On 23.12.2022 08:18:17, Stefan Althöfer wrote: > > Why do you have so many errors on can1? The device event went into bus-off. > > The device was not freshly started, and I just intended to share the > setup. No errors when the problem appeared this night. Ok > > Can you share your "scselftest" program? > > Sure, attached as extracted from the build system for my test tools. > Hope this is not the bad guy. Included my ancient copy of > libsocketcan. > > Do you get the same error if the test is only running on a single > > mcp2518fd? > > I'm pretty sure I had cases where I started the test on both, one > failed sooner and one failed later. However, currently the second test > is still running (after 9h). I want to know if there's an influence between the 2 interfaces. > > Do you get the same error if only a single mcp2518fd is attached to the raspi? > > Only tested this with the CM4 Would be interesting if you only attache 1 mcp2518fd to the board and then re-run the test. > > Can you share your kernel, including the modifications? > > Yes if it turns out to be necessary, but I just cloned it from here: > git clone --depth=1 --branch <branch> > https://github.com/raspberrypi/linux You said you made some modifications to the kernel, also it would be good to use the _extact_ version you using to reproduce the error. > Following these instructions: https://www.raspberrypi.com/documentation/computers/linux_kernel.html#kernel-headers > Now this nights error log: Thanks for the text, much easier to handle. :) > Loop 4429 2022-12-23T03:07:45 last+4s total:4h299m > + ./sctestself -b -n 4 -l 999 -t 2 -v cmperr,logmsg -F refilldata,leastdots,allowintloopb,stoponerror -d 8 can1 > CAN selftest can1 .......idx 2 tx:2A5#00 02 6A 8F 6C 81 C3 B0 > rx:2A5#00 00 85 9A EF 54 34 5D > idx 3 tx:2A5#00 03 4F 8F 85 59 8E DA > rx:2A5#00 00 85 9A EF 54 34 5D > .c (700) fail > 0: RX (004/003) 2A5#00 00 85 9A EF 54 34 5D > 0: RX (004/002) 2A5#00 00 85 9A EF 54 34 5D 2 bad RX here > 0: TX (003/002) 2A5#00 03 4F 8F 85 59 8E DA > 0: RX (003/001) 2A5#00 01 7C 30 5C 7A 2D 1C > 0: TX (002/001) 2A5#00 02 6A 8F 6C 81 C3 B0 > 0: RX (002/000) 2A5#00 00 AC AF 70 9B 32 03 > 0: TX (001/000) 2A5#00 01 7C 30 5C 7A 2D 1C > 0: TX (000/000) 2A5#00 00 AC AF 70 9B 32 03 > 0: RX (004/003) 2A5#00 03 65 32 F8 E7 E6 F5 > 0: RX (004/002) 2A5#00 02 CC D5 09 FE 55 99 > 0: TX (003/002) 2A5#00 03 65 32 F8 E7 E6 F5 > 0: RX (003/001) 2A5#00 01 62 4F 17 34 44 CE > 0: TX (002/001) 2A5#00 02 CC D5 09 FE 55 99 > 0: RX (002/000) 2A5#00 00 FE 58 FE FD C0 EF > 0: TX (001/000) 2A5#00 01 62 4F 17 34 44 CE > 0: TX (000/000) 2A5#00 00 FE 58 FE FD C0 EF > 0: RX (004/003) 2A5#00 03 C9 7B EF 83 93 E0 > 0: RX (004/002) 2A5#00 02 DD 5E 82 B4 FD C4 > 0: TX (003/002) 2A5#00 03 C9 7B EF 83 93 E0 > 0: RX (003/001) 2A5#00 01 90 3B 32 80 90 67 > 0: TX (002/001) 2A5#00 02 DD 5E 82 B4 FD C4 > 0: RX (002/000) 2A5#00 00 8F 28 D2 F5 B6 98 > 0: TX (001/000) 2A5#00 01 90 3B 32 80 90 67 > 0: TX (000/000) 2A5#00 00 8F 28 D2 F5 B6 98 > 0: RX (004/003) 2A5#00 03 C5 1E FE 68 F1 65 > 0: RX (004/002) 2A5#00 02 7E 2B BA 18 F1 7B > 0: TX (003/002) 2A5#00 03 C5 1E FE 68 F1 65 > 0: RX (003/001) 2A5#00 01 CE 1B D7 9E 42 15 > 0: TX (002/001) 2A5#00 02 7E 2B BA 18 F1 7B > 0: RX (002/000) 2A5#00 00 85 9A EF 54 34 5D properly received > 0: TX (001/000) 2A5#00 01 CE 1B D7 9E 42 15 > 0: TX (000/000) 2A5#00 00 85 9A EF 54 34 5D send > 0: RX (004/003) 2A5#00 03 37 6D 66 8E C6 9B > 0: RX (004/002) 2A5#00 02 2D 83 A3 D3 66 27 > 0: TX (003/002) 2A5#00 03 37 6D 66 8E C6 9B > 0: RX (003/001) 2A5#00 01 E8 A5 9A C6 C1 AC > 0: TX (002/001) 2A5#00 02 2D 83 A3 D3 66 27 > 0: RX (002/000) 2A5#00 00 E6 7A D0 27 3E DF > 0: TX (001/000) 2A5#00 01 E8 A5 9A C6 C1 AC > 0: TX (000/000) 2A5#00 00 E6 7A D0 27 3E DF > 0: RX (004/003) 2A5#00 03 D4 54 CF 9A 6D 7A > 0: RX (004/002) 2A5#00 02 06 3F 57 59 35 F6 > 0: TX (003/002) 2A5#00 03 D4 54 CF 9A 6D 7A > 0: RX (003/001) 2A5#00 01 C1 F6 A6 E3 84 64 > 0: TX (002/001) 2A5#00 02 06 3F 57 59 35 F6 > Stopped. Ctrl-C to exit. > > And the register dump [...] > mcp251xfd-dump did not work for me, but It looks as if the correct data is in the rx fifo. With v5.18 the RAM layout changed, the mcp251xfd-dump tool was not adopted. yet. Marc -- Pengutronix e.K. | Marc Kleine-Budde | Embedded Linux | https://www.pengutronix.de | Vertretung West/Dortmund | Phone: +49-231-2826-924 | Amtsgericht Hildesheim, HRA 2686 | Fax: +49-5121-206917-5555 |
Attachment:
signature.asc
Description: PGP signature
