100% CPU load with current master about mumudvb HOT 9 CLOSED

I pulled the commit and verified that it works.
Many thanks to both of you for the fast debugging and fix!

from mumudvb.

Looks like this is maybe caused by the udp unicast/multicast input in commit f177c62
I'm looking into it

from mumudvb.

Many thanks!

Perhaps it also helps to see one of my config files:

freq=12110
pol=h
srate=27500
autoconfiguration=full
unicast=1
ip_http=127.0.0.1
store_eit=1
port_http=9099
card=0
multicast_ipv4=0
multicast_ipv6=0

I don't use multicast and set it to disabled, although I don't know whether that plays any role here.

from mumudvb.

Yeah, this is not related to multicast output, but input (instead of using a DVB card). So I think the issue should manifest with any config. Could you pull 3fe7a9c (the commit right before udp input) and see if that works (no 100% CPU), that's what I'm seeing during my testing, anyway.

from mumudvb.

I will try to find the time today, but cannot promise (meetings all day).

from mumudvb.

No worries, I think I have a general idea of what's causing it, and my TT-S3600 still sorta works enough for testing

from mumudvb.

@trollcop Thanks, this commit seems to work. I put the new binary on one of the tuners and there are no cpu spikes. I'll monitor the next slot to see if it also works when recording (the tuner is idle ATM).

from mumudvb.

I found the problem, its a simple != vs == compare deciding to poll for dvb data which was introduced when adding udp input.
I fixed it in my fork with commit 8e23eff, I'll try to figure out how to PR just that and submit to here.

from mumudvb.

from mumudvb.