Created attachment 17559 [details]
--2D driver: master f2ec3fd8f2c63b1bc0745802dcf737eda4bb42b6
--3D driver:master 2b2cb566563b9f1f9739327ef9874143af838850
--DRM: master 401f77a2ef1d55feaccc13cdae1e5379f4cfec7e
Bug detailed description:
VGA-monitor connected by SDVO ,it will display below error info:"(EE) intel(0): Unable to read from SDVOCTRL_E for SDVOB slave 0x70." and X cannot start .
using SDVO connected monitor
X starts failed with above info.
start X normally.
Created attachment 17560 [details]
xorg conf file
we connected two monitor,one is connected by VGA,and the other is connected by SDVO.and it also happened on xf86-video-intel-2.4-branch.
Details on the ADD2 card you use?
My Add2 card is: WINTEC ZDDVI8SI-A
a quick test, how about if you use another ADD2 card instead? this can help us quickly rule out if there is a HW issue here...thanks.
We have tested with other ADD2 card, also had this problem.
We used this ADD2 case on our Q965 machine before. So this is a regression.
(In reply to comment #6)
> We have tested with other ADD2 card, also had this problem.
> We used this ADD2 case on our Q965 machine before. So this is a regression.
could be a HW issue,too. here HW mean the motherboard...Can you confirm if this 965 the "Q965" that you said works before?
(In reply to comment #7)
> could be a HW issue,too. here HW mean the motherboard...Can you confirm if this
> 965 the "Q965" that you said works before?
I used to play the ADD2 card on the "Q965" to test dual head. It always worked.
*** Bug 17240 has been marked as a duplicate of this bug. ***
I'm increasing priority, since 2 users (from bug#17240) could confirm this is not hw issue.
Created attachment 18913 [details] [review]
sdvo probe debug patch
Please try if this patch restores the normal behavior back.
Yes, it does.
The patch works for me, too.
Haien, please try Zhenyu's patch.
Created attachment 18969 [details] [review]
Try detect bits before init HDMI
The updated patch also works for me. (I still get the error messages "(EE) intel(0): Unable to write to SDVOCTRL_E for SDVOB Slave 0x70." and "(EE) intel(0): Unable to write to SDVOCTRL_E for SDVOC Slave 0x72."), but they don't cause any loss of functionality.)
Created attachment 19230 [details] [review]
force sdvo detect patch
This is the one I'd like to push. Use an option to enable SDVO force detect, as it will cost much in X startup time, so normal user shouldn't be affected by this probe.
Pushed patch. Close.
Author: Zhenyu Wang <firstname.lastname@example.org>
Date: Sun Sep 28 10:08:26 2008 +0800
Bug #16631: add option for SDVO force detect
Some ADD2 card doesn't get SDVO detect status setup right,
which disabled outputs on those cards. This adds a new
option "ForceSDVODetect" to probe all SDVO ports anyway.