[RFC] removal of server generation mechanism

Keith Packard keithp at keithp.com
Sun Oct 25 14:33:57 PDT 2009


Excerpts from Tiago Vignatti's message of Sun Oct 25 14:14:28 -0700 2009:

> this is ... ugly! Do the server (or any kind of software) really needs 
> this kind of guarantee?

Well, yes -- you want to know that the server can be reset and return
to the initial configuration with no lingering effects from the
previous session.

> can be done. So, seems you've being favorable with the removal as well, 
> but want to keep this only due protocol spec. Do I get it right?

Yes, there's no reason to involve the DDX in the reset process; we
just need to make sure the server returns to the initial state at
reset time as seen through the protocol.

Reset was extended to involve the driver back in the R3(4?) time frame
so that memory leaks and other such things could be more easily
detected -- resetting everything left the number of objects allocated
very small at one point in the process, and we could hand-audit the
remaining objects to ensure that none of them were leaks.

-- 
keith.packard at intel.com
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 190 bytes
Desc: not available
Url : http://lists.x.org/archives/xorg-devel/attachments/20091025/4c5869bf/attachment.pgp 


More information about the xorg-devel mailing list