Bug 2192

Summary: RFE: Port the SolarisIA extension to the Xorg Xserver
Product: xorg Reporter: Roland Mainz <roland.mainz>
Component: * OtherAssignee: Alan Coopersmith <alan.coopersmith>
Status: RESOLVED INVALID QA Contact: Xorg Project Team <xorg-team>
Severity: enhancement    
Priority: high CC: alan.coopersmith, jeremyhu, stuart.kreitman
Version: unspecified   
Hardware: All   
OS: Solaris   
Whiteboard: 2011BRB_Reviewed
i915 platform: i915 features:
Attachments:
Description Flags
Server side of SolarisIA extension
none
xc/include/extensions/interactive.h (shared by client-side and server-side)
none
xc/include/extensions/interactiveCommon.h
none
xc/lib/X11/XInteractive.c (client side of extension protocol)
none
Patch to xc/lib/X11/OpenDis.c none

Description Roland Mainz 2005-01-02 10:38:35 UTC
RFE: Port the SolarisIA extension to the Xorg Xserver (=on Solaris as only this
OS has the neccesary kernel infratructure (IA scheduler class etc.)).
Right now this extension currently only exists in the Solaris Xsun server but
not for the stock Xorg server on Solaris, making interactive usage on multiuser
systems slightly more sluggish...
Comment 1 Alan Coopersmith 2005-01-02 10:58:47 UTC
I've already done the port for the Xorg server delivered in Solaris, just need
to work on getting it released back to the open source version.
Comment 2 Alan Coopersmith 2005-08-31 09:59:37 UTC
Created attachment 3135 [details]
Server side of SolarisIA extension
Comment 3 Alan Coopersmith 2005-08-31 10:04:03 UTC
Created attachment 3136 [details]
xc/include/extensions/interactive.h (shared by client-side and server-side)
Comment 4 Alan Coopersmith 2005-08-31 10:05:17 UTC
Created attachment 3137 [details]
xc/include/extensions/interactiveCommon.h
Comment 5 Alan Coopersmith 2005-08-31 10:09:12 UTC
Created attachment 3138 [details]
xc/lib/X11/XInteractive.c (client side of extension protocol)
Comment 6 Alan Coopersmith 2005-08-31 10:15:25 UTC
Created attachment 3139 [details] [review]
Patch to xc/lib/X11/OpenDis.c
Comment 7 Alan Coopersmith 2005-08-31 10:20:58 UTC
Okay - I think that's everything needed to use it in Xorg.   The original Usenix
paper describing the work that led to this is at:
http://www.usenix.org/publications/library/proceedings/cinci93/full_papers/evans.txt

A small bit of documentation on the IA priority control scheduler in Solaris
that this uses can be found at:
http://docs.sun.com/app/docs/doc/817-0689/6mgfkpd12?a=view#Inter-Active%20Class
and of course in the Solaris kernel sources at:
http://cvs.opensolaris.org/source/

Before we actually integrate it into Xorg though, I think we need to consider what
to do about non-Solaris platforms, including:
 - abstracting out the priocntl/etc. calls into a wrapper which can be implemented
   differently for each OS
 - renaming it to something more generic and more descriptive than "SolarisIA"
Comment 8 Jim Gettys 2005-08-31 10:45:41 UTC
(In reply to comment #7)
> Okay - I think that's everything needed to use it in Xorg.   The original Usenix
> paper describing the work that led to this is at:
>
http://www.usenix.org/publications/library/proceedings/cinci93/full_papers/evans.txt
> 
> A small bit of documentation on the IA priority control scheduler in Solaris
> that this uses can be found at:
> http://docs.sun.com/app/docs/doc/817-0689/6mgfkpd12?a=view#Inter-Active%20Class
> and of course in the Solaris kernel sources at:
> http://cvs.opensolaris.org/source/
> 
> Before we actually integrate it into Xorg though, I think we need to consider what
> to do about non-Solaris platforms, including:
>  - abstracting out the priocntl/etc. calls into a wrapper which can be implemented
>    differently for each OS
>  - renaming it to something more generic and more descriptive than "SolarisIA"


There are also some other interesting approaches that need exploration: in
particular, having scheduler credits traded back and forth between the X server
and clients as messages go back and forth, rather than the general schedular
behavior.

And, of course, we have to fix our broken drivers not to gratuitously use CPU
and thereby causing the X server to look like a CPU pig (which it is, at the
moment).
Comment 9 Daniel Stone 2007-02-27 01:24:57 UTC
Sorry about the phenomenal bug spam, guys.  Adding xorg-team@ to the QA contact so bugs don't get lost in future.
Comment 10 Jeremy Huddleston Sequoia 2011-09-24 17:24:49 UTC
Alan, is this still relevant?
Comment 11 Alan Coopersmith 2011-09-25 08:41:35 UTC
(In reply to comment #10)
> Alan, is this still relevant?

We still ship this extension as a local patch to Xorg and still haven't
gotten around to upstreaming it, so yes, I think it is.
Comment 12 Adam Jackson 2018-06-12 19:07:12 UTC
Mass closure: This bug has been untouched for more than six years, and is not
obviously still valid. Please reopen this bug or file a new report if you continue to experience issues with current releases.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.