Github user PSUdaemon commented on the pull request:
https://github.com/apache/trafficserver/pull/502#issuecomment-202990616
This kinda feels like it should be three separate PR's.
f304c50a2c00d7d711015c981b2007dacdf159b7 - seems probably ok, simply
because if it's wrong it won't affect many people. Ideally I think CK++ should
supersede this though.
9d13995ce3a5e26ef80e54b629daba538c0ec87f - I think I see what you are doing
here, but I don't know enough about the cache internals to know if this is OK.
@SolidWallOfCode thought it seemed ok, but was concerned about HostDB
interaction. Also seems like maybe we can say this should be rounded up to a
multiple of page size rather than this special case treatment.
96656663b07f6da8204f3edbb392bf3d55b05398 - Can you explain the need for
this? Seems like we should not write our own spinlock implementation. Perhaps
we can borrow one of the 8 implementations libck has. And this also seems like
a one off case that can be made more generic. I don't really like the #ifdef
around everything. I think it would be better to have a mutex macro that we can
override with whatever method we want. For example, cohort locks might be
better for us.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
