Bug 12784

Summary:	radeon driver misdetects lvds dpi resulting in huge font size
Product:	xorg	Reporter:	Chí-Thanh Christopher Nguyễn <chithanh>
Component:	Driver/Radeon	Assignee:	xf86-video-ati maintainers <xorg-driver-ati>
Status:	RESOLVED DUPLICATE	QA Contact:	Xorg Project Team <xorg-team>
Severity:	normal
Priority:	medium	Keywords:	regression
Version:	git
Hardware:	All
OS:	All
Whiteboard:
i915 platform:		i915 features:

Description Chí-Thanh Christopher Nguyễn 2007-10-11 15:54:23 UTC

Using an Acer Aspire 5022, Mobility Radeon X700 with LVDS panel and xserver 1.3.0.0

Since upgrading from xf86-video-ati-6.7.192 to current git head, the fonts have become very large. This is because the screen size is not detected correctly. Log can be found in attachment 11987 [details].

git bisect points to this checkin:

http://gitweb.freedesktop.org/?p=xorg/driver/xf86-video-ati.git;a=commit;h=fad9e7b00de051f632000eb2304ac23a524f7c8e
---
fad9e7b00de051f632000eb2304ac23a524f7c8e is first bad commit
commit fad9e7b00de051f632000eb2304ac23a524f7c8e
Author: Alex Deucher <alex@t41p.hsd1.va.comcast.net>
Date:   Sun Sep 23 23:43:26 2007 -0400

    RADEON: Rework LVDS mode validation

    The old validation code wasn't really well suited to randr.
    This fixes several issues:
    - missing display size for panels with edid
    - broken duplicate modes
---

The symptoms are the same as in bug 10304. As that is also an Acer notebook, (using the intel driver instead of radeon) the cause may be related.

Comment 1 Alex Deucher 2007-10-11 16:01:32 UTC

There's nothing wrong with the driver, the edid for you panel is wrong:

Image Size:  289 x 21 mm

I'm guessing it should be 289x210mm.  Looks like we need another edid quirk.

Comment 2 Chí-Thanh Christopher Nguyễn 2007-10-11 16:36:59 UTC

Yes, the image size seems wrong. However, the panel size is 330 mm x 210 mm, which is detected correctly:

(II) RADEON(0): Max H-Image Size [cm]: horiz.: 33  vert.: 21

Also, "hwinfo --vbe" will give 

  Size: 330x210 mm

for the panel. From what I understand about EDID specifications, some smart person at Acer got confused with this
---
21: Maximum Horizontal Image Size (in centimeters).
22: Maximum Vertical Image Size (in centimetres).
...
66: Horizontal Image Size (in mm)
67: Vertical Image Size (in mm)
68: high significant bits for Horizontal Image Size (4 upper bits)
    high significant bits for Vertical Image Size (4 lower bits)
---
and just copied from bytes 21-22 to 66-67 (setting byte 68 to 0x10 for some reason), resulting in (256+32) mm x (0+21) mm image size for the standard timing.
Exactly the same happens in bug 10304.

Comment 3 Alex Deucher 2007-10-11 16:41:59 UTC

this is definitely a dup of bug 10304.  even your panels are the same.

*** This bug has been marked as a duplicate of bug 10304 ***

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.