Simple PC <> Launchpad USB link fails

I've determined that the controller dies in this library function (usbringbuf.c)
because the value of ulSize is 0. Maybe there should be an ASSERT to verify
this on entry?

I followed this to USBRingBufRead, but I have to leave for the university now.
Will investigate further this evening. I was quite sure I called all the right
_inits though.

//
// Correct for wrap. We use a loop here since we don't want to use a
// modulus operation with interrupts off but we don't want to fail in
// case ulDelta is greater than ulSize (which is extremely unlikely but...)
//
while(*pulVal >= ulSize)
{
*pulVal -= ulSize;
}

These are the first few packets on the bus (taken from the usbmon system):

ffff880094176b40 1023713490 S Bi:6:046:1 -115 64 <
ffff880094176b40 1023713612 C Bi:6:046:1 0 64 = 00010203 04050607 08090a0b 0c0d0e0f 10111213 14151617 18191a1b 1c1d1e1f
ffff880094176b40 1023713754 S Bi:6:046:1 -115 64 <
ffff880094176b40 1024914144 C Bi:6:046:1 -2 0
ffff88020c840e40 1024914158 S Co:6:002:0 s 23 08 92e1 0001 0000 0
ffff88020d639300 1024914229 S Bi:6:046:1 -115 64 <
ffff88020c840e40 1024914442 C Co:6:002:0 0 0
ffff88020d639300 1026114542 C Bi:6:046:1 -2 0
ffff88010684d180 1026114573 S Co:6:002:0 s 23 08 92e1 0001 0000 0
ffff88020dbb5780 1026114669 S Bi:6:046:1 -115 64 <

It's a slow process. Usually, I quite like Code Composer, but sometimes it does get a little frustrating. For some reason, I could set a breakpoint at the function I mentioned above, but CCS refuses to set a breakpoint at USBRingBufRead, which is in the same source file, getting the message:

No code is associated with "C:\Driverlib/build/Driverlib.test/usblib/usbringbuf.c" ...

Note that:

- The slashes are as typed above: C:\ and then continuing with the unix-like '/'
- I am on Linux, so I've no idea where the C:\ keeps coming from. All path are defined correctly, else the code wouldn't compile.
- When I wanted to trace the ulSize problem the first time, a dialog asked me to give the path to the usbringbuf.c file, and that worked. It seems the association then got lost.
- The breakpoint gets set, but is disabled immediately with the above message.

I can't find if I can edit this path somewhere. The project 'Properties' doesn't seem to show it.

Hi Amit.

It seems that many of my problems were due to a hardware problem. In those cases the PC-side program reported the lauchpad to be unavailable (not enumerating). Though I thought there were problems in the software, in fact it was a port of the hub that stopped working. Today, I found a program which can selectively reset the hub's ports. Each time I hit a problem, a reset solved the problem. For anyone interested, this post describes a solution. It's a small C program which send a reset to the file corresponding to the hub's port (Linux only!).

I searched again through the USB library doc, and think that DMA is not available for the Bulk Device driver. I only found a reference to DMA use when in the Host mode, and then only for Audio. (It's also mentioned in Mass Storage, but with a warning about silicon errata) Is that correct (no DMA for USB Device Bulk mode using the USB library)?

John

>But isn't it interesting to know why the port on the hub stops working in the first place. Is this related to the PC in use or any other PC also has the same issue.

The problem is on the launchpad side, but probably due to errors in my code. The fact that the hub hangs makes debugging complicated. Before the reset program I was (un)plugging the launchpad (or later, the hub), but the frequent connections at times caused problems in the computer.

> The errata on TM4C123x for the uDMA is only when using multiple channel maps.

The errata was for Mass Storage, so that actually doesn't touch me. But I can't find documentation on how to enable uDMA on Bulk Transfers. From the Peripheral Library manual:

"When configuring the endpoint, there are additional DMA settings required"

How do I do that with the USB Library functions?

Can I use the low-level functions to connect uDMA to the USB controller (which is programmed with the high level library)? In the USB Library manual, there's a statement that Low level access shouldn't be used simultaneously.

BTW, I am using the launchpad to test a design for a low-cost data collection terminal for a research project, which possibly would be shared by several universities.