Bug 81015 - radeon driver refuses to use dotclocks < 20Mhz
Summary: radeon driver refuses to use dotclocks < 20Mhz
Keywords:
Status: CLOSED RAWHIDE
Alias: None
Product: Fedora
Classification: Fedora
Component: xorg-x11
Version: rawhide
Hardware: athlon
OS: Linux
medium
medium
Target Milestone: ---
Assignee: Mike A. Harris
QA Contact:
URL:
Whiteboard:
Depends On:
Blocks: 82776
TreeView+ depends on / blocked
 
Reported: 2003-01-03 14:49 UTC by Nils Philippsen
Modified: 2007-11-30 22:10 UTC (History)
0 users

Fixed In Version:
Doc Type: Enhancement
Doc Text:
Clone Of:
Environment:
Last Closed: 2004-04-20 13:18:04 UTC
Type: ---
Embargoed:


Attachments (Terms of Use)
Patch that ensures a minimum dotclock of >= 12.5 Mhz (1.53 KB, patch)
2003-02-10 15:40 UTC, Nils Philippsen
no flags Details | Diff
Patch implementing an option to force a specific minimum dotclock on Radeon cards (4.21 KB, patch)
2003-10-14 14:04 UTC, Nils Philippsen
no flags Details | Diff
Patch implementing an option to force a specific minimum dotclock on Radeon cards (2.52 KB, patch)
2003-10-27 14:03 UTC, Nils Philippsen
no flags Details | Diff

Description Nils Philippsen 2003-01-03 14:49:57 UTC
Description of problem:

When using the radeon driver, modelines with dotclocks of less than 20 Mhz are
ignored. Using a tweaked driver (from the .src.rpm) which ignores this
limitation, the card uses modelines with lower dotclocks without any hassles.

Version-Release number of selected component (if applicable):

XFree86-4.2.1-10.2

How reproducible:

Always

Steps to Reproduce:
1. Put a modeline with a dotclock < 20 Mhz into XF86Config, use it
    
Actual results:

Modeline gets ignored due to too low dotclock (according to /var/log/XFree86.*.log)

Expected results:

Modeline gets used or (if the card is lying about its capabilities) a
configuration option should exist which overrides the limitd determined by XFree86.

Comment 1 Mike A. Harris 2003-01-28 10:10:47 UTC
Sample modeline to reproduce?

Comment 2 Nils Philippsen 2003-01-28 11:14:59 UTC
Modeline "720x576" 13.875 720 744 808 888 576 582 588 625 -hsync -vsync
interlace # H 15625 [Hz], V 50 [Hz]

That's 720x576 PAL, let me know if you want an NTSC one -- I have some NTSC
modelines, but they're not tested.

Comment 3 Nils Philippsen 2003-02-10 15:40:56 UTC
Created attachment 89971 [details]
Patch that ensures a minimum dotclock of >= 12.5 Mhz

Do not use this at home kids. It might break your VGA card, your monitor or
even more.

Seriously: This patch forces the driver to use a minimum dotlock of 12.5 Mhz
(or less). I don't know if there are any Radeons that can't handle that (mine
does) and how they would react in that case. This should preferably be a
configuration option ('Option "EnsureTVDotClocks" "yes"'? -- we might have to
ensure that the max dotclock is at minimum 17.5 Mhz -- yeah right :o)), but I
don't know how to implement such options and I'm currently to busy/lazy to
figure out.

Comment 4 Mike A. Harris 2003-02-10 18:44:39 UTC
This type of thing would be a feature enhancement, and not really a bug
fix.  Due to the potentially hardware-damaging nature of such a change,
I won't consider something like this for our current development.  It's
IMHO something that should be requested upstream in order to be given
wider coverage.

Setting to DEFERRED

Comment 5 Nils Philippsen 2003-10-14 14:04:50 UTC
Created attachment 95161 [details]
Patch implementing an option to force a specific minimum dotclock on Radeon cards

This patch implements the "ForceMinDotclockenstein" option in the Radeon
driver. It accepts a frequency as its argument (like in "12.5Mhz"). This patch
documents the option in the radeon man page.

Comment 6 Nils Philippsen 2003-10-27 14:03:25 UTC
Created attachment 95511 [details]
Patch implementing an option to force a specific minimum dotclock on Radeon
cards

Revised patch as suggested by Kevin Martin:

- rename option to "ForceMinDotClock"
- fix code duplication

Comment 7 Nils Philippsen 2003-10-27 14:04:54 UTC
File this patch upstream, deferred.

Comment 8 Nils Philippsen 2004-04-20 13:17:29 UTC
Fixed in xorg-x11-6.7.0-0.5.


Note You need to log in before you can comment on or make changes to this bug.