Objective Development Forums

You can disable some functionality in hardware.h, just look at the first few defines.

The STK500v2 firmware version can be configured in stk500protocol.c, see defines for STK_VERSION_MAJOR and STK_VERSION_MINOR.

As far as I know, the green LED is not actually used by USBasp either. Since AVR-Doper does not have a status to display with this LED, we just leave it unconfigured.

Are you aware that you can't use the same hex file for ATMega8 and 48/88/168? You must specify the correct device in the compiler command line.

Thanks for the info about Timer 1 and the jumper! I have made a new release which fixes these issues. Please let me know how it works! Regarding RxD and TxD: These pins are swapped because they name the function on the target, not the programmer. RxD must be connected to TxD and vice versa. Since UA...

OK, you're right. My comment about passing zero length messages is wrong, this is explicitly handled in the driver.

However, a zero sized message may not be handled correctly by the host. This is actually a bug in the RemoteSensor example. You should send at least one dummy byte.

Sorry, I don't have a real USBasp board for testing, only something very similar. So I can't test all details.

Anybody out there who can comment or who has even found a bug?

usbSetInterrupt() requires a pointer as the first argument and a byte count (greater than zero!) as the second. Anything else may result in corruption of the internal data structures.

You can't use osccal.c anyway, because to code is too big for the ATTiny2313. The 12.8 MHz module is not suitable for the ATTiny2313.

Which size is this? The size of RAM plus FLASH, or the flash only? You get the flash size with avr-size main.hex With a minimum boot loader condition and BOOTLOADER_CAN_EXIT defined to 1 I get 2012 bytes for an ATMega8 with 12.8 MHz RC, including the following clock adjustment code in usbconfig.h: m...

String descriptors and the device descriptor (which contains the PID) are unique for the device, not a particular endpoint. They are always queried through endpoint 0. It won't help you to implement a second endpoint. Even a composite device has only one name. It's only one device (although composit...

I don't know. At least for bulk endpoints, it's done by the controller hardware. The polling is so fast that it would block the CPU otherwise. Interrupt endpoints might be implemented in the kernel driver's timer interrupt. This is still VERY efficient, compared to polling from user space.

As far as I know, this is done with avr-objcopy. Don't know the section name to extract, but I'm sure you can find it with a web search.

This must be a problem with the programming algorithm. I don't have a mega16 for testing, though. Has anybody else seen this problem?

The interrupt-in blocks until the device sends interrupt data. This is an advantage for the host side software because no CPU time is used for polling.

No. These devices require High Voltage Parallel Programming or ISP.