On 11/12/19 11:47 AM, Joakim Zhang wrote: >> I think we can explain case 2), but I think case 1) is different. Can you repeat >> testcase 1) several times and send the output? > > I do not quite understand. Bus bitrate 1Mbps, so the counter overrun every 65ms: > Case 1) default frame gap is 200ms, case 2) configured frame gap > 100ms, why we can explain case 2)? What is the essence of these two? Oh, I missed the fact that 1) uses the default gap of 200ms. The problem is wakeup latency from frame reception to IRQ handler is much bigger than the counter overflow. Then we can explain cases 1) and 2). The question remains, has the wakeup time increased compared to older kernels? Or was it pure luck that you have not seen this problem in older kernels? 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: OpenPGP digital signature
