Am 2005-01-20 21:32 +0200 schrieb Olav Kongas: > I have never used the chip with edge-triggered interrupts, > while it is working in two different designs here as level > triggered. If your processor supports level-triggered > interrupts, you should perhaps try using them.
Nice to know, ok. I try to use level triggered IRQs, which works like a charm with several devices, the ISP is the first one driving me mad or going into debugging IRQ stuff for the 1st time. > Dump all the processor's interrupt/gpio registers related to > the pin which must get interrupts from isp to make sure the > pin is indeed correctly configured. Assert/deassert the I am pretty sure the ISP IRQs reach the Kernel. > interrupt pin manually to see, whether it really changes the > relevant status bit in the raw interrupt register. > Achtung!!! Doing this may require disconnecting the > interrupt pin from ISP to avoid damage to the latter; > consult with your hardware guys. I _am_ the hardware guy, I designed the whole board including CPU, peripherals and ISP. Actually manually asserted IRQ lines get recocnized by the Kernel also. No Problem. The Problem is that I get interrupts, which the ISP driver seems to handle, but the kernel thinks not. When I plug in a USB device enumeration starts and a dozen of "irq 65: nobody cared" messages including stackdump are printed. After Enumeration I plug out, same thing. Second cycle, same or other device, same dozen of nobody cared prints. Third also. The fourth time I plug a device in, enumeration goes with _none_ nobody cared message. Plug out also. Device is funtional, works. No nobody cared message ever again. -> ? When I got the chip edge triggering this day, the nobody cared messages went away, but I made the chip loose its ability to generate edge triggered IRQs then, I don't know why, hence my initial question. > Then modprobe the isp116x-hcd and see whether a manual > assertion of the pin will increase the interrupt count in > /proc/interrupts. Dump again the contents of interrupt/gpio The /proc/interrupts looks very well! It increases and seems to be consistend, not to much or to less in count. If nobody cared messages are printed or not. > registers to see, whether they are still in the right state. May be a good idea to look into the registers to see whats happening there... I tried everything until now :( Konsti -- GPG KeyID EF62FCEF Fingerprint: 13C9 B16B 9844 EC15 CC2E A080 1E69 3FDA EF62 FCEF ------------------------------------------------------- This SF.Net email is sponsored by: IntelliVIEW -- Interactive Reporting Tool for open source databases. Create drag-&-drop reports. Save time by over 75%! Publish reports on the web. Export to DOC, XLS, RTF, etc. Download a FREE copy at http://www.intelliview.com/go/osdn_nl _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel
