On Tue, Jul 15, 2025 at 03:41:16PM -0000, Michael van Elst wrote: > >I understand kprintf operates at IPL_HIGH, while the USB subssytem > >processes I/O at IPL_SOFTCLOCK. Does that mean I just cannot rely > >on USB for console output? > > It can only work if USB is used in polling mode, which usually also means > that you must not use that USB controller for anything else.
I had great success with a kernel thread that takes care of the actual USB output, while being fed by cn_tab->cn_putc() through a ring buffer. That works well for displaying kernel messages. cn_tab->cn_getc() is another beast. It is called at splhigh and wants a char immediatly. I have red the polling code used by ukbd.c and uhidev.c I understand it needs to use usbd_open_pipe_intr() and that this uses an interrupt endpoint, not a bulk one. The USB to serial adapter has no interrupt endpoint, hence I understand I cannot use the polling mode. You suggest it could be set controller-wide, which is admissible when cn_tab->cn_getc() is used: root device prompt on boot -a, and in DDB. But how heavily should the current code be refactored to do that? -- Emmanuel Dreyfus m...@netbsd.org
