[quote user="Dwarakesh R"] Probably disabling composite and trying to enable only g_ether, would be a worth try. [/quote] I believe that I only have g_ether enabled in this kernel the "composite" driver is how the EVM SDK works: which is my only "successful" frame of reference [quote user="Dwarakesh R"] Can you check if the idVendor in drivers/usb/gadget/composite.c file is proper ? Looks like it is coming from there. [/quote] Thanks! I will try and insert some logging here to work out why the VID is empty; the macros for its definition appear to be included (here and in ether.c) Having connected the board to a PC live-booting in to Ubuntu I can see the following the in the logs of the host: [ 781.025966] usb 1-10: new high-speed USB device number 11 using xhci_hcd [ 781.138015] usb 1-10: Device not responding to setup address. [ 781.342081] usb 1-10: Device not responding to setup address. [ 781.545867] usb 1-10: device not accepting address 11, error -71 [ 781.657858] usb 1-10: new high-speed USB device number 12 using xhci_hcd [ 781.770190] usb 1-10: Device not responding to setup address. [ 781.974096] usb 1-10: Device not responding to setup address. [ 782.177938] usb 1-10: device not accepting address 12, error -71 [ 782.289955] usb 1-10: new high-speed USB device number 13 using xhci_hcd [ 782.290121] usb 1-10: Device not responding to setup address. [ 782.494072] usb 1-10: Device not responding to setup address. [ 782.697844] usb 1-10: device not accepting address 13, error -71 [ 782.810050] usb 1-10: new high-speed USB device number 14 using xhci_hcd [ 782.810210] usb 1-10: Device not responding to setup address. [ 783.014039] usb 1-10: Device not responding to setup address. [ 783.217926] usb 1-10: device not accepting address 14, error -71 [ 783.217999] usb usb1-port10: unable to enumerate USB device Which at least gives an error code though it looks a bit generic "#define EPROTO 71 /* Protocol error */" there is no entry reported if I "lsusb" after insertion so the fact of the VID reported is hidden Time to keep digging! All the best, Richard
↧
Forum Post: RE: Linux/AM3352: Problems enabling USB Ethernet
↧
Forum Post: RE: Linux/PROCESSOR-SDK-AM335X: apt-get update failed
When i build the uboot ,i excute the apt-get update on my pc and then the error occurred . I did that because when i exucute : make CROSS_COMPILE=arm-linux-gnueabihf- O=am335x_evm am335x_evm_config all error: ../scripts/dtc-version.sh: line 17: dtc: command not found ../scripts/dtc-version.sh: line 18: dtc: command not found *** Your dtc is too old, please upgrade to dtc 1.4 or newer ../Makefile:1380: recipe for target 'checkdtc' failed It seems that i didn't install the device tree,so i excute apt-get install device-tree-compiler and it said that couldn't find the device tree and then i excute the apt-get update.
↧
↧
Forum Post: RE: Linux/AM3352: Problems enabling USB Ethernet
From what I can presently tell within ether.c static struct usb_device_descriptor device_desc; is properly populated during the eth_bind function. but the values that are entered in the struct usb_composite_dev *cdev variable passed to the eth_bind function look bogus. These are obtained from " coverwrite " declared via a MACRO ( USB_GADGET_COMPOSITE_OPTIONS() ;) I have not yet been able to ascertain where the values being used are intialised. From my reading of the the comments it is as though these are expected to be passed at runtime when loading the driver? In which case there should be an equivalent, or defaults, to be used in the case of a statically included driver? The start up of g_ether with some debugging to compare the contents of device_desc and cdev [ 1.580980] musb-hdrc: ConfigData=0xde (UTMI-8, dyn FIFOs, bulk combine, bulk split, HB-ISO Rx, HB-ISO Tx, SoftConn) [ 1.591218] musb-hdrc: MHDRC RTL version 2.0 [ 1.595422] musb-hdrc: setup fifo_mode 4 [ 1.599239] musb-hdrc: 28/31 max ep, 16384/16384 memory [ 1.613219] udc musb-hdrc.0.auto: registering UDC driver [g_ether] [ 1.619397] using random self ethernet address [ 1.623695] using random host ethernet address [ 1.628879] usb0: HOST MAC 2a:02:8f:d5:78:6e [ 1.633135] usb0: MAC a2:70:84:34:d0:8f [ 1.636844] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 375 ECM [ 1.642961] using random self ethernet address [ 1.647254] using random host ethernet address [ 1.651571] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 384 fi_rndis = cf408400 [ 1.659073] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 397 VID 0525 PID a4a2 [ 1.666296] g_ether gadget: adding config #2 'RNDIS'/c0940e9c [ 1.671937] rndis_register: configNr = 0 [ 1.675725] g_ether gadget: adding 'rndis'/cf482b40 to config 'RNDIS'/c0940e9c [ 1.682771] rndis_set_param_medium: 0 0 [ 1.686476] g_ether gadget: RNDIS: dual speed IN/ep1in OUT/ep1out NOTIFY/ep2in [ 1.693491] g_ether gadget: cfg 2/c0940e9c speeds: high full [ 1.698971] g_ether gadget: interface 0 = rndis/cf482b40 [ 1.704263] g_ether gadget: interface 1 = rndis/cf482b40 [ 1.709578] g_ether gadget: adding config #1 'CDC Ethernet (ECM)'/c0940df0 [ 1.716216] g_ether gadget: adding 'cdc_ethernet'/cf482c00 to config 'CDC Ethernet (ECM)'/c0940df0 [ 1.724898] g_ether gadget: CDC Ethernet: dual speed IN/ep1in OUT/ep1out NOTIFY/ep2in [ 1.732471] g_ether gadget: cfg 1/c0940df0 speeds: high full [ 1.737930] g_ether gadget: interface 0 = cdc_ethernet/cf482c00 [ 1.743823] g_ether gadget: interface 1 = cdc_ethernet/cf482c00 [ 1.749720] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 430 cf49aa80 CDEV VID 0000 PID 0000 0 0 [ 1.758446] RM - drivers/usb/gadget/composite.c:usb_composite_overwrite_options 2293: CDEV cf49aa80 COVR c0991804 [ 1.768442] RM - drivers/usb/gadget/composite.c:usb_composite_overwrite_options 2297: CDEV->vendor 0 [ 1.777265] RM - drivers/usb/gadget/composite.c:usb_composite_overwrite_options 2301: CDEV->product 0 [ 1.786299] RM - drivers/usb/gadget/composite.c:usb_composite_overwrite_options 2305: desc->bcdDevice 0 [ 1.795393] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 437 cf49aa80 CDEV VID 0000 PID 0000 1 0 [ 1.804133] g_ether gadget: Ethernet Gadget, version: Memorial Day 2008 [ 1.810530] RM - drivers/usb/gadget/legacy/ether.c:eth_bind 447 VID 0525 PID a4a2 1 2 [ 1.818082] g_ether gadget: g_ether ready [ 1.821972] g_ether musb-hdrc.0.auto: usb_gadget_udc_start Thanks for reading - I'll continue digging Richard
↧
Forum Post: RE: Linux/AM3352: Problems enabling USB Ethernet
good going! Yes these could be provided at the time of insertion of the module or from user space through some proc/sysfs entry, may be. Can you try hardcoding these values in usb_composite_overwrite_options to check if it is making progress ?
↧
Forum Post: RE: Linux: Recommendations for starter board for embedded linux
Hi Nathan, Here are links to both: www.ti.com/.../beaglebk www.ti.com/.../tmdssk3358 Both are good for embedded linux learning, but only Beaglebone Black match your price requirements. Regards, Pavel
↧
↧
Forum Post: RE: Linux/EVMK2EX: Booting Linux kernel using U-boot on K2E device
Hello, Have you checked this guide: processors.wiki.ti.com/.../Linux_Core_U-Boot_User's_Guide You also could use our Keystone forum: e2e.ti.com/.../639 Hope this helps. BR Margarita
↧
Forum Post: RE: Linux: DRA725，J6 ECO usb problem
Hi Pavel, From the SCH,have you find some hardware problem?Could you help check our DRA725 SCH?
↧
Forum Post: RE: Linux/AM5728: The processing of Input from USB keyboard on Linux
Dear Clemens-san, Thank you for your quick reply. Customer's new keyboard has special keys that does not have in general USB keyboard. Special keys are assigned to "Usage ID". "Usage ID" is not used in general USB keyboard. (Please refer to " www.usb.org/.../Hut1_12v2.pdf " for Usage ID.) Also customer uses USB HID protocol. BTW, you said "That description ignores all the processing done by the Windows driver.". What do you mean about this? Does this mean Linux USB HID driver can use customer's special keyboard without problem? Please let me know. I appreciate your quick reply. Best regards, Michi
↧
Forum Post: RE: Linux/AM3352: Problems enabling USB Ethernet
Thanks yes I tried that - well immediately after returning from the overwrite function I inserted cdev->desc.idVendor = device_desc.idVendor; cdev->desc.idProduct = device_desc.idProduct; cdev->desc.bcdDevice = device_desc.bcdDevice; as I had confirmed that the values in device_desc look believable This has made no difference though - the info reported by the HOST is still nonsense Device ID : USB\VID_0000&PID_0002 I can confirm that the HW itself is fine as I get a serial port if I try an boot with no code Device ID : USB\VID_0451&PID_6141 Looks like I need to crawl lower in to the USB handling to see what's lurking there Thanks again, Richard
↧
↧
Forum Post: RE: Linux: DRA725，J6 ECO usb problem
Judd, This is linux OS forum, and your question is related to HW schematics. What I mean is this is not the correct forum for your question. And I see you already double posted in the correct forum: e2e.ti.com/.../601691 The USB expert has been notified regarding your question in DRAx forum and will reply there. Regards, Pavel
↧
Forum Post: RE: Linux/AM3354: Linux DMAengine/EDMA Question
Alexandru, EDMA is the right direction to reduce GPMC latency. We don't quote latency number for interrupts due to the complexity of the Cortex-A8 and the many variations of sofware possible. Latencies can be very different from Linux to RTOS or bare metal solutions depending on many factors such as if the MMU is used and how it is used. Other major factors include any critical sections in the customers code base. Also depending on your latency requirements, there are methods available to limit latency. One example of this is to lock down the L2 cache with any code you may have for which you require low latency. Note that there are other options (instead Linux PSDK) that might reduce latency, options like StarterWare, Linux-RT, RTOS. www.ti.com/.../processor-sdk-am335x See also if the below e2e threads will be in help: e2e.ti.com/.../1056619 e2e.ti.com/.../36625 e2e.ti.com/.../143795 e2e.ti.com/.../356576 e2e.ti.com/.../573232 e2e.ti.com/.../477311 Regards, Pavel
↧
Forum Post: RE: Linux/PROCESSOR-SDK-AM335X: apt-get update failed
Nancy, Could you check your Device tree compiler version using the command "dtc -v". If it is too old try to install newer version manually downloading the tool from: launchpad.net/.../device-tree-compiler or check this: installion.co.uk/.../index.html Also check whether you are using the right tool chain as it is described in the U-boot user's guide: processors.wiki.ti.com/.../Linux_Core_U-Boot_User's_Guide BR Tsvetolin Shulev
↧
Forum Post: RE: Linux/AM5728: The processing of Input from USB keyboard on Linux
In Windows, the kbdhid.sys driver handles the HID protocol, and kbdclass.sys then handles all keyboard events and generates the messages to be sent to applications. If Windows already sends messages for the custom keys, then it's likely that Linux does the same without a separate driver. Connect the keyboard to some Linux PC. Check that it shows up in /dev/input/by-id/. Run the evtest tool on that input device to check whether you get events for the custom keys.
↧
↧
Forum Post: RE: Linux/DRA745: Ti-processor-sdk-linux-automotive-dra7xx-evm-03_02_00_03, issue during compilation on ubuntu 14.04
Hi Gaurav, Can you test this manually with git clone? Below is what I have on my side: $ git clone git://git.ti.com/glsdk/release-manifest.git Cloning into 'release-manifest'... remote: Counting objects: 547, done. remote: Compressing objects: 100% (541/541), done. remote: Total 547 (delta 356), reused 0 (delta 0) Receiving objects: 100% (547/547), 71.47 KiB | 0 bytes/s, done. Resolving deltas: 100% (356/356), done. Checking connectivity... done. See also below e2e thread: e2e.ti.com/.../597501 Regards, Pavel
↧
Forum Post: RE: Linux/AM4376: Software development on Linux host system for custom board
Hi Pavel, I downloaded the SDK "PROCESSOR-SDK-LINUX-AM4376 03_03_00_04" and found the header files in the path you mentioned. I also referred the TI threads you shared, but these are for TI AM335X processors used on beagle bone black board. It uses the ADAfruit libraries for this processor which consist hardware description for AM335x processor ( github.com/.../Adafruit_GPIO ) If python libraries are available for AM4376 , then I can merge with the rest of the codes.. Regards, Dhara
↧
Forum Post: RE: Linux/AM5728: Migrating C code from BBB to AM572x
Hi Sajeevan, Which AM335x PSDK you are using for BBB UART programming? On which UART instance (UART0 to UART5) you are working on AM335x BBB? Do you use also IRQ or/and DMA in AM335x BBB UART program? Which AM572x PSDK you are using for AM572x EVM UART programming? I see you are using first UART here (UART1), from 10 UARTs available (UART1 to UART10). On AM572x EVM, when linux boot up, check if you made the correct pinmux, make that check in user space with omapconf tool or devmem2 tool. Regards, Pavel
↧
Forum Post: RE: Linux/am3352: mcasp aplay problem
Hello Pavel: I have dump the register of the new kernel ,it is almost the same . i print the value at the tail of static void mcasp_start_tx(struct davinci_mcasp *mcasp) [ 407.155260] regs_offset: ac,val:101 [ 407.158775] regs_offset: 6c,val:11 [ 407.162199] regs_offset: a8,val:18074 [ 407.165886] regs_offset: 68,val:10070 [ 407.169571] regs_offset: b0,val:80 [ 407.172995] regs_offset: 70,val:180080 [ 407.176767] regs_offset: b4,val:8000 [ 407.180364] regs_offset: 74,val:8000 [ 407.183962] regs_offset: 14,val:8000001 [ 407.187821] regs_offset: 64,val:ffff [ 407.191417] regs_offset: a4,val:ffff [ 407.195014] regs_offset: 78,val:0 [ 407.198351] regs_offset: b8,val:3 The xstat register always 0x10c,it report Transmit clock failure. But it is also 0x10c in the old kernel. It is strange that why I must add mcasp_set_bits(dev->base + DAVINCI_MCASP_PDIR_REG, AHCLKX); in the davinci_mcasp_probe(). If I do not add this line ,the register of PDIR also show val 8000001 in the old kernel,there is no sound .
↧
↧
Forum Post: RE: Linux/LINUXSDK-OMAPL138: OMAPL138 with wifi usb adapter
Hi, You should be able to cross-compile the driver against the 3.3 kernel running on your platform. Then you should be able to insmod the driver. Best Regards, Yordan
↧
Forum Post: RE: Linux/AM5728: TIOCL Fatal error seen while using OpenCV based application on PSDK v3.03.00.04 (Linux)
Hi, [quote]The application runs well on PSDK 3.02.00.05 (Linux). The application does not make use of DSP acceleration[/quote] In this case you should move it from the reserved cmem region: reserved-memory { #address-cells = ; #size-cells = ; ranges; cmem_block_mem_0: cmem_block_mem@a0000000 { reg = ; no-map; status = "okay"; }; cmem_block_mem_1_ocmc3: cmem_block_mem@40500000 { reg = ; no-map; status = "okay"; }; }; As you can see from your log the OpenCV app runs from within the reserved region 0xa0000000. Best Regards, Yordan
↧
Forum Post: RE: Linux/PROCESSOR-SDK-DRA7X: Build issue : path is getting pre appended twice
How do you setup your environment? Best Regards, Yordan
↧