Confused about rtnl routing mutex scalability

Confused about rtnl routing mutex scalability

[nick]

Greetings All,
I am curious as it after reading the network code for a while it seems that to me
at least the rtnl mutex lock is a bottleneck. Is there any reason that we are still
using a global mutex for all routing changes in the networking stack for the Linux
kernel or does it actually scale only, seems to me that it is causing nothing but
problems with deadlocks related to it or issues with locking timeouts on separate
netdev routing tables for two different networks. Is this just me or does this lock
actually cause serious issues or can do so.
Regards,
Nick

Confused about rtnl routing mutex scalability

[Valdis.Kletnieks at vt.edu]

On Wed, 07 Sep 2016 23:28:54 -0400, nick said:
> I am curious as it after reading the network code for a while it seems that to me
> at least the rtnl mutex lock is a bottleneck.

Given your proven inability to read and understand kernel code, nobody's going to
bother investigating unless you demonstrate other evidence that lock is a bottleneck,
such as actual benchmarks showing contention.

>  Is this just me or does this lock actually cause serious issues or can do so.

It's probably just you.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 830 bytes
Desc: not available
Url : http://lists.kernelnewbies.org/pipermail/kernelnewbies/attachments/20160910/f2f727a3/attachment.bin