url
stringlengths 33
62
| title
stringlengths 10
115
| author
stringclasses 2
values | content
stringlengths 61
41k
| comments_count
int64 0
1.68k
| scraped_at
float64 1.76B
1.76B
|
|---|---|---|---|---|---|
https://www.phoronix.com/news/MTAxMzM
|
An OpenChrome VIA Driver Status Update
|
Michael Larabel
|
For any unfortunate Linux souls still stuck with VIA hardware, here's a status update about the OpenChrome driver project for providing an open-source graphics driver for VIA integrated graphics processors.
James Simmons, one of the few open-source developers left that's daring enough to touch VIA hardware and the one who's single-handedly been porting OpenChrome to work with kernel mode-setting (KMS), wrote a brief status update.
Here's a few notes:
- The VIA kernel mode-setting code is still being worked on by James. It's still not done yet (it's been in-development for about one year) nor ready for merging. The soonest it could be merged now is the Linux 3.3 kernel next year, but I'll be surprised if it's ready then in the next few months. There's still several outstanding bugs, including completely broken DisplayPort detection.
- There's still a separate branch of the xf86-video-openchrome DDX for use with the OpenChrome DRM driver, since the mainline OpenChrome DDX isn't yet capable of playing nicely with the KMS support.
- The goals James has for the X.Org/DDX side of the project is to fix an EXA (2D acceleration) bug on the OLPC and to then drop XAA support, followed by other EXA code fixes so that it works with both TTM and non-TTM kernel memory managed environments. Simmons still needs to work out a buffer manager too that can handle TTM/non-TTM cases, if wishing to still provide user-space mode-setting support for OpenChrome. After all of that, there's still some other KMS work to do to round out the support for the X driver portion.
- To the OpenChrome DRM code itself, the goals here are to implement proper fencing, hardware-accelerated blitting, and supporting LVDS KMS. James has initial KMS code for LVDS panels, but it's not finished.
His status update can be found on the OpenChrome mailing list.
In terms of other work within the OpenChrome community, there isn't much. Only a handful of messages have been going by on the openchrome-devel mailing list each month, not a lot of activity in the SVN driver repository, etc.
If you're hoping to see a VIA Gallium3D driver in the near future, keep dreaming. There's still a lack of documentation from VIA, VIA itself basically abandoned their open-source strategy, and a severe lack of skilled development manpower.
| 5
| 1,760,738,368.674288
|
https://www.phoronix.com/news/MTAxNDU
|
A Change For Glamor X.Org Acceleration
|
Michael Larabel
|
The plan is for Glamor, an OpenGL-based acceleration method for X.Org drivers and common driver, to ship as a Glamor rendering library and a Glamor EGL support library.
Yesterday I mentioned there was patches to enable Glamor Acceleration within the Intel X.Org driver. On Friday, Intel's Zhigang Gong also sent a Glamor update to xorg-devel.
For those unfamiliar with the Glamor project, see this original posting from September. "Glamor is a 2D rendering acceleration implementation that's based upon OpenGL. The three parts of Glamor for X.Org are: a 2D rendering acceleration implementation, integration of Glamor into Xephyr, and a stand-alone DDX driver. The standalone Glamor driver uses this 2D rendering acceleration over OpenGL while leveraging KMS and EGL with Mesa."
The change in Glamor is to build a separate Glamor library that can be used by any possible DDX driver. The library will be built incrementally and merge Glamor into the Intel DDX driver. A new Git repository provides "glamor", the rendering library that provides all rendering functions, and then "glamor-egl", which is the EGL support library for creating and initializing the OpenGL/EGL context.
Zhigang Gong also notes that the current Intel driver patches for Glamor only migrate some of the UXA code-paths to using Glamor, but the remainder of the functions will soon be converted.
| 0
| 1,760,738,369.223616
|
https://www.phoronix.com/news/MTAxNTA
|
Replacing The Mac OS X XQuartz DDX In X.Org
|
Michael Larabel
|
There's a new project to replace the XQuartz DDX, as used by Apple Mac OS X, with the stock XFree86 DDX using a dummy video driver.
Jamey Sharp and other faculty at Portland State University are facilitating another X-related capstone project with the focus this time of moving rootless (in terms of no root window, not as running X.Org Server as an unprivileged user) out of the xorg-server. The goal is to replace the functionality offered right now by the X Server's root-less layer and the XQuartz input/rendering. Part of this work entails writing a compositing window manager that can grab window contents out of the server and then blit them into native windows on the host.
While XQuartz is really only relevant on Mac OS X / Darwin, they hope the resulting client code will be of use on other operating systems too. Part of the hope is that it could replace XWin with a Windows version of this compositing window manager. There's also thinking that it may be of use as well to the Wayland Display Server.
This project is still early, but they're hoping to have it completed by next March. Based upon the track record of previous X.Org projects at Portland State University, it will hopefully be another success. More details can be found in this xorg-devel message.
| 1
| 1,760,738,369.231828
|
https://www.phoronix.com/news/MTAxMDk
|
Experimental Code Published For Virtual CRTCs
|
Michael Larabel
|
If you're interested in multi-GPU rendering, capabilities for DisplayLink-like devices, or NVIDIA Optimus / MUX-less hybrid graphics switching, here's some news worth reading about virtual CRTCs.
Annoucned today to DRI and fbdev developers is a patch-set that a few unknown developers have been working on for several months. This code allows DRM GPU drivers to have an arbitrary number of CRTCs, including virtual CRTCs, rather than being limited by the number of physical CRTCs on the graphics hardware.
A CRTC (Cathode Ray Tube Controller) deals with video timings and reads the frame-buffer from video RAM so that it can be outputted. This code allows virtual CRTCs to be created in the kernel drivers so that they can be used for a variety of tasks. A virtual CRTC with this code is treated just like a real hardware CRTC. This experimental code allows these virtual CRTCs to be attached to "CTD devices" (Compression Transmission and Display). Then finishing off the equation is a new kernel module, the VCRTCM (Virtual CRTC Manager), that bridges the traffic between GPUs and CTDs.
With being able to basically stream rendered data from one GPU/display to another, there's interesting possible use-cases. Here's a description from the announcement, "In one example, we use AMD/ATI Radeon GPU to do 3D rendering (accelerated, of course) and we use our code to add additional monitor heads using DisplayLink devices. In other words, we achieve accelerated 3D rendering on a DisplayLink monitor. In another example we funnel rendered pixels to userland by emulating a Video-for-Linux
device (and then userland can do whatever it wants with it). While doing all this, GPU has no idea that we are doing this, the entire DRI "thinks" that it is just dealing with a GPU that has a few "extra" connectors and CRTCs. So everything ports without the need to modify anything in the userland."
As was confirmed by the developer announcing this work, this could help in the NVIDIA Optimus / MUX-less hybrid graphics situations. The discrete GPU can be used for rendering the data and then passing the contents to the integrated graphics processor.
These developers have also written a fake driver to represent a V4L2 device as a CTD device. Drivers wishing to take advantage of virtual CRTCs require slight modifications. The developers also seem to have some other interesting plans for this work too, assuming the upstream DRM developers welcome the contributions.
Find more information in the mailing list announcement for Virtual CRTCs.
| 38
| 1,760,738,369.837301
|
https://www.phoronix.com/news/MTAxMTk
|
X.Org Server 1.11.2 Released
|
Michael Larabel
|
The second point release of X.Org Server 1.11 is now available.
While all major development work is now happening on X.Org Server 1.12, the 1.11 series is still being maintained with bug-fixes and other minor work. This is important since the X.Org Server 1.11 series is likely what will end up being used by Ubuntu 12.04 LTS. Among the distributions shipping with X.Org Server 1.11, which was originally released in August, is the soon-to-be-released Fedora 16.
There's still several known issues, but overall this release has many fixes, in the areas of DIX, XQuartz, etc.
The release announcement of xorg-server 1.11.2 is available here.
Meanwhile the AMD Catalyst driver still doesn't support the X.Org Server 1.11 series, but it will later in the month.
| 0
| 1,760,738,369.845528
|
https://www.phoronix.com/news/MTAwNDY
|
First Phase Of X.Org GPU Hot-Plugging Works
|
Michael Larabel
|
David Airlie has demonstrated success in the first phase of his X.Org GPU/driver hot-plugging work, which eventually may lead to proper dynamic GPU switching under X.
On David Airlie's blog he has written briefly about his X.Org hot-plug work that he's been doing on and off for the past year. He's also included a video (embedded at the end of this post) that demonstrates this video hot-plugging work.
What's going on is using the xf86-video-modesetting driver with his custom X.Org Server work and xterm and GNOME Metacity. When he plugs in a USB-based DisplayLink adapter that's using his custom Linux kernel DRM driver, the USB-driven display lights up and is functional.
What goes on is that his custom X.Org Server code is started up with the respective drivers and David's major X.Org video driver ABI re-work comes into play here. One X protocol screen is exported and then when the GPU hot-plug happens, the xorg-server initializes another screen for the new device and plugs it into the single protocol screen. The driver-level pixmaps/pictures/GCs are copied to the new driver. The single protocol screen is then multi-plexed across the plugged in driver screens. David explains it as being "like Xinerama pushed down a lot further in the stack, so instead of doing iterations at the protocol level, we do it down at the acceleration level. Also I have randr code hooked up so all the outputs appear no matter what GPU they are from."
This demonstration is from a DisplayLink USB adapter, which doesn't have 3D hardware acceleration abilities. What Airlie eventually wants to get to is getting the main GPU to do the rendering and then only send the scan-out buffer to the USB device. This rendering-on-one-GPU-and-send-to-another could also be used for multi-GPU desktop systems to offload rendering work for one intensive OpenGL application/game to a secondary GPU and then piping the contents back to the primary graphics processor, among other possibilities.
While good progress has been made, David has been dabbling with this code for a year and there's still much more work ahead before we might see the code merged into mainline and deployed. "The real solution is a still a long ways off, but this is just a small light in a long tunnel, I've been hacking on this on/off for over a year now, so its nice to see something visible for the first time."
In terms of dynamic GPU switching for multi-GPU setups (such as NVIDIA Optimus and other notebooks with dual graphics processors), David says in a comment on his blog that the KMS drivers would need to be modified to work with the new X.Org Server for switching and such. Sharing of more kernel-level objects via PRIME or dmabuf is also needed.
| 9
| 1,760,738,370.370817
|
https://www.phoronix.com/news/OTk0OA
|
Glamor Project Is Called To Be Merged Into X.Org Server
|
Michael Larabel
|
Glamor, an open-source project that up until now has received little community attention or public acknowledgement outside of its small development group, has now been called to be merged into the X.Org Server. But what is Glamor?
Glamor is a 2D rendering acceleration implementation that's based upon OpenGL. The three parts of Glamor for X.Org are: a 2D rendering acceleration implementation, integration of Glamor into Xephyr, and a stand-alone DDX driver. The standalone Glamor driver uses this 2D rendering acceleration over OpenGL while leveraging KMS and EGL with Mesa.
While many are still interested in Xephyr, the Glamor DDX is interesting as it allows another generic (hardware-independent) X.Org driver that accelerates 2D using the 3D engine via the EGL interface with Mesa and without any native window system. As long as the graphics hardware has Mesa support and complies with the Linux kernel mode-setting (KMS) interfaces, it should work.
Current limitations though are not full GLX support, DRI2 support still isn't fully complete, and currently only Intel graphics hardware is supported. Intel support is only there due to the KMS, Mesa/EGL, and GBM requirements.
Glamor has been in development for a long time, but up to now it's not been heard much outside of its small development group. Glamor consists of more than 10,000 lines of new X code spread over 200 commits. Among the well known X.Org developers who have worked on Glamor include Eric Anholt and Kristian Høgsberg.
The Glamor DDX is faintly similar to the Xorg/XA state tracker that allows 2D acceleration via the GPU's 3D engine in a fairly generic manner. Both depend upon kernel mode-setting support, but Glamor also depends upon EGL support while the Xorg/XA state tracker targets the Gallium3D hardware drivers. The Gallium3D state trackers for X.Org are also independent of any specific xorg-server changes.
See more in the Glamor pull request announcement. Glamor has already been criticized due to little public review prior to its pull request and not being well known within the community, but it's largely independent of the rest of X.Org and its developers hope the code will now be examined by some new eyes.
| 8
| 1,760,738,370.379168
|
https://www.phoronix.com/news/MTAwMTc
|
Drafting Plans For X12, The X11 Successor
|
Michael Larabel
|
While it's likely in the next two to three years that the Wayland Display Server will play a pivotal role on the Linux desktop, the X11 Server isn't going away immediately. There's still legacy X applications that must run, uncertainty about what Solaris/BSD will do for their display server due to their graphics driver shortcomings, and other uncharted issues. It's possible that X12, an improved version of the X11 protocol, could even be developed.
The last time the undeveloped X12 protocol was talked about on Phoronix was back in June of 2010, but to some surprise this morning when waking up there was an update to the X12 Wiki page on X.Org. The Wiki X12 update was by Edward Cullen and it expanded the X12 introduction and rationale while outlining some of the X12 requirements.
X11 was defined at the start of the 1990's. Since then, computers have changed ''almost'' beyond recognition; the simple framebuffer model used by computers at the time has been replaced programmable graphics hardware that is both complex and powerful. Not only has graphics hardware changed, but the basic processing model has and continues to change; parallelism in core system design is becoming the norm, rather than a special case for 'large' systems.
And that's just talking about desktop systems; now there are smart phones, netbooks, tablets and probably will shortly be other device types that this author can't imagine (else I'd be working on them...).
In short, X11 was designed for a different era of computing. Among the proposed requirements for X12 is security designed from the start, multiple platform support (in terms of hardware platforms from mobile phones and tablets to desktop workstations), maintaining network transparency, support for modern graphics hardware and rendering, maintaining the frame-buffer concept, being as efficient as possible, and thinking on parallel.
There's a lot more information about current X11 issues and other X12 ideas on the full X12 page. So far though there's been no concerted efforts to begin work on X12, nor is it likely we will see such actions in the near future, if ever should Wayland manage to be a smashing success.
| 83
| 1,760,738,370.961514
|
https://www.phoronix.com/news/OTkzMg
|
Last Call For The X.Org/Linux Graphics Survey
|
Michael Larabel
|
It's your last chance to participate in the 2011 Linux Graphics Survey.
Originally this annual survey was set to end on 20 September, but due to being busy with Oktoberfest, that deadline was forgotten about. As a result, there's still time to participate.
The new end date for the 2011 graphics survey will be Monday, 26 September.
More information and participate in this quick but important survey via this Phoronix.com page.
| 1
| 1,760,738,370.969686
|
https://www.phoronix.com/news/OTk1OA
|
New, Generic X.Org KMS Driver Work
|
Michael Larabel
|
David Airlie has announced new work on the xf86-video-modesetting driver, which aims to be a generic X.Org (DDX) driver that will take advantage of the generic parts of the Linux KMS (kernel mode-setting) APIs so that any GPU should be supported.
While most hardware with kernel mode-setting support already has a matching X.Org driver that is KMS compliant or even still provides user-space mode-setting, this driver would be a simple, un-accelerated fall-back driver. It's similar to the VESA driver, but uses Linux KMS.
The features of the xf86-video-modesetting "restart" driver branch include ARGB cursor support, RandR 1.2 support, and dirty tracking ioctl support. What's changed from the original -modesetting work is that it drops all DRI2/EXA knowledge, uses the kernel dumb interface directly for buffer object creation and mapping, the mode-setting logic is derived from the Radeon driver, and probing is based upon the fbdev driver.
You can see David's email to dri-devel about this recent driver work.
While this driver is meant to be universal for hardware that has proper KMS support, there isn't any hardware acceleration. If you're looking for a universal X.Org driver with acceleration, and don't want to use the driver specifically targeted for your hardware, there is the Xorg state tracker (and now the XA state tracker too) for Gallium3D that relies upon KMS and then offers shader-based EXA/X-Video acceleration. There's also the recently announced Glamor work, which does have a generic driver for KMS environments while depending upon Mesa/EGL.
| 4
| 1,760,738,371.482027
|
https://www.phoronix.com/news/OTkyMw
|
TI Proposes A Low-Level Linux Display Framework
|
Michael Larabel
|
Texas Instruments has proposed a new low-level display framework for Linux.
An OMAP driver developer at Texas Instruments wrote a rather lengthy post about a new low-level display framework.
As Keith replies, the TI proposal isn't radically different from what's being worked on now within the DRM world to make overlay support within KMS a suitable framework. Hit up the already lengthy mailing list thread if you're interested in the details.
| 2
| 1,760,738,372.159252
|
https://www.phoronix.com/news/OTkzMA
|
A Major Rework To The X.Org Video Driver ABI
|
Michael Larabel
|
One of the mailing list threads I've been trying to catch up on this week while at Oktoberfest is the heated discussion about merging video/input drivers back into the X.Org Servers. This discussion was started at the XDC2011 conference, but there's many e-mails being exchanged from more parties not in favor of merging the drivers into the xorg-server tree or wishing to see other developmental process changes.
David Airlie of Red Hat mentions during this discussion that he has a major overhaul coming to the X.Org Server video driver ABI, which would require significant changes to the DDX display drivers.
At the moment I have a large project in the works to redo the driver ABI, its pretty much an impossible project no matter what happens, its a major ABI change and will pretty much require major reworks off all the drivers.
Now I've no idea how to do this with the current model, even with the drivers merged its going to be a real pain to land, thats if I ever even get it to a useable stage.
Dave.
The mailing list message appeared here.
We'll see if this "impossible project" materializes or what comes about. Lets hope that if any major video driver ABI breaks are coming, that AMD will be more prepared to support them in a timely manner, unlike their notorious support time right now for most major X.Org Server and Linux kernel revisions.
| 27
| 1,760,738,372.725663
|
https://www.phoronix.com/news/OTkzNQ
|
X.Org Server 1.11.1 Brings Two Brown-Bag Fixes
|
Michael Larabel
|
Jeremy Huddleston released xorg-server 1.11.1 a few hours ago. This release was done since two "brown-bag" issues were found in X.Org Server 1.11.0, which was released just one month ago.
Due to two high-priority issues, this new release isn't coming six months after 1.11.0, which is normally the timed cycle for the point releases. Assuming nothing major comes up, xorg-server 1.11.2 will go back to the six-week release cycle.
The two committed changes to this release come from NVIDIA's Aaron Plattner. It turns out the extension ABI had to be bumped, but wasn't done so after the respective breaking commit. The other is an fb change.
Read more in the X.Org Server 1.11.1 release announcement.
| 2
| 1,760,738,372.734005
|
https://www.phoronix.com/news/OTkwNw
|
The X.Org Code-Base Of The Future
|
Michael Larabel
|
On Monday at XDC2011, Jamey Sharp talked about what he sees as "the codebase of the future" for X.Org and was an open discussion with the three dozen other developers at the Chicago event.
Jamey acknowledges there's much code duplication in the X.Org Server code-base and wishes it would be cleaned up. Among the work he wants to see happen is merging together the DDX (Device Depoendent X) and DIX (Device Independent X) portions, merging togetheer Xephyr and Xnest (or converting them to drivers for the X.Org Server), provide support for video card hot-plugging (including USB display adapters or virtual GPUs), RandR GPU objects, and a separation of scan-out and render.
That was about it for the X.Org code-base of the future. Not much about the Wayland Display Server has been said at this conference. However, it looks like there might be an X@FOSDEM 2012 event in Brussels in February. Keith Packard has said he would like to have at least a half-day of talks strictly about Wayland.
There's also some more photos and notes from day one of XDC2011 Chicago in this news posting.
| 1
| 1,760,738,373.35968
|
https://www.phoronix.com/news/OTkxOQ
|
The Fight Over Merging Drivers Back Into X Server
|
Michael Larabel
|
The debate that started back up again this week at XDC2011 Chicago about merging drivers back into the X.Org Server has now moved online. Jesse Barnes has published the pros / cons that were mentioned at the X.Org Developers' Conference this week to the X.Org development mailing list for developers to now debate the idea online. This has been a hotly disputed matter for the past two years.
Jesse Barnes, who is in support of merging the video and input drivers back into the X.Org Server, wrote the following message to xorg-devel with the pros and cons.
At XDC this week we discussed merging drivers back into the server tree. One thing I found frustrating about the discussion was that we didn't have a whiteboard nor a list of the pros & cons of such a change. So I'd like to capture that here (from memory) to let us continue the discussion about whether it's worth it or not.
Luc, I think you're the most vocal opponent of this move, so I've cc'd you so you can enumerate any issues I've forgotten.
Anyway, as I recall, the issues are as follows:
Pros:
1) easier to propagate API changes across drivers (just like Linux)
1a) thus easier to change ABI 2) developers focused on driver development now have more incentive to make sure the server works well so regular releases can still happen (i.e. more people working on blockers whether driver or not as releases approach) 3) allows removal of driver compat code for various server versions
3a) thus removes combinations of driver+server that developers have to support & test 4) increased test coverage for the server as users wanting current driver code will be building new servers too
Cons:
1) more work for distros to backport just driver changes to old servers (especially if people follow through on (3) above)
1a) if backporting is harder, new hardware support will be more difficult to land in "enterprise" level distros
2) harder for users to just upgrade drivers independently, now they'll have to build the whole server
2a) thus less testing of current driver code from technical users
I've already made my views pretty clear; I'd prefer merging the drivers back in. But I don't do as much work on the DIX or DDX as I used to, and lots of others would be affected as well, so I'd like to hear what people think. Have I captured the pros & cons fully? What to distro maintainers think? And driver developers, both input & output?
Thanks,
Jesse Michel Dänzer, who is now at AMD but was not at XDC2011, was quick to respond saying that out-of-tree drivers would become second class citizens. He brings up that out-of-tree drivers would just not be the binary blobs (i.e. NVIDIA and AMD proprietary drivers), but also the Gallium3D Xorg state tracker. Michel also says, "Speaking as a radeon driver developer, merging the driver into the server tree would be unworkable at this point because since the "new development model" has been in effect, it's not possible to get even trivial changes into the server tree without a ridiculous amount of time/effort."
Alex Deucher also expressed he doesn't see much advantage in moving the X.Org drivers back into the server.
Plus there's several other email comments from developers, both those that were in attendance at XDC2011 and not. Continue the heated discussion thread here. (I just landed in Munich and will have more time to go through the read later.)
| 26
| 1,760,738,373.36824
|
https://www.phoronix.com/news/OTkxNA
|
X.Org Server: Merging Drivers Back In, Avoiding Regressions
|
Michael Larabel
|
Ending XDC2011 Chicago on Wednesday afternoon was a discussion led by Apple's Jeremy Huddleston, Intel's Keith Packard, and Oracle's Alan Coopersmith. The discussion was about X.Org Server release schedules. The two main points brought up is merging the drivers back into the X.Org Server tree as well as aiming for a regression-free X.Org Server by reverting any commits to the server Git tree that are regressions that aren't fixed within one week's time.
Any regression without a credible fix within one week's time is now going to be reverted by Keith Packard. The user simply needs to bring it to his attention and if there isn't any fix within a week (or signs of an imminent fix), the commit will be dropped until it can be reworked. This will hopefully lead to a less-buggy X11 Server and also make it easier to stick to the xorg-server release schedules.
The other big item is that the X.Org drivers are likely to be merged back into the X.Org Server. This has been a hot discussion item now for two years at the X.Org conferences and on the mailing list, and there is still some opposition. However, this time it looks like merging the drivers back in has the majority support now that the servers are being released routinely on-time and other concerns addressed. Luc Verhaegen and enterprise distribution vendors largely remain the ones in stiff opposition.
A final discussion will likely take place on the mailing list with possible voting to finalize the decision. Maintainers of video and input drivers not wishing to push the driver into the xorg-server tree can still let them live externally, but they'll be responsible for handling X.Org Server API/ABI breaks.
The developers want to merge the drivers back in so that they can clean up the API and it also makes it easier to bisect regressions when the drivers are being built at the same time as the xorg-server and should always be compliant with the server in question.
That's the short-end of the story for what was brought up during the release schedules discussion.
| 0
| 1,760,738,373.874479
|
https://www.phoronix.com/news/OTg5NQ
|
XDC2011 Chicago Kicks Off Tomorrow
|
Michael Larabel
|
The 2011 X.Org Developers' Conference is set to begin tomorrow morning in Chicago, Illinois, United States. This three-day event will focus on what's coming down the pipe for the X.Org Server, Mesa, the Linux kernel DRM, and related areas of open-source graphics and input.
The line-up for Monday is mostly about the X.Org Server in general and input related fun. Currently the agenda is:
- X.Org Foundation Session
- X.org Release Schedules Discussion
- The Code-base Of The Future
- XKB 1.1
- Multi-Touch BoF / Wacom BoF
- Driver Support BoF
On Tuesday the discussions are more graphics related. Tuesday on the schedulke there is:
- Compiler Stacks
- Nouveau
- TimeGraph / GPU Scheduling
- LLVM In Mesa
- Graphics Memory Management
- Contributing To X.Org & Open-Source Panel Discussion
Tuesday afternoon there is also free beer to X.Org attendees, sponsored by Phoronix Media. I'm not too pleased with the beer selection that the catering company has offered, but there's a few "drinkable" options. The last session of the day is the Q&A panel discussion where any interested developers in the Chicago area are welcome to attend if they're possibly wanting to contribute to upstream open-source projects.
On Wednesday, the last day of this conference, there is:
- DRM Overlay Plane
- OpenGL 3.0
- GPGPU
- Auto Regression Analysis
- GPU Kernel & User-Space Border
There's more information on the X.Org Wiki.
Today there's an X.Org Developers meet-up at a local bar (Snickers) in downtown Chicago beginning at 2:00PM.
I'll be recording videos for Phoronix, posting interesting bits of information on Phoronix, and Tweeting the rest. On Thursday it's then off to Munich for Oktoberfest 2011 and the Phoronix meet-ups.
| 0
| 1,760,738,373.924739
|
https://www.phoronix.com/news/OTkxMw
|
XDC2011 Chicago - Caption Contest
|
Michael Larabel
|
For those not in Chicago for the 2011 X.Org Developers' Conference (XDC2011), here are a few photos from the Phoronix beer event yesterday. The full photo-set of XDC2011 will be published on Phoronix in the next few days, but here's a few highlights aside from the ones published already.
Tuesday was the Phoronix-sponsored beer afternoon at XDC2011, which happened prior to the Q&A panel, so it was fairly entertaining. Here's some of the interesting/comical photos. Feel free to suggest captions for these images (or guess what was going on) using the discussion link for the Phoronix Forums at the bottom of this entry.
Luc Verhaegen _______________. What did Keith Packard do to his Intel Sandy Bridge notebook? Bud Light? Lederhosens? AMD's John Bridgman is telling a story.
| 2
| 1,760,738,374.473683
|
https://www.phoronix.com/news/OTg5OQ
|
Photos From XDC2011 Chicago (Day 1)
|
Michael Larabel
|
Here's some photographs from the first day of XDC2011, the annual X.Org Developers' Conference, which this year is taking place in Chicago, United States. XDC2011 is running through Wednesday and is being hosted at the Illinois Institute of Technology.
Jerome Glisse, John Bridgman, and others getting ready for the Chicago technology conference.
Jamey Sharp was talking about the future code-base of X.Org. This involves the possibility of merging the DDX (Device Dependent X) and DIX (Device Independent X) code, output hot-plugging, RandR GPU objects, and other work.
Matthias Hopf talked about recruiting new developers to X.Org and Mesa. He talked about Mesa being way behind in OpenGL support, the lacking of X documentation, the X.Org Board of Directors could use better marketing, existing developers are mostly employed long-term contributors, development on X/Mesa can be intimidating, and that X.Org could use better advertising.
Some of the attendees. Yep, lederhosens and toe shoes too.
Lunch with (from left to right) Jerome Glisse, Egbert Eich, myself, and Luc Verhaegen. Matthias Hopf took the photograph.
Tomorrow at the X.Org Developers' Conference is the Q&A new contributors panel and a number of graphics related talks.
| 3
| 1,760,738,374.482649
|
https://www.phoronix.com/news/OTg4MQ
|
Two More Developers Join The Chicago Linux Panel
|
Michael Larabel
|
Next Tuesday during XDC2011 Chicago at the Illinois Institute of Technology I am hosting a panel about contributing to Linux and open-source projects, in particular, X.Org, Mesa, and the Linux kernel, but the information should be largely relevant to any free software project. This discussion panel is largely targeted towards university students and others that aren't yet contributing to upstream projects, with most of the panel participants having begun their Linux contributions prior to graduating from university and then most of them being poached by major open-source companies.
Peter Hutterer, Martin Peres, Corbin Simpson, and Keith Packard were the original participants and now they will be joined by Matt Dew and Kenneth Graunke. Matt works on X.Org documentation, showing that all contributions aren't limited to code. Kenneth began contributing to graphics drivers two years ago while still being a graduate student and was then hired by Intel. I'll be the moderator of this panel.
Any interested individuals in the Chicago area are invited to attend this Q&A panel during the X.Org Developers' Conference. This panel is taking place at 4:00PM on Tuesday, 13 September, at the IIT McCormick Tribune Campus Center ballroom. If you're not from the area, you can propose questions in advance using this forum thread. A video recording will be available afterwards.
More details about this unique panel can be found in this news posting.
| 0
| 1,760,738,375.203466
|
https://www.phoronix.com/news/OTg3Nw
|
X.Org Smooth Scrolling Prepped For Merging
|
Michael Larabel
|
One of the features part of the new X Input extension to be included in X.Org Server 1.12 is smooth scrolling support.
Patches for smooth scrolling in the X.Org Server originally appeared back in June by Peter Hutterer, but the smooth scrolling patches have now appeared on the mailing list once more. With X.Org Server 1.11 being out the door, the merge window for X.Org Server 1.12 is opening and so Peter has pushed out the patches for one final review before calling for them to be pulled into the mainline tree by Keith Packard.
More information about smooth scrolling support can be read on Peter's blog.
| 0
| 1,760,738,375.914531
|
https://www.phoronix.com/news/OTg3NQ
|
A Q&A Panel About Contributing To X.Org & Open-Source
|
Michael Larabel
|
Next week at XDC2011 Chicago there will be a rather unique discussion taking place that's quite different from what normally goes on at this annual X.Org Developers' Conference. There is going to be a moderated panel discussion (tentatively titled "Contributing to X.Org and Open-Source") about contributing to X.Org, the Linux kernel, Mesa, and open-source software in general. For those not residents of the Chicago area, this session will be broadcast on the Internet.
With XDC2011 being hosted at the Illinois Institute of Technology (IIT), and there being some interest in this event from the computer science department and their computing group, this panel is largely centered around spurring a new generation of open-source developers and how university students can become involved. It doesn't matter whether you're a student, as much of the information should be relevant to anyone interested in the open-source ways and the opportunities it presents.
What should make this talk particularly interesting to students is that the panel is composed largely of open-source developers who began contributing when they were at university. Some of them are still at university while others have since been hired by key open-source companies and/or handling both simultaneously. This should provide a unique perspective about challenges they have had in beginning their contributions to open-source projects, recommendations they may have for other beginners, different ways that individuals can contribute and become involved with various projects, etc.
Recent X.Org Events: XDS2008 Edinburgh, X@FOSDEM 2009 Brussels, XDS2010 Toulouse
I'm the organizer and panel moderator of this Tuesday last-of-the-day session. The panel members include:
Peter Hutterer: Peter began contributing to X.Org while at university. For his PhD work at the University of South Australia, Peter masterminded Multi-Pointer X (MPX), which provides support for multiple independent pointers in the X11 Server. Peter began this work in 2005 and it was successfully completed in 2008. After finishing MPX, he's continued contributing to Linux input drivers, the X Input 2 extension, and other areas. He was hired by Red Hat Australia following the completion of his PhD.
Martin Peres: Martin Peres is a French engineering student that began contributing to X.Org for improving the ATI Radeon graphics driver support. Most recently he's been a core developer of the Nouveau GPU driver that aims to provide an open-source NVIDIA Linux driver that is created via clean-room reverse-engineering of NVIDIA's proprietary software. Through his work on Nouveau, he's also done work for the PathScale compiler company and their graphics driver interests. Martin has also been involved with other open-source software projects.
Corbin Simpson: Corbin, a programmer at the Oregon State University Open-Source Lab, began contributing to X.Org and open-source via Google's Summer of Code. He participated in the 2009 Google Summer of Code project where he led the efforts on writing the open-source Gallium3D driver for ATI R300 (through ATI R500) graphics processors. Since then he's contributed to other areas of Mesa / Gallium3D and kernel mode-setting work. In 2011 he was also a mentor to a new Google Summer of Code student.
Keith Packard: Keith Packard hasn't been at university in decades, but as the release manager for the X.Org Server, one of the X.Org founding members, former core member of XFree86, and being involved with X11 since the 1980s, he has valuable information to share. Keith is also a board member for the X.Org Foundation, an employee of Intel's Open-Source Technology Center, and has created other core Linux software such as Cairo.
Matt Dew: Not all open-source contributions are limited to code, but there's also other ways to contribute. Matt Dew's contributions to X.Org are known for being one of the few that actually works on documentation.
Kenneth Graunke: Kenneth attended an X.Org Developers' Conference in Oregon two years ago while a graduate student of Portland State University. Graunke began contributing to graphics drivers and ultimately was hired by Intel's OSTC Portland team to work on their open-source graphics driver.
These panel members should be able to provide some insight on how to become better involved with X.Org and other open-source projects, challenges they've had in dealing with open-source, recommendations, etc. You are invited to ask questions. If you will not be attending the event in person, you can prepare a question in advance. Simply use the comments and discussion link below to post your question(s) in the forums thread. I'll have a list of the best questions prepared.
Ideally this session will lead to new X.Org / Linux contributors, or at the very least, inspiring new individuals about contributing to open-source projects.
There may be a live audio/video stream of this session, but otherwise it will appear on Phoronix and YouTube following the event. If you are from the Chicago area and are interested in contributing to open-source software or learning how you can become involved with this key open-source desktop project, regardless if you're a university student or not, you are welcome to attend this talk.
This Q&A panel is scheduled to take place beginning at 4:00PM on Tuesday, 13 September. The X.Org Developers' Conference is taking place at the ballroom of the McCormick Tribune Campus Center at the Illinois Institute of Technology. This venue can be quickly and easily reached by public transportation. There is no cost to attend this session.
If you are a member of the public wishing to attend, please be there promptly by 4:00PM. In fact, ideally by 3:50PM, as this talk will start promptly on time and potentially a few minutes before hand if there's a hard deadline of being out of the ballroom by 5:00PM (hence why this session cannot be later in the day). This Q&A panel will be approximately one hour in length and is the last session of the day.
Free beer! To make this session more interesting and to entertain the existing X.Org developers at the conference during this session, Phoronix.com is set to sponsor free beer on Tuesday afternoon for the hour. That is assuming the university's catering department has everything in order. (There's an alcohol service request form and other American legal Scheiße that needs to be processed and approved by the university. I'm also still waiting on hearing back if the catering department actually has any good beers to serve.)
More information on the location and any scheduling changes will be posted to the XDC2011 Wiki. If you are from the public and just plan to attend this general hour-long Q&A session about contributing to X.Org / Linux kernel / Mesa or other open-source work, there isn't a need to add yourself to the attendee page on the Wiki. Just be sure to show up to the MTCC on-time Tuesday afternoon. If you have any other questions about the event, contact me.
If you are unable to make the event, you can ask any questions of these panel members using the forum discussion thread. Check back on Phoronix where I'll have video recordings of this session and the other talks following the event.
| 6
| 1,760,738,376.430677
|
https://www.phoronix.com/news/OTg3Mw
|
2011 Linux Graphics Survey Reminder
|
Michael Larabel
|
As a reminder to those that haven't yet participated, the 2011 Linux Graphics Survey is currently taking place and your feedback is requested to help X.Org / Linux graphics developers understand the current driver market-share and to help the desktop community at large.
Click this link for more information and to participate in the Phoronix survey that's running through the 20th of September. It should just take a few minutes to complete and it's for finding information about the graphics card and drivers you're using, features that are important to you, and other basic information. Join the thousands of others that have already submitted their results.
Unfortunately, since the survey started late this year, the results won't be out in time for any discussion at XDC2011 Chicago conference. However, I happen to be doing some last-minute planning for something else that may prove to be interesting and of great interest to Phoronix readers and many within the Linux community that are interested in contributing to open-source projects. And yes, it even involves free, good beer too! Stay tuned for details hopefully in the next few days.
| 0
| 1,760,738,376.438788
|
https://www.phoronix.com/news/OTg2OA
|
What To Expect From X Input 2.1; Multi-Touch Is Delayed
|
Michael Larabel
|
While there has been an X.Org multi-touch implementation for X Input 2.1 dating back about a year, there's still no formal X Input 2.1 release (X Input 2.0 came in 2009). When X Input 2.1 is released, likely as part of X.Org Server 1.12, it won't even offer up the multi-touch support. Here's some of what you can expect to see from X Input 2.1 and then the X Input 2.2 milestone.
X.Org input expert Peter Hutterer has begun to blog about what's new in X Input 2.1 (Xi2.1) and a bit about X Input 2.2. "Some minor changes may still come to X Input 2.1, but this release isn't going to be exciting. XI 2.1 is a small incremental change with two bugfixes and one new feature. These are uncontroversial and more importantly, they are ready now. So XI 2.1 is released as a small change, with bigger changes coming for XI 2.2." The plan is to have X Input 2.2 in X.Org Server 1.12, which will be released in early 2012. Don't be surprised though if it's delayed to X.Org Server 1.13, which would mean it's still about a year away. (X.Org Server 1.11 was released one week ago.)
The X Input 2.1 changes include new raw events behavior, smooth scrolling, and XI2 defines. Smooth scrolling for X.Org is talked about in this article. Lacking is multi-touch support, even after the multi-touch support fell apart and then came back together back in December with several revisions. Canonical has been the one interested in this and they were the ones who originally envisioned an X.Org Gestures Extension. For now Canonical has been patching their xorg-server with their own input support.
Those wishing to learn more about X Input 2.1, Peter's blog series begins here and the second part is already available and it's about X Input 2.1 defines. Peter (and Daniel Stone) will also be at XDC2011 Chicago to talk about the future of X.Org input support.
| 9
| 1,760,738,377.031148
|
https://www.phoronix.com/news/OTg0MA
|
X.Org Server 1.11 Has Officially Been Released
|
Michael Larabel
|
X.Org Server 1.11 was officially released this Friday evening.
X.Org Server 1.11 was originally planned for released in mid August, but then the unfortunate passing of Keith Packard's mother (the X.Org release manager) led to a one-week delay. Our condolences go out to Keith Packard and his family.
After numerous belated releases from X.Org in the past, a one week delay is nothing to complain about, especially considering the sad circumstance.
While this is a new major X.Org Server release, it's mostly about bug-fixing. X Input 2.1 was delayed (with its touch-related features) to the next X.Org Server release (or later), there isn't any RandR extensions (after RandR 1.4 was restarted), and just nothing to get too excited over, besides addressing outstanding issues. Regardless, it's an improvement that incorporates six months of enhancements.
X.Org Server 1.11 does break the ABI for graphics drivers, but the latest NVIDIA binary driver already supports it and the AMD Catalyst blob will quickly (if it doesn't already, I haven't checked on my latest development copies).
The next release, X.Org Server 1.12, should be here in approximately six months. It will likely be discussed next month at XDC Chicago 2011 when the developers will next meet. One of the targets for the 1.12 release is smooth scrolling support.
There's also a question when X.Org 7.7 will be released. X.Org 7.6 was released back in late 2010, the X.Org Wiki says that "X11R7.7 development is underway for release in the second half of 2011," but there's been no discussions recently about tagging the 7.7 katamari. It's quite possible that X.Org 7.7 won't be released until early 2012 when X.Org Server 1.12 is here, but it's not too much of a big deal considering the modularization of X.Org and most Linux distributions already shipping the latest components.
X.Org Server 1.11 isn't being shipped in the upcoming Ubuntu 11.10 release, but it will be found in Fedora 16 and other distributions.
The official xorg-server 1.11 announcement is available from the mailing list. All it really says though is "Here's the final 1.11 release. Thanks, as always, to everyone who contributed code, testing, documentation and review."For those wondering about the Wayland Display Server, see this Wayland status update when looking at it earlier in the week under Ubuntu 11.10.Be sure to participate in the 2011 X.Org/Linux Graphics Survey.
| 5
| 1,760,738,377.040489
|
https://www.phoronix.com/news/OTgyOQ
|
X.Org Server 1.11 Set For Release On Friday
|
Michael Larabel
|
X.Org Server 1.11 was originally planned for release on the 19th of August, but following a one-week delay, it should be officially released this Friday. This is another significant update to the X.Org Server.
X.Org Server 1.11 was originally planned for release last Friday, but last week was met by the unexpected passing of Keith Packard's mother. With Keith serving as the release manager and being out of the game last week, the xorg-server release was obviously postponed.
Since the X.Org Server 1.11 RC2 release, there's just been the merging of four relatively small fixes. Assuming there's nothing else that pops up in the next three days, X.Org Server 1.11 will be released on the 26th of August. This was mentioned on the X mailing list.
This release though isn't a huge deal like some past releases since it still lacks some of the input fun and doesn't deliver on really any prominent features. The xorg-server 1.11 release isn't even being picked up by Ubuntu 11.10 but they will keep shipping X.Org Server 1.10 series to avoid a video driver ABI break.
X.Org Server 1.12 should be released in about six months per the X.Org Server development process. Expect more details to emerge next month during the X.Org Developers' Conference 2011 in Chicago.
Lastly, be sure to participate in our 2011 Linux Graphics Survey to collect some valuable user data for open-source developers.
| 2
| 1,760,738,377.541482
|
https://www.phoronix.com/news/OTgyMg
|
X.Org Developers' Conference 2011 Approaches
|
Michael Larabel
|
Three weeks from today, the X.Org Developers' Conference, the annual meet-up of Linux graphics driver developers, will be taking place in Chicago. Here's the latest developments for this conference.
As most Phoronix readers are already aware, I am the organizer of the XDC2011 conference that is taking place in Chicago on the campus of Illinois Institute of Technology. The X.Org conference is running from the 12th to 14th of September where the future of open-source Linux graphics drivers, the X.Org Server, and possibly Wayland will be discussed.
There will be up to 60 X.Org developers at this Chicago event. If you are a developer and are interested, be sure to visit the Wiki page to register for the event. As good news to Phoronix readers, over the weekend it was confirmed that AMD's Alex Deucher and John Bridgman will be participating in the event. Here's a breakdown of some of the other known attendees that are registered for this free development summit.
Intel: Keith Packard, Jesse Barnes, Ken Graunke, Shuang He, Ian Romanick
AMD: Alex Deucher, John Bridgman
SUSE / Former RadeonHD Developers: Matthias Hopf, Luc Verhaegen, Egbert Eich
Canonical: Bryce Harrington, Chris Halse Rogers, Chase Douglas
Red Hat: Jerome Glisse, Peter Hutterer, Ben Skeggs
Nouveau: Martin Peres (Ben Skeggs also commonly works on Nouveau at Red Hat)
Alan Coopersmith and Stuart Kreithman of Sun / Oracle will also be there along with Daniel Stone, among other attendees. For those that can't make the event, there will surely be live chatter on IRC, I will be tweeting, and Phoronix should also be providing live (or slightly delayed) audio/video recordings of the event. At XDC2011 I might also be sharing a special announcement.
Details on the X.Org social event / formal dinner are still being worked out as I am currently waiting back on information from the X.Org Foundation board of directors. With AMD's confirmed attendance at the event, there may also be a special Phoronix party to celebrate the four year anniversary of AMD's open-source strategy. Those details are now quickly being worked out.
Expect plenty of Phoronix coverage from XDC2011, which is taking place immediately before Oktoberfest 2011 where there will be another annual Phoronix meet-up.
| 2
| 1,760,738,377.550943
|
https://www.phoronix.com/news/OTgxOQ
|
Texas Instruments Proposes Extending DRI2 For Video
|
Michael Larabel
|
Rob Clark of Texas Instruments has proposed to X.Org developers that the DRI2 protocol be extended so that overlays can display video content, as a possible replacement to X-Video or for a client-side X-Video API on top of DRI2.
A few extra parameters are needed in order for Direct Rendering Infrastructure 2 to handle displaying video content. While DRI2 is mostly about 3D and not video, some processes are similar and there is a lot of infrastructure connection between the two, which is why Rob Clark is looking to extend the DRI2 protocol. However, if developers prefer not doing so, he's willing to create a new protocol instead. Whatever the outcome, Texas Instruments is willing to make the necessary X Server infrastructure changes.
Why do this work? "Current solutions use the GPU to do a scaled/colorconvert into a DRI2 buffer from the client context. The goal of this protocol change is to push the decision to use overlay or GPU blit to the xorg driver. Eventually this should replace Xv. With a few additions, like attributes, it could perhaps be possible to implement the client side Xv API on top of dri2."
More information is available from this mailing list message.
| 3
| 1,760,738,378.170312
|
https://www.phoronix.com/news/OTc4MQ
|
One Month Until The X.Org Meeting In Chicago
|
Michael Larabel
|
We're now just one month away from the 2011 X.Org Developers' Conference (XDC2011) that's being hosted in Chicago, Illinois, United States.
XDC2011 is being organized by your's truly and is running from the 12th to 14th of September. The venue is the Illinois Institute of Technology on the southern-edge of downtown Chicago.
The program (see the Wiki page) is still very much a work-in-progress, but at the moment it looks like there will be talks on multi-touch, XKB1.1, auto-regression analysis, Nouveau, and a DRM overlay-plane API. There will also surely be other talks over the three-day event, but that's waiting on the X.Org developers to submit their session plans. I'll also likely be announcing some driver automated testing / performance monitoring work at the event as well (that comes out of some Phoronix Test Suite / OpenBenchmarking.org research work with Intel supplying the hardware).
Additional details on the XDC2011 event can be found on this Wiki page. I'll be posting live to Phoronix any interesting bits of information from the event. The day after XDC2011 ends, it's then on to Oktoberfest 2011.
| 3
| 1,760,738,378.178641
|
https://www.phoronix.com/news/OTczNQ
|
Multiple X Servers For One Graphics Card, Again
|
Michael Larabel
|
One of the long sought after features of X.Org and the Linux graphics stack has been the ability to run multiple X Servers from a single graphics card. While this wouldn't be used by many, there are still many interested in seeing this feature request become a reality.
For instance, being able to run multiple X.Org Servers from one graphics card (basically one X.Org Server per output port) would be ideal for many educational institutions and enterprise environments where one PC/server could be easily shared by many heads.
Dr. Klaus Kusche, a German professor, is the latest individual seeking out this feature for Linux / X and is even willing to put money behind seeing it become a reality. "Mainboards with 4 or even up to 7 PCIe x8 or x16 slots are available today, and graphics cards with 3 to 5 outputs are also common. Such a combination would make a very attractive classroom server for up to 25 pupils/students, cheaper than any other solution, much easier to administrate than a network of single-seat PC's, and much faster w.r.t. graphics than a server with thin clients. We would be highly interested in such a solution (I'm a professor for computer science at a technical high school and university of applied sciences), and I'm currently investigating the possibilities for that."
Xephyr, a KDrive-based nested X Server, would work in this scenario as it targets a host X.Org Server. However, the Xephyr method is un-accelerated (it just uses shared memory) so it doesn't make sense for anything but the most basic tasks.
There was previously some proof-of-concept code by David Airlie to be able to run multiple X.Org Servers on a single graphics card, but that code never matured beyond experimental form and he doesn't have the time to see that materialize in mainline form in the foreseeable future.
See this mailing list thread for the discussion surrounding the feature request of multiple X Servers per output/graphics card. Due to the mailing list software, here's the jump to the August responses.
| 22
| 1,760,738,378.681202
|
https://www.phoronix.com/news/OTc0NQ
|
X.Org Server 1.11 RC2 Is Released
|
Michael Larabel
|
Keith Packard has tagged X.Org Server 1.11 RC2. The X.Org Server 1.11 release is imminent and the RC2 marks the end of development except for critical fixes.
The X.Org Server 1.11 RC2 tag was made via this Git commit. Only critical fixes will be accepted until the release later in the month.
Compared to X.Org Server 1.10, the 1.11 release doesn't add too much to the table besides fixes and other small enhancements. Ubuntu 11.10 isn't even bothering to upgrade to the xorg-server 1.11 release.
Following the X.Org Server 1.11 release, there is then the X.Org Developers' Conference in Chicago.
| 0
| 1,760,738,379.374691
|
https://www.phoronix.com/news/OTY3MQ
|
Two Months To The X.Org Chicago Conference
|
Michael Larabel
|
As I mentioned earlier this week on the X.Org mailing lists, there's just two months left until XDC2011, this year's X.Org Developers' Conference that I've been organizing. XDC2011 is taking place from the 12th to 14th of September in Chicago, Illinois, United States.
Details about this year's event can be found on the X.Org Wiki. If you are an X.Org developer or otherwise involved with and/or interested in X, the sign-up sheet is here.
So far there's a collection of the usual names that Phoronix readers are used to seeing mentioned in regards to Mesa or X.Org: Daniel Stone, Alex Deucher, Peter Hutterer, Jerome Glisse, Martin Peres, Ian Romanick, Jesse Barnes, Keith Packard, etc. I think David Airlie as well (at least a second Red Hat employee from Australia, besides Peter Hutterer), but no word if John Bridgman will be attending to further represent AMD's open-source team. The overall attendance level is expected to be around 60.
If you're not able to attend in person, I'm hoping to offer a Ustream live video stream of the event. However, it might turn into just uploading the recorded audio/videos later on. Regardless, it should be good though as my home/office is only 20 kilometers away from the Chicago venue at the Illinois Institute of Technology.
The program for this year's event will eventually be published on this page.
Details for the evening entertainment (possibly at the Argonne National Laboratory) and recommended beer establishments is still being determined.
If anyone has any questions about the event, please feel free to contact me or the X.Org Foundation board.
The day after XDC2011 ends, it's then onto Munich for Oktoberfest. For the Phoronix readers wondering, in August I will also be at the Desktop Summit in Berlin along with LinuxCon in Vancouver.
| 0
| 1,760,738,379.898609
|
https://www.phoronix.com/news/OTY4MA
|
SystemD/Udev Multi-Seat Support For X.Org
|
Michael Larabel
|
Lennart Poettering has published patches this Sunday to provide support for udev/systemd multi-seat input device hot-plugging support for the X.Org Server.
Details about the multi-seat systemd implementation can be found in this FreeDesktop.org specification. Adding in this support to the X.Org Server presents a new "-seat" X Server switch.
Those wishing to learn about this new X.Org Server multi-seat support can see this mailing list message.
| 31
| 1,760,738,379.907036
|
https://www.phoronix.com/news/OTY2Ng
|
VIA Kernel Mode-Setting Almost Handles LVDS
|
Michael Larabel
|
While the Intel, Radeon, and Nouveau DRM drivers are now mature with kernel mode-setting -- and the Intel Poulsbo KMS driver is even getting ready to leave the staging area -- there's still one fairly common name missing from the desktop Linux KMS scene: VIA Technologies. While VIA defenestrated their open-source efforts and completely blew their TODO list, James Simmons, an independent developer, has basically been the community VIA development source.
There's been some work by the OLPC engineers on a new VIA UMS driver and some dabbling by the OpenChrome developers on their own X.Org driver, among a couple of other individuals, but James Simmons has been effectively the only one working on kernel mode-setting support for the VIA DRM.
He's been making some nice progress on the VIA KMS support and it's nearly done after starting work on it more than six months ago. Today he's written a new status update to the OpenChrome mailing list.
In today's message, James says that he has expanded the VGA/analog support for the VIA CX700M hardware with dual VGA outputs, fixed a kernel module issue, and now has LVDS being detected and nearly functioning. The LVDS display is still blank, but it's nearly working. Simmons has also begun on basic PCI Express TTM handling code, but neither that, nor the LVDS code has yet been committed to his VIA DRM kernel repository.
After the LVDS is working and other code fixed up, James still has to make the OpenChrome DDX driver play with the KMS driver so that it respects the mode-setting being done within the kernel when running an X.Org Server. Needless to say, the VIA kernel mode-setting code still will not be ready for the mainline Linux 3.1 kernel.
It's nice to see though that James is still committed to the VIA kernel mode-setting support upbringing. James was the developer who last year wrote the 3Dfx Linux kernel mode-setting driver, but it was never merged into the mainline tree.
| 3
| 1,760,738,380.56411
|
https://www.phoronix.com/news/OTYwOQ
|
A Number Of X.Org Input Drivers Get Updated
|
Michael Larabel
|
For anyone still relying upon non-kernel non-evdev non-synaptics input drivers for X.Org on Linux, there's a number of new releases. Peter Hutterer has released six updated xf86-input drivers.
Here's the new releases as of this morning:
xf86-input-fpit 1.4.0 - The xf86-input-fpit driver is for legacy hardware, namely the Fujitsu Stylistic Tablet PCs. This new Fujitsu input driver release just brings a bug fix, compatibility with the latest X.Org driver, and various clean-ups.
xf86-input-elographics 1.3.0 - The xf86-input-elographics is for old Elo touch-screen hardware. This X.Org input driver is just for their older touch computers -- i.e. their older VIA Eden systems that are quite slow but physically built quite well and popular in point-of-sale environments (I have lots of Elo experience from one of my other companies). The Elo X.Org driver update just has a few bug-fixes and support for the X Server's input ABI 12.
xf86-input-hyperpen 1.4.0 - The xf86-input-hyperpen driver is for very old HyperPen input devices, which you'll be very hard pressed to find these days. "This driver is one of the legacy input drivers and if you know users of this driver you might want to put them under a display cabinet in a museum because they must have rarity value by now. There has been exactly one fix (in 2007) to this driver since the modularisation in 2003." This release just fixes up support for the latest input ABI and some other minor changes.
xf86-input-penmount 1.5.0 - The driver for old PenMount input devices has also been updated with new ABI support and other minor clean-ups like the other drivers.
xf86-input-mutouch 1.3.0 - The xf86-input-mutouch driver is for select MicroTouch devices. This driver too now works with X.Org input ABI 12 and carries other minor fixes.
xf86-input-synaptics 1.4.1 - Of the driver releases this week by Peter, the xf86-input-synaptics driver should at least see some downloads since it's at least still used and works with some modern Synaptics touch-pads. Still though, this point release isn't particularly exciting. The Synaptics 1.4.1 driver update carries two fixes to syndaemon, an include update, and fixes a manual typo.
| 0
| 1,760,738,380.572208
|
https://www.phoronix.com/news/OTU4NA
|
Say Hello To XWayland & A New X.Org Video Driver
|
Michael Larabel
|
Corentin Chary has announced the release of XWayland and the new xf86-video-wlshm driver.
The XWayland component is what previously was known as X.Org "Hosted" for running an X.Org Server as a Wayland client. WIth the new XWayland system, it's been re-based on X.Org master, updated against the latest Wayland changes, provides new input code that supports the latest xf86-input-* drivers, and improves overall stability.
The new xf86-video-wlshm driver provides a "fake" video driver that doesn't depend upon any video driver-specific functionality but works without any "real video hardware" to provide for easy testing.
Read more in this mailing list message.
| 4
| 1,760,738,381.087016
|
https://www.phoronix.com/news/OTUzNQ
|
VIA KMS Linux Driver Still Far From Being Ready
|
Michael Larabel
|
In the KMS (kernel mode-setting) world there is not only news today to report on a new open-source Freescale KMS driver, but on the state of VIA's kernel mode-setting driver. VIA Technologies may have killed off their open-source strategy, but for the past number of months there's been a developer writing a VIA KMS/TTM DRM driver that would work with the OpenChrome user-space X.Org driver.
The last time really mentioning this VIA KMS work by James Simmons was back in March when the driver was working on the VIA OLPC notebook, but there's been not much more since then nor was there any news when it came time for the Linux 3.0 kernel merge window to be opened and the code could have been pushed.
Following a user inquiry on the OpenChrome mailing list, James Simmons has provided an update on the driver. Simply put, lots of work on this code is left before it will be merged into the Linux 3.x kernel tree. "Going mainline will not be for a while. The main line driver has been in a state of decay for some time. The driver we work on is currently growing since it attempts covers several generations of VIA hardware. So alot of work needs to be done. KMS is being implemented but it's currently turned off by default. Unfortunely turning on KMS gives you the black screen of death. PLL needs to be programmed as well as the display fifo. TTM is mostly implemented. Fencing needs to be done. I plan to finish the TTM layer soon and start working on the xorg driver to use this layer."
James also confirms he hasn't been doing any VIA Mesa work nor hooking in the previous VIA Mesa/DRM work done before by Thomas Hellström, a Tungsten / VMware developer.
| 0
| 1,760,738,381.096383
|
https://www.phoronix.com/news/OTU4Mw
|
VIA OpenChrome KMS Support Is Nearly Done
|
Michael Larabel
|
James Simmons has written a status update to the OpenChrome development list concerning his ongoing work towards enabling kernel mode-setting (KMS) support for VIA hardware with this community-maintained VIA Linux project.
James says that the Linux KMS support for his VIA work is basically done; all of the core code support is in place and VIA KMS is working over VGA. LVDS and other connectors aren't yet working in the VIA KMS world, but James says that it will be easier to do now that the core changes have been made.
Additionally, the OpenChrome X.Org driver still hasn't been updated to play well in a KMS world and to respect the mode-setting being done in the kernel. James right now is deciding whether or not to tackle the output support for other connectors or to first get the xf86-video-openchrome driver playing well.
James's brief status update can be found in this mailing list message. It looks like it will still be a few more Linux 3.x kernel releases before we could see the VIA KMS support merged into the mainline kernel.
| 7
| 1,760,738,381.588166
|
https://www.phoronix.com/news/OTU1MA
|
Synaptics Gets Predictable Motion, Smooth Scrolling
|
Michael Larabel
|
Last week there was the release of a set of patches for X.Org that implement smooth-scrolling support for the X Server. These patches were released by Daniel Stone and today he's now released a set of patches that target the X.Org Synaptics input driver for implementing smooth scrolling, predictable motion, and better acceleration.
For the past several months at Collabora, Daniel has been working on Chromium OS related matters for Google. It turns out that this new input work was done for the Google operating system and is now being released freely for mainline inclusion.
These patches try to make motion as predictable as possible by using hardware time, motion estimation, a better acceleration function, improved palm detection, smooth scrolling, and support for removing erroneous motion using strict filters. This work is applicable to all Synaptics hardware.
There's also a second part of work to the xf86-input-synaptics driver that is said to be more invasive, but he hasn't commented on that work as of yet.
More information can be found in this mailing list message.
| 8
| 1,760,738,381.611908
|
https://www.phoronix.com/news/OTUyNw
|
Smooth Scrolling Targeted For X.Org Server 1.12
|
Michael Larabel
|
In early 2012 we can expect to see the release of X.Org Server 1.12 with various touch improvements and other input related work. Also being queued up for the 1.12 release is support for smooth scrolling.
Daniel Stone has pushed out a set of 27 patches to the X.Org mailing list this morning. These 27 patches all pertain to X.Org input support and are slated to be pulled in for the xorg-server 1.12 release. Coming up in August will be xorg-server 1.11, but we just passed that merge window and so this input work will not land until the 1.12 development begins in late August or early September.
Most of the patches aren't terribly exciting alone, but with providing support for smooth scrolling it becomes interesting. Smooth scrolling, a feature already supported by other operating systems, attempts to remove the "jumpy" movement feel when scrolling down or up a page.
The patch queue starts here.
| 23
| 1,760,738,382.145152
|
https://www.phoronix.com/news/OTUxNQ
|
X.Org Server 1.11 RC1 Has Been Released
|
Michael Larabel
|
Just returning back from a brief holiday, Keith Packard has tagged X.Org Server 1.11 Release Candidate 1.
This release candidate is also coming on time with the merge window having just closed (it actually closed two days late, but that was due to his holiday). The 1.11 release schedule puts the final copy going out in mid-August (19 August), which is just a few weeks before the X.Org conference in Chicago that I am organizing.
Since the release of X.Org Server 1.10 earlier in the year, the 1.11 RC1 adds 30,588 lines of new code while dropping some 35,662 lines of code. Overall though, the 1.11 release isn't particularly exciting. It doesn't have X Input 2.1 with all of its touch features as that's not coming until X.Org Server 1.12 now. On the input side though, support for pointer barriers was added. Peter Hutterer has also proposed merging some of the input work and releasing that as X Input 2.1 to go into this release, but as of RC1 it's not been merged. Xres 1.2 patches might also be merged if they can get cleaned up by the xorg-server 1.11 RC2 release.
RandR 1.4 also hasn't returned. There is though a whole heap of XQuartz changes from Apple.
Also to be found is a lot of bug-fixes and other work, as talked about in the release announcement.
| 4
| 1,760,738,382.867337
|
https://www.phoronix.com/news/OTQ5OA
|
X.Org Server 1.10.2 Brings A Bunch Of Bug-Fixes
|
Michael Larabel
|
Apple's Jeremy Huddleston has just released X.Org Server 1.10.2. This second point release was set to be released yesterday, but then there was fear of a regression causing a delay (turns out it's no longer reproducible), so now we have a holiday weekend release of xorg-server 1.10.2.
X.Org Server 1.10.2 had went through two release candidates and offers a plethora of fixes spanning the GLX, XQuartz, and X Input areas in particular. Apple's using X.Org Server 1.10.x in the forthcoming Mac OS X 10.7 "Lion" operating system, which led them to stepping up to the plate to maintain the stable point releases and for working out the various XQuartz bugs. Mac OS X Lion is set to be released within the next month or two.
The release announcement for X.Org Server 1.10.2 can be found on the mailing list. The next major release of the X.Org Server, version 1.11, is expected to be available in August. The merge window for X.Org Server 1.11 is closing next week. Meanwhile, an X.Org Server 1.10.3 point release will be out in a few weeks (RC1 of 1.10.3 is planned to be pushed in three weeks).
| 5
| 1,760,738,383.435921
|
https://www.phoronix.com/news/OTQ4OA
|
Northern Islands & Fermi Busted On Open-Source
|
Michael Larabel
|
Even with the likely release of the Linux 3.0 kernel, open-source graphics drivers continue to be a big problem for the Linux desktop. While they have improved a lot in recent years, for many Linux users they can cause horrific headaches. Recently it was mentioned on Phoronix that Intel Sandy Bridge is in bad shape for Ubuntu 11.04 and that it even broke upstream in Linux 2.6.39, but Intel's far from being the only driver experiencing a choppy boat ride.
For those that don't follow my Twitter feed, there were going to be another round of AMD and NVIDIA tests in the next couple of days based upon recent advancements in their respective open-source drivers, but there's show-stopping bugs. Tests were planned of the Radeon HD 6000 "Northern Islands" support using the very latest Linux kernel Radeon DRM from the Linus tree, which includes the Gardenshed DRM pull request, and also Git of the xf86-video-ati DDX and Mesa. However, that support continues to be badly broken.
As you can see from the above video, compositing works, but when running any OpenGL program, there's soft lock-ups. This has been broken for months on my end, including today's Git and the Linux 2.6.39 stable kernel. There's this bug report I've written previously for this troubling Northern Islands support, but it's been to no avail. The same problem occurs with both of the Radeon HD 6000 series GPUs I've been testing. Granted, even if the OpenGL acceleration was fully working, the fan on the Radeon HD 6870 is still screaming constantly under the open-source driver. At least there is the AMD Catalyst proprietary driver support and it's in good shape.
When testing out the Nouveau kernel tree for its brand new Fermi FUC microcode replacement, that too was bugged up. As of this morning, on the GeForce GTX 460 it's not working right and there's just lots of PRUNK140 messages being emitted. For better or worse, at least this Nouveau code isn't being primed until the Linux 2.6.41/3.1 release.
What shall attempted to be tested next?
More stories from the Phoronix testing adventures to continue on my Twitter feed.
| 21
| 1,760,738,383.445187
|
https://www.phoronix.com/news/OTQ3NA
|
X Server 1.11 Breaks The Video Driver ABI
|
Michael Larabel
|
While not exactly uncommon for a major X.Org Server update, the video driver ABI for the DDX drivers will break with the forthcoming xorg-server 1.11 release. This means that for those using the proprietary graphics drivers, namely the AMD Catalyst driver, you may be stuck waiting a couple of months for support.
Committed (in this Git commit) to the xorg-server tree this afternoon is a bump to the video driver ABI (ABI_VIDEODRV_VERSION) after developers noticed the video drivers built against the ABI of X.Org Server 1.10 are now producing a segmentation fault when running the latest server code. Developers have tracked down the change causing the seg-fault and it's intentional.
While the open-source drivers are quick to adopt to major version changes of the X Server, it's with the proprietary drivers where it can be a problem, even with a small change such as this break to ABI version 11. Fortunately, at least on the NVIDIA side, they are generally quick to at least put out a beta or pre-release driver to support the new ABI version. In the case of AMD, however, that's where it usually takes a couple of months.
In fact, it's usually not until the next Ubuntu Linux release with the new xorg-server that AMD puts out a supportive driver of that kernel / xorg-server. Even then it's usually not on time to the point where AMD almost always has to pre-release Catalyst drivers to Canonical so their binary blob works in time.
In the case of Ubuntu 11.10, Canonical is planning to still ship X.Org Server 1.10 and not the forthcoming 1.11 series. This means there won't be this video ABI break to worry about this time around, but there may be kernel breaks to deal with as Ubuntu 11.10 will be shipping with the Linux 2.6.40 kernel or potentially even Linux 2.6.41.
Still, I'd not expect AMD to officially support X.Org Server 1.11 until at least this October or November. The upstream xorg-server 1.11 release is slated to occur in August. It would certainly be a pleasant surprise though to see AMD support the newest kernels and X Server releases in a more timely manner.
| 19
| 1,760,738,383.944248
|
https://www.phoronix.com/news/OTQ0Mg
|
X Input 2.1 Not Coming Until X.Org Server 1.12
|
Michael Larabel
|
X Input 2.1 was originally talked about for X.Org Server 1.10 with its initial multi-touch implementation having been published back in late 2010.
After this version of the X Input extension missed the 1.10 cycle, it was getting back on track for a xorg-server 1.11 merge. The multi-touch work has already went through several revisions by Daniel Stone and Canonical.
This work was looking like it would finally land for X.Org Server 1.11 when it's released in August, but it doesn't look like it will make the merge window closing in a few weeks. One of the problems causing a delay in the merge deals with touchpads and where touch/mouse events are delivered to different windows.
Canonical is still working on the tool-kit support and gestures with X Input 2.1 and will continue to provide early back-ported patches into the Ubuntu 11.10 environment. Canonical had also provided early "prototype" patches for their multi-touch / gestures work into Ubuntu 11.04. Daniel Stone is the one driving the upstream work and actual development.
With Ubuntu 12.04 they may not even carry X.Org Server 1.12 depending upon when it's released and its quality, due to 12.04 being a Long-Term Support release. For even Ubuntu 12.04 this may mean back-porting the X Input 2.1 support once again if they will be shipping X.Org Server 1.11 instead.
Wayland was also brought up, but further investigation whether evdev events from the kernel will be enough there. An example Wayland client is needed to better determine its input needs.
| 3
| 1,760,738,383.985534
|
https://www.phoronix.com/news/OTQ0MA
|
Nouveau Gallium3D, LLVMpipe In Ubuntu 11.10?
|
Michael Larabel
|
Here's the next chapter of the X.Org / Mesa plans for Ubuntu 11.10, in continuation of the earlier X.Org / Mesa talks at UDS Budapest.
My key comments from the meeting that just ended include:
- Nouveau Gallium3D will finally be enabled by default, hopefully. For the past few releases it's been optional in the package repository, but now it's finally ready to enter the limelight. Why? Largely because upstream Nouveau developers are willing to look at Gallium3D bug reports, according to Canonical. There's still some concerns by the Ubuntu X developers over the state of the OpenGL driver, but following my comments -- and noting that the Nouveau support can be like a game of Russian Roulette depending upon the kernels -- they'll still likely move forward. In enabling this open-source NVIDIA driver, users could then use the new Unity (3D) desktop without the NVIDIA binary driver. The enabling will likely occur soon for Oneiric but if there's too much fall-out around the time of Ubuntu 11.10 Alpha 3, the feature could be reverted.
- Updates to the proprietary AMD Catalyst and NVIDIA drivers post-release will finally be available! No longer will you be bound to the latest NVIDIA / AMD driver release at the time of the distribution's release. Canonical is looking to provide optional packages via Jockey that user's can opt-in to using that would enable using the latest monthly / stable blobs. This though isn't for open-source driver users since those are harder to update in a sane manner, but there is also xorg-edgers for that as well.
- Mesa 7.11 will be the targeted version for Ubuntu 11.10.
- X.Org Server 1.10 will be used in Ubuntu 11.10, just like in Ubuntu 11.04. Canonical's just being conservative and concerned about the time that it takes NVIDIA and AMD (but particularly the latter) to provide updated xorg-server support if they were to go with the 1.11 release. Granted, updating to the Linux 2.6.40 kernel in the Ubuntu 11.10 driver will likely hose the proprietary driver support anyways. Canonical plans to back-port some bug-fixes and features (e.g. pointer barriers) from 1.11 to their 1.10.x package.
- "Testing days" is being discussed as a possibility for the Ubuntu Oneiric drivers in a similar manner to the Fedora Graphics Test Week. [Granted, upstream may be less interest considering Ubuntu's graphics stack will be out-of-date in relation.]
- I brought forward the matter of replacing Mesa's classic software rasterizer with the LLVMpipe driver. Canonical's position on the matter of using this Low-Level Virtual Machine optimized CPU software driver is to see what other distributions are doing and how it's handled. Red Hat is using LLVMpipe with Fedora 15, and if the Red Hat engineers fix up any outstanding issues in time, Canonical may enable LLVMpipe for Ubuntu 11.10. Otherwise it's still with the (largely useless) Mesa swrast.
- I also brought up whether the Poulsbo KMS driver will be enabled in Ubuntu 11.10. They're waiting to see whether it leaves staging in Linux 2.6.40 kernel and if they have any commercial deployment contracts on Poulsbo, which would use the proprietary driver.
| 9
| 1,760,738,384.55898
|
https://www.phoronix.com/news/OTQwNw
|
There's An X.Org Driver For Nested X Servers
|
Michael Larabel
|
Announced just hours ago on the X.Org development mailing list is recent work to create the xf86-video-nested driver. As implied by the name of the driver and the title of this news post, this is an X.Org video driver designed to run nested X.Org servers. In other words, X.Org on top of X.Org.
When using the xf86-video-nested driver, it's possible to run a new X.Org Server within a program window, similar to running a xorg-server nested within Wayland, but this is still on top of pure X.
With this driver you can also do nesting to multiple levels and the version of the X.Org Server being run isn't too relevant.
Beyond the video working, keyboard/input should work fine too.
The xf86-video-nested driver was previously hosted over at GitHub, but now it's available on FreeDesktop.org. The recent push of it is around 85,000 lines of code.
There's also the mailing list announcement.
| 11
| 1,760,738,384.568348
|
https://www.phoronix.com/news/OTM5OQ
|
X.Org Developers Conference In Chicago
|
Michael Larabel
|
As I just announced to the X.Org mailing lists (I'm the organizer of this year's event), the 2011 X.Org Developers' Conference is taking place in Chicago, Illinois from the 12 to 15th of September.
Initial details can be found on the X.Org Wiki. For easy convenience of developers, it's falling the week after the Linux Plumbers Conference in Santa Rosa. (It's also falling right before Oktoberfest, in fact it ends the night before the start of the yearly Phoronix pilgrimage.)
As always, I'll be posting my notes from the X.Org conference on Phoronix. There will also hopefully be audio/video streaming.
| 8
| 1,760,738,385.088194
|
https://www.phoronix.com/news/OTMzOA
|
X.Org Server 1.10.1 Released
|
Michael Larabel
|
Jeremy Huddleston has tagged the first point release in the X.Org Server 1.10 series.
X.Org Server 1.10 was released in late February after RandR 1.4 was pulled from the release. X Server point releases don't add in any new features, however, but just correct outstanding bugs.
The xorg-server 1.10.1 release has bug-fixes for XQuartz, X Input, XKB, and various other areas, but no single change jumps out as being too prominent.
Jeremy Huddleston, an employee of Apple, is again managing the X.Org Server 1.10 stable series like he had stepped up to do with X.Org Server 1.9. He stepped up to the plate for xorg-server 1.9.x since Apple is shipping it with Mac OS X 10.7 "Lion". Right now with their Mac OS X 10.7 Developer Preview 2 build (benchmarks coming out on Monday), they are using X.Org Server 1.9.5 (after upgrading from 1.9.4 in DP1), but it's unlikely they'll be jumping to xorg-server 1.10.x for Mac OS X Lion.
The release announcement and change-log for X.Org Server 1.10 can be found on the mailing list.
Also in X.Org news, I'll hopefully have new details to announce next week to the X.Org developers for XDC2011 in Chicago. Just waiting back on one of the universities to provide the needed information. Ideally it's going to be 12~14 September (the week following the Linux Plumbers' Conference) in Chicago at the Illinois Institute of Technology, but wait for an official announcement on the X.Org mailing list to know when it's been confirmed and then can begin making travel arrangements.
| 6
| 1,760,738,385.096445
|
https://www.phoronix.com/news/OTI5Ng
|
The SiS X.Org Driver Gets Some Improvements
|
Michael Larabel
|
Patches have been presented on the X.Org development mailing list that provide a number of fixes and improvements to the ill-loved xf86-video-sis, the SiS X.Org driver.
The open-source SiS X.Org driver provides support for the Silicon Integrated Systems IGPs still out there from the SiS 300 through the SiS 760 and SiS 5598, among other ASICs. While there is also the xf86-video-xgi driver out there for the old XGI Tech graphics cards out there, this SiS driver also supports the XGI Volari Z7 and Volari V3/V5/V8 AGP graphics cards (XGI is derived from SiS).
The work presented today is a set of 15 patches published by Ubuntu's Timo Aaltonen. This work is trying to narrow down the difference between the upstream xf86-video-sis and the "sisfree" driver and then the various forks maintained by Intel, SiS, and Mandriva developers.
None of this work presented today isn't particularly exciting, but is more of cleaning up this open-source graphics driver by removing dead X-Video code, among other items and fixes.
This SiS Linux work is also all happening in the SiS DDX driver and there isn't any yet SiS/XGI kernel mode-setting (KMS) support, a Gallium3D driver, or any new and exciting work happening at this point.
The patch set can be found on the X.Org development list until they end up being merged into the upstream xf86-video-sis Git tree, for those still using this hardware.
| 2
| 1,760,738,385.6826
|
https://www.phoronix.com/news/OTIzMA
|
There's More Hope For Mesa & X This Summer
|
Michael Larabel
|
While student registration for this year's Google of Summer of Code (GSoC) has not yet commenced, it's looking quite hopeful for the X.Org / Mesa work this summer. There was an OpenGL 4.1 state tracker that was proposed and some developers are calling this too ambitious. Just days ago there was then a multi-GPU PRIME and hot-switching proposal. This though is not the end of the list.
In terms of the multi-GPU PRIME and GPU hot-switching proposals, David Airlie, who previously has worked on each of these items, is a bit weary. He's not sure if he will be able to mentor this summer and "I was going to drop these from the project list this year as I'm not sure I can wait."
David goes on to add in his message that the PRIME work is being redone and he's already been able to play OpenArena on an NV50 while the display is connected to an Intel Ironlake graphics processor. David is hoping that the Nouveau developers will be able to enable the NVIDIA PCOPY engine and that he'll be able to upstream all of this rewritten PRIME work.
Dave is also concerned about the amount of work that would be needed for GPU hot-switching within the X Server and compositing managers. He also points out that GPU hot-switching on Linux will become more important with USB-based GPU DRM support now set to appear in the Linux 2.6.39 kernel.
There's also a brand new GSoC proposal this morning. An Indian student developer has proposed XCB keyboard support in this e-mail.
| 0
| 1,760,738,386.385912
|
https://www.phoronix.com/news/OTIyMQ
|
Multi-GPU PRIME & GPU Hot-Switching Proposal
|
Michael Larabel
|
Last week a student developer from Belgium had proposed an OpenGL 4.1 state tracker for Gallium3D to be developed this summer as part of the X.Org / Mesa involvement with the annual Google Summer of Code. Under this proposal, OpenGL 4.1 would be implemented from scratch (Mesa / Gallium3D are currently only supportive of OpenGL 2.1 with limited support for OpenGL 3.0 extensions) without any dependence on Mesa; some of the well-known Mesa developers called this too ambitious, but it's unclear if the Belgian developer will still attempt this workload. Meanwhile, a Russian student developer has just voiced two ambitious proposals: Multi-GPU PRIME support and GPU hot-switching.
The Russian student developer, Антонов Николай, is interested in either open-source PRIME multi-GPU support or multi-graphics card hot-switching support to be worked on as this year's Google Summer of Code.
Open-source GPU PRIME support came about a few days over a year ago as an attempt to provide multi-vendor graphics processor offloading / multi-GPU rendering. The PRIME name comes from David Airlie, the author of the original code, dubbing it off NVIDIA's Optimus Technology that was introduced a month prior. Unlike Optimus, PRIME could theoretically work with any open-source graphics driver regardless of hardware vendor. However, the only active work on PRIME lasted for a matter of days and so David looked for someone else to take over this work. Now there may be that chance with the 2011 Google Summer of Code.
The other alternative project that Antonov has expressed interest in is graphics card hot-switching for X.Org. This would be interesting for being able to pop-in a second GPU without blowing out an existing X.Org Server or simply for dual-GPU notebooks to flip from the integrated to discrete graphics seamlessly. It's along the lines of last year's switcheroo work, but more integrated into the X.Org Server for seamless switching.
With these two features, however, there is some display-server-specific work, so any X.Org Server code wouldn't necessarily provide direct benefit to the Wayland Display Server.
There are also other Mesa / X GSoC projects that the open-source developers are hoping student developers will tackle this summer, thanks in large part to Google's sponsorship.
| 7
| 1,760,738,386.984701
|
https://www.phoronix.com/news/OTIwMQ
|
What's Cooking For Mesa & X.Org This Summer?
|
Michael Larabel
|
Summer is quickly approaching in the northern hemisphere so that means it's time for yet another year of Google's Summer of Code. Once again, X.Org / Mesa should be participating, so it's now time to submit ideas for areas where potential student developers could focus their summer work. Here's a few of the possibilities.
One of the GSoC ideas for X.Org this year is to actually focus on Wayland. In particular, the Wayland idea that's been brought up is to work on the remote display capabilities for this promising display server. One of the possibilities for handling remote displays in Wayland is by having a proxy compositor on a remote machine, which would forward the Wayland protocol and buffer updates to the real compositor. Clients running within Wayland would submit where there's damage/changes to the visible surface so that the system is efficient as possible with minimal bandwidth usage. Wayland could be big in 2012.
Ideas for this years summer of code are being collected on the X.Org Wiki but working on this major Wayland feature is the only new idea for 2011 so far this week.
Other ideas proposed in previous years that haven't yet been achieved include an OpenGL 3.2 Gallium3D state tracker, a Cairo state tracker, Gallium3D H.264 video decoding with VDPAU (using shaders, of course), Gallium3D Radeon improvements, automated testing, PRIME multi-GPU support, and even VT switching support for GNU/Hurd and porting DRM to this free software platform.
Let's hope this year's Google Summer of Code for X will be plenty successful (a brief summary of last year's work) with a number of useful projects comign to fruition. Notable X GSoC projects in the past have included Corbin Simpson's work on the Radeon "R300g" Gallium3D driver, Gallium3D XvMC video decoding for Nouveau, R300 GLSL compiler improvements, and kernel mode-setting on vintage hardware.
| 41
| 1,760,738,386.99252
|
https://www.phoronix.com/news/OTE4OA
|
The xf86-video-chrome Driver Is Now "Stable"
|
Michael Larabel
|
The xf86-video-chrome driver, which is the open-source VIA Linux X.Org driver from the One Laptop Per Child project, now says their driver is "stable" after fixing up some rendering bugs.
The xf86-video-chrome driver was forked from the xf86-video-via driver a few months back and it basically strips out all of the old crud (ShadowFB, pre-RandR-1.2 support, etc) and focuses just upon the graphics chipset for the OLPC devices bearing VIA graphics. It also brings its own kernel DRM and does not utilize the recent work for kernel mode-setting and TTM/GEM on VIA, as the VIA Linux community is just badly fragmented with each segment having its own driver it seems.
This driver is now used in OLPC's latest software images and there is now a "stable1" snapshot from their Git repository at Laptop.org. This week they also went ahead and stripped out XAA 2D acceleration support so that only EXA is left.
This message was made on the OpenChrome mailing list. Meanwhile, the VIA TTM/GEM and KMS efforts by James Simmons continues to move along. This much-improved DRM driver can be used in combination with the OpenChrome driver, but the bits that play well with the in-kernel mode-setting and memory management aren't yet mainline.
| 2
| 1,760,738,387.601261
|
https://www.phoronix.com/news/OTE1Nw
|
A Restart For The Botched RandR 1.4
|
Michael Larabel
|
Besides laying out the plans for releasing X.Org Server 1.11 in August, Keith Packard has restarted the discussion surrounding RandR 1.4 so that it will hopefully be readied for integration into this next X Server release. It was part of X.Org Server 1.10 until the last minute when it was pulled from the server and caused a last minute video ABI break.
This is the second failed attempt to pull version 1.4 of the Resize and Rotate extension into the X.Org Server. Originally, RandR 1.4 was supposed to be in X.Org Server 1.9. RandR 1.4 was merged into X.Org Server 1.10 with per-CRTC pixmaps support and other changes, but just last week it was dropped.
In trying to finally get RandR 1.4 out the door with X.Org Server 1.11, Keith Packard initiated this mailing list thread.
Besides bringing the much talked about per-CRTC pixmaps for providing multiple scan-out buffers, RandR 1.4 also adds support for sprite position and image transforms and a request change for CRTC configurations, which can reduce screen flickering and make it easier for NVIDIA's developers to finally implement RandR 1.2+ support within their binary driver. This version though does not provide per-output DPMS controls or any multi-GPU improvements.
| 3
| 1,760,738,387.609877
|
https://www.phoronix.com/news/OTE1NA
|
X.Org Server 1.11 Release Planned For Mid-August
|
Michael Larabel
|
It seems as if the X.Org project has finally formed a habit of wanting to release on time. In years past, even point releases have been more than 200 days late and there hasn't been much to their release schedules that were actually executed on time. It's something I had long pointed out and have received jabs back in turn, but the past few X.Org Server releases have been tagged on time, plus or minus a few days. It looks like X.Org Server 1.11 may be another on-time release, it's at least being planned right out of the starting gate.
Keith Packard has written to the xorg-devel mailing list about the X.Org Server 1.11 release process. Again, he volunteers himself to be the release manager for the major releases.
He also asks how the merge rules for the xorg-server are going and whether any policies need to be tightened or relaxed. Obviously with X.Org Server 1.10 there was the big mistake of pulling in version 1.4 of the Resize and Rotate (RandR) extension code only to realize at the very last minute that it needed to be pulled. This also resulted in a very last minute ABI break for the X.Org Server video drivers.
One of the changes for X.Org Server 1.11 proposed by Keith is to have an official protocol tree maintainer. For this position, Keith has nominated Daniel Stone based upon his X protocol work and voicing interest in the past for handling such work.
In sticking with the same six-month release cycle for the X.Org Server, Keith has laid out some tentative dates for delivering X Server 1.11 on the 19th of August. With this release schedule, the merge window for 1.11 would close on the 27th of May, the non-critical bug deadline would be the 29th of July, and a few release candidates would be sprinkled prior to the middle of August.
In a follow up message, Apple's Jeremy Huddleston has also volunteered to take over the X.Org Server 1.10.x stable point release management. Jeremy initially volunteered for this role with X.Org Server 1.9 stable series as Apple is using xorg-server 1.9 in Mac OS X 10.7, but he has decided to continue with xorg-server 1.10 although we don't suspect Apple will use any 1.10+ release until a "Mac OS X 10.8" release.
Some developers have also asked Keith Packard to communicate better while the merge window is open and he is not responding to emails on the status of patches.
Look for another news posting shortly about X.Org Server 1.11 features. This current release schedule, if met, should allow X.Org Server 1.11 to be in Ubuntu 11.10 and Fedora 16.
| 2
| 1,760,738,388.259682
|
https://www.phoronix.com/news/OTE1Ng
|
VIA DRM Kernel Mode-Setting On The OLPC
|
Michael Larabel
|
While the group behind the One Laptop Per Child (OLPC) child ended up writing their own VIA Linux graphics driver, which is further fragmenting VIA's nasty Linux situation, James Simmons now has his OpenChrome-based VIA DRM kernel mode-setting driver working from the OLPC hardware.
The developer who had been working on 3Dfx kernel mode-setting support for this vintage graphics hardware had turned his attention to VIA TTM memory management and kernel mode-setting a few months back with VIA Technologies failing to deliver any code in this area and the other VIA Linux communities (e.g. OpenChrome) not actively pursuing the work either.
The VIA TTM/GEM kernel memory code is nearly ready and kernel mode-setting has moved along. None of this code has been merged into the mainline Linux kernel (or the drm-next tree for that matter), but now James has this code working on the VIA-based OLPC notebook too.
James is reporting from his new blog that his drm-openchrome kernel is working with the OLPC VIA graphics. However, the OpenChrome X.Org DDX driver doesn't yet recognize the KMS support properly so the kernel and user-space drivers are currently colliding. This though may not be of any direct benefit to the OLPC crew since their xf86-video-chrome driver is derived from VIA's xf86-video-via DDX and not the OpenChrome driver for which James is targeting.
James has also shared he has resurrected work on his 3Dfx kernel mode-setting tree.
| 0
| 1,760,738,388.26821
|
https://www.phoronix.com/news/OTE1MQ
|
Who Did The Most For X.Org Server 1.10? Oracle?
|
Michael Larabel
|
Tiago Vignatti has put out some statistics on the contributions to X.Org development during the X.Org Server 1.10 development cycle, to the xorg-server itself as well as the open-source drivers.
The results can be found on his blog.
Oracle's (former Sun engineer) Alan Coopersmith led with the most change-sets, sign-offs, and reviews. Overall this put Oracle in first place for the most change-set contributions by employer, even beating out Red Hat, Nokia, and Intel.
There were 70 employers involved during this process. When it came to the most changes lines overall, coming in first was actually Matthew Dew, who has been working on cleaning up and organizing the X documentation.
When it came to X input drivers, to no surprise at all, Peter Hutterer had led that work. For user-space video drivers, the work was led by VMware's Brian Paul followed by Intel's Eric Anholt, Vinson Lee (VMware), and David Airlie (Red Hat). In terms of lines changed when classified by employer, the work was led by Intel followed by VMware and then Red Hat.
Ttiago, an X developer at Nokia, then went to comment on the Microsoft-Nokia deal. "I’m sure MeeGo is not dead by any chance though… Nevertheless, Nokia’s contribution to X11 development will be obviously diminishing. It’s sad. Our Graphics Team were just feeling the first effects of the new introduced culture for pushing whatever work (well the ones we are allowed) to upstream and now all was cracked down. So,unfortunately this won’t happen with the same volume anymore and the collected numbers of 1.10 is definitely a mark for Nokia."
| 8
| 1,760,738,388.785462
|
https://www.phoronix.com/news/OTE0OQ
|
Mac OS X 10.7 Lion Is Using X.Org Server 1.9
|
Michael Larabel
|
Last August we reported that Apple's lead X.Org engineer, Jeremy Huddleston, had stepped up to take over stable release management of the X.Org Server 1.9 series. The assumption was that Apple was planning to make X.Org Server 1.9.x available within their forthcoming Mac OS X 10.7 "Lion" operating system rather than the old X.Org Server 1.4 found with Mac OS X 10.6 "Snow Leopard". After trying out the latest Mac OS X 10.7 build, this is indeed the case.
Jeremy keeps pushing out new X.Org Server 1.9 stable updates to bring XQuartz fixes for Apple and various other bug-fixes for users of the xorg-server under Mac OS X, Linux, and other operating systems. With the Mac OS X 10.7 developer preview that Apple began seeding to developers last week, X.Org Server 1.9 can in fact be found.
For X applications on Apple's operating system, this developer preview build is X.Org Release 7.5 with X.Org Server 1.9.4. The build date is 2011-02-19.
Can anyone guess where Apple's Lion will next be found? Yes, Mac OS X 10.7 is making its way to OpenBenchmarking.org.
| 5
| 1,760,738,388.793899
|
https://www.phoronix.com/news/OTEzOA
|
That Was Quick, X Server 1.10 Officially Released
|
Michael Larabel
|
X.Org Server 1.10 RC3 was barely out for 24 hours, but X.Org Server 1.10 was officially released on Friday evening by Keith Packard.
This final build carries just a couple of fixes over RC3, but compared to X.Org Server 1.9 there is a whole lot of work under the hood. Unfortunately it's lacking RandR 1.4 support due to pulling that support out at the last minute. X.Org Server 1.10 doesn't have all of the multi-touch / gestures work either as part of X Input 2.1, but that will come with X.Org Server 1.11 later in the year. To many developers and users pleasure, X.Org Server 1.10 also goes without merging the drivers back into the X Server, as was once proposed by Keith Packard. There is also not threaded input events. What there is though is NVIDIA's first shot at fence sync object support.
At least though this release was delivered nearly on time (just a few days off).
The X.Org Server 1.10 release announcement is available on the X mailing list with a full list of changes since xorg-server 1.9. Petter Hutterer and Adam Jackson dominate the changes this time around.
X.Org Server 1.10 will be found in Ubuntu 11.04, Fedora 15, and many other distributions shipping in the coming months.
| 1
| 1,760,738,389.314797
|
https://www.phoronix.com/news/OTEzMA
|
Scheiße! RandR 1.4 Gets Yanked From X Server 1.10
|
Michael Larabel
|
Only a few days have passed since the release of X.Org Server 1.10 RC2, but another release candidate has now arrived. Given the short turnaround time since the previous release candidate and now being days away from the final release, it's a mundane release candidate, right? Actually, no. RandR 1.4 was just pulled in its entirety from xorg-server 1.10, which also caused the server's video ABI to now be bumped again.
In announcing xorg-server 1.9.99.903, Keith Packard mentions, "RandR 1.4 has been entirely removed from the server. The client interface just wasn't what we wanted, and it wasn't going to be fixed in time for 1.10. Good thing the main RandR 1.4 developer was OK with pending to the next release."
The X.Org Server 1.10 release cycle is a week behind schedule right now and there are still a few more patches that need to be pulled for X.Org Server 1.10 final. This newest release candidate bumps the video driver ABI due to the removal of RandR 1.4 and pushing it back to RandR 1.3. "This prevents drivers built in the last month from running against the 1.10 server. I don't frankly know if the server is back to the full 1.9 ABI or not." The GPU vendors with proprietary Linux drivers will also need to put out new driver releases that are rebuilt for this final 1.10 ABI.
Version 1.4 of the Resize and Rotate extension provided per-CRTC pixmaps, among other changes. It also provided hope that NVIDIA might now support RandR 1.2+ due to these other changes, but now those NVIDIA customers will be waiting until at least X.Org Server 1.11.
The list of other changes in this release candidate can be found in the aforementioned release announcement link. The final release should be out in a week or two.
| 0
| 1,760,738,390.05935
|
https://www.phoronix.com/news/OTExNA
|
X.Org Server 1.10 Is Getting Close
|
Michael Larabel
|
Keith Packard has just announced the release of the X.Org Server 1.10 release candidate. "RC2 has all of the reviewed fixes merged in. The only fixes I've got pending at this point are some RandR DIX fixes that haven't been reviewed yet. Other than that, I think we're pretty much ready for the 1.10 release. Let me know if you've got pending fixes that you want to see merged in 1.10."
X.Org Server 1.10 is expected to be released this month. The RC2 release announcement can be found on the mailing list.
| 0
| 1,760,738,390.634418
|
https://www.phoronix.com/news/OTEwOQ
|
2011 X.Org Foundation Election Candidates
|
Michael Larabel
|
To all registered X.Org members, the 2011 elections are beginning next week for replacing four board member seats this year. The elections will end on the 28th of February for selecting four of six members.
The list of candidates is included below.
- Eric Anholt
- Alan Coopersmith
- Stuart Kreitman
- Bart Massey
- Tiago Vignatti
- Carl Worth
Their personal statements why they are pursuant of becoming a board member is documented on this Wiki page. Two of the candidates are from Intel, two from Oracle, one from Portland State University, and one from Nokia. Though under current board rules, no more than two board members can be associated with one company. Keith Packard of Intel is already sitting on the board, so if these rules stand (Daniel Stone has questioned on the private members mailing list whether this rule is still relevant), one of the Intel candidates may not be qualified. The others serving on the board of directors right now include Alex Deucher, Matthieu Herrb, and Matthias Hopf.
On this Wiki page is a question and answer form that all of these board candidates have been asked to fill out.
| 0
| 1,760,738,390.64252
|
https://www.phoronix.com/news/OTA4NQ
|
X.Org Server 1.9.4 Released; 1.9.5 Expected
|
Michael Larabel
|
Before ending out last week, Jeremy Huddleston released X.Org Server 1.9.4. At least one more release, X.Org Server 1.9.5, is also expected before this branch is retired in favor of X Server 1.10, which will be released in the coming weeks.
X.Org Server 1.9.4 isn't too exciting as it just pulls in about two dozen bug-fixes. There's a couple fixes to EXA, DRI2, EDID, and RandR, but nothing too noteworthy. A bulk of the fixes are by Jeremy himself for XQuartz.
The XQuartz fixes shouldn't be much of a surprise as earlier xorg-server 1.9 point releases were also dominated by XQuartz fixes. Jeremy Huddleston is also an Apple employee and its the first time he has taken over xorg-server release management for the stable branch. It's expected the reasoning behind this is that with Mac OS X 10.7 "Lion", due out later this year, they will provide support for X.Org Server 1.9. With Mac OS X 10.6, their X Server support is still living in the 1.4 series.
The X.Org Server 1.9.4 release announcement can be found in this mailing list message.
| 0
| 1,760,738,391.146522
|
https://www.phoronix.com/news/OTA4Nw
|
xf86-video-v4l Driver Ported To V4L2
|
Michael Larabel
|
One of the seldom mentioned X.Org DDX drivers is xf86-video-v4l. This driver is not for some obscure or vintage hardware like many other drivers in the X.Org tree (e.g. -savage, -s3virge, -tseng, -trident) but it targets no specific graphics card and instead provides an X-Video extension port for video overlay. Just load the module and it works with any hardware supported by a Video 4 Linux (V4L) driver. While this driver isn't often worked on (the most recent commits in Git are from July), it's in the process of being revised.
Red Hat's Mauro Carvalho Chehab has provided a large patch to the X.Org mailing list that migrate the xf86-video-v4l driver from using the V4L1 API to using V4L2. With the forthcoming Linux 2.6.38 kernel, the V4L1 API is removed entirely, which with the latest X.Org V4L DDX means it's broken. To fix this, Mauro has ported the driver to using the V4L2 API.
Video 4 Linux 1 had been around for quite a while (Linux 2.1 development, Linux 2.2) and Video 4 Linux 2 was then officially introduced with the Linux 2.6 kernel while bearing a compatibility mode for older Video 4 Linux 1 software. As Video 4 Linux 2 has been around long enough now, V4L1 support is now being ripped out of the kernel.
Besides just bringing this driver over to the V4L2, Red Hat's work also brings some other improvements to this driver as a result. It's considered a "Major rewrite, as driver got ported to V4L2 API." Looking at the size of the patch, this is indeed a major rewrite.
The patch can be found here.
This isn't all of the Video 4 Linux driver work that Mauro has been up to, but according to this message, there's still more. The xf86-video-v4l driver still uses some deprecated overlays in the X-Video extension and so Mauro is looking at porting the driver now to support Textured Video. This clean-up would result in xf86-video-v4l working with more drivers.
| 16
| 1,760,738,391.164361
|
https://www.phoronix.com/news/OTA4MQ
|
Yet Another VIA Linux Driver Has Arrived
|
Michael Larabel
|
VIA's small Linux development community is badly fragmented; there is yet another group of developers creating their own VIA driver. I wish it was a joke, seeing as there are already a number of drivers for the same VIA chipsets and none of them are feature complete or in really great condition, but a new driver has been released. This time the new driver comes from the OLPC (One Laptop Per Child) crew and it's just being dubbed xf86-video-chrome. Not only though is there yet another X.Org driver, but it's bringing its own kernel DRM.
While VIA Technologies has officially killed their Linux / open-source strategy, there are bits of documentation available and there's a few people out there like James Simmons who are working on open-source VIA support in a useful manner, by doing things like hacking on GEM/TTM and KMS support. This new driver is based upon the work reported on last December, but now there's more to it, including its own Direct Rendering Module for the Linux kernel.
This work was announced today by OLPC's Daniel Drake on the OpenChrome mailing list with the following title: Announcing OLPC's work on new "chrome" VIA video driver.
Though at least this driver does a bit more than some of the other VIA drivers out there, but it doesn't have any TTM/GEM memory management support nor for kernel mode-setting. What this driver does ship with is Chrome 9 hardware acceleration, support for multiple X-Video surfaces, RandR 1.2 support, and accelerated rotation support. There is also still no Mesa / Gallium3D driver.
While it looks like this driver is in better shape than some of the VIA driver alternatives, the developers are just concerned about supporting the VIA chipset found in the OLPC laptops. "Our resources are limited, and we focus on our platform so we can not be too attentive to bug reports from outside our platform." The driver should have just been called xf86-video-olpc-chrome or something to that effect. The code is also being hosted over on OLPC's Laptop.org infrastructure.
The code for the xf86-video-chrome DDX is based upon that of code previously released by VIA and then Jon Nettleton gutted out as much code as possible for older features like ShadowFB, non-RandR support, and other code before releasing it as a new driver.
The VIA DRM kernel code that OLPC is releasing seems to be mostly a re-branded version of the DRM code VIA previously release and unsuccessfully has tried pushing into the Linux kernel. Meanwhile, James Simmons ongoing GEM/TTM memory management and KMS work is going into a different DRM module and the DDX driver being hacked on to play with the memory management and kernel mode-setting is the OpenChrome driver. There's also the xf86-video-unichrome, xf86-video-openvia, and xf86-video-via drivers floating around elsewhere in the VIA Linux pit.
| 17
| 1,760,738,391.839728
|
https://www.phoronix.com/news/OTA3Mg
|
New Version Of VIA DRM TTM/GEM Patches
|
Michael Larabel
|
Thanks to James Simmons, an independent developer in the open-source community, last month a patch was published that adds TTM/GEM memory management support to the VIA DRM Linux kernel driver. This was after VIA basically admitted defeat for their Linux / open-source strategy. Over the weekend the second version of this TTM/GEM patch was published by James.
This TTM/GEM patch for VIA hardware in the second revision now works when unloading the VIA kernel module or restarting the X Server, moves all resource handling over to TTM (Translation Table Maps), utilizes the GEM creation ioctl, reworks the detection of resources, and reworks the video memory (VRAM) detection to use the VIA north-bridge for detecting the memory and capacity.
Still lacking from this open-source VIA Linux code is accelerated bmoves, fence object handling (needed for 3D), and support for VIA PCI DMA GART. James and the OpenChrome developers also need to decide how to proceed on the DDX driver side, which is described in this mailing list post.
It looks like this work will actually move forward and should reach the mainline Linux kernel at some point.
| 4
| 1,760,738,391.848299
|
https://www.phoronix.com/news/OTA2MA
|
Videos Of The Graphics Talks From Linux.Conf.Au 2011
|
Michael Larabel
|
The videos from this year's Linux.Conf.Au conference are now available online. There were a few graphics talks by Keith Packard, Eric Anholt, and Adam Jackson, which can be found embedded below.
Keith Packard's talk was about X and the future of Linux graphics. It basically went over the history of X, kernel mode-setting, the Graphics Execution Manager, etc. No major breakthroughs from this talk if you've been keeping up with all of the Phoronix news items, but for everyone else, the nearly 40 minute video is embedded below.
Eric Anholt talked about the GLSL (GL Shading Language) compiler work they have been doing to Mesa over the past year.
Red Hat's Adam Jackson talked about doing remote graphics on Linux with X, SPICE, VNC, etc.
All of the Linux.Conf.Au talks can be viewed at Blip.TV.
| 16
| 1,760,738,392.354458
|
https://www.phoronix.com/news/OTA0OA
|
The New X Stack Is Going Into Ubuntu 11.04
|
Michael Larabel
|
For those following closely the development of Ubuntu 11.04, in the process of going into the Ubuntu Natty repository this week is its new X stack. This means the latest snapshot of the unreleased X Server 1.10 and Mesa 7.10 for the open-source DRI / Gallium3D drivers. Due to the usual API/ABI breaks, this also means updated X.Org drivers are also going into the "Natty Narwhal" repository at the same time.
This updated Linux graphics / input stack has been in testing the past few weeks via a Personal Package Archive (PPA) and is now deemed ready for the mainline repository so that all of the new bits can be found in Ubuntu 11.04 Alpha 2 when released next week.
Canonical's Bryce Harrington warns of turbulence ahead in this mailing list message for the Natty repository over the next few days due to merging these X packages and the breaks it may cause.
The main X.Org driver updates are of course the xf86-video-intel, xf86-video-ati, and and xf86-video-nouveau drivers along with xf86-input-synaptics, xf86-input-evdev, and the rest. Canonical has also made the decision to strip away several vintage X.Org drivers from the main repository, including the APM, Ark, Chips, Glide, i740, Tseng, and Voodoo drivers. These drivers aren't too well maintained any more and are rarely used by Ubuntu users. For right now these dropped drivers will be available for manual installation via the Natty Universe repository.
For the time being this package pull also means the ATI Catalyst driver will not work with Ubuntu 11.04 but only the open-source ATI/AMD Radeon driver. The Linux 2.6.37 kernel support is now in place with today's Catalyst 11.1 release, but the X.Org Server 1.10 isn't yet in the public driver. It will also need to support the Linux 2.6.38 changes too if Ubuntu Natty begins tracking that upcoming kernel. NVIDIA's new beta driver meanwhile should play fine with these newest bits.
Look for Ubuntu 11.04 Alpha 2 with these changes by the end of next week while the stable release will be out in April.
| 10
| 1,760,738,392.362853
|
https://www.phoronix.com/news/OTAzMg
|
A Guide To How Graphics Cards Work
|
Michael Larabel
|
This entry on the X.Org Wiki isn't brand new, but for those that have yet to see it, there is a development guide to how graphics cards work on this Wiki page. There was just a trivial update to the guide today and I had then realized it hasn't been mentioned before on Phoronix.
This technical guide is intended for those interested in getting into graphics driver development with Linux / X.Org. The key sections include the video RAM, display control (CRTCs, PLLs, outputs), the 2D engine (solid, blit, Xorg acceleration), 3D engine (overview and buffers), the overlay, hardware sprites, PCI, AGP, PCI Express, and apertures coverage. There are also driver examples by referencing functions within the open-source X.Org drivers.
Beyond the X.Org guide, there's also been other documentation in the past that's been scattered around the web in hopes of easing the process in getting into open-source GPU driver development on Linux. There's also this Radeon driver guide by Alex Deucher and this DRM development documentation covering DRI2, KMS (kernel mode-setting), GEM (Graphics Execution Manager), etc.
| 2
| 1,760,738,392.873496
|
https://www.phoronix.com/news/OTA0Ng
|
A New QEMU QXL Driver Uses UXA
|
Michael Larabel
|
Soren Sandmann Pedersen has announced the release of the xf86-video-qxl 0.0.13 driver. The QXL X.Org driver isn't commonly talked about at Phoronix like the ATI/AMD, Intel, and Nouveau Linux drivers, but this is the driver used for the QEMU para-virtualized guests with the QXL Virtual GPU that is found in SPICE, the Red Hat Enterprise Virtualization System. This driver brings some semi-interesting changes.
The xf86-video-qxl 0.0.13 driver begins hooking in UXA acceleration, the EXA-derived GEM-based acceleration architecture conceived by Intel for their driver. This QXL driver update also has a faster memory allocation (malloc) implementation, better debugging support code, and a number of bug fixes.
As UXA, the UMA Acceleration Architecture, isn't integrated into the mainline X.Org Server, the xf86-video-qxl driver adds its own copy of the UXA files to the driver, which tacks on around 6,400 lines of replicated code. This acceleration architecture was originally introduced by Intel in mid-2008 after they abandoned plans to advance EXA so they kept the EXA API while gutting it internally to hook into their Graphics Execution Manager API for pixmap memory management, etc. The Radeon and Nouveau drivers continue to use EXA for their 2D acceleration architecture.
The xf86-video-qxl 0.0.13 driver release announcement with a full list of changes can be found on the mailing list.
| 2
| 1,760,738,393.5707
|
https://www.phoronix.com/news/OTAyMg
|
The VIA TTM/GEM Patch Appears Ready
|
Michael Larabel
|
Just one month ago an independent developer began working on VIA TTM/GEM support for the VIA kernel DRM driver along with VIA kernel mode-setting support, even while VIA's open-source Linux strategy is dead. Just a few weeks later, James Simmons' VIA TTM/GEM memory management patches are now ready.
James has written to the OpenChrome mailing list and has sent along a patch that provides this readied TTM (Translation Table Maps) and GEM (Graphics Execution Manager) support for the limited DRM (Direct Rendering Manager) driver living within the Linux kernel.
This patch is set to go against David Airlie's drm-next branch of the kernel. This patch that plugs-in the new memory management back-end will still work with existing OpenChrome UMS (user-space mode-setting) installations. The driver works, but there's a few problems left to be solved, such as a kernel oops when unloading this DRM module. It can also be tried with the KMS branch to benefit from kernel mode-setting.
The patch can currently be found here. In just about one month, with one developer getting TTM/GEM support hooked in, it's sad that VIA couldn't get it done back when they had Linux plans in 2009 and still the TTM/GEM support was months out or that it couldn't be tackled by any of the OpenChrome community developers. Let's see though if this patch will be fully ironed out and make it into the Linux 2.6.39 kernel or another release in the near future.
| 6
| 1,760,738,394.070723
|
https://www.phoronix.com/news/OTAyOQ
|
Pixman Continues With ARM Performance Improvements
|
Michael Larabel
|
The October release of Pixman 0.20 brought performance improvements, particularly when using this pixel manipulation library for X and Cairo under ARM platforms. A month later there were more ARM optimizations in the first development snapshot for Pixman 0.22. The second development snapshot for Pixman 0.22 has arrived this morning, and guess what? It brings more ARM performance improvements.
Siarhei Siamashka, a senior software engineer at Nokia in Finland, continues to viciously focus on ARM performance improvements, particularly for the ARM NEON. There's also bug-fixes, but the ARM optimizations from Siarhei come in the form of adding various fast paths that are specific to the ARM NEON and taking available of its available CPU features.
Aside from the ongoing ARM work from Nokia, there are also performance improvements for bit-map filling in Pixman 0.21.4. There's also improvements to their test suite and also refactoring of the image access code.
The Pixman 0.21.4 release announcement can be found on xorg-announce for this library that plays a critical role to X and Cairo.
At the same time as delivering this new development snapshot, Soeren Sandmann also announced the stable release of Pixman 0.20.2. The release announcement can be found here, but it's not particularly exciting unless you're impacted by one of the bugs that this point release fixes.
| 7
| 1,760,738,394.089889
|
https://www.phoronix.com/news/OTAwMw
|
The VIA Kernel Mode-Setting Code Progresses
|
Michael Larabel
|
While VIA defenestrated its open-source Linux graphics driver strategy, there has been some recent work under-way on providing a GEM/TTM + KMS driver for VIA's integrated graphics processors by the community. In particular, this work is being done by James Simmons, the former Google Summer of Code student developer who was working on 3Dfx kernel mode-setting support a few months back.
We reported on the progress as of last week, but some more work has been accomplished since. James is now sharing that he finished his KMS / UMS (user-space mode-setting) code reordering and so far has found no fall-out from these changes. However, the actual KMS support has yet to be added. Once the kernel mode-setting support is found within the DRM driver, the OpenChrome-derived X.Org driver will be updated to properly play with the VIA Linux KMS support. This KMS support though for VIA hardware will not be enabled by default as this would break things for users of newer VIA-KMS-capable Linux kernels but with an outdated DDX driver.
This update was shared on the OpenChrome mailing list. There is not yet any word on when we may see the improved VIA DRM driver with GEM/TTM memory management and kernel mode-setting support enter the mainline Linux kernel.
| 2
| 1,760,738,394.655684
|
https://www.phoronix.com/news/ODk4Nw
|
VIA's Open Linux Graphics Driver Has Been Defenestrated
|
Michael Larabel
|
For those that were hoping that VIA Technologies would pull through in providing their open-source graphics driver support like they had promised with kernel mode-setting, a Gallium3D driver, and being Linux friendly, kiss those thoughts goodbye as they've been basically thrown out the window. Sadly, it's not happening. I had a very productive conversation with VIA's Stewart Haston, who is their international marketing specialist, and their Linux outlook is extremely dark.
Stewart was showing off the new dual-core VIA Nano X2 CPU (under Windows), but he saw my badge and was extremely upfront about VIA's Linux matters. In fact, he can't really object to the recent Phoronix VIA articles. VIA Technologies as a company has shrunk and they simply don't have the resources or need to follow through with their once hopeful Linux support. About the only official Linux support that we may see out of VIA going forward is if any of their embedded customers specifically need support for Linux in a given configuration. Some of their motherboard chipsets and CPUs may work fine with Linux, fortunately, but in terms of graphics enablement, there won't be much to see.
Stewart did mention that Bruce Chang is still with the company (he's the one that's basically been the face of VIA's Linux strategy since it was announced in 2008) but he said that he's basically "thrown in the towel" and it's not really clear that there is any strategy left.
This though shouldn't come as a huge surprise since we've said that the Linux support is basically dead and what's left of the VIA Linux community is fragmented. There has though been some interesting community work in recent days in particular with kernel mode-setting and GEM/TTM support being written by a student developer, but without VIA's support for either code or documentation on new hardware and for features currently undocumented on existing ASICs, this support won't exactly be able to grow going forward.
If there is some good news out of the situation, VIA has promised that he will get VIA Taiwan to answer some questions. So if any of you have any questions left about VIA's remaining thoughts on Linux, pleas for them to open-source their S3 Graphics driver, ideas for how they can provide better Linux support without allocating any additional resources, or anything else to say, use the comment link below so we can get this Q&A going.
| 39
| 1,760,738,394.664133
|
https://www.phoronix.com/news/ODk4MA
|
VIA Releases The Dual-Core Nano X2 CPU
|
Michael Larabel
|
While VIA isn't working on their Linux graphics support, they are continuing to design new hardware. For the Consumer Electronics Show in Las Vegas, VIA Technologies has just released their Nano X2 Dual-Core CPU.
The VIA Nano X2 features two x86 cores with support for x86_64 instructions, SSE4, VT CPU virtualization, PadLock hardware security, and is also pin-to-pin compatible with VIA's older Nano, C7, C7-M, and Eden processors.
The press release and other details on their newest CPU can be found from the VIA Press Room. This also comes right after they launched their first discrete graphics card.
| 4
| 1,760,738,395.147954
|
https://www.phoronix.com/news/ODk3OA
|
VIA KMS + TTM/GEM Driver Moves Along Without VIA
|
Michael Larabel
|
At the end of December we reported on the 3Dfx KMS Linux developer working on VIA code to provide kernel mode-setting support for VIA's IGPs in the Linux kernel and thus TTM/GEM memory management support too. This is after VIA had promised to deliver this support (along with Gallium3D support) in 2010, but failed miserably. This code though is now moving along but without any support for VIA.
James Simmons is the developer that has been working on this code and he's now saying it needs a home for hosting it in Git along with the other X.Org drivers rather than the current OpenChrome infrastructure of using SVN and on their own servers. James is making progress on the kernel side with this DRM driver and then also with his OpenChrome DDX driver that supports both KMS (kernel mode-setting) and UMS (user-space mode-setting) paths.
Surprisingly, VIA's Bruce Chang has not commented on the matter yet of the community coming up with this support after VIA's failed attempts. Bruce still hasn't responded to my inquiries from months ago either, which begs the question if he's even with the company any longer. Fortunately, I'll see VIA at lunch tomorrow.
| 4
| 1,760,738,395.182726
|
https://www.phoronix.com/news/ODk1OQ
|
The Fourth Version Of X.Org Multi-Touch Support
|
Michael Larabel
|
Earlier this month we reported that the X.Org multi-touch work was nearing completion and now this work is getting even more readied for X.Org Server 1.11 once its merge window opens in February. Daniel Stone has today put out a fourth version of these X patches that provide proper multi-touch support to Linux and other operating systems running X.Org.
The fourth version of these multi-touch support patches clear up feedback generated from earlier revisions and touch not only the X.Org Server but inputproto, libXi, X Input, and xf86-input-evdev too. The multi-touch work had stalled for a bit, but Daniel Stone is back working on this code.
Multi-Touch in X.Org is to be part of X Input 2.1 and should be officially introduced with X.Org Server 1.11 in the summer of 2011. There are though some stopgap solutions for multi-touch to be found in Ubuntu 11.04 and other Linux distributions until the proper support lands upstream.
Those interested in the Git tree information for this work and the technical discussions among the open-source input developers, see this mailing list thread.
| 0
| 1,760,738,395.693064
|
https://www.phoronix.com/news/ODk1OA
|
The 3Dfx KMS Developer Now Working On VIA TTM
|
Michael Larabel
|
With VIA not really doing anything for open-source and Linux as all of their efforts seemed to have stalled, the small open-source development community centered around VIA has become quite fragmented as we have talked about multiple times now. There's multiple X.Org drivers for VIA, with not a single one clearly dominating or being feature-complete and well maintained, while other areas like the DRM/KMS and Mesa/Gallium3D support are just in shambles.
VIA was supposed to have their Linux kernel DRM driver with TTM (Translation Table Maps) / GEM (Graphics Execution Manager) memory management support done more than six months ago, according to the VIA Linux TODO list. But that hasn't happened nor are there any signs of that officially happening by them. Two years ago there was work by Tungsten Graphics (now part of VMware) on a VIA TTM/DRM driver with a Mesa 3D component too, but that was never finished. Now there's a new developer, James Simmons, beginning work on VIA DRM code in the form of providing TTM memory management support in the kernel for these VIA IGPs.
James Simmons is the student developer who over the summer began work on creating a 3Dfx KMS driver. While vintage 3Dfx graphics cards still in use are very rare, James worked on the 3Dfx DRM/KMS driver with TTM support as a way to experiment with and learn about kernel mode-setting on Linux and to be able to write proper documentation concerning the process.
The 3Dfx KMS driver has not been merged into the mainline Linux kernel, but recently he picked up a new motherboard with an AGP slot to continue work on this driver. It just so happens this AGP motherboard also has an integrated VIA Chrome IGP chipset too, which has now drawn his attention.
"This board comes with a built in VIA based IGP. Well that chipset is poorly supported so I started to work on the via drm driver. Currently I'm porting it to the TTM infrastructure," he wrote to the xorg-devel mailing list.
James mentions this VIA TTM memory management work as he had some questions/comments about this in-kernel GPU memory management interface used by the open-source Linux drivers aside from Intel's driver. Hopefully some of this VIA TTM work will end up going upstream at some point.
| 8
| 1,760,738,395.701469
|
https://www.phoronix.com/news/ODk0MQ
|
Oh Hey, X.Org 7.6 Is Finally Released For Christmas!
|
Michael Larabel
|
The past few days I've been wondering whether or not X.Org 7.6 would make it out in 2010 or not. After all, this X.Org katamari update was supposed to be here in August after X.Org 7.5 was released in October of 2009 and the 7.6 release was delayed to November. The release of X.Org 7.6 didn't come in November, but there was one release candidate but not much information since. This afternoon, however, Alan Coopersmith has announced the final release.
While the X.Org 7.6 release is late, fortunately it's not too important since most of the individual packages making up the X.Org 7.6 katamari have been available for months. X.Org Server 1.9 has been available since August along with many other updated X packages.
X.Org 7.5 was released in 2009 with X.Org Server 1.7 and since that point there was the X.Org Server 1.8 release at the beginning of the year too. Thus in X.Org 7.6 is now officially support for the input class with the xorg.conf configuration, xorg.conf.d support, new input hot-plugging, DRI2 improvements, documentation updates, and much more.
X.Org 7.6 is also the first katamari collection that includes XCB (the X C Bindings) by default as it's required by the updated libX11 and other packages.
Dropped from X.Org 7.6 is Xsdl, which was a KDrive-based server using SDL that was never completed, and also stripped away was frame-buffer support in XF86DGA and multi-buffer extensions in the X Server.
Additional information is on the X.Org Wiki.
X.Org 7.7 will be the next release and should arrive in approximately one year's time with X.Org Server 1.10 or more likely would be X.Org Server 1.11. X.Org Server 1.10 is already schedule for release in February with many new features.
| 5
| 1,760,738,396.261053
|
https://www.phoronix.com/news/ODkzMQ
|
VIA Fails With KMS/3D, But Has Yet Another X Driver
|
Michael Larabel
|
One year ago VIA came out with their Linux TODO list, which was disappointing. This list had a VIA TTM/GEM memory manager module for Q2'2010, a kernel mode-setting driver in the works for H2'2010, and a Gallium3D driver in-development for Q4'2010. Even meeting this TODO list would be bad as the support most Linux customers are after (3D and KMS to a lesser extent) would not be arriving until three years after VIA announced this newest Linux strategy. But, VIA has failed miserably in accomplishing any of these mile-stones for KMS and open-source 3D acceleration support. Though resulting in VIA's Linux community being fragmented even more, new VIA X.Org (DDX) drivers seem to keep popping up. If there wasn't already enough of these not-fully-working and rarely-touched open-source drivers, another VIA Chrome X.Org driver has been started recently that's a fork of another open-source VIA driver.
It was bad enough on the ATI/AMD side when there were just three drivers: the proprietary Catalyst driver and then the two open-source X.Org drivers: xf86-video-ati and xf86-video-radeonhd. Here at least there were different approaches in the open-source drivers (i.e. AtomBIOS vs. not), different hardware support (xf86-video-ati supporting pre-R500 ASICs where RadeonHD did not), and a semi-different feature-set. The RadeonHD driver ultimately ended up having a short life, but at least it was worked on, both groups of developers ended up learning things from each other, and the importance of both drivers were reduced when the Radeon kernel mode-setting driver was merged into the mainline Linux kernel. The open-source VIA situation only seems to degrade from being a horrible mess to a support catastrophe.
Earlier this year there was work on another open-source VIA driver being called xf86-video-openvia. It seems work on the OpenVIA driver has now been abandoned by its lead developer in favor of another one simply being called xf86-video-chrome. The xf86-video-chrome driver is just the latest, but it's not simply a rename or an evolution of the OpenVIA driver, but rather it's a restart.
The xf86-video-chrome repository (at GitHub) only goes back to the end of October when the work began. The xf86-video-chrome driver is based upon VIA's official open-source code (xf86-video-via), in particular the via-xserver-87a-55727_src package (it can be found on the VIA Linux web-site). This is the driver package with VIA and S3 Graphics copyrighted code going back to 1998 and 2001, respectively, with contributions as well from VA Linux Systems and Precision Insight. The open-source driver supports the CN700, CX700, VN896, VX800, and VX855 chipsets.
So what does this new xf86-video-chrome driver provide on top of this? Well, not much. In fact, at this point, it just removes more than it adds. The xf86-video-chrome driver removes support for the following features: MergedFB, ShadowFB, Xinerama, non-RandR, legacy RandR, down-scaling, excessive debug outputs, and other legacy functions. What the xf86-video-chrome driver then provides is a cleaner code-base by stripping out a lot of code and then doing some "general clean-up" to the open-source driver. Aside from that, it just adds 1200 x 900 LCD mode-line support and support for PLL quirks, which amounts to a few new lines of code. That really summarizes the xf86-video-chrome driver at this point.
The xf86-video-chrome driver now joins xf86-video-unichrome, xf86-video-openvia, xf86-video-via, and xf86-video-openchrome as all being X.Org display drivers for the VIA hardware (there may be other forks/branches too), none of which drivers are full-featured and all while there's still no proper OpenGL/3D acceleration under Linux for VIA hardware going into 2011. Intel's Poulsbo Linux situation may too be a bloody mess and involve binary blobs, but at least there the efforts are more concerted and there's working configurations complete with 3D.
For a while, VIA appeared to be tossing its limited Linux efforts behind the OpenChrome driver, but its development list remains eerily quiet and its SVN repository is far from lively. The OpenChrome driver is expected to provide its DDX support for the new OLPC XO 1.5 device, but that isn't merged yet. VIA's Linux cheerleader, Bruce Chang, also no longer seems to be active with their Linux support efforts.
Will 2011 be any different for VIA Technologies on Linux? We'd like to think so, but more than likely it won't be any better. The Linux 2.6.38 kernel merge window is soon opening up and with VIA (or any others) not yet producing any TTM/GEM patches, which are needed to foster any kernel mode-setting and Gallium3D driver work, this pushes back VIA's Q2'2010 TODO list back at least to being Q2'2011 before there could possibly be mainline VIA kernel memory management. Based upon how long it's taken the open-source Radeon and Nouveau (for NVIDIA hardware) driver teams to get where their at, it really is unlikely there could be any viable open-source KMS/3D support for VIA prior to 2012 and that's being optimistic. By that time, if the company is still around, they'll probably need to begin work on bringing up support for their newest generation of chipsets.
| 11
| 1,760,738,396.998996
|
https://www.phoronix.com/news/ODg3Mg
|
RandR 1.4 Brings Per-CRTC Pixmaps; NVIDIA Support?
|
Michael Larabel
|
While version 1.4 of the Resize and Rotate (RandR) extension was supposed to be introduced with X.Org Server 1.9 (to the point that it delayed closing their merge window), it never made it. RandR 1.4 and its per-CRTC pixmaps once again delayed closing the merge window, but this time it's for X.Org Server 1.10 and its promised to only be a few days. Fortunately, it looks like Keith Packard was right this time and RandR 1.4 is about to land.
Intel's Keith Packard wrote a few emails to the X.Org developers over the night commenting on his per-CRTC pixmap implementation for RandR 1.4 in xorg-server 1.10. For those unfamiliar, this support basically provides, "multiple scan-out buffers which applications can create and assign to arbitrary collections of CRTCs. These pixmaps can be associated with a window for use with OpenGL or drawn to directly." This feature really becomes useful when dealing with display setups where the screen layout exceeds the maximum size of the rendering/scan-out engines, provides the abilities for integrating compositing and project transformation into one step, and eliminating visual artifacts during screen rotation.
The technical details behind this xorg-server/RandR implementation can be found in this message. As part of this work, Keith also added support to RandR 1.4 for sprite transforms This will make per-CRTC pixmaps more useful to extended and/or rotated desktops. The technical side of this implementation is documented in this email.
As another change done by Keith last night, with RandR 1.4, NVIDIA may finally be more prone to supporting RandR 1.2+ in their proprietary Linux/Unix graphics driver. While NVIDIA promised proper RandR 1.2 support for their binary graphics driver two years ago, it has yet to come with any screen manipulation still really needing to be done through NVIDIA's extensions where implemented like with their NVIDIA Settings panel. However, last night Keith commented on a discussion (the message) he had with one of NVIDIA's Linux engineers some years ago.
NVIDIA's position was that they could not support the Resize and Rotate extension since their driver couldn't access the mode-setting instructions provided by RandR as a whole, but rather each mode-setting step is submitted to the driver individually. Keith has changed this in RandR 1.4 by adding a new RandR request (RRSetCrtcConfigs) that allows the sending of all instructions over to the X driver in one step. Hopefully now we will see RandR 1.2/1.3/1.4 support finally coming to NVIDIA's proprietary Linux driver in the near future.
| 7
| 1,760,738,397.530998
|
https://www.phoronix.com/news/ODkyOQ
|
X.Org Multi-Touch Nears Completion
|
Michael Larabel
|
For X.Org Server 1.11, to be released likely in late summer of 2011, there will be proper multi-touch support. For quite a while now there's been work under-way largely by Daniel Stone and then more recently by Canonical. A few weeks back Daniel Stone returned to the multi-touch party to get X Input 2.1 with multi-touch support beaten into shape. It's now nearly ready.
Daniel has published his third revision of these patches to the X mailing list. He also writes on his blog, "[I] have finally posted the third patch series, which I think should be pretty close to final." In that blog post he goes into details how touch events with X Input 2.1 are handled and compared to standard X Input events.
The major difference here is obviously the event delivery: instead of delivering only to the first applicable target, we now deliver to allm grabbing clients, and the first selecting client. An 'owner' flag denotes whether or not the client currently owns the stream: grabbing clients may accept the touch (which ceases delivery to all other clients), or reject it (removing them from the delivery list and passing ownership on to the next client). X Input 2.1 Multi-Touch is the new solution rather than Canonical's proposed X Gesture Extension. Nokia developers are already working on supporting the new multi-touch capabilities within the Qt tool-kit and there is a GTK multi-touch branch on the GNOME side, but it's currently based upon an earlier specification of X multi-touch.
This work not only touches the xorg-server itself, but also the inputproto, libXi, and X Input components. It will not be in X.Org Server 1.10 due out for release in February, but all indications at this point are that it will be more than ready for its successor, X.Org Server 1.11.
| 4
| 1,760,738,397.540299
|
https://www.phoronix.com/news/ODg3OQ
|
The First X.Org Server 1.10 Snapshot Brings Some Fun
|
Michael Larabel
|
Following the hiatus last week with the X.Org Server 1.10 merge window being kept open to allow time for finishing up RandR 1.4 with per-CRTC pixmaps and then NVIDIA pushing for fence sync support in this release, the work has now settled and the merge window has closed. Keith Packard has also announced the first development snapshot of X.Org Server 1.10.
Up until the beginning of last week it looked like xorg-server 1.10 would contain just a lot of bug-fixes with no appearance of X Input 2.1 / Multi-Touch or any other major features, but then a lot of work ended up being pushed at the last minute. Keith wanted to get the per-CRTC pixmap support for the X Server finally finished after it was previously delayed from X Server 1.9. The per-CRTC pixmap support did hit the mainline tree this weekend as part of version 1.4 of the Resize and Rotate (RandR) X extension.
What per-CRTC pixmap support means and its benefits along with other RandR 1.4 changes are talked about in this Phoronix news post. RandR 1.4 also brings sprite transforms and changes that should make it easier for NVIDIA to properly support RandR 1.2+ functionality within their proprietary graphics driver for Linux/Solaris/BSD.
NVIDIA's Fence Sync support was also pulled into the X.Org Server tree yesterday before the feature merge window finally closed. While these NVIDIA-supplied open-source patches have been available since early this summer, it was not until the last minute that there was a technical discussion and some debate (where NVIDIA working on a new driver architecture was also brought up) over the fence sync design. In the end, however, these patches were merged in a way to please all parties the best as possible.
These are the major improvements, besides the bug-fixes, to be found in X Server 1.10. There's also been patches available for months that provide threaded input events for the X.Org Server and are in a state nearly ready to be merged, but alas it hasn't been merged. Tiago Vignatti, the principal developer behind the threaded input work, has been burned out from it (or is still in Toulouse partying ;)) and is no longer pursing mainline integration. Keith Packard though would be willing to accept this work for xorg-server 1.10 even after this feature freeze if there is some developer willing to finish off the input work.
The release announcement for xorg-server 1.9.99.901 can be found here. The official, stable release of X.Org Server 1.10 is expected in mid-February and will begin appearing with Q2'2011 Linux distributions like Fedora 15 and Ubuntu 11.04.
| 1
| 1,760,738,398.045324
|
https://www.phoronix.com/news/ODkwNw
|
X Server 1.9.3 Has Now Arrived
|
Michael Larabel
|
Coming just as anticipated, Apple's Jeremy Huddleston has announced the release of X.Org Server 1.9.3. This is the third maintenance release in the 1.9 series, which was originally introduced in August.
The xorg-server 1.9.3 release though isn't particularly exciting, so even if you are in one of the snow-filled parts of the world right now with nothing better to do, it's not necessarily worth expediting your build process or package update to the 1.9.3 release. At least though it will continue to work with the just-released Catalyst Linux 10.12 driver.
Officially there are 52 fixes found in this third maintenance release. These bug fixes concern performance, stability, and correctness, but none of them are too dramatic. There's even a couple KDrive fixes. Most of the fixes though are just for Apple's XQuartz.
The release announcement for X.Org Server 1.9.3 can be found on the X mailing list. X.Org 7.6 still hasn't been released (it was supposed to come in October, but the release candidate didn't arrive until November), so unless X.Org Server 1.10 (or X.Org Server 1.9.4) makes it out before hand, this next X.Org katamari release should ship the 1.9.3 package.
For those interested in something a bit more exciting as you drink your winter Glühwein, Gløgg, or Jägertee, there is the first X.Org Server 1.10 development snapshot that arrived last week. Due to some last minute additions, the 1.10 release due out in February will bring RandR 1.4 with per-CRTC pixmap support and spirte transforms, X Fence Sync support, and a whole lot of bug-fixes. Fortunately it's without merging the drivers back into the X Server.
| 2
| 1,760,738,398.127643
|
https://www.phoronix.com/news/ODg3MQ
|
X.Org Server 1.9.3 May Come Next Week
|
Michael Larabel
|
Apple's Jeremy Huddleston has announced the second release candidate for the forthcoming X.Org Server 1.9.3 point release. This point release in the stable 1.9 series delivers on more bug-fixes, with a handful of them for Apple's XQuartz, which is important especially as it looks the 1.9 series will be used by Mac OS X 10.7.
Assuming no regressions are to be found in this xorg-server 1.9.3 release candidate (tagged as v1.9.2.902), Jeremy intends to issue the official 1.9.3 point release in one week's time.
Meanwhile, all major development work has been going into X.Org Server 1.10 which should have its merge window closed next week while the final release is expected in February.
The X.Org Server 1.9.3 RC2 announcement can be found on the mailing list.
| 0
| 1,760,738,398.609423
|
https://www.phoronix.com/news/ODg1OQ
|
X Input 2.1 Multi-Touch Getting Back On Track
|
Michael Larabel
|
While the X Server 1.10 merge window should soon close, the patches for implementing X Input 2.1 / Multi-Touch support are on track for the next release, X Server 1.11. Ubuntu is also expected to ship these patches with their X.Org Server in Ubuntu 11.04.
The most recent Multi-Touch / X Input 2.1 patches were published nearly one month ago and Canonical's Chase Douglas had based this work upon Daniel Stone's earlier patches. Chase though did not pursue getting this work into X.Org Server 1.10 but is rather waiting around for the 1.10 merge window to close and then will focus upon re-basing the work.
In recent days there's also been increased discussion on the X.Org development list about this input work and other input driver happenings.
Daniel Stone has also commented on the patches, "I'm back from 2.5 months of holidays and am slowly working my way through the email backlog ... I'm working on multitouch fulltime though, so hopefully we can get this sorted and merged fairly soon."
| 0
| 1,760,738,398.630784
|
https://www.phoronix.com/news/ODg1Nw
|
X.Org Server 1.10 Merge Window Remains Open
|
Michael Larabel
|
While the merge window was supposed to close yesterday for X.Org Server 1.10, which is supposed to be released in February, it looks like Keith Packard will keep it open for a few more days. Keith Packard, who is continuing to serve as the X.Org Server release manager, wants to keep the 1.10 merge window open until at least next Monday so he can pull in some new code he has been developing.
Keith is hoping on or around Monday he will be able to land the work he has been doing on per-CRTC pixmaps. It was this per-CRTC pixmap work that also kept the X.Org Server 1.9 merge window open longer, but it never ended up being merged in that release. Per-CRTC pixmap provides, "multiple scan-out buffers which applications can create and assign to arbitrary collections of CRTCs. These pixmaps can be associated with a window for use with OpenGL or drawn to directly." This would also lead to version RandR 1.4 of this X extension.
Those interested in more technical details about per-CRTC pixmaps can find the details in Keith's proposal, which also describes its benefits as driving multiple monitors where the screen geometry exceeds the capacity of the rendering/scan-out engines. integrating compositing and projective transformation into one operation, eliminating visual artifacts during rotation, and reducing the number of visual operations required to achieve a desired configuration.
Keith's code for this implementation is not yet complete, so let's hope he can get the patches battened away and reviewed by other X.Org developers within the next week before it potentially jeopardizes the February release of xorg-server 1.10. Fortunately, all the ABI driver changes are source compatible. Keith's message calling for the extension of the 1.10 merge window can be found on the X.Org development list.
Shortly after Keith brought this up, Tiago Vignatti shared that his threaded input events is also ready to be merged. He also has other patches ready that would make libpciaccess optional again for the X.Org Server. Unfortunately, Tiago is no longer motivated to land his threaded input or libpciaccess work into the mainline X.Org Server. See here.
| 0
| 1,760,738,399.117079
|
https://www.phoronix.com/news/ODg1MA
|
Supporting Old Hardware In X Gets Brought Up Again
|
Michael Larabel
|
It's long been a topic of what parts of X.Org should be killed with fire. There's plenty of dated and obscure X.Org and Mesa drivers around for hardware that hasn't even been manufactured in years and are rarely used. At XDS Toulose and on other occasions it's been decided not to do a massive purge of all these legacy graphics drivers for Linux. Old hardware support by the X Server has once again been brought up, but this time it's about monitors.
In the middle of November there was a series of patches published by Peter Huttrerer and one of these patches changed the X classic driver default resolution to 1024 x 768 (mailing list message) and also changed the default pixel clock to match accordingly. This would seem like something trivial -- and to most people it won't even impact them at all -- since these days it's hard even finding anyone with a monitor resolution less than 1024 x 768 with most people (aside from netbook users) running at much greater resolutions. Our annual graphics survey confirmed this that there's not many running a resolution less than 1024 x 768 at least as far as non-mobile devices are concerned.
Luc Verhaegen commented though, "Why change this, are we suddenly forgetting the whole of the 90s ever happened?" AMD's Alex Deucher responded, "At some point you have to move things forward. If you happen to have one of those old monitors that only support 1024x768i chances are you know what you are doing. The vast majority of people who will hit this case have a monitor that is capable of 1024x768p which is a much better user experience than an interlaced mode or 640/800 progressive." Additionally he said, "Most people that hit this now will be using a KVM or have a busted EDID or something and 1024x768p for most people "just works", no xorg.conf editing needed. In fact on newer monitors, it probably has a better chance of working than 1024x768i or some other archaic mode. If you have an old vga fishbowl, then you can edit your xorg.conf." Of course, a series of email exchanges between Luc and other developers has since ensued. Here's some quotes below.
We don't handle serial mice without an xorg.conf either. If your hardware's old enough to have reached the age of consent in most of the world, and if you're then running a modern X.org on it anyway, writing a static configuration seems like a sensible thing to do. KMS defaults to 1024x768 without KMS, so if you can't manage that you've failed long before you get near X.
Matthew Garrett
We provide pretty much no support for hardware that's the same vintage as the monitors you're talking about. Why would the people using these monitors be running current versions of X? Why would they not be able to write an xorg.conf? Why are they more important than the people who have to deal with the more common cases of missing EDID?
Matthew Garrett
If this really is how you feel about the thing, then be consistent, and rm -Rf everything except -intel, -ati and -nouveau. Because in your world, no-one would be running anything else.
Luc Verhaegen
People using old monitors probably still have an xorg.conf that they've been using with it for years, unless they just dug it out of a closet or rescued it from the recycling pile.
X.Org has discontinued a lot of the old input drivers already, and I wouldn't be surprised to see older video drivers like rendition and i740 join them eventually. There are always going to be limits to mixing and matching components of vastly different ages in your system, whether it be trying to run 2010 Xorg on a 1995 kernel or a 1995 monitor.
Alan Coopersmith
I wasn't going to say anything, but I have an XGI Volari hanging on my wall with a Dx9 sticker on it. I've just been too lazy/incompetent/indifferent/much of an AMD whore to get started on KMS for it. My Didj and Ego share a similar fate, along with other various video cards. Modern and rare things do exist.
Corbin Simpson
And just how certain are you that 1024x768 is going to help all those cases with missing EDID compared to 800x600?
This makes me want to start arguing for making 640x480 the default.
Luc Verhaegen.
Perhaps this is why certain vendors don't want to work with Xorg. We target 1990 hardware to the detriment of modern desktops.
Alex
[Editor's note: In reference to, Why More Companies Don't Contribute To X.Org]
The 1024 x 768 default though was merged yesterday into X.Org Server 1.10 in time for its merge window closing.
| 20
| 1,760,738,399.158005
|
https://www.phoronix.com/news/ODg0Mg
|
X Server 1.10 Merge Window Closing; Ubuntu 11.04 To Use
|
Michael Larabel
|
The X.Org Server 1.10 merge window closes tomorrow for for its release schedule so that it can ideally be released on time in mid-February (hopefully we will finally see X.Org 7.6 before then too). This next X.Org Server release isn't particularly interesting but is more along the lines of a big maintenance update with some new fun on the input side.
The plans with X.Org Server 1.10 were to merge the graphics and input drivers back into the X.Org Server to make it one monolithic build, but most developers didn't like the idea of regressing the modularized X.Org system to tie the drivers to a particular X.Org Server as it would also mean the X.Org Server release cycle would need to be sped up (circa three months per release). On the plus side it would have meant they could have killed off a legacy driver code, spruced up the API, and made some other changes, but fortunately it was decided not to merge all of the drivers back in at this time.
There is also no X Gestures Extension or X Input 2.1 for the xorg-server 1.10 release, but there is some other input-related improvements. Found though in this release are also some GLX improvements.
With X.Org Server 1.10 looking rather non-volatile, Canonical is hoping to be able to ship this release in Ubuntu 11.04. Christopher James Halse Rogers, the one responsible for Ubuntu's X stack, has written an email message casting his vote for including xorg-server 1.10 in the Ubuntu "Natty Narwhal" release.
The merge window closes tomorrow, and there doesn't seem to be anything particularly flammable. There's lots of cleanup, the new input stuff that we'll be getting anyway, and some extra GLX infrastructure which drivers may want to hook into.
This all looks fairly benign, so I suggest we go with 1.10 for Natty. For more on the rest of the X.Org stack for Ubuntu 11.04, see here and here. While Wayland may be getting easy to build, it's still not yet ready for prime-time and will not be making its formal debut with Ubuntu 11.04.
| 3
| 1,760,738,399.642973
|
https://www.phoronix.com/news/ODgzNg
|
VA-API Support For Google Android Platform
|
Michael Larabel
|
We have been tipped off that a few VA-API patches have hit the upstream libva tree for furthering along Google's Android support for this video acceleration API. VA-API is arguably the second best video playback acceleration API available to Linux users, after the NVIDIA-created VDPAU.
In particular, this patch and that patch make a few refinements to VA-API while at the same time making updates to the Google Android support. The work was done by Austin Yuan and signed off by Intel engineers; Intel is the principal backer of VA-API. It was also at this point that libva 1.0.6 was tagged in Git.
Of course, simply supporting the Video Acceleration API on Android isn't much good unless it's supported within the multimedia players, but this is becoming less of an issue with nearly every player either now supporting VDPAU and/or VA-API. The leading exception right now though is Adobe's Linux Flash Player, but better GPU acceleration support for Flash on Linux is on the way. More of a problem is the embedded graphics drivers that need to support VA-API, so hopefully next up when it comes to Android and video/GPU news, we can report that there are improvements in this area.
| 12
| 1,760,738,400.386796
|
https://www.phoronix.com/news/ODgzMw
|
Why More Companies Don't Contribute To X.Org
|
Michael Larabel
|
Being brought up from the discussion surrounding the RadeonHD driver being vandalized, which wound up just being a prank by two X.Org developers to torment one of the former RadeonHD developers, was a discussion why more companies don't contribute back to X.Org. Do companies think the X.Org code is too hard? That it's not worth the time? Is it all politics?
Here's the beginning of this new thread. Reasons expressed by other developers range from Microsoft F.U.D. to vendors just wishing X.Org would go away, provide the vendors with a competitive advantage by not pushing their patches upstream, to their code just being in a poor and unreliable state. There's also the matter that with the modularized X.Org state, it's easier to keep and maintain an out-of-tree DDX driver than it is maintaining an out-of-tree kernel driver on Linux.
In an attempt to increase the X.Org developer base, there are efforts underway to increase the X.Org documentation pool so that the barrier to entry is lower, but still there is a whole lot of areas of the X stack that are currently left undocumented. There's also reported to be a few days prior to the 2011 X Developers' meeting where it will focus upon improving the documentation at large.
Besides the lack of documentation and higher barrier to entry in general than other areas of the Linux stack, what's stopping you or your company from contributing back to X.Org in one form or another? It will be interesting to see if the situation with Wayland turns out any better since its code-base is significantly smaller and cleaner (though it's now LGPLv2 licensed) than the aging X11 Server.
| 91
| 1,760,738,400.895593
|
https://www.phoronix.com/news/ODgxOA
|
If You Forgot, S3 Graphics Does Linux Drivers Too
|
Michael Larabel
|
Last night when checking to see if VIA has made any open-source / Linux progress that went unnoticed (they haven't), that also led me to see what S3 Graphics is up to these days. S3 Graphics doesn't back any open-source driver strategy and they don't have many GPUs on the market, but their binary Linux driver claims to support OpenGL 3, VDPAU, and even kernel mode-setting since last year.
We haven't been able to test the S3 Graphics Linux drivers due to troubles actually finding S3 Chrome 400 / 500 hardware, but we have been talking about their "magical driver" for two years since they mentioned OpenGL 3 support and even Blu-ray and Microsoft DirectX 10.1 support for Linux. It turns out that S3 was indeed working on a new Linux driver, but it was missing in action even when the Chrome 540 GTX launched, but it ended up coming months later than expected.
There was no Blu-ray / DirectX support (likely this was mentioned due to uninformed S3 PR folks), but it was eventually released with support for OpenGL 3.1, VDPAU H.264/VC-1/MPEG-2 video acceleration, RandR, and even kernel mode-setting. This came by October of 2009 when S3 Graphics also brought x86_64 Linux support as up to that point their binary blob only supported the 32-bit kernel.
While this gave us some hope for seeing better S3 Graphics support on Linux, since that last 2009 driver release there has only been one driver update from them in 2010. This most recent driver update brought support for Fedora 12 and bug-fixes, but no major feature advancements. At that time, Fedora 12 was already succeeded by Fedora 13, but S3 ended up just hitting F12 support at that time. The bug-fixes in this release included Xinerama failing under certain conditions, system hangs, video playback problems, and other issues.
Unfortunately with the lack of recent driver updates, there is no S3 Graphics Linux driver support yet for Ubuntu 10.10, Fedora 13/14, Debian 6.0, or any other recent Linux distributions. This driver is designed to support the S3 Chrome 530 ULP, Chrome 540 ULP, and Chrome 540 GTX. There's also embedded S3 ASICs including the Chrome 5300E and Chrome 5400E.
It's too bad that S3 Graphics doesn't put forth more effort in their binary Linux driver or provide any open-source support (or be friendly to the open driver community like with the S3TC patent situation), but at least they are doing something. If there is actually anyone with S3 Graphics hardware on Linux, let us know in the forums how the driver and its advertised features are actually working.
| 27
| 1,760,738,400.903985
|
https://www.phoronix.com/news/ODgxNg
|
VIA Graphics Still Lack Any Real Linux Progress
|
Michael Larabel
|
While other hardware vendors are constantly improving their open-source support, this isn't the case for all vendors. VIA's open-source Linux support is still in very bad shape -- two and a half years after they had envisioned themselves becoming open-source friendly.
Last month it was reported that with the Linux 2.6.37 kernel there still is no VIA DRM that should be there according to their Linux graphics TODO list. VIA planned to have a TTM/GEM module done in the second quarter of this year, kernel mode-setting in the second half of this year, and a Gallium3D driver in development this quarter. They and their community developers have failed in making much headway towards their open-source goals. In the past month, nothing has really changed.
VIA's Bruce Chang has been the one largely seen as organizing these open-source efforts and communicating with the Linux community, but even he has not been heard from publicly or privately in months. There's been no mailing list announcements from him and (unlike in the past) he has not responded to our inquiries seeking a status update. Besides his lack of communication, here's updates on some of the other components that make up VIA's Linux offering.
- There's been no activity on openchrome-devel (the development mailing list for the OpenChrome driver that VIA had partnered with) since April of 2010.
- VIA's hidden Linux driver has not been touched in its Git repository since 15 May of this year.
- The OpenChrome driver has at least received a few commits recently. Their SVN repository was quiet for sometime, but there's been a couple code pushes this month. However, all of this work is relatively minor: xalloc to malloc changes, replacing deprecated X Server functions, addressing TV-Out flickering, typos in code comments, etc.
- There's been no VIA Mesa or DRM activity, at least publicly.
- VIA's Linux portal (linux.via.com.tw) is still vastly out of date. There's no Ubuntu 10.10, openSUSE releases post-11.0, etc. Their forum and bug tracking sections are also also still "under construction" as it has been for the past two and a half years since their most recent open-source announcement. From their main web-site, their Linux support hasn't been updated since the Fedora 11 era as their latest on the Red Hat side.
What an unfortunate situation. Perhaps VIA had all of their Linux work backed up on a single consumer-grade hard drive too?
| 11
| 1,760,738,401.409601
|
https://www.phoronix.com/news/ODgxNw
|
Vandalizing Open-Source Drivers?
|
Michael Larabel
|
Somebody with root access to the FreeDesktop.org server decided to vandalize the RadeonHD graphics driver in this Git commit. The make files were deleted and replaced with "It's dead, Jim" and a Git commit message line of "PERHAPS BONGHITS WILL FIX MY MAKEFILE." The supplied email address was "[email protected]."
While the RadeonHD driver is no longer actively developed, obviously this can come across as offensive to those who vigorously worked on this open-source Linux driver. More importantly though it puts into question the security of the FreeDesktop.org infrastructure. It took three weeks to spot as the Git notification email was not sent to the appropriate mailing list. Luc has already written to the mailing list.
"It is clear that this is not a normal security breach, as this commit is
fully in line with the naming scheme used by fd.o. Plus, given the
history of radeonhd, combined with who i think have root access, makes
it seem quite likely that this was simply one of the people with regular
root access," says Luc Verhaegen.
Fortunately, with Git such childish actions can be easily reverted.
| 51
| 1,760,738,401.423155
|
https://www.phoronix.com/news/ODc5Ng
|
Pixman 0.22 Is Getting More Performance Optimizations
|
Michael Larabel
|
It was less than one month ago that Pixman 0.20 was released with faster image scaling, faster affine transformations, ARM NEON improvements, and SSE2 back-end optimizations, among other new features. The first development release towards Pixman 0.22 has now been released and it too brings more performance optimizations.
This pixel manipulation library for the X Server and Cairo how has more ARM performance improvements when it comes to image scaling and separately there are general performance improvements for affine transformations. Besides those two areas with optimizations since Pixman 0.20, the Pixman 0.21.2 development release just carries other bug-fixes and improvements.
These ARM optimizations for Pixman were contributed by Siarhei Siamashka who works as a software engineer for Nokia Finland. The improvements to affine transformations come via another Nordic developer, but this time by Søren Sandmann Pedersen of Red Hat.
The release announcement for Pixman 0.21.2 can be found on the X.Org mailing list.
| 0
| 1,760,738,402.020064
|
https://www.phoronix.com/news/ODc4NA
|
The X Input 2.1 Multi-Touch Implementation Is Here
|
Michael Larabel
|
Canonical's Chase Douglas has corralled Daniel Stone's X Input 2.1 Multi-Touch patches and have readied them for integration into the X.Org Server and related software components.
The patches for the xorg-server, protocol, and input drivers are now available on the mailing list. There's also an X Input 2.1 Multi-Touch PPA for Ubuntu users seeing as all of the interest Canonical has had in multi-touch (with their own multi-touch framework and proposing an X Gesture Extension) and their plans to ship greater multi-touch support in Ubuntu 11.04 regardless of what's pushed upstream in time.
Unfortunately, it doesn't look like X Input 2.1 with this multi-touch support will be pushed in time for X.Org Server 1.10 due out early next year. There's a few weeks left before the xorg-server 1.10 merge window closes, but it doesn't look like Chase at least is actively trying to get this work completed in time. "The xserver source code is essentially 1.9 plus a backport of the X input ABI 12 work. I believe it should be trivial to rebase the xserver patches on top of master, and I will likely do so after the merge window closes for 1.10 on Dec 1st."
So for X.Org Server 1.11 it looks like we have to look forward to greater multi-touch support with the bringing of the X Input 2.1 extension.
| 3
| 1,760,738,402.029573
|
https://www.phoronix.com/news/ODc4MQ
|
X.Org 7.6 Release Candidate 1 Is Finally Here
|
Michael Larabel
|
Alan Coopersmith has announced the first release candidate of X.Org 7.6. Originally the X.Org 7.6 release was supposed to come in October, but that didn't happen and now into November we are finally seeing the first test katamari.
Seeing as X.Org 7.6 is just a collection of mostly existing X.Org packages, there isn't too much to do before the final release. "At this point, only some final bug fixes, documentation updates, and janitorial cleanups are expected between now and the final release of X11R7.6, including final releases of several modules for which release candidates are included in this set (including libX11 1.4.0 and xorg-docs 1.6.0)."
This release uses X.Org Server 1.9.
The X11R76-RC1 announcement with a list of changed module versions since X.Org 7.5 can be found on the xorg-announce mailing list.
| 0
| 1,760,738,402.522844
|
https://www.phoronix.com/news/ODc3OA
|
X.Org GSoC 2010 With KMS, Cairo, Etc. What Was Done?
|
Michael Larabel
|
Two days ago there were some ATI R300 GLSL compiler optimizations made to the open-source Radeon driver stack by Tom Stellar as part of his Google Summer of Code project as a student developer to improve the R300 GLSL compiler support. This code is still living in a branch outside of Mesa, but some progress has been made. Though the R300 GLSL compiler optimization work was just one of five projects that were part of the 2010 Google Summer of Code for X.Org. Here's a few words on the other student projects.
While most of the X.Org GSoC 2010 projects are centered around graphics (and that's where our interest happens to be too), one of the input-related projects was improving input within XCB (the X-protocol C-language Bindings) by Christoph Reimann. Christoph wanted to add XKB protocol bindings and other utility functions to XCB via the XML protocol description and writing a Python code generator. Unfortunately, this work hasn't made its way into any mainline code-bases for X.Org and the code appears to have not been touched in around three months.
The last XCB mailing list message we found by Reimann was in August when getting ready to test new libxcb and xcb-proto code that seemed to be working at least somewhat for XKB access via XCB. There wasn't much of a discussion that ensued from that announcement and the code hasn't really been touched since then. His Google Summer of Code work can be found in this xcb_proto branch and in this libxcb branch. The last code commits made were from mid-August and we haven't been able to find any activity since that point.
Another one of the 2010 projects for X.Org was creating a Cairo Gallium3D state tracker. This state tracker by Igor Trindade Oliveira would accelerate Cairo operations directly atop the Gallium3D driver architecture rather than needing to use Cairo with a DRM or OpenGL back-end (among others) that in turn would be indirectly communicating with the graphics hardware. Igor wrote a few times on his blog about progress on the Cairo state tracker with the last update being in August. The Cairo state tracker is working atop Gallium3D for some operations and there is Git code available. This code, however, has not been touched since the end of August and there's no sign of it going into the mainline code-base anytime soon.
Lucas Ferreira, another student developer, took a different look at things by trying to use upstream software components to create a fully plug-and-play USB multi-seat solution on the X.Org Server. On his blog are a few posts about creating a setup with Plug-able Dock Stations, GDM, ConsoleKit, and X.Org Server 1.8 using Ubuntu 10.04 LTS. For this, many components from Git need to be pulled manually and the upstream code isn't quite yet ready to make such USB multi-seat solutions an "out of the box" process. The last blog post on the matter was in late July.
One of the other graphics projects this summer was creating a kernel mode-setting driver for the Permedia 3/4 graphics processor and then writing documentation for others to utilize on how to write a KMS-enabled Linux kernel driver. The Permedia GPU came out from 3Dlabs as the first low-cost OpenGL GPU more than a decade ago. The goal wasn't to make the KMS driver itself useful seeing as there aren't many 3Dlabs Permedia GPUs still around and being used, but more importantly to produce some clear and proper documentation for other developers so they can more easily create their own KMS driver for other graphics hardware. The last update we talked about with the KMS/DRM Glit driver was in June where the code was beginning to function but kernel memory management support had to be added. In Matt's personal Git repository is where this KMS driver continues to live and was last worked on in mid-August. Only a handful of commits were made since June when pushing out some initial work on the CRTC support, VGA connector, DAC encoder, and video memory initialization. A very basic TTM memory manager was added along with frame-buffer support to the Glint driver. There, however, appears to be no recent activity on this KMS driver or in producing the KMS driver documentation (we haven't found any yet) nor does the DRM Glint driver appear ready for any upstream integration in the Linux kernel.
Like in past years, the X.Org Foundation should eventually issue a report on whether the Google Summer of Code involvement was once again successful. At future X.Org Developer Summits there will also hopefully be greater involvement by these student developers. As part of my XDS Chicago proposal, I suggested getting the GSoC developers actually involved with XDS. "On a separate but related matter, another suggestion I would like to make – that others have expressed would be a good idea as well – is better inviting active X.Org Google Summer of Code developers to participate in future events. As far as I know, it's currently not part of the standard procedure to pursue these Google Summer of Code developers to attend the X.Org event to talk about their progress or work completed. The X.Org Foundation has expressed there are funds available for sponsoring a greater number of developers to attend these events and inviting these active Google Summer of Code developers to attend could result in a more vibrant selection of talks during such events with the Google Summer of Code projects often focusing upon experimental areas and other topics not generally covered by normal discussions. While these student developers already have a vested interest in successfully completing their work due to the financial incentive from Google, if active Google Summer of Code developers were at these events they may be further motivated to work on their projects and could benefit from the knowledge learned via in-person discussions. It could potentially lead to greater motivation by these student developers to participate further in X.Org upon their formal completion with Google Summer of Code thereby expanding the X.Org development community." Perhaps then we can possibly see more student developers interested in continuing to contribute and finish their work past August.Update: There's now more information in our forums.
| 9
| 1,760,738,402.562573
|
https://www.phoronix.com/news/ODc2NQ
|
Some Small Progress On Linux GPU Laptop Switching
|
Michael Larabel
|
A few weeks ago we reported that notebook hybrid graphics switching on Linux still sucks. For these newer laptops that boast dual GPUs -- an integrated low-power IGP and a more performance-oriented discrete GPU for demanding environments with switching between the two being done "seamlessly" in real-time based upon usage or via a hot-key -- the support under Linux is still virtually nonexistent. There is a crude form of Linux GPU switching, but for the most part it's not nearly up to par for what's available in Microsoft Windows 7 or Apple Mac OS X. The situation remains that way, but some small progress has been made.
Red Hat's David Airlie is reporting on some work done by Matthew Garrett and himself on exploring the situation for those notebooks with combined Intel and NVIDIA graphics. They discovered a WMI driver would need to be written for Linux for passing the ACPI output IDs to the WMI MXDS method for switching the MUX. This at least would support some Intel-NVIDIA notebooks, but evidently Apple notebooks are handled differently without WMI/ACPI, but some Nouveau developers are reverse-engineering that side.
There is yet no working support, but at least some progress is being made in learning how these switch-able notebook graphics processors actually work. Read the blog post here.
| 6
| 1,760,738,403.122159
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.