro said:

the HOST side ,

I load

modprobe am335x

what is the am335x module? is it supposed to be musb_am335x?

Marco Destro said:

FROM E THE DEVICE side

modprobe libcomposite
modprobe musb_hdrc
modprobe musb_am335x
sleep 1
modprobe usbnet
modprobe u_ether
sleep 1
modprobe g_cdc
modprobe u_rndis

Why you have to load such many gadget modules? I think you only have to do

# modprobe musb_am335x

# modprobe g_ether

Hi Bin :)

Thanks a lot for your answers!

i've typed wrong. You're right it's musb_am335x module.

You're right too on g_ether.

The effects on the host are the same.

Please, could you tell me which files need to be changed?  if the changes are too much , it doesn't matter. I'll clone the repository  of 3.12.15 version and recompile all.

THANKS A LOT Bin and Titus

I will let you know as soon as possible if it solves the problem

Best regards

Marco

Marco,

I intended to want you to clone the repo, but I found the tree does not have all the MUSB driver fixes I am looking for.

So before clone the tree, please try the 2nd patch mention in the wiki http://processors.wiki.ti.com/index.php/Sitara_Linux_SDK_MUSB_Issues.

Let me know if the patch does not fix your issue, I will try to give you a list of MUSB patches from kernel 3.16/3.17.

thanks a lot Bin! Your help is very valuable

Now i'll try to apply the patch and see the changes.

Again, thanks a lot .

Hi All, 

oops i accidentally edited  the file with + at 333 line in copy paste operation.

I but I've already tried without that error. The results are as reported above

marco

Marco,

Have you tried to disable CPPI DMA to isolate if the issue is in the CPPI driver?

Please give me a couple days, I will collect all the latest MUSB/CPPI related patches from the linux-usb mailing list for you.

Hi Bin Liu

Yes i've tried it. I've disabled the support for cppi41 DMA from menuconfig.

Unfortunatley the driver musb_am335x doesn't recognize the usb connection.

Regards

Marco 

Hi bin,

I got this results: on the sheel board

root@localhost:~ > zcat /proc/config.gz | grep CPPI
CONFIG_USB_TI_CPPI41_DMA=y
# CONFIG_TI_CPPI41 is not set

sorry , i suppose it should be

# CONFIG_USB_TI_CPPI41_DMA is not set
# CONFIG_TI_CPPI41 is not set

I'll correct it and try again.

Thanks a lot bin

Marco

Hi Bin

i'm sorry, but I've no success disabling the cppi_DMA in that way :(

I've a doubt. This could be a trivial question but... i've to recompile some modules too, in order the changes take place?

have you got some news about the source files?

best regards 

Marco

Marco,

Marco Destro said:

i'm sorry, but I've no success disabling the cppi_DMA in that way :(

If you still want to go this path, please provide the details of how you did the change, and what dmesg error you got, I will take a look. Disabling CPPI is very trivial, I guess you just missed a simple step.

Marco Destro said:

I've a doubt. This could be a trivial question but... i've to recompile some modules too, in order the changes take place?

After you disabled the CPPI options in .config or menuconfig, you have to rebuild kernel and modules and install them into the filesystem, "make zImage", and "make modules".

Marco Destro said:

have you got some news about the source files?

I am still working on it.

Hi Bin Liu

Problem solved :) removing the DMA support of CPPI41 works.  I've forgot to install the modules properly.

I repeat. it makes problem only in host mode. In device mode it works fine even with the dma enabled.

Thanks a lot for the forum support.

Best regards

Marco    

Marco,

Thanks for the update.

If the throughput with DMA disabled does not meet your requirement, here is the list of important patches I collected from mainline kernel releases for SDK 7.0 kernel.

http://processors.wiki.ti.com/index.php/Sitara_Linux_SDK_MUSB_Issues