Run pitonstream TIMEOUT

[Cong Tran Van]

Hi everyone, 

I run pitonstream to test my IP on FGPA (KC705) with baremetal by code C, but it prints "TIMEOUT" on baremetal
How to increase timeout or disable timeout? 
Which file do I need to edit at repo https://github.com/tianrui-wei/openpiton/tree/kc705 ?

Thanks,
Cong

[Jonathan Balkind]

Hi Cong,

Did you set the DIP switch to the right direction such that if you open the UART in e.g. putty/screen/tio, it prints "DONE" every time the board is resetted? It could be that the board itself isn't in pitonstream mode.

Thanks,

Jon

--
You received this message because you are subscribed to the Google Groups "OpenPiton Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpiton+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openpiton/5db01bc1-ccb0-48e3-aca0-5fed1bab817cn%40googlegroups.com.

[Andreas Wrisley]

Hi,

I have a similar problem, but I can run my application with a few iterations (so the DIP switch is correctly positioned). When I increase the number of iterations and as a result print a lot more data to the serial port I always get TIMEOUT. How can I change this?

Regards,

/Andreas

[Jonathan Balkind]

If the issue is specifically that you get a pass for something short and a timeout for something longer, I believe the following macro might help: https://github.com/PrincetonUniversity/openpiton/blob/openpiton/piton/design/chipset/include/chipset_define.vh#L92

Although it's set to 5B cycles by default so I don't know for sure if that's going to be your issue. It's also possible that your program gets stuck in some other way which causes the timeout. Give it a shot and see.

Also just to note because I know Google Groups doesn't communicate this to users: Due to spam, I have to manually approve all messages. You don't need to re-send.

Thanks,

Jon

To view this discussion, visit https://groups.google.com/d/msgid/openpiton/53cb5e0a-ddcd-4fce-ab46-a0e40db022b4n%40googlegroups.com.

[Andreas Wrisley]

Hi,

Thanks for the link to chipset_define.vh file, changing that macro did help. My logfile capturing the uart (by piping the output from pitonstream) was almost 60 MB with about 160,000 samples. I will look at my application as well to try to minimize the output. At the moment it captures all the samples and then prints them to the uart.

This took forever to capture, is there another way to transfer data to/from the memory? I have tried getting OpenOCD working without success, the manufacturer field during identification is invalid.

Sorry about resending the message. I have seen the message saying approval is needed, but didn't see one this time so I wrote another thinking I did cancel instead of send. Is it better to use Zulip? What is the purpose of each communication platform?

Regards,

Andreas

[Jonathan Balkind]

Hi Andreas,

Yeah, UART sucks for doing any meaningful amount of data movement. My recollection is that when Michael got the OpenOCD+GDB stuff working, it was using a patched version of openocd because it didn't have full RISC-V support yet (or something along those lines). I could see two possibilities: one is that OpenOCD has different expectations now than how the system was setup in the first place and the merged/later edited code doesn't align - this could perhaps be fixed by using an old patched openocd? Not sure. The other is that the ariane openocd config file changed at some point. It could be worth digging around a little in the repo for older/newer versions of the config to see if they could help? Would probably be easier than the former idea.

Regarding the group vs zulip, I don't really have a system. It's just often easier to have back and forth by messaging rather than by email. Either way, things get documented. The other thing I offer on zulip is for anything sensitive/non-public we can have a private channel or discuss by DM.

Thanks,

Jon

To view this discussion, visit https://groups.google.com/d/msgid/openpiton/78777bf7-4fc1-4029-9584-3f44084a4f00n%40googlegroups.com.