[Openchrome-devel] Reply: Openchrome-devel Digest, Vol 30, Issue 8

[黄然]

Hi Kevin,

	Thanks for your detailed information below. Please see my comments below.

Thanks,
Frank


> 
> I am pretty much the only person who has done any substantial coding related
> to this project for the past year or so.
> Regarding VX11 chipset, VIA Technologies has not released any hardware
> programming documents related to its integrated graphics portion.
> I would imagine that it is a descendant of Destination core (DirectX 10
> compliant) they developed around 2007.
> 
> http://www.techpowerup.com/42118/s3-launching-directx-10-graphics-chips-s
> oon.html
[HuangRan] For VX11 chipset, I know it is a chipset which is released around 2010-2011. And it is widely used now for VIA's products.


> 
>     Regarding the current state of OpenChrome, this is where things stand,
> best to my knowledge.
> 
> - DRM / KMS code for OpenChrome
> 
> This is where it is located.
> 
> http://cgit.freedesktop.org/openchrome/drm-openchrome/
> 
> In particular, here is the source code for DRM / KMS code.
> 
> http://cgit.freedesktop.org/openchrome/drm-openchrome/tree/drivers/gpu/dr
> m/via
[HuangRan] Okay. I have "git cloned" the kernel DRM code and see it is a pretty integrated driver which(fence, crtc, ttm support, i2c) can support CLE266/CX700/VX800/VX855/VX900. It is very similar as old RADEON driver's architecture.


> 
> At a glance, it appears that CLE266, VX800, VX855, and VX900 are supported in
> the code, but I feel like the support for other devices are missing.
> I could be wrong on this, but James Simmons (the sole developer) has
> disappeared, so that is what I know for now.
> Until all known devices up to VX900 chipset (Chrome 9 HD) are supported and
> validated, I do not think it should be in the Linux kernel.
> Personally, James Simmons is far better of a developer than myself, but I try
> whatever I can (I am not really a professional graphics device driver developer
> myself.).
[HuangRan] So I have a question here. If I want to use KMS instead of UMS, I think I still need DDX module xf86-video-openchrome-0.3.3, right? 
And how about Xserver version I need use? And how about MESA for 3D?
By the way, which Linux OS distribution version you are using? Ubuntu 12.04 is okay?


> 
> 
> - Existing OpenChrome
> 
> As you know, I have been working on trying to fix Christopher's VX855 chipset
> screen detection bug.
> I believe it is the an I2C bus selection issue when probing for VGA.
> DVI's I2C bus needs to be used to probe the VGA.
> I did come up with a temporary workaround that should be good enough in the
> short term, and did observe that it did not break the device driver at least with
> my own laptop.
[HuangRan] I thought VGA and DVI should have their specific I2C pair lines to probe EDID information(not shared).


> I still have to create a patch for Christopher to try, but I have not done so yet
> (worked on porting ATI Technologies r128 device driver to an old Ubuntu
> release instead).
>     Regarding the existing OpenChrome code base, these are the known bugs
> / issues other than the VX855 DVI / VGA display detection issue.
> 
> * Occasional VT crash when logging in for the first time in Lubuntu 12.04 (low
> priority for now)
> * xorg X server crash when FP (Flat Panel) and VGA are turned off and on,
> respectively, simultaneously (low priority for now)
> * VX800 chipset ACPI S3 State resume hang (medium priority)
> * No mouse cursor on the screen with CLE266 chipset in Lubuntu 10.04
> (medium priority)
> * LVDS FP (Flat Panel) gets lost when resuming from ACPI S3 State or STR
> (Suspend to RAM) (high priority; have a partial workaround that restores the FP,
> but mouse cursor appears on FP in many cases)
> 
> 
> Other than these issues, these are the high priority items to work on.
> 
> * Reducing or deemphasizing the use of predefined display output table (This is
> easy to implement, but then we need to redo the testing of various devices
> since display output devices now have to be probed via I2C bus all the time.)
> * Completely rewriting display detection and resource allocation code
> * Support for dual head (support stretched screen)
> * Adding support for missing devices in the DRM / KMS code so that updated
> DRM / KMS code is included in the Linux kernel (more coding and a lot of
> validation work)
[HuangRan] Okay, so do those bugs are recorded in bugzilla(https://bugs.freedesktop.org/enter_bug.cgi?product=xorg&component=Driver/openchrome)?


> 
> 
> That is my take on what needs to be done.
> There is practically no support from VIA Technologies itself.
> They have released some documents, but it will be highly appreciated if they
> can release more documents previously that were available under NDA
> (Non-Disclosure Agreement).
> I live 30 km from their U.S. headquarters in Fremont, CA, in case they want to
> see me in person.
[HuangRan] Okay. Right now I am trying to get more documents for VX11 because we have some relationship/resources with VIA to do that. So for graphics driver, I am not sure if you are mentioning S3 Graphics department or VIA company you are near with?


> Of course, if they have their old hardware laying around their office, I will gladly
> take possession of them since I have purchased some of their hardware out of
> my own pocket in the past year or so.
> I still do not have Git repository access, and obtaining one and becoming the
> official maintainer is another item on my DOTO list.
[HuangRan] Okay. I hope I can do an official maintainer work too for OpenChrome.


> 
> Regards,
> 
> Kevin Brace
> 
> 
> 
> ------------------------------
> 
> Subject: Digest Footer
> 
> _______________________________________________
> Openchrome-devel mailing list
> Openchrome-devel at lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/openchrome-devel
> 
> 
> ------------------------------
> 
> End of Openchrome-devel Digest, Vol 30, Issue 8
> ***********************************************