1430050 - IPC to the socket process for DNS resolution

Status: RESOLVED FIXED

(Core :: Networking, enhancement, P2)

Description

[Shih-Chiang Chien [:schien] (UTC+8) (use ni? plz)]

We need initialize DNS resolver on socket process and provide corresponding IPC interface to both chrome process and content process.

Comment 1

[Honza Bambas (:mayhemer)]

I think this reduces only to IPC the implementation of nsIDNSService to proxy to the socket process.  We already run an instance of the dns service on the socket process (AFAIK).  This can also be seen as a mid priority work only and more as an optimization/cleanup.

We need to verify that running dns service and resolver on both the parent and socket processes is fine and not seriously breaking anything.

P3 based on the above.

Comment 2

[Dragana Damjanovic [:dragana]]

(In reply to Honza Bambas (:mayhemer) from comment #1)
> I think this reduces only to IPC the implementation of nsIDNSService to
> proxy to the socket process.  We already run an instance of the dns service
> on the socket process (AFAIK).  This can also be seen as a mid priority work
> only and more as an optimization/cleanup.
> 
> We need to verify that running dns service and resolver on both the parent
> and socket processes is fine and not seriously breaking anything.
> 
> P3 based on the above.

We do not want to have 2 resolvers, that would mean having 2 dns caches so prefetching does not have any effect.
I think we should finish this with all http work.

Comment 3

[Honza Bambas (:mayhemer)]

Yeah, good point.  However, I'm prioritizing by stability and only then by performance impact.  This is though one of the first cleanups for perf to do for me.  Let's keep it as P2 then to not loose track.

Comment 4

[Kershaw Chang [:kershaw]]

Attachment: Bug 1430050 - Do DNS resolution in socket process

In order to reuse the current code of DNSRequest, this patch mainly separates the logic of sending/processing DNS request from DNSRequest IPC actors (DNSRequestChild and DNSRequestParent).

Comment 5

[BugBot [:suhaib / :marco/ :calixte]]

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:kershaw, could you have a look please?

Comment 6

[Kershaw Chang [:kershaw]]

(In reply to Release mgmt bot [:sylvestre / :calixte] from comment #5)

There's a r+ patch which didn't land and no activity in this bug for 2 weeks.
:kershaw, could you have a look please?

Sorry that I forgot to close this bug.
This patch is only landed to larch for now.

https://hg.mozilla.org/projects/larch/rev/b1dabf31a7ff