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/ODc2Mw
|
NVIDIA Puts Out New Fence Sync Patches
|
Michael Larabel
|
While NVIDIA may not be focusing upon supporting the Wayland Display Server at this time, they are continuing to focus upon improving the experience of their driver with the X.Org Server. Three months after putting out their most recent X Sync object patches, they are now out with a revised set.
NVIDIA's James Jones has published a new set of nine patches that go against the xorg-server for implementing the X Sync Fence Objects support. These latest patches add a DDX driver interface to allow the individual hardware drivers to redirect the operations to hardware rendering back-ends, such as what you'll likely see from NVIDIA's proprietary driver in the future once these patches have been merged to master.
For those wondering what these patches are about:
Adds support for binary sync objects. Objects are set to "triggered" using X commands that are executed relative to X rendering commands. Clients can wait for fence sync objcts to reach the triggered state using XSync APIs or using interop functionality in other APIs. This allows for efficient cross-API synchronization with X rendering operations. The latest patch-set can be found here.
| 0
| 1,760,738,403.821717
|
https://www.phoronix.com/news/ODcyNw
|
Pixman 0.20.0 Is Here With Performance Improvements
|
Michael Larabel
|
Just over a week ago we reported on Pixmain gaining improved gradients and is rendering much faster in the project's latest development release. Now this free software project that provides pixel manipulation capabilities for the X Server and Cairo, has reached its version 0.20.0 stable milestone.
As aforementioned in this news posting and further reiterated in the 0.20.0 release announcement, this major release focuses upon gradient and performance improvements. There's also a great deal of bug-fixes and portability improvements.
When it comes to gradients in Pixman 0.20, radial gradients now follow the PDF specification, large linear gradients are rendered much more accurately, and conical gradients work much better.
In terms of Pixman's 0.20 performance, there is faster image scaling, faster affine transformations, improvements on the ARM NEON platform, and improvements to the SSE2 back-end.
This release is timed quite appropriately for the pending release of X.Org 7.6.
| 0
| 1,760,738,404.331342
|
https://www.phoronix.com/news/ODc1NQ
|
Intel Still Wants Xephyr, Willing To Maintain It
|
Michael Larabel
|
Haitao Feng, a software engineer at Intel China believed to be working on MeeGo, has come forward seeking clarification on the future of Xephyr and KDrive. Intel is still interested in Xephyr and Haitao is willing to maintain it, especially after doing work to add OpenGL acceleration to this KDrive-based X Server that targets a host on another X Server as it's frame-buffer.
As outlined on the FreeDesktop.org page, Xephyr was written a number of years ago as a possible Xnest replacement and to be used for tool-kit debugging, X.Org Server development, and for creating a multi-seat setup.
Haitao has written to the X.Org mailing list asking about the future of Xephyr and KDrive. Xephyr hasn't been maintained well lately, but Haitao is interested in taking over the duties after he and Robert Bragg, a fellow Intel engineer, wrote the "Design and Implementation of Xephyr with OpenGL Acceleration" design document earlier this year (it was made available on MeeGo-SDK in September).
However, if Xephyr is deemed as having a little future ahead, Feng will cease development on it. Meanwhile, the Intel engineers working on MeeGo Touch are busy looking forward to the Wayland Display Server.
| 6
| 1,760,738,404.340065
|
https://www.phoronix.com/news/ODcxNQ
|
Intel, Radeon DRM Get Precise VBlank Timestamps
|
Michael Larabel
|
Mario Kleiner has published patches over the weekend that introduce precise vblank time-stamping support within the Linux kernel's DRM core and has implemented this support already within the Radeon and Intel kernel drivers too. The precise vblank timestamps and counting is needed by the DRI2 sync and swap extensions and in particular to conform with the OML_sync_control extension.
The existing infrastructure within the kernel DRM to generate timestamps was not reliable due to interrupt handling delays and thus were not conformant since they were not actually synchronized to a vblank. The DRM core patch adds support for vblank timestamp caching, calculating the precise vblank timestamps, and filtering of redundant vblank IRQs and removing of some race-conditions in the vblank IRQ disable/enable code path. This work was previously discussed at XDS Toulouse so it really shouldn't have a problem being accepted and hopefully is in a state where it can be pushed into the mainline Linux kernel this week prior to the Linux 2.6.37 kernel merge window closing prior to next Sunday.
For those not familiar with GLX_OML_sync_control, it's a nine year old extension that is described by the registry specification as: "This extension provides the control necessary to ensure synchronization between events on the graphics card (such as vertical retrace) and other parts of the system. It provides support for applications with real-time rendering requirements by providing precise synchronization between graphics and streaming video or audio." Right now this code is just hooked up for the Intel and Radeon DRM drivers, but we imagine hooking in the Nouveau driver will not be too far behind.
Mario has described this vblank timestamp endeavor in this mailing list message that he sent prior to mailing of the actual three-piece patch series, with additional comments inside that e-mail as well.
| 8
| 1,760,738,404.863226
|
https://www.phoronix.com/news/ODcxOQ
|
X.Org Server 1.9.2 Should Be Here By Christmas
|
Michael Larabel
|
Apple's Jeremy Huddleston released X.Org Server 1.9.1 this past weekend, and now he has laid out plans for the next point release in the xorg-server 1.9 series. This release will carry the same type and quality of fixes as presented in X.Org Server 1.9.1 and should be released in December.
Jeremy's current plans for X.Org Server 1.9.1 have the first release candidate coming on 12 November, a second release candidate on 3 December, and then either an RC3 or final release on 10 December.
This news was announced by Huddleston on xorg-devel.
On a related note, it's been decided by the X.Org Board of Directors to participate in Google's Code-In program, which provides students between the ages of 13 and 18 years old various open-source work to participate in during their winter holiday. X.Org has long participated in Google's Summer of Code, which is a much bigger program and runs the duration of the summer holiday, but the Google Code In is designed for smaller tasks.
On this X Wiki page are a few ideas for any able students interested in participating. Among the ideas are small code tasks like adding X-Video attributes to GNOME/KDE display applets, support for more RandR output attributes in GNOME/KDE display applets, and exporting KMS connector attributes via sysfs. There's also non-code work like improving the documentation, helping out with the testing process, and research.
| 0
| 1,760,738,404.871274
|
https://www.phoronix.com/news/ODcwOA
|
X.Org Server 1.9.1 Released By Apple's Huddleston
|
Michael Larabel
|
As we mentioned earlier, Apple's Jeremy Huddleston took over release management of the X.Org Server 1.9 series now that it's stable and will only receive bug-fixes from this point on. Meanwhile, Keith Packard and the gang of X.Org developers are focusing on X.Org Server 1.10 to have that ready by early next year. Jeremy Huddleston on this Saturday night has just made his first point release, X.Org Server 1.9.1.
This release though is not exactly a surprise considering there's been release candidates for a few weeks and it was expected to make its debut in October so that it can be released as part of the X.Org 7.6 Katamari, but all has been quiet on that front, so it may be delayed or has just been held up until xorg-server 1.9.1.
Introduced in X.Org Server 1.9.1 are a few stability and performance fixes that have come since the release of X.Org Server 1.9.0 back in August. There's also some fixes to the XQuartz DDX and other areas.
The X.Org Server 1.9.1 release announcement can be read on xorg-announce.
When looking to take over management of the xorg-server 1.9 series, Jeremy had mentioned "[Apple is] probably going to pick 1.9.x as our next long-term support branch for XQuartz (replacing 1.4.x)." From this week's announcement of Mac OS X 10.7 "Lion" coming next summer, this is where we will likely see their X.Org Server 1.9 deployment.
On a related note, yes, we are looking forward to benchmarking Mac OS X 10.7 vs. Linux vs. FreeBSD vs. Windows 7 next year. There will be rock-solid support in the Phoronix Test Suite for Mac OS X 10.7 "Lion", but until it's released you can look over our Mac OS X vs. Linux OpenCL benchmarks, tests of Apple's enhanced OpenGL stack, and Mac OS X 10.6 vs. Windows 7 vs. Ubuntu 10.04 benchmarks.
| 0
| 1,760,738,405.373903
|
https://www.phoronix.com/news/ODcxMQ
|
Coming Soon: X.Org 7.6 Release Candidate
|
Michael Larabel
|
X.Org Server 1.9.1 was released yesterday, which led us to wonder about the status of the X.Org 7.6 katamari that was supposed to come at around the same time as this first stable point release. We hadn't heard anything about the X.Org 7.6 release schedule in weeks and before that the target to ship this grouped collection of X packages was mid-October, which has clearly passed.
The good news, however, is that spinning off from a discussion on the X.Org mailing list is a bit of a time-frame. Oracle's Alan Coopersmith has written the following on the mailing list.
I'd like to get the RC out by the end of the week if we can - getting the libX11 doc build situation resolved so that we aren't breaking the build for many people has been the primary blocker (aside from general lack of time) - I can't think of any other major changes needed that aren't in git yet, though there's still a bunch of modules that need tarballs released from the latest git. So hopefully we will see the X.Org 7.6 release candidate come out by the end of this coming week and the final release come shortly thereafter. Granted, with X.Org's current modular design, tagging X.Org 7.6 isn't as important as before with many distributions like Ubuntu 10.10 already shipping with the X.Org Server 1.9 and the updated display/input drivers and other X packages. Nevertheless, a year has already passed since X.Org 7.5, which came months late with X.Org Server 1.7.
| 0
| 1,760,738,405.389068
|
https://www.phoronix.com/news/ODY5Ng
|
Pixman Has Improved Gradients, Is Much Faster
|
Michael Larabel
|
While Cairo is frequently mentioned on Phoronix, mentioned less but used by Cairo (as well as the X.Org Server) for pixel manipulation is the Pixman library. Soeren Sandmann announced a new release candidate of Pixman 0.19.6 this afternoon and it has a few interesting changes worth noting.
New to Pixman are much improved radial and linear gradients, performance improvements, enhanced performance specific to the SSE2 back-end, and a number of bug-fixes and test suite enhancements. The performance improvements should be rather noticeable, especially on ARM platforms.
Pixman 0.19.6 is a development snapshot leading up to the stable Pixman 0.20.0 release. The release announcement and download links for this version of Pixman can be found on xorg-announce.
| 4
| 1,760,738,405.860134
|
https://www.phoronix.com/news/ODY2Mw
|
Multi-Touch For The X.Org Synaptics Driver
|
Michael Larabel
|
Takashi Iwai of Novell/SuSE has just published a series of 18 patches for the X.Org Synaptics input driver that primarily provides multi-touch support. There's also some fixes and other changes to this Synaptics driver for X, but the primary feature is the addition of multi-touch support under Linux (and other X.Org-using operating systems) for supported hardware.
It turns out that Takashi actually wrote these patches some months ago, but that Novell's legal department didn't allow him to publish the code for upstream integration over concerns of possible patent infringements. This was mentioned on the mailing list.
However, after this work was independently discovered, Takashi has now went ahead and mailed out the patches for hopeful integration.
| 11
| 1,760,738,405.885211
|
https://www.phoronix.com/news/ODY0OA
|
X.Org Server 1.9.1 Is Approaching; RC1 Released
|
Michael Larabel
|
Under the release management of Apple's Jeremy Huddleston, the first release candidate for the first point release in the X.Org Server 1.9 series is now available. X.Org Server 1.9.1 will be the first update to X.Org Server 1.9, which was released in August, and should make it out around the same time as the X.Org 7.6 katamari release later in October.
The X.Org Server 1.9.1 release isn't limited to bug-fixes only, but Jeremy has also pulled in RandR (the Resize and Rotate extension) support for XQuartz, which is used by Apple on Mac OS X. There's also now 18bpp support in XFree86 and NDS32 architecture support in XFree86.
Besides these minor new features, there are a handful of bug-fixes, which are laid out in the xorg-server 1.9.1 RC1 release announcement. All new development work meanwhile is going into X.Org Server 1.10, which is scheduled to be released early next year.
| 0
| 1,760,738,406.493878
|
https://www.phoronix.com/news/ODYzMw
|
Running The X.Org Server Without Root Privileges
|
Michael Larabel
|
Besides talking about the X.Org Server development process and the X.Org Server 1.10 plans, Keith Packard also talked about not running the X.Org Server as root (also sometimes referred to as a "root-less X Server") during XDS Toulouse.
Thanks to kernel mode-setting and other improvements to the X.Org Server, we are basically there, but it's something that has been said for over a year. Moblin though has been running the X.Org Server as a user since their Moblin 2.0 release and now with the MeeGo operating system while others are now looking to get in on the root-free action.
Canonical hoped to not run the X.Org Server as root with their forthcoming Ubuntu 10.10 release, but they didn't make that goal. Though the Ubuntu developers are just about there so it should be a candidate for Ubuntu 11.04.
As talked about at the X.Org Developers' Summit in Toulouse, one of the last development changes to overcome is how to handle input devices when you have multiple X.Org Servers not running as the root user so as to know which input to direct to the appropriate X.Org Server. For those more interested in this topic, below you can see/hear our compressed recordings from this XDS talk. There aren't any slides or other visuals really expressed so the audio feed is more important than anything else.
Due to YouTube limitations, the video is available in three parts.
| 5
| 1,760,738,407.173273
|
https://www.phoronix.com/news/ODYzMQ
|
Keith & Peter Talking About X.Org Development
|
Michael Larabel
|
The 2010 X.Org Developers' Summit in Toulouse has been over for a week, but the disappointing weather in Munich today Oktoberfest finally made it sound more enticing to take care of the remaining XDS 2010 coverage rather than drinking Augustiner in a wet pair of lederhosens. With that said, below are the video recordings of when Keith Packard and Peter Hutterer were talking about the X.Org Server development process changes that have resulted in surprisingly on-time releases.
Though there are no slides or other visuals during this presentation and their talk was quite casual (hence why the video quality ended up being reduced and compressed), but it's there for those who are interested.
For those wanting a written synopsis of what was talked about during their informal thirty-minute talk, read this Phoronix news posting where we share the key highlights.
| 0
| 1,760,738,407.775907
|
https://www.phoronix.com/news/ODYyOQ
|
2010 Linux Graphics Survey Reminder
|
Michael Larabel
|
As a reminder to those interested in participating in our annual Linux Graphics Survey, there's now less than a week before this survey officially is over at month's end. This survey is meant to provides some quantitative data about what the Linux community is most interested in when it comes to graphics driver features and the hardware and drivers currently being most utilized, etc.
Earlier this month we provided an initial look at this year's numbers prior to the X.Org Developers' Summit in Toulouse, but since then the survey results have continued to grow by a few more thousand submissions.
If you want to participate in this survey, you can participate here. Results will be published in early October.
| 1
| 1,760,738,407.784088
|
https://www.phoronix.com/news/ODYxNw
|
Most Drivers Won't Be Merged Into X Server 1.10
|
Michael Larabel
|
The last talk of the 2010 X.Org Developers' Summit was regarding X.Org Server 1.10. The good news is that nearly every X.Org graphics driver will not be merged back into the xorg-server repository.
The release schedule for X.Org Server 1.10 was talked about, which has the final release set to arrive in February. Some of the features for this next major X.Org Server release include libxkb, RandR 1.4, input clean-ups, threaded input events, and other clean-ups. "It's pretty much our job right now to remove system-level code out of the server and into a share-able environment." Such work also directly benefits the Wayland Display Server, like the XKB common library that was talked about.
Merging drivers back into the X.Org Server was a hotly debated topic to the point that yesterday it was said merging the graphics drivers back in would effectively kill the Linux desktop. The good news, however, is that this won't happen.
As predicted, the xf86-input- drivers will likely be integrated into the X.Org Server, but the xf86-video- drivers will remain outside the xorg-server unless an individual driver maintainer wants to move in the driver. So far none of the open-source driver maintainers have come forward to say they will go ahead and integrate their code into the X.Org Server, not even Keith Packard who proposed this plan to begin with in order to clean-up the API/ABI and force the X.Org Server to receive better testing. Alex Deucher with the ATI/AMD Radeon driver is also not interested in moving his driver back into the server.
Though at least one video driver will be merged into the xorg-server and that would be the dummy DDX driver, which actually isn't useful for anyone besides learning to write an X.Org driver. Merging the driver into the X.Org Server may actually make it easier to track when there is ABI breakage of the xorg-server.
| 8
| 1,760,738,408.416408
|
https://www.phoronix.com/news/ODYyMA
|
Talking About EGL In Mesa On Linux
|
Michael Larabel
|
A few days back I reported on the first operating system where you may see the Wayland Display Server used rather than an X.Org Server after talking with Kristian Høgsberg while in Toulouse. At the X.Org Developer Summit' he talked to everyone about EGL in Mesa, which also plays an important role with Wayland.
For those unfamiliar with EGL, it's an API that's maintained by the Khronos Group and serves as a binding API to OpenGL, OpenGL ES, and other rendering APIs. EGL is described by the Khronos Group as:
EGL is an interface between Khronos rendering APIs such as OpenGL ES or OpenVG and the underlying native platform window system. It handles graphics context management, surface/buffer binding, and rendering synchronization and enables high-performance, accelerated, mixed-mode 2D and 3D rendering using other Khronos APIs. During Kristian's XDS talk he talked about using EGL with sharing resources across different APIs (using the EGL image extension), his work to run EGL on the KMS frame-buffer directly, a Khronos API for sharing EGL images between processes that is similar to the extension created for DRM/Mesa, and the state of the EGL API within Mesa and Gallium3D.
It was after adding the EGL_MESA_DRM_image extension to Mesa that it became possible to run Wayland off mainline Mesa. Wayland uses EGL and previously this support was provided by a side-project of Kristian's known as Eagle before it was merged into Mesa.
Below is Kristian's XDS 2010 EGL talk in two parts.
| 31
| 1,760,738,408.425233
|
https://www.phoronix.com/news/ODYxMw
|
Luc Calls For A Dead Linux Desktop If Keith Gets His Way
|
Michael Larabel
|
While X Server 1.10 is not being discussed at length until tomorrow (the final day of XDS Toulouse), besides today's notes, Luc Verhaegen who formerly was with Novell working on the RadeonHD driver and has also worked on the open-source VIA Unichrome driver and a few other X related projects, is preparing for another heated battle.
Back in February at FOSDEM in Brussels, Luc made a presentation on modularizing Mesa and DRI drivers, which ended up in a very heated discussion but ultimately his ideas fell on deaf ears. With X.Org Server 1.10, Keith Packard of Intel has expressed interest in merging the drivers back into the server, or in other words de-modularizing the X.Org Server after it was modularized a few years ago as being a feature.
Luc is very much in starch opposition to moving X.Org drivers back into the X Server to the point that he believes if the drivers are pulled back into the server that the Linux desktop will die. Those in favor of merging the drivers back into the xorg-server want this done so that they can more easily break the API/ABI and make such changes without worrying of backwards compatibility. Merging the drivers back in would also cause much more upstream testing of the xorg-server since most GPU developers will now be forced to rebuild the xorg-server quite often.
Those in opposition to pulling the drivers back in are upset as the X.Org Server can be a pain in the ass (and more time consuming) to build, will really not lead to major benefits, etc. It also means that a new X.Org Server needs to be released for providing a new driver, which may mean upgrading the xorg-server just to have new hardware support or to receive GPU bug-fixes. This could be problematic for not only end-users but also distribution vendors in how they push out stable updates.
Luc firmly believes that pulling the input and video drivers back into the xorg-server will put the system into a position where it will be far from bug-free and never in a constant "useful" state. Each time you want to upgrade one component, you'll need to upgrade the entire stack, which is likely to introduce other bugs. Mesa and libdrm are already tightly coupled together and major Linux kernel upgrades are already required for upping the DRM/KMS. "No normal person can then run a free software desktop system, and expect to use it, because an arbitrary mix of hardware cannot possibly work together acceptably, at least not for a measurable amount of time...Looking further, by shutting out our own users, we will take away the breeding ground that free software is based on."
Luc also praises the proprietary NVIDIA driver for its compatibility with a wide range of kernel and X.Org Servers and wants the free software graphics drivers to be more like this Santa Clara company's work. Many more thoughts are shared in Luc's blog post about killing the Linux desktop.
While most developers are not as polarized as this open-source enthusiast, there are other X.Org developers here who also are not in favor of their DDX drivers being pulled in for xorg-server 1.10. Tomorrow there should be much more information to report, but I'd likely suspect that X.Org input drivers will be merged back into X.Org Server 1.10 but not the xf86-video- drivers.
The proposed future direction for graphics drivers is to create graphics driver stacks. If not, we, the developers, might just as well stop working on free software graphics drivers altogether.
And while the current situation currently is bad, it is not impossible to fix. The problems are known and clear, a path to the solution should by now also be clear, but the willingness to put in the bit of extra thought is simply lacking.
So guys, if you really want to move into the wrong direction, please state the real reasons for doing so, state the consequences to your users; and know what the end result will be.
| 103
| 1,760,738,408.93153
|
https://www.phoronix.com/news/ODYxNQ
|
The Next X.Org Developers Summit?
|
Michael Larabel
|
The X.Org Developers' Summit in Toulouse, France just ended and it's time in the morning to head to Oktoberfest to meet with many Phoronix readers at this annual outing. XDS 2010 turned out to be a wonderful event and more organized than some X.Org events in the past. Thanks to the wonderful organization by Matthieu Herrb, the venue itself was nice, the social event last night was terrific, the Internet and power at the event was plenty, etc. Stay tuned for Phoronix notes and some audio/video recordings to be published in the coming days, beyond what has already been reported. At XDS 2010 it was also brought up where to host XDS 2011.
It was brought up whether to host the 2011 X.Org Developers' Summit in Brazil, simply on the basis of the X.Org events usually being in the United States or Europe, even though that's where a vast majority of the X.Org developers are located. No real reasons in favor of an XDS Brazil event were provided and there isn't even any X.Org developers presently living in Brazil that could organize such an event. There were plenty of concerns though regarding the cost of transportation, the time needed to fly to Brazil for both Americans and Europeans, and just the overall location being inconvenient for everyone.
No suitable alternatives came about during this X.Org Foundation board meeting being held at the event. However, to provide a possible alternative for the next X.Org Developers' Summit, I have submitted a proposal to the X.Org Foundation Board. Under my proposal, a future X.Org Developers' Summit is to be held in Chicago, Illinois and will be organized by your's truly.
After plenty of drinking last night at the X.Org parties, most of the night was somehow spent constructing a rather elaborate proposal for XDS Chicago under my organization. Finished now and submitted is a five-page proposal why Chicago could serve as a suitable venue for an X.Org Developers' Summit and would be rather convenient for most of the developers regardless of their location -- it's much more than that of others just saying "let's have it here!" I have also brought up some other possible changes for future X.Org events.
Those interested in reading this proposal that was submitted to the X.Org board, the Adobe PDF file can be found at MichaelLarabel.com. Most of the content is also relevant to any other Linux or technology groups that may be looking to host a conference or summit in Chicago. If there is a Chicago event, there is also likely to be live audio/video streaming feeds on Phoronix.
| 0
| 1,760,738,408.939989
|
https://www.phoronix.com/news/ODYxMA
|
What Parts Of X.Org Should Be Killed With Fire?
|
Michael Larabel
|
Originally at the X.Org Developers' Summit here in Toulouse this week there was going to be a talk entitled "Kill It With Fire" where Corbin Simpson (mostly known for his work on the ATI R300 Gallium3D driver) was going to be speaking about what drivers or parts of X.Org should be eliminated from the stack. This talk though is no longer occurring, in part as Corbin is no longer in attendance; he washed his US passport in the laundry.
Corbin has wrote a summary of what he was going to talk about on the X.Org Wiki. One of the items that has been talked about recently has been XAA vs. EXA and whether to eliminate the prior since EXA/UXA is superior to that of the ancient XFree86 Acceleration Architecture. However, XAA will not be killed off anytime soon since there still are old drivers (the niche ones outside of Intel/Nouveau/Radeon) and hardware that aren't a good fit for EXA. In particular, there's the older hardware that doesn't work well for heavily variable-stride allocations, compositing with alpha, and using 3D engines. As a result, XAA or EXA will not be killed off anytime soon. Even if EXA support was stripped away from the DDX drivers, it still can be accelerated on GPUs via the 3D engine with shaders by the X.Org state tracker in Gallium3D that implements EXA and X-Video.
Living within Mesa are a few obscure drivers that could die with little or few people noticing, like the Mach64, MGA, TdFX, and Savage DRI drivers, but since they at least receive build system fixes (but really no further development), Corbin sees no reason to let them die.
Corbin would like to see the R300 and R600 classic DRI drivers stripped from Mesa though as soon as the R300g and R600g drivers are mature. The ATI R300g (Gallium3D) driver is pretty much mature and is on-par with the classic driver but many in better cases, faster in some OpenGL tests, and more feature-rich thanks to the growing number of Gallium3D state trackers. The R300g and R300 Mesa drivers support up through the ATI Radeon X1000 series. The R600g driver is still not yet on-par with the R600 classic driver or comparable to the R300g driver, but it's quickly advancing and soon should be there with support through the ATI Radeon HD 5000 "Evergreen" hardware.
Legacy code from Mesa should also be stripped as soon as possible (such as GLSL parsing code that's no longer used by Intel's new GL Shading Language compiler). On the opposite end, Corbin wants to pull in Mesa / Gallium3D patches for adding in Haiku, AROS, and Syllable support patches.
On the kernel side of the Linux graphics stack, there are many frame-buffer drivers that this student developer would like removed, but first they need to be ported over to KMS. However, there's not a lot of hardware for these obscure drivers in use and these FB drivers vary already in code quality state. Much activity on this side isn't expected in the near future.
| 20
| 1,760,738,409.521664
|
https://www.phoronix.com/news/ODYxMQ
|
A Few Notes From Day 2 Of XDS Toulouse
|
Michael Larabel
|
More details will come later along with the audio/video recordings that ended out the X.Org Developers' Summit in Toulouse, but here are a few random bits from so far today:
- For those that have become interested in coming up with a new logo for X.Org, Alan Coopersmith issued this mailing list message today. Coming up with a new logo for the X.Org Foundation has been on their agenda for many years, but now it may finally materialize thanks to Phoronix readers.
- Canonical introduced the X.Org Gestures Extension last month as part of their own multi-touch framework, but there's been some dissenting opinions about this extension proposed by Canonical's Chase Douglas. One of the controversial items was this extension moving the handling of gesture recognition inside the X.Org Server itself rather than the window manager, a separate library, or elsewhere. It has now been agreed upon that gesture recognition will not be entering the X.Org Server.
- X.Org multi-touch support won't be done by the time X.Org Server 1.10 is released nor is there any real time-line for its completion.
- From last night: The French ENAC/CENA was visited in Toulouse where they had two interesting demonstrations and their advanced input and graphics work done. What X.Org driver do they use? The proprietary NVIDIA driver, it was their only possible option for handling their OpenGL needs.
Additionally, read What Parts Of X.Org Should Be Killed With Fire? or yesterday's coverage of Recapping The New X.Org Development Process or A Newbie Getting Involved With X.
| 0
| 1,760,738,409.530081
|
https://www.phoronix.com/news/ODYwNw
|
Recapping The New X.Org Development Process
|
Michael Larabel
|
Scheduling issues had plagued X.Org Server development for the past few years: to the point that even delivering a point release had come more than a year late and major X Server releases were never delivered on time. This though has fortunately changed.
With X.Org Server 1.8 though it was proposed to make some fundamental development changes and better refining the X.Org development process to be more like the Linux kernel -- though not the same -- where there is an official release manager, timed releases, and a defined process for requesting changes/patches be pulled into a given release. Since that point, the X.Org Server has basically been released on time. X.Org Server 1.9 was released on time just last month.
At XDS 2010 in Toulouse, France, Keith Packard and Peter Hutterer just finished talking about this development process. No major development process changes were proposed or altered. Again, the three phases of the server development process as described by Keith is the "free for all" process when anything new can enter the server (similar to the Linux kernel "merge window" during each cycle), then the stabilization period, and lastly is the API/ABI freeze for the release.
Here are a few other random notes from this discussion:
- Using a BugZilla release blocker bug for X.Org Server development has worked out really well. "If you bring an X.Org bug up to being a release blocker status, we'll look at it."
- Unless it's a security issue, Keith would not hold-up an X.Org Server release for no blocker bugs. "It's hard to say we have any new security holes, but lots of old ones."
- There's a need for X.Org and distribution/OS vendors to work together in a more efficient and effective manner. For example, there's many OpenBSD patches that still have not been pulled upstream.
- "Is the review process stifling change?" Adam Jackson called for more developers to become involved in the review/patch acceptance process beyond Peter, Keith, and him. Adam thinks the rate of code change since this new development process has decreased as a result of this new development process. Keith says this is because of the complexity of the code and that the current X.Org server "does what it does" and is better advancing than in the past.
- In the future when patches are pulled in by Keith Packard for changes to the X.Org Server, he will now reply and acknowledge the commit via e-mail rather than leaving developer to wonder about its state, monitor the commit list, or be frequently pulling the xorg-server Git code.
- Red Hat's Adam Jackson and Apple's Jeremy Huddelston are taking over management of the X.Org Server 1.9.x stable series.
| 8
| 1,760,738,410.02905
|
https://www.phoronix.com/news/ODYwNg
|
A Newbie In X; Creating X.Org Documentation
|
Michael Larabel
|
Matt Dew, a self-proclaimed "X newbie", just finished talking about his experiences as just entering the world of X.Org development and hopes to contribute to the X.Org world by gathering up and improving X.Org documentation.
As most know who have ever investigated X.Org, traditionally the documentation covering the software stack has been rather fragmented or even nonexistent in many places. This lack of reliable X.Org documentation can lead to a steep learning curve for new developers and can be a deterrent when coupled with the fact X.Org is complicated, but Matt hopes to work towards addressing this longstanding problem by rounding up the existing documentation from various sources and then eventually to write documentation for the missing pieces (particularly the newer areas, but also for areas like libdrm).
As part of this clean-up process, Matt hopes to move all X.Org documentation into one format (right now there's around five different document formats used by X components), clean-up this documentation, and then create a table of contents / index, creating a CSS style, cross-linking, and other items. The initial documentation conversion should be complete by the end of September while more of the work he hopes to have completed by January.
By summer of 2012, Matt Dew hopes to have X.Org documented to a state that it's considered "good", or even possibly "great."
The audio recordings are still going on, although the quality of some of the talks may vary as the audio line feed from the theater's audio system seems to have died.
| 2
| 1,760,738,410.69727
|
https://www.phoronix.com/news/ODU5MA
|
X.Org Server 1.10 Release Plans; Drivers May Still Go In
|
Michael Larabel
|
While there are only a few days left until the 2010 X Developers' Summit, Keith Packard has laid out his plans for the development of X.Org Server 1.10.
With X.Org Server 1.10, as talked about before, the X stack may be de-modularized to the point that X drivers would be merged back into the X.Org Server. At least the protocol headers should be merged into a single package and the input drivers are likely to be moved into the X Server too, but moving back in the GPU drivers is a matter that's still hotly debated and will certainly be talked about at the X Developers' Summit. At this point Keith is asking, "Anyone want to volunteer to have "their" driver get merged into the server for 1.10?" Keith's interest in moving the drivers back into the server is so that they can be re-factored so that mode-setting code can be dropped for hardware where there's already kernel mode-setting support, etc.
It's also been talked about before by Keith to bring the X.Org Server development down to a three-month release cycle, but for now it's standing at a six-month release cycle. If this release cycle remains, X.Org Server 1.10 is set to be released on the 18th of February.
Other details from Keith and information for those looking to have their xorg-server patches merged, read this mailing list post.
More information is sure to come next week in Toulouse, which we will be writing about. Also be sure to participate in our 2010 Linux Graphics Survey.
| 4
| 1,760,738,411.201445
|
https://www.phoronix.com/news/ODYwMw
|
X.Org Is Looking For A New Logo
|
Michael Larabel
|
XDS 2010 has just begun in Toulouse, France. Well, besides yesterday's pre-event where we were discussing Wayland and other topics. At the moment just the X.Org Foundation itself is being discussed.
There is an audio feed of the event that I am recording directly from the microphone feeds and those will be published on Phoronix in the coming days. I will also be recording videos of select talks in HD, but it will be from an integrated microphone source.
During this initial X.Org Foundation talk one item was brought up worth sharing: the X.Org Foundation is looking for a new logo. It's actually been something under consideration for a number of years, but it's now been discovered a Brazilian cable company is using a logo just like that of X.Org. With the fact that the current X.Org logo is rather simplistic and old (25 years old), the foundation is looking for a new logo that can be a registered trademarked in the US and Europe.
As everyone at XDS is a developer and not graphics designers, perhaps any of the artistic members of the Phoronix community can propose a new logo to the X.Org Foundation.
| 54
| 1,760,738,411.216552
|
https://www.phoronix.com/news/ODU3OA
|
A Tentative Schedule For XDS 2010
|
Michael Larabel
|
For those interested in the X Developers' Summit (XDS) that is taking place next week at a tobacco factory in France, a tentative schedule has now been published by Matthieu Herrb for the 50 or so people that will be participating in the summit.
The key sessions for the 2010 XDS include:
- X.Org foundation session
- Review of latest DRI2 protocol additions (i.e. sync and swap extensions)
- Documentation or: how a newbie got involved
- Development process recap (to the point of releasing on time)
- Multi-touch session (such as the multi-touch protocol specification)
- Gestures session (Canonical's X Gestures Extension)
- XCB session
- Xephyr future
- EGL in Mesa (including the ability to run Wayland from Mesa)
- Wacom input driver
- OpenGL - which functions can developers rely on?
- Kill it with fire? (what drivers should be deprecated)
- libxkbcommon
- Non-Linux DRI/KMS support
- Handling input events: threads or not? (threaded inputs)
- X Server 1.10 planning (and whether to merge the drivers back into the X Server)
Though some of these sessions may not occur or be handled differently as Matthieu assumed everyone who made a proposal would be in attendance, but for instance AMD's John Bridgman will not be there for "OpenGL - which functions can developers rely on?"
There will be Phoronix coverage (along with audio/video recordings) of all of these X talks that are taking place right before the Phoronix Oktoberfest outing.
The latest XDS2010 schedule can be found on the Wiki in its latest form.
| 19
| 1,760,738,411.51847
|
https://www.phoronix.com/news/ODU2OA
|
XDS 2010 Has Been Moved To A Tobacco Factory
|
Michael Larabel
|
While there is Oktoberfest in two weeks, in just a week and a half there is the annual X.Org Developers' Summit. This year's summit for these developers is taking place in Toulouse, France. The event was going to be hosted at a conference room at the University of Toulouse, but due to delays in renovating that room, this X.Org summit has been moved to an ex-tobacco factory.
Last year the X.Org Summit was hosted at Portland State University in the United States, but the year before that we met up at the Edinburgh Zoo for XDS 2008. With X developers meeting at the zoo and even needing to stop the event for a Gentoo penguin parade (see video below), X developers are used to strange venues. Let's just hope at this ex-tobacco factory they actually have a reliable Internet connection -- unlike at the Edinburgh Zoo when 30~40 individuals were leaching off a mobile phone connection.
So far the attendance is looking to be about 50 people for XDS2010.
Among the topics to be discussed at this ex-tobacco factory event are whether X drivers should be merged back into the X.Org Server for the 1.10 release, DRM/KMS support on non-Linux platforms, what hardware should continue to be supported and what drivers should be eliminated, a review of the latest DRI2 protocol additions, a development process recap, a multi-touch session, a session on X Gestures, handling input events, board of directors chat, documentation / how to get more individuals involved, EGL in Mesa, and libxkbcommon maintainer-ship. Birds of feather sessions for Xephyr, XCB, and the Wacom support are also planned.
XDS 2010 is planned to last from the 16th to 18th of September in Toulouse at this Manufacture des Tabacs. There will, of course, be live coverage on this event on Phoronix with any interesting notes. We will also be providing audio and/or video recordings too.
Speaking of X.Org, be sure to participate in our 2010 Linux Graphics Survey. While this survey lasts all month, prior to the X Developers' Summit I will likely publish some current results from this survey around the 15th of this month. They may prove to be interesting during the "Should We Kill It With Fire?" session in terms of what hardware/drivers to continue supporting.
| 3
| 1,760,738,411.738048
|
https://www.phoronix.com/news/ODU2Mg
|
Who Contributed The Most During X Server 1.9?
|
Michael Larabel
|
Two years ago we compiled a list of the top contributors to the X Server over the years and that was followed by compiling a similar list of the developers behind Mesa. Tiago Vignatti has now compiled some statistics surrounding the top contributors to X.Org Server 1.9 and related X components just looking at this most recent development cycle. There's also numbers for the input, video, and Mesa components too.
Below are various bits of information gathered from Tiago's statistics.
- Alan Coopersmith was the developer with the most change-sets to X, which Tiago counts as the X Server, proto, library, and XCB repositories. Alan accounted for 15.3% of these changes through his work at Oracle, after being a long-time X developer at Sun Microsystems. While Oracle may be doing nasty things to other open-source projects they hold stakes in and doing things like killing off OpenSolaris, they are at least so far letting Alan continue to contribute back to upstream X. The 2nd through 5th most active developers were Jamey Sharp, Gaetan Nadon, Keith Packard, and Tiago Vignatti, respectively.
- While Alan had the greatest number of change-sets to the X implementation, when it comes to the number of lines of code that were changed, his work fell into second place (at 23.4%) with Matt Dew accounting for more than 50% of the changed lines!
- Intel's Keith Packard had the most commit sign-offs and reviews during the 1.9 cycle, which should not come as a surprise considering he served as the X.Org Server 1.9 release manager.
- When it comes to X input with the xf86-input drivers, xkbcomp, and xkeyboard-config repositories, the developer with the most change-sets for 1.9 was Gaetan Nadon at 40.4% while Peter Hutterer who was the mastermind behind Multi-Pointer X (MPX) came in at second with 21.8% of the change-sets. Work from Sergey Udaltsov accounted though for 74.3% of the total changed lines of code for the X input side.
- When looking at the changes to libdrm, Mesa, and xf86-video drivers during this time, there were commits from 107 different developers but the developer with the most change-sets was Brian Paul, the creator of Mesa and now a VMware employee (following the Tungsten Graphics acquisition) who continues to contribute mostly to Mesa's Gallium3D architecture.
- The second most change-sets came from Eric Anholt at just 10.6% (Brian accounted for 10.7% of the work) followed by Gaetan Nadon, Vinson Lee, and Marek Olšák. Overall, Intel's Eric Anholt accounted for the most new lines of code as well as the number of removed lines of code.
- When looking at the video work done by employer in terms of change-sets, the top ten list comes down to: VMware (33.3%; mostly Gallium3D and core Mesa work), Intel (27.7%; their Intel code obviously plus the new GLSL compiler and other contributions), Videotron (7.7%), Marek Olšák (7.4%; a hobbyist most often working on the ATI Mesa / Gallium3D drivers), Kenneth Graunke (5.8%; hobbyist contributions to Intel, etc), LunarG (3.5%; the modern-day consulting equivalent of what Tungsten Graphics was, led by Jens Owen and Alan Ward), Red Hat (3.3%; mostly the ATI drivers plus Nouveau and various other contributions), Luca Barbieri (2.3%; Nouveau driver), Corbin Simpson (1.3%; ATI Gallium3D), and AMD (1.1%; the ATI code).
- Among other interesting facts that can be gathered from this X video data, community developers accounted for more work to the open-source ATI graphics drivers than was done during this development cycle by AMD itself. Granted, most of the past few months Alex Deucher and Richard Li at AMD were busy working on the Evergreen acceleration code that was only recently released to the public after clearing legal review. The AMD developers are now going back to work in private to begin supporting the yet-to-be-released AMD Radeon HD 6000 series graphics cards.
- While NVIDIA doesn't contribute to any open-source graphics drivers, when it comes to the X Conformance Test, they are the largest contributor with more than 90% of the changed lines of code being attributed to them -- largely NVIDIA's Aaron Plattner.
- Missing from any of these lists showing off contribution data for everything from X documentation to graphics drivers is Canonical. Canonical (Ubuntu) only had one sign-off during this entire time that was measured, as showed by this list.
All of these statistics are available in raw form on Tiago's blog.
| 16
| 1,760,738,412.067367
|
https://www.phoronix.com/news/ODU1MQ
|
Canonical's X Gesture Extension Being Re-Evaluated
|
Michael Larabel
|
Earlier this month Canonical introduced its own multi-touch framework for Ubuntu that is set to premiere with Ubuntu 10.10 "Maverick Meerkat" and it's called UTouch and is joined by their own gesture/touch language. That same day as announcing UTouch for Ubuntu that will support devices like the Apple Magic TrackPad and Dell XT2, Canonical proposed the X.Org Gesture Extension to the X.Org development community. While it's good to see Canonical making more contributions to upstream projects that it depends upon for Ubuntu Linux, the X.Org Gesture Extension is already being re-evaluated and may in fact not be needed.
Since Canonical's Chase Douglas published his xorg-devel email in which he laid out their proposed specification for X.Org Gesture Extension protocol v1.0, there's been some dissenting opinions by X.Org developers about how the gesture recognition should be handled. Part of this proposed gesture extension provides an interface for X clients to register and receive gesture primitive events and X clients to then act as a gesture engine. Canonical is of the belief that the gesture handling should be handled server-side by X while the X.Org developers -- such as Kristian Høgsberg and MPX-creator Peter Hutterer -- want this input gesture recognition to be handled client-side out of the X.Org Server.
In a new email by Chase, Canonical justifies the gesture recognition per the X.Org Gesture Extension proposal to be handled within the server for the following reasons: 1. With handling of input events (touches) that may span multiple windows, but the recognized gesture event is supposed to only affect the first window, the X.Org Server shouldn't be firing off the input events to any other windows. 2. If the handling is done by the client, there's also the possibility the recognition would occur twice -- once by the window manager and then again by any window clients that may want to recognize their own input events. Canonical admits though that if this task were to be performed twice there would be little in the way of latency problems. 3. Lastly, multi-touch events are not exposed to the client in the X.Org Server used by Ubuntu 10.10, where Canonical wants this first cut of UTouch / X.Org Gesture Extension to be deployed.
The common position against having the gesture recognition engine in the X Server and instead on the client-side (such as within the window manager instead), can be summarized by Carsten Haitzler with "I absolutely agree with peter on this. Frankly the problem is that input from a mouse, or a tablet or multiple touch-points is ambiguous. You may be painting in GIMP - not trying to "swipe to scroll". I can go on with examples (dragging a slider inside a scrollable area as opposed to wanting to scroll with a drag). Only the client has enough detailed knowledge of the window contents, application mode etc. To possibly make a reliable guess as to which one to do. It's X's job to provide as much device input to the client as it needs in a sane digestible way to make such a decision, but... that's [in my honest opinion] where the server's job ends."
Chase does acknowledge, "It's true that the logic behind point one may be perfectly fine, but having recognition through X inserts a lot of extra code into the server. If we are content with touches being split up into window regions before recognition occurs, then we may be able to get around the need for the X Gesture extension completely. The window manager use case could be supplied through input grabbing and event replaying."
We will see how this ongoing architectural debate ends, but for Ubuntu 10.10 at least it will be handled server-side.
| 10
| 1,760,738,412.277005
|
https://www.phoronix.com/news/ODU0MQ
|
Threaded Input Events On The Way For X Server 1.10
|
Michael Larabel
|
While X.Org Servger 1.9 was released less than a week ago (heck, it's only been four days since releasing om time), the first interesting patch for X.Org Server 1.10 is already queued up and on the X.Org development mailing list for discussion. This patch, which was written by Tiago Vignatti and Fernando Carrijo, provides the "thread-ification" of the X Server input event generation code. Rather than being bound to the same thread as the X Server, the input event code with this patch is now running in its own thread, but this may only be the start of things to come with finally multi-threading the X.Org Server.
Nokia's Tiago Vignatti originally mentioned his and Fernando's threaded input work on the X.Org mailing list earlier this month during another input-related discussion and mentioned their input thread branch of the X Server where their work was happening. Though in terms of moving the input event generation to its own thread, this isn't being done for performance reasons but rather to better the code by avoiding the use of SIGIO/SIGPOLL with signal handling. This threading work is being targeted at X Servers running on Linux, but it may end up benefiting other platforms too. More interesting, Tiago looks at this initial input work as "opening space to other thread works" like moving X rendering or other tasks to an independent thread as well. Basically there would be one thread for dispatch and the core object model, one thread for input, and another thread for the GPU. Threading the X Server is something that has interested Tiago since at least 2008 when he talked about it. There was also an effort to multi-thread the X11R6 Server more than a decade ago, but those attempts had failed.
The current X input event generation thread patches can be found in this mailing list. The patch right now is actually only a few hundred lines of code to implement, but more threading work will hopefully come for other areas of the X Server if other developers provide positive feedback. This patch at least should be in X.Org Server 1.10 and for now can be enabled by configuring the xorg-server build process with a new --enable-input-thread build option. Another major change that may also happen with the xorg-server 1.10 release is moving the drivers back into the X Server, but that's a heated issue and something that's to be talked about next month at the 2010 X Developers' Summit in Toulouse, which we will providing coverage.
| 10
| 1,760,738,412.739854
|
https://www.phoronix.com/news/ODUzMg
|
X.Org Server 1.9 Makes It Out On Time!
|
Michael Larabel
|
Keith Packard has announced the official release of X.Org Server 1.9 this evening. Besides this release offering up various bug-fixes and minor improvements, this X Server release is noteworthy as it happens to be the first release in recent times to actually make it out on time. X.Org Server 1.9 hit its original release schedule of being released on the 20th of August.
This is significant as some earlier releases were delayed by many months and such setbacks had become a fairly common occurrence with X.Org releases. In fact, there was more than a six month delay in just delivering a point release. Fortunately the new X Server development process is working out in everybody's favor and is good news for Ubuntu 10.10, Fedora 14, and other Linux vendors looking to ship updates with this new X.Org Server.
Among the changes for X.Org Server 1.9 are DRI2 improvements (a new authentication mechanism, buffer invalidation, etc), RandR 1.4 with per-CRTC pixmap support, code clean-ups, documentation updates, and much more.
Meanwhile, the X.Org 7.6 katamari release should be here by October. The xorg-server 1.9.0 release announcement can be found on the X.Org mailing list.
| 13
| 1,760,738,412.80724
|
https://www.phoronix.com/news/ODUyMA
|
The High-Profile X.Org / Linux Kernel Security Bug
|
Michael Larabel
|
As many learned today, there's been a rather critical bug living within the Linux kernel for several years (as possibly far back as the original Linux 2.6 kernel release) that was finally fixed and this "high priority" bug is now publicly detailed. This issue (CVE-2010-2240), which allows arbitrary code to be executed as root, is easily exploitable by most current Linux desktops via simply running any compromised GUI application that has access to the running X.Org Server.
This security vulnerability that can be easily reproduced with most X.Org Servers on Linux was discovered by The Invisible Lab (a computer security research firm) and after privately reporting it to the X.Org team back in June was now detailed today in the company's blog and this formal security paper. Below is the summary provided within this paper entitled "Exploiting large memory management vulnerabilities in Xorg server running on Linux."
A malicious authenticated client can force Xorg server to exhaust (or fragment) its address space. If running on Linux, this may result in the process stack top being in an unexpected region and execution of arbitrary code with server privileges (root). x86 32 and x86 64 platforms are affected, others most probably are affected, too.
Note that depending on the system configuration, by default local unprivileged users may be able to start an instance of Xorg server that requires no authentication and exploit it. Also if a remote attacker exploits a (unrelated) vulnerability in a GUI application (e.g. web browser), he will have ability to attack X server.
In case of a local attacker that can use MIT-SHM extension (which is the most likely scenario), the exploit is very reliable. Identifier CVE-2010-2240 has been reserved for the underlying issue (Linux kernel not providing stack and heap separation). This issue has been known for at least five years.
The good news is that this issue is now corrected in the stable 2.6.32.19, 2.6.34.4, and 2.6.35.2 Linux kernel releases (along with the upstream Linux 2.6.36 kernel code) after it was corrected in the past few days via a number of patches. For those running a non-patched kernel it's also possible to make the system less vulnerable to the exploit by disabling the MIT-SHM (MIT Shared Memory) X extension, which when enabled allows the X Server to exchange data between the client and server using shared memory.
Of course, if the X Server wasn't running as root but rather an unprivileged local user, the vulnerability of a GUI application taking advantage of this bug could possibly have been adverted -- at least in terms of exploiting the fundamental kernel memory bug through X. Fortunately, thanks to kernel mode-setting this is now becoming a possibility. The X Server has traditionally had to run as root since X drivers have had to bang the hardware directly, but with these activities being moved into the Linux kernel DRM drivers, there isn't a real need for X to be root any longer. The mainline Linux kernel supports kernel mode-setting for the open-source Intel, ATI, and Nouveau (NVIDIA) drivers and is used by default on most recent desktop Linux distributions.
While most distributions now leverage KMS when using the open-source Intel/ATI/NVIDIA drivers, not many of them at this point are taking advantage of a root-less (in the sense of not not running as the root user) X Server. There's still a few items for the distribution vendors to address such as handling multi-user sessions while still having the X Server running as a non-root user and permission issues with a few areas of X. Though if you are using one of the more obscure open-source graphics drivers (Poulsbo, VIA, XGI, etc) or the proprietary drivers (mainly ATI) running the X Server as root is also not a possibility since they don't currently support KMS, and thus these non-kernel drivers need to still talk to the hardware directly. The proprietary NVIDIA driver doesn't implement KMS support, but it can allow the X Server to not run as root assuming the /dev/nvidiaX files have appropriate permissions.
There's been work towards making the X Server non-root since the kernel mode-setting work was integrated into the mainline Linux kernel. Moblin 2.0 was the first major distribution to abandon running the X Server as root, which was easy in their case since they just aim for Intel hardware support where using KMS is the only choice, and this non-root usage continues to be the case with the MeeGo distribution. Ubuntu is working towards a root-less X Server for Ubuntu 10.10 (though it looks like some bits may now not land until Ubuntu 11.04), as with other Linux distributions, when a KMS driver is utilized. The security-oriented OpenBSD developers have also been interested in porting the Linux KMS code to their BSD kernel for the less of a security threat created by the X Server not having root privileges.
| 15
| 1,760,738,413.359632
|
https://www.phoronix.com/news/ODUxMg
|
Introducing The X.Org Gesture Extension
|
Michael Larabel
|
Earlier this morning Canonical had announced the UTouch Framework, which is their multi-touch framework to be formally introduced with Ubuntu 10.10. Canonical developers crafted up their own multi-touch solution and even their own gesture language for Ubuntu, rather than leveraging any similar free software projects, but -- to some surprise -- it turns out they are now going to try to engage with upstream developers to at least have a formalized extension to the X.Org Server for gestures.
Chase Douglas of Canonical has just announced the X Gesture Extension to the X.Org development community and he hopes this will become part of the X Input 2.1 extension. The X Gesture Extension is to work hand-in-hand with the recently-drafted X.Org Multi-Touch Protocol Specification that was written by input-expert Peter Hutterer.
A draft of the X Gesture Extension specification can be found on xorg-devel. The X Gesture Extension provides an interface for X clients to register and receive primitive gesture events and an interface for these clients to act as a gesture engine.
The X Gesture Extension will be found in Ubuntu 10.10, which is using X.Org Server 1.9 but with Canonical's patches. The Multi-Touch and Gesture protocols for X.Org are planned for X Input 2.1, which could come with X.Org Server 1.10 due out early next year.
An introduction to this proposed X extension that's written by Chase Douglas can be found in this mailing list message. At the moment there have been no comments by any of the X.Org developers surrounding this X.Org Gesture Extension.
| 10
| 1,760,738,414.039708
|
https://www.phoronix.com/news/ODUwMw
|
X.Org Server 1.9 Release Candidate 6
|
Michael Larabel
|
Last night there was the sixth and potentially last release candidate of X.Org Server 1.9. The xorg-server 1.8.99.906 test release contains various fixes and documentation updates reflecting X Server 1.9 API changes. The final release of xorg-server 1.9 should be just around the corner and may come as soon as next week.
There's still a few bugs that Keith Packard would like to see addressed, but Keith Packard feels the release is in good shape and is on track for releasing on the 20th of August.
The xorg-server 1.8.99.906 release announcement can be read on xorg-announce.
| 1
| 1,760,738,414.530606
|
https://www.phoronix.com/news/ODQ5Nw
|
The Updated XGI Open-Source Graphics Driver Released
|
Michael Larabel
|
Even the nearly-defunct XGI Technology Inc is able to produce open-source graphics driver code for Linux while VIA continues on with their Linux mess, even with XGI developers working from Windows. In preparations for the X.Org 7.6 Katamari and this month's release of X Server 1.9, a new release of the XGI DDX driver has been made available.
Red Hat's Adam Jackson has tagged the xf86-video-xgi 1.6.0 release today along with a horde of other katamari packages and those needing updates anyways for xorg-server 1.9 compatibility. Besides a bunch of configure/build fixes, there is lots of work contributed by XGI Technology into this XGI 1.6 driver release.
Among the new features introduced in this driver update are support for the ARM architecture, EXA 2D acceleration support, I2C EDID, support for custom modes, and quite a number of bug-fixes. The ARM support and other improvements is coming after XGI recently inked a deal with Marvell (press release) to use the Volari Z11 in ARM-based systems using the Kirkwood platform. The Marvell Kirkwood with XGI Volari Z11 is targeted for industrial computing, thin-clients, point-of-sale terminals, and kiosks.
The xf86-video-xgi 1.6.0 driver release announcement can be read on the mailing list. This XGI DDX driver supports the following PCI and AGP hardware: XGI5597/5598, XGI530/620, XGI6326/AGP/DVD, XGI300/305, XGI540, XGI630/730, XGI315/H/PRO, XGI550/551/552, XGI650/651/M650/661FX/M661FX/M661MX/740/741, XGI330 (Xabre), and XGI760.
Other X.Org packages updated today include glproto 1.4.12, kbproto 1.0.5, renderproto 0.11.1, videoproto 2.3.1, xproto 7.0.18, xextproto 7.1.2, xf86-input-aiptek 1.3.1, xf86-input-vmmouse 12.6.10, xf86-video-ast 0.91.10, xf86-video-i128 1.3.4, xf86-video-mga 1.4.13, xf86-video-vmware 11.0.2, and xf86-video-xgixp 1.8.0.
| 3
| 1,760,738,414.549752
|
https://www.phoronix.com/news/ODQ5Mw
|
Working On The X.Org 7.6 Katamari
|
Michael Larabel
|
If all goes according to plan, X.Org Server 1.9 will be released in about two weeks, but after that there still is the X.Org 7.6 release "katamari" to be done. While X.Org releases themselves aren't as important any longer with the X Server releases being done at different points and the rest of the X.Org package collection being modular, the X.Org 7.6 release is expected in October.
Alan Coopersmith, the X.Org wrangler at Sun/Oracle, has done some X.Org 7.6 planning on the xorg-devel mailing list. On the date of the xorg-server 1.9 release, which is scheduled on the 20th of August but could potentially slip by a few days, all driver and protocol updates needed for the 1.9 release should also have been released. At this point, Alan would also like the first release candidate of Xlib 1.4.0 to be tagged.
Around the 31st of August, there should be a feature freeze for all X.Org katamari modules and the module maintainers should begin releasing source packages they would like tested for inclusion into the official 7.6 package set. At the end of the X Developers' Summit in France, which is taking place in the middle of September, there should be X.Org 7.6 RC1 and Xlib 1.4.0 final should be released.
Last but not least, Alan is planning for the X11R7.6 final release to come in early-to-mid October and at the same time there should be an X.Org Server 1.9.1 release to incorporate any bug/regression fixes that made it into the 1.9.0 build.
Some of the changes since the release of the X.Org 7.5 katamari that occurred last October is the inclusion of XCB (the X C Bindings), documentation updates, updated drivers, major configure/Makefile clean-ups and improvements, and most prominently the features that have been delivered with the X Server 1.8 and forthcoming X Server 1.9 releases. Many distributions are already shipping with the individually updated X.Org packages, so again the 7.6 katamari itself isn't too important from at least the end-user perspective. This is also likely to be the first katamari release that would be done on time (X.Org 7.5 was supposed to come last April) after beginning a good track record for releasing close to on time with the xorg-server 1.8 and now 1.9 releases, following a new development process.
| 29
| 1,760,738,415.060788
|
https://www.phoronix.com/news/ODQ5Mg
|
Missing From Linux 2.6.36: VIA's TTM/GEM DRM
|
Michael Larabel
|
The Linux 2.6.36 kernel is set to carry some much-anticipated changes like AppArmor integration, major VFS performance improvements, a likely Btrfs performance regression fix, and quite a few changes when it comes to the graphics Direct Rendering Manager code. Just when it comes to the Linux DRM code in this next kernel release there is quite a bit of fun with the Intel kernel driver better supporting Embedded DisplayPort and tracepoints for page-flipping and vblank, the ATI Radeon kernel driver gaining R600/700 tiling support / support for reading R600 thermal sensors / R300 through R500 Hyper-Z support, and the Nouveau driver supporting KMS with the GeForce GTX 400 series and better DisplayPort capabilities. While the Intel, ATI/AMD, and NVIDIA/Nouveau DRM driver improvements will excite a majority of the desktop Linux user-base caring about open-source graphics drivers for their hardware, it's not enough to excite everyone. In particular, VIA is still missing from the field -- more than two years after they announced a new open-source initiative and promised to better engage within the Linux ecosystem.
VIA's pledged a lot to the open-source Linux community since their April 2008 announcement at the Linux Foundation Collaboration Summit in Texas, but they haven't actually delivered on much. VIA's driver shortcomings can not even be summarized within a single article, so for those not familiar with the situation we recommend reading VIA's Open-Source Efforts A Bluff?, VIA Publishes 2D/3D Documentation & Partners With OpenChrome, VIA Will Not Provide An OSS Chrome 9 3D Driver, and VIA's Linux Dreams Are Not Materializing, among many other Phoronix articles.
At the end of last year, VIA published a Linux TODO list for their Linux graphics driver and its first action item was to have a TTM/GEM module for the second quarter of 2010. Well, we are now into the third quarter and even with the Linux 2.6.36 merge window now being open, there is no VIA TTM/GEM code ready for integration. It will not be until at least the Linux 2.6.37 kernel that we could see this code emerge, but that would a Q4'2010 roll-out. VIA was hoping to have kernel mode-setting support going in the second half of 2010, but first they must get proper kernel memory management support integrated into the kernel. Unless they manage to push a large swoop of code into 2.6.37, this is pushing back their road-map into 2011 for KMS and that's not even counting their Gallium3D driver that was supposed to be in development for Q4'2010. For end-users with the VIA DRM code not being in the Linux 2.6.35 or 2.6.36 kernels, this is already making it not an option for distribution vendors updating their operating systems anytime this year.
It was back in March when we learned that VIA was working on a hidden open-source driver and supposedly handed off some memory management code for Translation Table Maps (TTM) and the Graphics Execution Manager (GEM) to OpenChrome driver developers, but nothing has yet materialized for mainline integration. This hidden DDX driver (xf86-video-openvia) hasn't even been touched in Git since the middle of May, the OpenChrome development mailing list hasn't even been used since April, no code changes have been made to the OpenChrome driver in SVN in weeks, and most importantly there have been no pull requests by any of the open-source VIA driver developers asking David Airlie (the Linux kernel DRM subsystem maintainer) to pull in any VIA DRM code so that Linus Torvalds in turn could pull in this code from David's tree. There's not even been any recent discussions on the DRI development list or Linux kernel mailing list by VIA's representatives about getting to such a point.
VIA Technologies has attempted to push their DRM code into the mainline Linux kernel code multiple times in the past, but it's been largely held up by the fact that their DRM code is dependent upon using a closed-source user-space X.Org driver, but this is supposed to be fixed with the VIA and OpenChrome "partnership" and also this new OpenVIA driver. This older DRM code for the Chrome 9 hardware has also lacked any form of proper kernel memory management, which is ultimately holding back the rest of the driver stack. Even if VIA were to publish their new DRM code tonight, while the 2.6.36 merge window is still open until the first release candidate, it's likely that it still wouldn't even make the cut for this next kernel release. VIA should start pushing out code soon and engaging in discussion with the X.Org developers soon if they hope for even Linux 2.6.37 integration.
With the pace that VIA developers and their "friends" (as described by VIA's Bruce Chang) are working on this open-source Linux support, it will truly be a surprise if VIA is even able to deliver on an open-source Linux driver stack with kernel mode-setting and a modest Gallium3D driver for any of their hardware generations before Ubuntu 12.04 LTS in April of 2012, which would mark the four-year anniversary of this most recent VIA open-source initiative. There is also the possibility that we may never even see this code materialize if VIA Technologies were to go bust or they just abandon their failed open-source efforts like they have done in the past.
| 11
| 1,760,738,415.070868
|
https://www.phoronix.com/news/ODQ4MQ
|
The X.Org Multi-Touch Protocol Specification Is Drafted
|
Michael Larabel
|
Peter Hutterer, one of the few X.Org input developers and the developer behind Multi-Pointer X when he was a student in Australia, has now published the first public draft of the multi-touch protocol specification. The multi-touch protocol specification is a low-level spec to be integrated with the X Input Extension for further enriching the multi-touch capabilities on Linux and other operating systems using the X.Org Server.
The multi-touch protocol specification adds backwards-compatible pointer emulation for single-touch interaction, supports a dynamic number of simultaneous touch points, and provides support for devices that are both multi-touch and traditional pointer devices. Peter intends for this new specification to be integrated with the X Input Extension for version 2.1.
Those interested in reading this low-level specification and providing any feedback on it can find the text on the xorg-devel mailing list. While X.Org Server 1.9 is expected for release before month's end, X Input Extension 2.1 with this specification update for better the multi-touch protocol may be on the table for X.Org Server 1.10. X Server 1.10 will likely make it out the door in Q1'2011.
| 1
| 1,760,738,415.614369
|
https://www.phoronix.com/news/ODQxNA
|
X.Org Server 1.9 RC5 Released
|
Michael Larabel
|
X.Org Server 1.9 is due to be released in just about one month, so as such, Keith Packard the release manager has pushed out the fifth release candidate of this forthcoming X.Org update. As said by Keith, "Not a huge number of changes this time around, but a couple of useful bug fixes."
Packard also said in the release announcement, "We've got four more weeks before the release; keep the changes coming!" The changes found in 1.9 RC5 range from XKB to DIX and DRI2 changes.
Keith has also provided an update on the 1.9 release blocking bugs that need to be cleared prior to the August release of X.Org Server 1.9.
| 1
| 1,760,738,415.622589
|
https://www.phoronix.com/news/ODM4Nw
|
X.Org Server 1.9 RC5 Released For Testing
|
Michael Larabel
|
There's just about a month and a half left until Keith Packard hopes to ship X Server 1.9, so as such he has put out another release candidate today of this updated xorg-server -- on the same-day that the X Server 1.8.2 point release made it out.
This release candidate contains mostly bug-fixes and other work as Keith is trying to keep X.Org developers from committing any new code that doesn't deal with correcting the outstanding bugs. For the X.Org Server 1.9 blocker bug there are currently 17 outstanding bugs that need to be addressed.
Today's 1.9 RC5 release announcement can be found on the announcement mailing list. While X.Org Server 1.9 doesn't bring as many features or major changes as some of the past releases, this time around it's on a tighter release schedule than normal, but there is RandR 1.4 and some other enhancements. X.Org Server 1.9.0 is expected for release in mid-August.
| 6
| 1,760,738,416.124561
|
https://www.phoronix.com/news/ODQxMQ
|
More Of What To Look Forward To At XDS2010
|
Michael Larabel
|
There is just over two months left until the 2010 X.Org Developers' Summit taking place in Toulouse, France. There's just over 30 registered participants at this time and not much has changed since the reminder a few weeks back, but a few more of the sessions have been expressed.
Besides determining whether to merge the drivers back into the X Server as part of the X.Org Server 1.10 release and planning its other changes as well as discussing DRM/KMS support on non-Linux platforms (primarily *BSD), four other sessions have come about since the last Phoronix update.
One of the talks is being called "Kill It With Fire" and it's about deciding what graphics hardware to actively support by X.Org with KMS, EXA, and other key Linux graphics features. Obviously the strong focus is on supporting Intel, ATI/AMD, and NVIDIA (via Nouveau) hardware, but there's other hardware out there and there's even work on a KMS driver for old 3Dlabs graphics cards that are more than a decade old. There's also Linux users with other old or obscure graphics hardware looking for modern graphics support, such as was just expressed a few hours ago in our forums.
Another talk added recently for XDS2010 is reviewing the latest DRI2 protocol additions, such as the protocol requests and swap events. This DRI2 talk is being led by Intel's Jesse Barnes.
Similar to the X.Org Server 1.10 planning session, there are also plans in the XDS2010 program to discuss the new X.Org development model. For X.Org Server 1.8 and X.Org Server 1.9 this new development model -- that makes it more like how Linus Torvalds maintains the Linux kernel -- appears to be working better than their previous efforts and at least they are closer to releasing on-time. Also going to be discussed is on branching the X.Org Server with master versus stable and the possibility of creating a xorg-server-next branch, similar to what's done with the linux-next tree as a place for patches to land prior to their inclusion in the next release cycle.
Last but not least, a multi-touch session is planned. This is a three day summit so more sessions will certainly come up like usual as the event nears.
The current XDS2010 program can be found on the X.Org Wiki along with the attendee list. There will be coverage on Phoronix of this summit taking place at the University of Toulouse (no developers will be raiding the zoo this time) from the 16th to 18th of September.
| 8
| 1,760,738,416.141334
|
https://www.phoronix.com/news/ODM3Mg
|
A Reminder Of The 2010 X.Org Developer Summit
|
Michael Larabel
|
Matthieu Herrb is sending out reminders this morning that the 2010 X.Org Developers' Summit is taking place this September in Toulouse, France. From the 16th through the 18th of September at the University of Toulouse, the X.Org developers and interested parties will be discussing various topics related to Linux graphics and input.
Planned for discussion so far is X.Org Server 1.10 planning (X.Org Server 1.9 should be released the month prior) and the heated topic of whether the drivers should be merged back into the X Server or not, which would bring it back to being a more monolithic package. The other listed discussion topic at this time is on *BSD support for the DRM/KMS support, which has been struggling (along with OpenSolaris) behind the Linux DRM/KMS level of support. Other discussion topics will crop up shortly for this three-day event with the deadline for talks being the 16th of August.
So far there is just about two dozen listed attendees (with only about a dozen being confirmed), but that list too should grow as the September event approaches. I will be at the XDS2010 event providing live coverage on Phoronix of the X.Org discussions and announcement. For anyone else considering this event, the summit this year aligns rather well with the start of Oktoberfest in Munich where it's just a short train or plane ride away. After a few days of X.Org talks, it's perfect for then enjoying a few Maß of Augustiner and Franziskaner with some Weißwurst and other wonderful dishes.
Additional details on XDS2010 in Toulouse, France can be found on the X.Org Wiki or in today's e-mail reminder.
| 4
| 1,760,738,416.7413
|
https://www.phoronix.com/news/ODM4NA
|
X.Org Server 1.8.2 Officially Released
|
Michael Larabel
|
Peter Hutterer has announced the official release of X.Org Server 1.8.2, after it has went through two release candidates over the past month. No new changes have been introduced since X.Org Server 1.8.2 RC2 and since the release of X.Org Server 1.8.1 in mid-May it just carries more bug-fixes, a couple DRI2 updates, and addresses other maintenance items.
The xorg-server 1.8.2 release announcement can be read on xorg-announce. This is the last planned X.Org Server 1.8.x series point release unless someone is interested in taking over all of the responsibilities of being the release manager. X.Org Server 1.9 is due out in just over one month.
| 5
| 1,760,738,417.438208
|
https://www.phoronix.com/news/ODM2Ng
|
Ubuntu Nearing X Server Not Running As Root
|
Michael Larabel
|
Based upon a recent email to the X.Org developers' mailing list, Canonical is nearing the point of one of their goals for Ubuntu 10.10 of a rootless X Server, or being able to run the X.Org Server without root privileges.
All that's left to accomplish within the Ubuntu land according to Canonical's Christopher James Rogers is working out a /dev/backlight device interface that udev would set the appropriate permissions on for the user. The /proc/mtrr may also need to be handled too, but Rogers doesn't believe any of the drivers (at least the main KMS drivers) are using this interface. With all of the necessary prerequisites addressed, when starting the X Server they will have a check to see if kernel mode-setting is being used, if /dev/backlight exists, and if /dev/input/* has appropriate user permissions. If all conditions are true, the X.Org Server would not be run as the root user, which leads to better security. Of course, this feat has already been achieved by other Linux distributions such as Moblin and now MeeGo.
This would largely help out those with the open-source ATI, Intel, and Nouveau drivers that use kernel mode-setting while those using non-KMS drivers, including the binary drivers from ATI and NVIDIA, would still be running their X Server as root.
The mailing list thread discussing this can be found on xorg-devel. There is also the Maverick blueprint discussing this likely feature of Ubuntu 10.10. Other details can also be found on the Ubuntu Wiki.
| 2
| 1,760,738,417.982587
|
https://www.phoronix.com/news/ODM3MA
|
A Bunch Of Interesting Cairo-Trace Benchmarks
|
Michael Larabel
|
GNOME developer Benjamin Otte has published quite a few benchmarks on his blog showing off the Cairo performance with the Xlib and OpenGL back-ends on Intel hardware and also NVIDIA hardware with the open-source Nouveau driver.
He isn't done with his testing yet and there are no ATI test results yet of Cairo, but if you're at all interested in open-source graphics or care about the Cairo rendering performance with different back-ends, the results are worth checking out.
Hopefully this will result in greater optimizations within the graphics drivers and Cairo.
| 4
| 1,760,738,417.99062
|
https://www.phoronix.com/news/ODM1Ng
|
Here Comes X.Org Server 1.9 Release Candidate 2
|
Michael Larabel
|
Keith Packard has just announced the release of X.Org Server 1.9 RC2, which is coming just days after the first release candidate. This second release candidate for the updated X Server pulls in most of the remaining patches that were requested prior to the RC1 release last week.
Still though there is an Xfont patch that needs to be pulled in for X.Org Server 1.9 along with the per-CRTC pixmap patches that are to make up RandR 1.4. Keith hopes these patches will be merged soon.
In total there are 20 patches making up the X Server 1.9 RC2 release. The changes can be found in the release announcement for xorg-server 1.8.99.902 as it's known properly.
| 1
| 1,760,738,418.48205
|
https://www.phoronix.com/news/ODM1Nw
|
Whoops, X.Org Server 1.9 Gets Another RC Today
|
Michael Larabel
|
The second X.Org Server 1.9 release candidate was released earlier today after the first RC making it out just last week, but already the third release candidate is available to interested parties.
X.Org Server 1.9 RC3 doesn't bring much to the table, but it takes care of two issues that were moments too late to make it into RC2. The first change reverts an earlier commit to X.Org Server 1.9 that dropped mioverlay support on the basis of it being "unused", but the NVIDIA binary driver happens to use the mioverlay support (the open-source drivers no longer do), so this 2,000 lines of code was added back into the X Server for NVIDIA's benefit. Instead just the XAA overlay hooks will be dropped from the server, which is no longer used by any video drivers.
The other change is just Kristian Høgsberg signing off on a fix for a simple macro.
The release announcement for X.Org Server 1.9 RC3 (xorg-server 1.8.99.903) can be found on the X.Org mailing list.
| 39
| 1,760,738,418.490486
|
https://www.phoronix.com/news/ODM1Mg
|
X.Org Server 1.8.2 Release Candidate 2 Is Here
|
Michael Larabel
|
Earlier this month there was the release of X.Org Server 1.8.2 RC1, but last night the second release candidate was pushed out by Peter Hutterer, who has been managing the X Server point releases. The 1.8.2 RC2 release should be quite representative of the final release barring any last minute problems.
There's only a handful of more bug-fixes and back-ports in the X Server 1.8.2 Release Candidate 2 version, but some tricky DRI2 back-ports did get pulled in. The release announcement with change details can be found on xorg-announce. The final version of X.Org Server 1.8.2 should come soon, which is expected to be the final point release in the 1.8 series.
Meanwhile, X.Org Server 1.9 is progressing well and is still on target for release in August. Its first release candidate came last week.
| 1
| 1,760,738,418.81656
|
https://www.phoronix.com/news/ODM0OQ
|
Introducing The Cairo Gallium3D State Tracker
|
Michael Larabel
|
In past days we have reported on the work being done at the moment for improving the ATI R300 GLSL compiler and kernel mode-setting support for old 3Dlabs GPUs by students participating in the X.Org project with Google's Summer of Code. Igor Trindade Oliveira, another GSoC student developer, has blogged about the work he is doing this summer on creating a Gallium3D state tracker for Cairo.
Like the other state trackers with accelerating OpenGL, OpenGL ES, OpenVG, OpenCL, and other APIs on the GPU hardware in a rather neutral way using the Gallium3D driver architecture, Igor aims to do this for Cairo too. Nearly all Linux users should be familiar with Cairo, but for the non-technical users it's a 2D vector graphics API that's found uses within GTK+, by Mozilla within the Gecko web layout engine, Mono and the Moonlight projects, WebKit also uses Cairo, and many other projects use it in some form as well.
Besides directly supporting X, Cairo supports multiple back-ends, including OpenGL and OpenVG, so it can be accelerated indirectly atop the hardware already using the other state trackers, but the Cairo state tracker would accelerate it without any middle layer.
Igor Trindade Oliveira is hoping this state tracker for Cairo will make things go significantly faster, particularly with the more advanced operations. Igor's current code-base of this Cairo-Gallium3D work can be found at FreeDesktop.org. So far Igor's state tracker is a few hundred lines of code that has initial implementations for Cairo paths and painting, but obviously the work has just begun.
Right now it also looks like he's just focusing on making it work with the Softpipe driver, but not any of the hardware drivers yet or even the interesting LLVMpipe.
While his first blog post is a bit lax on details, it can be found here.
In other news related to the X.Org GSoC happenings this summer, Tom Stellard who has been working on the R300 GLSL improvements has just pushed a new branch of his code that enables hardware loops and branching for ATI R500 (Radeon X1000) graphics cards. Information regarding this branch can be found on mesa-dev.
| 8
| 1,760,738,419.019631
|
https://www.phoronix.com/news/ODM0NQ
|
Kernel Mode-Setting With The Glint Driver
|
Michael Larabel
|
Earlier this month we reported on the ATI R300 GLSL compiler improvements being worked on as part of a Google Summer of Code (GSoC) project for X.Org, but how are their other 2010 projects progressing by these student developers? There's five GSoC X.Org projects in total this year and last night we received a bit of an update on the kernel mode-setting (KMS) efforts for porting the Glint driver.
For those that don't recall the scope of this KMS summer project or what the Glint driver is, the goal is to provide kernel mode-setting support for the Permedia 3 and Permedia 4 graphics cards. Right now the X.Org support is provided via user-space mode-setting in the Glint DDX driver. These old OpenGL-capable graphics processors from 3Dlabs are very rare to find these days seeing as they have been out for more than a decade, but under Linux they shall have KMS support for anyone who still uses such a GPU on an antiquated system.
The kernel mode-setting work is being done on the 3Dlabs Permedia 3/4 since writing a kernel driver for a very old graphics processor is much simpler than writing one for a modern GPU, especially over the course of a summer. While this KMS Glint support will be of direct benefit to almost no individuals, the GSoC developer hopes to better document the KMS driver writing process, which could help other developers down the road in targeting new hardware. Of course, with there already being KMS support for Intel, ATI/AMD, and NVIDIA hardware, even then new KMS drivers will really be targeting minority users.
Matt Turner, the student developer working on this 3Dlabs Glint KMS support, has all of his DRM-side code found in this personal kernel repository. Matt has started filling in some of the needed functions for enabling kernel-space mode-setting support, but is a bit uncertain regarding the next step(s) with his KMS support upbringing, so here is his mailing list message with where he is at on this DRM driver. With kernel mode-setting support, Matt needs to add kernel memory management support (via GEM/TTM is to be expected), which still needs to be tackled too. After the Glint KMS driver is up and running, Matt will turn his attention to documenting the process and providing some KMS driver writing tutorials.
If you are interested in more on the history of kernel mode-setting for Linux, this week we posted the slides by Egbert Eich from the Berlin LinuxTag.
| 7
| 1,760,738,419.34505
|
https://www.phoronix.com/news/ODM0MQ
|
X.Org Server 1.9 Release Candidate 1 Has Arrived
|
Michael Larabel
|
More than a week ago we reported on the X.Org Server 1.9 release status and how it was still planned for release in August, the merge window would be closing after some RandR 1.4 code got pulled in, and the first release candidate was supposed to come that day. Well, finally, that first release candidate has arrived.
The first release candidate for X Server 1.9 was delayed because of the Intel Linux graphics driver (xf86-video-intel) not building against the latest 1.9 code due to a driver issue, which is now fixed-up so Keith Packard went ahead and created this first release candidate: xorg-server 1.8.99.901.
There isn't anything too incredible to be found in X.Org Server 1.9 for end-users, but there is the RandR 1.4 support with per-CRTC pixmap support. As was said in that last status update, "This support, as described in the Resize and Rotate extension documentation, 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." A RRSetCrtcConfigs request has also been added to the RandR 1.4 specification. Intel has a branch of their xf86-video-intel driver with RandR 1.4 support while the other open-source drivers are likely to follow soon."
There's also bug fixes and a variety of other work that's been collected since the X.Org Server 1.8 release earlier this year. Expect a few more release candidates to come before the August 1.9 release.
The 1.9 RC1 release announcement can be found on xorg-announce.
| 0
| 1,760,738,419.554451
|
https://www.phoronix.com/news/ODMyNg
|
X Server 1.9 Window Closing After RandR 1.4 Pull
|
Michael Larabel
|
There's good news for the Ubuntu camp and others releasing in the September-October time-frame: development work on X.Org Server 1.9 is still going as planned for an August release and its merge window is about to be closed. In the past it's been tough for the X.Org project to release server updates in a timely manner that's on schedule, but continuing from their X.Org Server 1.8 success, 1.9 is shaping up nicely too.
Keith Packard hoped to close the 1.9 merge window by the end of last week, but a few last minute changes have held it up (particularly with the devPrivate changes). Still pending though are some Quartz changes that will be merged and then also the RandR 1.4 branch. The RandR 1.4 branch isn't as exciting as RandR 1.2, but it's a small update similar to what was RandR 1.3. The RandR 1.4 branch provides per-CRTC pixmap support.
This support, as described in the Resize and Rotate extension documentation, 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." A RRSetCrtcConfigs request has also been added to the RandR 1.4 specification. Intel has a branch of their xf86-video-intel driver with RandR 1.4 support while the other open-source drivers are likely to follow soon.
Beyond RandR 1.4, there isn't anything too major (in terms of new features) to be found in the X.Org Server 1.9 code-base beyond a lot of miscellaneous work.
Keith is also planning to release X.Org Server 1.9 Release Candidate 1 (RC1) today, but your's truly now has a flight to catch for LinuxTag this week in Berlin. The update regarding the 1.9 merge window closing can be found on xorg-devel.
| 8
| 1,760,738,419.938271
|
https://www.phoronix.com/news/ODMxMw
|
Multi-Touch Support For X.Org Evdev Driver
|
Michael Larabel
|
While there is multi-touch support in Qt and also Clutter along with MPX capabilities in GTK+ 3.0, the widely-used xf86-input-evdev driver up to this point has not supported multi-touch. Developers, however, have been working on the said support for evdev, which is the generic Linux input driver.
A new set of patches have been published by Benjamin Tissoires that provide preliminary support for multi-touch within the X.Org evdev driver. The patches can be found here, but of course you need a multi-touch device for it to work.
While these patches have been worked on for a while by Tissoires, Carlos Garnacho has today announced another approach to multi-touch handling. In his lengthy mailing list message he describes what he views as the problems with the current EvDev implementation and how it could impair multi-touch adoption on Linux. With his proposal, Carlos also has a branch of the xf86-input-evdev driver that has his changed code along with a small patch that goes against GTK+ to work more according to his ideal way with multi-touch support.
Nothing is yet committed to the mainline code-bases, but we'll be monitoring the situation to see how multi-touch support evolves.
| 1
| 1,760,738,420.706476
|
https://www.phoronix.com/news/ODMwOA
|
If Or When Will X12 Actually Materialize?
|
Michael Larabel
|
The first version of the X protocol for the X Window System emerged in 1984 and just three years later we were at version 11. However, for the past 23 years, we have been stuck with X11 with no signs of the twelfth revision being in sight, even though there is a whole list of X12 plans and hopes on the FreeDesktop.org Wiki. Julien Danjou, an XCB developer, has written a lengthy blog post looking at the situation and the prospects for the X protocol.
Julien's post briefly recaps the history of the X protocol and then more closely looks at the current X11 protocol. While X11 has been living for more than two decades, numerous extensions have been introduced since then (i.e. X Render) while parts of the X11 protocol have been lost (i.e. Zaphod mode) and other features like server-side fonts are no longer used. Over the course of time, various tool-kits and higher-level software has had to work around the shortcomings of X11, which adds another level of mess and complexity to the situation, which Julien also covers in his blog. Also covered is the lack of interest by GTK+ and Qt developers in targeting the XCB API where as Cairo being one of the few libraries that targets the X C Bindings and then Enlightenment's EFL only has an unmaintained port.
X12 could clean-up some of this mess, but this FreeDesktop.org developer thinks, "To me, it does not seem X12 will happen in the next decade neither." It's also possible that X12 could also never come, if the Wayland Display Server proves to be a success or some alternative solution emerges.
Julien's post that shares his thoughts and ramblings on the X protocol can be found on this page. You can share your thoughts on X11 or hopes for X's future in our forums.
| 103
| 1,760,738,421.239934
|
https://www.phoronix.com/news/ODMxMg
|
X.Org Server 1.8.2 To Come About Soon
|
Michael Larabel
|
X.Org Server 1.8.1 was released just one month ago, but Peter Hutterer is now preparing the release of X.Org Server 1.8.2. Due to time constraints, this is also the last planned point release he plans to put out.
An X.Org Server 1.8.2 release candidate (officially it's known as xorg-server 1.8.1.901) was put out last night and it offers up nearly three dozen bug-fixes since the first X Server 1.8 point release. The most notable set of fixes to be found in the X.Org Server 1.8.2 code include a number of DRI2 fixes by Intel's Jesse Barnes.
While the earlier X.Org Server 1.7 series had seven point releases, Peter is planning to end his X.Org Server 1.8 point releases after this second release in June. Peter is running low on time, but will allow anyone else to take over this role if interested. Fortunately, however, X.Org Server 1.9 should already be coming soon if it meets its August deadline.
This first release candidate for X.Org Server 1.8.2 along with its change-log can be found on the X.Org mailing list.
| 0
| 1,760,738,421.253466
|
https://www.phoronix.com/news/ODI4Mw
|
How VIA Could Have Not Screwed Its Linux Chances
|
Michael Larabel
|
Yesterday we reported on how VIA's open-source Linux dreams are not materializing and then this morning reported on a new secret driver being developed (xf86-video-openvia) between VIA Technologies, some VIA "friends", and the OpenChrome developers. However, as I said in this morning's VIA article, the situation is still a murky mess, this new OpenVIA driver isn't receiving much work, the DRM code with TTM/GEM memory management is still missing, and these open-source efforts by VIA are not very well organized.
VIA's Linux TODO list doesn't even offer hope for having kernel mode-setting support until later this year and any Gallium3D driver for offering OpenGL acceleration will not come until late this year at the earliest. If such code is delivered on time, it will likely not be until H1'2011 that this code is found in most Linux distributions like Ubuntu 11.04, which will be three years since VIA announced its most recent open-source driver strategy.
How though could have VIA avoided this mess when figuring out their open-source role? Well, here's a few things VIA could (and should) have done.
Improve Its Web-Site / Community Portal: Back in April of 2008 when VIA announced its most recent open-source initiative they had launched linux.via.com.tw. A couple files have been added since then, but for the most part there isn't a whole lot to look at. The bug tracking and forum sections of the VIA Linux web-site have been "under construction" since its inception more than two years ago.
The VIA Linux site really doesn't offer much value beyond offering download links for a couple of open- and closed-source graphics drivers and old DRM source code. They don't even reference any Git repositories, the OpenChrome driver they even partnered with, the original X.Org Unichrome driver, or this new xf86-video-openvia driver.
Beyond offering better download information or mentioning what VIA packages can be found in what distribution repositories, they should setup a Wiki or provide information similar to what's available with the Intel / ATI / Nouveau content available via the X.Org Wiki. Via the X.Org Wiki you can easily find information for other drivers on how to build it from Git, what games/apps work with a given driver, and the current status of the driver, among other valuable information. Right now all of this information for VIA's Linux offerings is just scattered around on Phoronix, among other locations on various web-sites. This just makes it a pain for any customers or potential customers to figure out what VIA hardware works under Linux, to what extent, and what hurdles may be encountered in getting the VIA IGPs to even mode-set with their favorite Linux distribution.
For VIA's bug tracking, they could be like all the other open-source Linux graphics drivers could use the FreeDesktop.org BugZilla. If they are after community forums, we are more than happy to help them out on the Phoronix Forums. In fact, we already have an active VIA Linux forum.
VIA also doesn't use FreeDesktop.org for hosting any of their source-code (like the other X.Org components), but the OpenChrome project uses their own SVN server and this new OpenVIA driver is hosted over on an unadvertised GitHub account.
The status quo for them now just leads to a higher barrier of entry for trying out and running the latest VIA Linux code and confusion amongst its potential customers to what is actually supported or working under Linux. VIA has no centralized source.
Hire A Professional: Shortly after making their 2008 Linux announcement, VIA Technologies announced that it contracted Harald Welte to be their open-source liaison. Harald has done good things for Linux and free software communities within the Linux kernel, creating an organization to track down GPL violations, and working with the OpenMoko project, but he hasn't played much of a role within the X.Org community. He really didn't do much while serving as this liaison for VIA and he no longer is even in this role.
VIA's own staff also isn't too involved within the X.Org community and its development processes (as can be seen from many mailing list messages where they are asking questions). VIA also goes unrepresented at X.Org conferences like X@FOSDEM and the annual X Developers' Summit. VIA could have hired or at least contracted an X.Org / Linux veteran to become involved, such as Luc Verhaegen.
Luc may have ambitious ideas and beliefs that go against the ways of other key X.Org developers, which often result in mailing list arguments and disagreements with other X.Org contributors, but he at least is familiar with the graphics driver development process and would have been a fitting candidate to work on community documentation and code. Luc is the maintainer of the original X.Org Unichrome graphics driver that he has worked on in his spare time for some years, was one of the key members of the RadeonHD driver and the open-source AMD proposal, and has other qualifications. This would have been similar to AMD picking up Alex Deucher when they began with their official open-source support a year earlier, in addition to AMD partnering with Novell for work on the RadeonHD driver.
It's not that VIA didn't know of Luc, but they had contacted him along the way to hopefully ride on some of AMD's open-source success. Luc was let go from Novell in 2009 during a round of German layoffs, but VIA Technologies still didn't offer up anything even when he had approached the company. The opportunity is now gone, but they could have easily fetched him to move their open-source efforts along. If they didn't want to work with Luc, they could have approached Red Hat for a business deal where they have many skilled developers with a forte for Linux graphics.
Seed Hardware To Developers: If VIA Technologies doesn't want to hire any employees or contractors (or doesn't have it within their budget), they can at least offer up VIA hardware to active open-source developers. This would certainly motivate such developers and make them more likely to work on VIA support with greater hardware availability. Some of the most notable open-source VIA work to date has been from community developers, such as VMware's Thomas Hellström getting TTM to work with OpenChrome in a branched version.
From Bruce's email to Phoronix last night, it appears that VIA does send some hardware to their "friends" in Taiwan, but this situation isn't clear and it appears those community developers are bound by Non-Disclosure Agreements.
Kill, Open The Blobs: While they may not be able to open-source their entire binary driver code-base due to third-party licensing, not wanting to cause any Digital Rights Management problems, or exposing some sensitive areas, they could at least open-source a majority of their currently closed-source code. Even if it's not in a usable, it should prove to be of use in jump-starting any kernel mode-setting (KMS) or Gallium3D driver efforts within the community. If they diverted their limited development resources away from working on their binary drivers, which are also of very poor quality, their open-source drivers could get greater attention.
With the xf86-video-openvia driver and new but unreleased DRM code it looks like they may be doing this somewhat, but it's years too late to do real good especially when their community developers are limited and not too active.
Actually Do Something: VIA's open-source / Linux work really seem to be a part-time effort at the moment. Documentation and code drops are far from frequent and there really isn't even much communication with the community or other X.Org developers. We do know that VIA Technologies does have more documentation ready than what has been publicly released, but through an NDA they have been funneling some of these documents to OpenChrome developers.
It's a rare occurrence to see any VIA employees write on the OpenChrome, X.Org, or Mesa mailing lists. You also don't see them engaging with their user-base or community, such as responding to questions on message boards such as ours. This is while, AMD for example, has John Bridgman very frequently answering questions of our readers along with Alex Deucher and they are both active on the ATI/X.Org/Mesa mailing lists too. Alex also occasionally writes about graphics driver programming and other topics on his personal blog. VIA also does not appear or represent itself at any X.Org conferences or other Linux development events.
Of course, these suggestions just hit the tip of the iceberg for how VIA Technologies could better improve their open-source Linux support, but sadly we doubt VIA Technologies will finally learn (even though they have failed in the past) and implement any striking changes to make it a really viable option for those consumers interested in open-source Linux graphics drivers. You can share your thoughts for how VIA could improve its Linux image via the discussion link below.
| 21
| 1,760,738,421.839681
|
https://www.phoronix.com/news/ODI3OA
|
VIA's Linux Dreams Are Not Materializing
|
Michael Larabel
|
Back in 2008 there was the announcement from the Linux Foundation Collaboration Summit in Texas that VIA had joined the open-source driver bandwagon after having abandoned previous open-source attempts. However, for the past two years, this has largely been a media bluff. VIA Technologies did things like appoint an open-source liaison (Harald Welte, who is actually no longer contracted by VIA and didn't even do much for their efforts), launch a VIA Linux web-site (that is ill-maintained and two years later there are still portions of the site "under construction"), but they have done some things like put out some code and republishing old documentation. We're almost half-way through 2010 and it doesn't look like VIA will be doing much this year for their open-source graphics drivers.
In late 2009 an article was published entitled VIA's Linux TODO List... Maybe Look Forward To 2011? where we looked at their actual efforts since their 2008 open-source announcement and then at their newest TODO list they had shared with the open-source community. In that TODO list, VIA's Bruce Chang talked about "VIA has been having a dream to host VIA's developing source in public" and that they plan for a TTM/GEM module in Q2'2010, kernel mode-setting for VIA in H2'2010, and a VIA Gallium3D driver in Q4'2010. At least publicly, however, there is not much work going forward and this is far from going as planned where back in 2008 VIA had said in their original announcement they would be providing quarterly driver updates.
We're more than half-way through the second quarter and there are no signs of any TTM/GEM memory management support being imminent. Even if something were to materialize before the end of June (but we wouldn't bet on it), there may be waiting until the Linux 2.6.36 kernel for the DRM code with memory management support to be accepted into the mainline Linux kernel unless it's accepted late in the 2.6.35 cycle -- the merge window is open right now but not for much longer.
If VIA does publish new Direct Rendering Manager (DRM) code that offers Graphics Execution Manager / Translation Table Map support for in-kernel memory management on VIA hardware, it may be another challenge just to get the code accepted by the open-source developers. VIA's previous efforts to inject their DRM into the Linux kernel have been rejected multiple times for various reasons from code problems to there not being open-source drivers (but rather VIA blobs) that actually take advantage of the DRM.
Having DRM code with proper memory management support is a prerequisite for having kernel mode-setting (KMS) support and a Gallium3D driver, so this may in turn end up delaying the rest of VIA's open-source schedule too. This is all while VIA has said they will not provide an open-source Chrome 9 3D driver and that they were looking for the community to make them a Gallium3D driver. On top of this, VIA still has yet to provide any new code or documentation for their newest chipset -- even just for basic DDX mode-setting support -- and code/documentation is still missing for older ASICs.
VIA's open-source efforts are still lousy at best as they just continue dreaming of being "open-source friendly" and playing the Linux community. They really aren't much better than NVIDIA's open-source strategy, which consists of recommending the VESA driver until their proprietary driver can be installed. At least with NVIDIA there is an active community (Nouveau) that's doing the reverse-engineering work and open-source enablement.
We'll keep monitoring the situation and also see if VIA has anything to say on the matter. So far this is just looking like VIA's open-source attempts from years earlier where they just ended up dropping open-source support.
| 15
| 1,760,738,421.848567
|
https://www.phoronix.com/news/ODI3MA
|
The State Of The X.Org Server 1.9 Release
|
Michael Larabel
|
Version 1.8 of the X.Org Server was just released at the start of April, but Intel's Keith Packard who's been serving as the release manager called for an even tighter release schedule with X.Org Server 1.9. Keith pushed plans for an August release of X Server 1.9. With that said, to meet that deadline, the merge window for the 1.9 release is closing at the start of June.
There isn't really any extremely exciting work to be found in the 1.9 release, but there is a lot of bug-fixes and general improvements. Keith has written a message to the X.Org developers to fill them in on the 1.9 status with there being some outstanding patches that he hasn't yet pulled and the 11 bugs that are officially deemed blocker bugs for the 1.9 release.
Due to travel plans, Keith plans to push back the merge window closing by a few days to the 4th of June. The non-critical bug deadline is set for the first of August and the official release is still scheduled for the 20th of August.
It looks like X.Org Server 1.9 will be yet another release that's lacking XKB2, which has been talked about for years, but at least it's looking like it will be the second consecutive X.Org Server release to make it out on time -- rather than months late like has been a frequent occurrence in the past.
X.Org Server 1.9 will be included as part of X.Org 7.6, which is expected for release in September or October. X.Org Server 1.9 will most certainly be found in Fedora 14 and with the likelihood this release will be out in time and not with any aggressive changes, it should be in Ubuntu 10.10.
The state of the 1.9 release can be read about on the xorg-devel list.
| 6
| 1,760,738,422.494737
|
https://www.phoronix.com/news/ODI0Nw
|
The X.Org, Mesa Plans For Ubuntu 10.10
|
Michael Larabel
|
The talk at the Ubuntu Developer Summit surrounding the X.Org plans for Ubuntu 10.10 just wrapped up. Compared to the Ubuntu 10.04 LTS release, Ubuntu 10.10 should provide a much more recent and up-to-date graphics experience.
Plans for Ubuntu 10.10 "Maverick Meerkat" already include shipping with the Linux 2.6.35 kernel, which is wonderful on many different fronts except for the time it may take to support within the ATI Catalyst Linux driver. This means Ubuntu Maverick users will get all of the DRM advancements made in the Linux 2.6.34 kernel like initial Radeon KMS power management and also some of the DRM improvements that are expected to me merged into the Linux 2.6.35 kernel when its development commences in a few weeks. Some of the Direct Rendering Manager improvements that are likely to be integrated into the Linux 2.6.35 kernel include H.264 VA-API support for Intel Clarkdale/Arrandale ASICs, Radeon HD 5000 / Evergreen improvements, Radeon DRI2 swap and sync support, and better ATI Radeon power management support.
With Ubuntu 10.10 not being a Long-Term Support (LTS) release for Canonical, they are open to shipping X.Org Server 1.9 assuming it is released in August. However, should this release be delayed, they will be either shipping an X.Org Server 1.8 point release or a 1.9 snapshot / release candidate. Ubuntu 10.04 LTS ships with X.Org Server 1.7.6.
When it comes to the Mesa stack, they are considering Mesa 7.9 if it is looking like it will be quite stable or if the 7.9.1 release will be made available with adequate time prior to Maverick's October release. Otherwise a Mesa 7.8 point release will be used. Mesa 7.9 may also mark the point where the ATI R300 "Classic Mesa" driver is swapped out by default to the R300g Gallium3D driver. If Ubuntu 10.10 is shipping with Mesa 7.9 and that change occurs, Ubuntu developers will be going through with the switch to Gallium3D.
In terms of Nouveau support in Mesa on Ubuntu, Fedora developers began making this an option with Fedora 13 and our Nouveau benchmarks already show it running fairly well. However, for Ubuntu 10.10 they are not looking at providing Nouveau 3D support by default. A separate Mesa driver package may be available from the Ubuntu Universe repository that would offer the option of installing the Nouveau OpenGL support. It was just with Ubuntu 10.04 LTS that they even enabled the Nouveau KMS driver by default for NVIDIA hardware.
While the X.Org DDX drivers aren't as imporant anymore with much of the work going on within the DRM/KMS drivers, these versions too were discussed during this X.Org meeting. Ubuntu 10.10 will likely continue to receive an updated xf86-video-ati Git snapshot, xf86-video-nouveau should have a Git snapshot too unless a stable release is made available, and xf86-video-intel will be at version 2.11 or potentially 2.12. The Intel DDX driver update means that they will be saying farewell to any user-space mode-setting support from this driver so those with ancient Intel i8xx hardware will be dropped down to using the VESA driver in poorly supported configurations or a xf86-video-intel 2.9 release bearing UMS support may be available through the universe repository.
| 24
| 1,760,738,422.503762
|
https://www.phoronix.com/news/ODIxMA
|
X.Org Server 1.8.1 Gets Into Shape With RC2
|
Michael Larabel
|
The first release candidate for X.Org Server 1.8.1 was only released seven days ago, but since then a good amount of fixes have already worked their way into the X.Org 1.8 series branch for this first point release.
There are just under two dozen additional commits in X.Org Server 1.8.1 RC2, with notable changes being the RECORD extension is now enabled by default and the SHA1 implementation is picked in a slightly different order. There's also a variety of other fixes in RC2.
"Please test this, I don't think I'll merge much this week but really crucial issues. So expect this one to be quite close to 1.8.1," said Peter Hutterer in the xorg-server 1.8.0.902 announcement. Meanwhile all new development activities are going towards X Server 1.9, which should arrive in August.
| 6
| 1,760,738,422.999048
|
https://www.phoronix.com/news/ODI0Mg
|
X.Org Server 1.8.1 Released To The Wild
|
Michael Larabel
|
X Server 1.8 was released in early April, but the first point release to this major X.Org update has now been pushed out by Peter Hutterer.
X.Org Server 1.8.1 boasts a variety of fixes and minor improvements while all major work is already focused on delivering X Server 1.9 that should make its debut in August.
X.Org Server 1.8 introduced support for xorg.conf.d, marked the removal of HAL support in favor of OS-specific input libraries (such as libudev on Linux), and various other bug-fixes and improvements.
X.Org Server 1.8.1 final isn't very different from the last RC that made the RECORD extension work and received a slightly different SHA1 implementation. There are just two new patches atop X.Org Server 1.8.1 RC2.
In total X Server 1.8.1 provides a couple dozen patches over X Server 1.8.0, which is listed in this mailing list message.
| 0
| 1,760,738,423.015651
|
https://www.phoronix.com/news/ODE5MQ
|
X.Org Server 1.8.1 Gets Ready For May Release
|
Michael Larabel
|
X.Org Server 1.9 is the key interest now among X developers working on the graphics and input stacks with its release coming as soon as August, but Peter Hutterer once again is taking over the role of maintaining the stable X.Org Server branch. With X.Org Server 1.7.7 being nearly out of the way, the first point release for X.Org Server 1.8 is being prepared for release.
Peter has put out a X.Org Server 1.8.1 release candidate, which pulls in quite a number of bug-fix patches from the current X.Org Server master code-base that's being built-up for the 1.9 release. For distribution vendors, X.Org Server 1.8.1 now offers two directories for carrying distribution-specific patches and the other for custom, user-installed patches.
The rest of the 43 patches that make up the X.Org Server 1.8.1 RC1 release range from XQuartz fixes to GLX and XKB fixes. A second release candidate for X.Org Server 1.8.1 is planned for early May and after that the first (of a couple) point releases will make it out into the world.
The X.Org Server 1.8.1 release candidate announcement can be read on the X.Org mailing list.
| 1
| 1,760,738,423.613028
|
https://www.phoronix.com/news/ODE4Ng
|
X.Org Project Has Five New Summer Projects
|
Michael Larabel
|
Back in March we talked about the possible X.Org projects this year during Google's Summer of Code, for which X.Org is a veteran participant (in the past items like the ATI R300 Gallium3D driver and generic GPU video decoding have been tackled), but the list of accepted projects for this summer have now been announced. Gallium3D H.264 video decoding, an OpenGL 3.2 state tracker, and porting of the DRM code to GNU/Hurd were among the talked about possibilities, but none of those will be addressed as part of GSoC 2010.
Instead, being worked on this summer within the X.Org project thanks to Google is improving input support for XCB, a Cairo state tracker for Gallium3D, creating a fully plug-and-play USB multi-seat solution using upstream software components, KMS support for the Permedia 3/4 graphics cards, and improving the GLSL compiler back-end for the ATI R300 driver.
Seeing a Gallium3D state tracker for Cairo will be particularly exciting for this 2D graphics API as it's used by GTK+, Mono, Mozilla Firefox / Gecko, WebKit, and many other free software projects. The Cairo state tracker will be joining the OpenVG and Xorg state trackers as other accelerated 2D components for this graphics driver architecture.
It's also exciting to see kernel mode-setting support being worked on for more hardware, but the Permedia 3/4 hardware comes as a bit surprising, if you even recognize the name. The Permedia 3 and Permedia 4 are products of 3Dlabs. These were low-end OpenGL-capable graphics cards that were offered by 3Dlabs prior to their acquisition by ATI. The 3Dlabs Permedia graphics cards that have been around for more than a decade are rightfully rare these days, but it's picking up kernel mode-setting support. The purpose of this GSoC project though is to document the KMS driver writing process rather than expecting this driver to have a useful life. Going with an older, simpler graphics processor should make it easier to implement a KMS driver over the course of a summer than a modern GPU that is much more advanced.
Improving the GLSL (GL Shading Language) compiler for the Gallium3D R300 driver will also be much appreciated by many users and should provide noticeable results once complete.
Christoph Reimann, Igor Trindade Oliveira, Lucas Ferreira, Matt Turner, and Thomas Stellard are the five new X.Org student developers that will take on these new projects this summer. Best of luck in successfully completing their projects. The announcement of these accepted projects was made on the X.Org mailing list.
| 19
| 1,760,738,424.341673
|
https://www.phoronix.com/news/ODE0Mw
|
X Server 1.7.7 Will Be Released, RC Arrives
|
Michael Larabel
|
While X Server 1.8 was released last week, Peter Hutterer is still planning to put out a final point release in the X Server 1.7 series before moving on to maintain the stable 1.8 series.
There were five point releases planned for X Server 1.7, but a sixth came about that fixed the RECORD extension, and now a seventh is also emerging. X.Org Server 1.7.7 brings a few fixes to XQuartz, DIX, KDrive, and other areas. Nothing really exciting though is contained within the xorg-server 1.7.7. change-log.
Peter has just announced the release candidate for X.Org Server 1.7.7 on the Xorg mailing list. This is the only planned test build before releasing the final version at some point in the near future.
The X Server 1.8 point releases will too be maintained by Peter and handled in the same way as he has done with the 1.7 series. Meanwhile, all new development efforts for X.Org have been going into X Server 1.9, which is scheduled to be released in August.
| 1
| 1,760,738,424.875074
|
https://www.phoronix.com/news/ODE2Nw
|
X.Org Server 1.7.7 To Be Released Soon
|
Michael Larabel
|
While X Server 1.8 has been around for almost one month, Peter Hutterer for now is still maintaining the X.Org Server 1.7 series before turning over to the 1.8 branch in the coming weeks. He's now in the process of preparing to release X.Org Server 1.7.7.
For those that are not yet migrating to X.Org Server 1.8 with its xorg.conf.d support, libudev support replacing HAL, and other changes, the 1.7.7 release should be out soon. Peter released a X Server 1.7.7 release candidate earlier this month, but coming this afternoon is a second and potentially final release candidate.
This second release candidate for the X.Org Server 1.7.7 series drops the XGE extension event to workout a problem experienced by some NVIDIA users that ran out of event opcodes within the X Server. There are also a couple crash fixes and for Mac OS X users there are XQuartz updates.
Peter has also been pushing patches into X.Org Server 1.7.7 that get the X Test Suite working, but at least one more patch is on the way for this, which after that point he will likely be releasing 1.7.7 final. An X.Org Server 1.7.8 release is not planned at this point.
The X.Org Server 1.7.7 RC2 release announcement with its change-log from RC1 can be found on the X.Org mailing list.
| 0
| 1,760,738,424.889834
|
https://www.phoronix.com/news/ODEyOA
|
X Server 1.9 Is Scheduled To Be Here In August
|
Michael Larabel
|
X Server 1.8 was only released a few days ago, but Keith Packard is quickly focusing his attention to X Server 1.9, where it looks like he is interested in being the release manager once again.
X Server 1.8 was a success in terms of getting more patch review done and getting the release out (nearly) on time for once, so X Server 1.9 will follow much the same process. However, Keith has called for a few more changes with X Server 1.9.
In preparation of the X.Org graphics drivers moving back into the X.Org Server at some point (likely X Server 1.10 or later), Keith will first be merging all of the X.Org protocol headers into a single repository. This should happen in the coming days so that all of these source header files can be obtained from a single, central location.
Also with the hopes of moving the graphics drivers back into the X.Org Server -- something that Luc Verhaegen is in stiff opposition to -- the X.Org Server release process may be even shorter. Keith is calling now for a three month or so release cycle. This is in contrast to X Server 1.8 that was done on a six month release schedule, but for older releases we have seen that stretch out much further.
Quicker releases of the X.Org Server would be needed to deliver new hardware support and other changes since the X.Org graphics drivers would no longer release updates individually. Daniel Stone has already publicly questioned what about delivering such driver updates in X.Org Server point releases rather than new full-blown server versions, but this has yet to be discussed on the mailing list.
For the first even shorter release schedule, Keith Packard is calling for X Server 1.9 to be released in August. Releasing in August would allow time for X Server 1.9 to make it into Fedora 14, Ubuntu 10.10, and a later release of MeeGo, for which Keith is now interested in as well due to Intel's involvement.
This 1.9 release schedule that was published today would result in the merge window closing on the 1st of June, a non-critical bug deadline going into effect on the 1st of August, and the official X Server 1.9 release to come on the 20th of August. There would also be snapshots / release candidates in between those targeted milestones.
Packard's X.Org Server 1.9 thoughts can be read on the xorg-devel mailing list. As of right now there isn't too much discussion surrounding these proposed changes.
Nothing too exciting is planned for integration with X Server 1.9 besides maybe seeing XKB2 finally.
| 8
| 1,760,738,425.525526
|
https://www.phoronix.com/news/ODEzNQ
|
Concerns Over Merging Drivers Back Into The X Server
|
Michael Larabel
|
While development efforts within the X.Org community are now ramping up for the release of X Server 1.9 that should arrive in August, there is an ongoing discussion concerning a planned long-term change for the X Server: pulling the drivers back in.
With X.Org came the modularization of drivers compared to XFree86, which allow drivers to be built independently of the X Server. However, it was proposed a few months back that with X Server 1.10 that the drivers get pulled back into the X Server. This would result in X.Org becoming a more monolithic package, but would allow driver developers to more easily break the API/ABI and other invasive changes.
To users, moving the drivers back into the X Server means that you must build the entire server + drivers whenever you want to update your lone DDX driver. Building the X Server takes longer and requires more dependencies than building a single driver like xf86-video-ati or xf86-video-intel. This cause even fewer people to test Git snapshots, which is a problem for a project that already could use more Linux desktop users testing out the latest code -- not less people.
Moving the drivers back in would also mean that the driver developers could no longer release updates independent of the X Server, but that the X Server would need to be released more frequently to provide support for new graphics cards and features. Intel's Keith Packard has already said this would mean a three to four month release schedule for the X Server, where as in the past the X Server has been released every six months, but with delays it's often months longer.
As a step towards moving at least the video drivers back into the X Server, Keith Packard has made a single protocol headers package for X Server 1.9 and that's what has set off this most recent discussion about where X.Org video drivers belong.
Some driver developers like Alex Deucher (AMD) have voiced concerns on moving the graphics drivers back into the X Server in that it will hamper inexperienced end-users in testing out the new code even more. As Daniel Stone has also pointed out, it's harder to test out new drivers if the state of the X Server Git code is in a poor state where one looking to just test out new graphics driver code could be stopped by bugs in say the input driver. It's been mostly Intel's developers like Keith Packard and Jesse Barnes voicing support for this structural move.
Also having been discussed is whether support for new graphics cards could be introduced in an X Server point release or whether it would wait until the next major X.Org release. Well, it could go either way, but providing it as a point release would mean that the few X.Org driver developers are taxed even further by now needing to back-port the code to an older X Server code-base.
Stephane Marchesin has also asked what gains this would really be when the APIs for EXA, RandR, and others have already been worked through mostly. Jesse Barnes agrees with this partially, but merging them back in would allow Intel to drop some conditional code from their driver that is targeted at older X Server releases and also to share more code amongst DDX drivers by placing common code within the X Server.
Alan Coopersmith of Sun/Oracle is also concerned about back-porting driver fixes under this proposed X Server model particularly with the enterprise operating systems / distributions. Pulling an entirely new X Server into an already released and supported distribution could cause troubles.
Another issue touched on by Keith is that in some cases building a new X Server would require updating to new libdrm code that's released quite frequently. However, in the end Keith says, "Let's see what we think in a few months when we're starting to do planning for 1.10; we'll have had some experience with the merged protocol headers by that point (I hope), perhaps that's all we need to do?"
Right now the views are pretty mixed but we will have to wait until August after X Server 1.9 is released to see what the final verdict is on moving the X.Org drivers back into the server area. If this change does go into X Server 1.10, it will likely impact users of Ubuntu 11.04, Fedora 15, and other H1'2011 Linux distributions.
| 67
| 1,760,738,425.534719
|
https://www.phoronix.com/news/ODA4OQ
|
X.Org Server 1.8 Release Candidate 2
|
Michael Larabel
|
X Server 1.8 may be the first X.Org release in recent times where it's released on time or at least close to being on schedule. Back in October, X Server 1.8 was given a release date of the 31st of March. This is just a little over a week away, but it looks like this next major update to the X Server that brings udev input handling, DRI2 updates, xorg.conf.d support, and other changes.
The first 1.8 snapshot came in late October followed by a second snapshot just before the winter holidays. The first X Server 1.8 release candidate came in mid-February. Over the night, Keith Packard has now put out the second release candidate.
X Server 1.8 Release Candidate 2 has arrived and its fixes are scattered across the board from XQuartz to DIX. The complete list of changes between RC1 and RC2 can be found in Keith's announcement.
According to the published schedule, we're supposed to be closing in on the release within the next couple of weeks. I'm pretty satisfied with the current state of the code, but I'd love to hear about regressions that people are finding so that we can clean things up before the release. At least it won't be as late as X.Org 7.5, which was supposed to come in April, but actually didn't come until late October, and these sort of delays have been the case with other recent releases too. The difference this time around is a greater commitment to delivering releases on a timely basis and there's also development process changes.
| 2
| 1,760,738,426.175416
|
https://www.phoronix.com/news/ODExOA
|
X.Org Server 1.8.0 Is Here
|
Michael Larabel
|
X Server 1.8 is here. Being developed under a new process, this new X Server from the X.Org project was given a 31 March release date. It wasn't released on Wednesday as planned, but it's coming today -- just two days later. Based upon the sizable delays in earlier X.Org Server releases, this is not bad at all.
The X.Org Server 1.8 release offers udev support to replace HAL on Linux, xorg.conf.d support, and much more. All of the X Server 1.8 chances are talked about in our Coming Soon: X Server 1.8 article from earlier this week.
The next release will be X Server 1.9 with X.Org 7.6 and it's expected to be released in late September or October. Kudos to Keith Packard (who has served as the new release manager) and the X.Org developers for getting this work done and pretty much on time.
At the time of publishing, an official X Server 1.8 release announcement has yet to appear, but you can find this new release tagged in Git.
| 21
| 1,760,738,426.183546
|
https://www.phoronix.com/news/ODA2MQ
|
Luc Modularizes Mesa, DRI Drivers
|
Michael Larabel
|
Last month at the X@FOSDEM meeting in Belgium, Luc Verhaegen gave a talk on cleaning up the Linux graphics driver stack. This talk was met by some that agreed with his views and, well, others that didn't exactly see eye-to-eye with him. He shared with everyone his views on changes that should be made to the Mesa/X.Org/DRM stack, including some greater modularization, in an effort to make the testing / build process easier, drive greater maintainability of drivers, and unifying components by providing more shared libraries and formalizing different APIs.
Luc showcased his xf86-video-unichrome driver that he modularized according to his proposed changes, but that didn't satisfy everyone since the VIA Unichrome driver is not nearly as complex as the Radeon, Nouveau, or Intel Linux drivers. As a result, Luc has begun modularizing everything himself in an attempt to prove the feasibility of this design.
Over on his FreeDesktop.org Cgit page Luc is now hosting branched Mesa DRI drivers for the Intel i810, Intel i9xx, Mach64, MGA, R128, Savage, SiS, Tdfx, and Unichrome that are done according to his philosophy. There's also the Mesa DRI SDK headers that he has created. The modularized Mesa Radeon driver will also be there once the FreeDesktop.org server has updated. Luc also has other changes pending that should be committed shortly.
| 26
| 1,760,738,426.795378
|
https://www.phoronix.com/news/ODA3MQ
|
With KMS, Now Run Two X Servers Off One GPU
|
Michael Larabel
|
Over the past several weeks there have been a number of new Linux graphics features introduced by David Airlie, a Red Hat employee and long-time X.Org contributer. Last month David began on a project rampage by bringing hybrid graphics to Linux via code he called "vga_switcheroo" to switch between ATI/NVIDIA/Intel GPUs without rebooting the system (though restarting the X.Org Server is needed at this time) that that code has now made its way into the mainline Linux kernel. Last week another David Airlie project was multi-GPU rendering support for Linux that was written as a proof of concept to show a second GPU could render 3D applications onto the screen of the first GPU, regardless of the hardware vendor. This week we now have the ability to run two X.Org Servers for a multi-head setup off a single graphics card.
This code that allows two X.Org Servers to run off one graphics card is still very much a prototype like the multi-GPU rendering support, but it's working for David and patches are available. The patches go against libdrm and the Linux kernel and depend upon the graphics hardware driver using kernel mode-setting (KMS) for all of this to work out. In essence you can now easily have a two seat setup off one graphics card thanks to these DRM mapping changes and the introduction of render device nodes.
There's still a few items left on Airlie's to-do list for this feature and he's hoping someone else will pickup the work so he can move onto prototyping his next feature, but those interested can read about it on his blog.
| 12
| 1,760,738,426.80383
|
https://www.phoronix.com/news/ODA1OQ
|
Proof Of Concept: Open-Source Multi-GPU Rendering!
|
Michael Larabel
|
Now that David Airlie's vga_switcheroo has went upstream in the Linux 2.6.34 kernel that provides hybrid graphics support and delayed GPU switching, David went on to look for something new to work on in his downtime when not busy with tasks at Red Hat. This new work is on GPU offloading / multi-GPU rendering.
Last month NVIDIA introduced Optimus as a way for dual-GPU notebooks to seamlessly switch between the two GPUs but also to offload the rendering workload to the other graphics processor. This is somewhat similar to NVIDIA's SLI and ATI/AMD's CrossFire for splitting the rendering workload across multiple GPUs, but it has its differences. David ended up developing a proof-of-concept similar to NVIDIA's Optimus that he is calling "Prime" and it works with Intel and ATI GPUs.
David's goals with Prime are to allow a second GPU to render 3D applications onto the screen of the first GPU, with it being configurable by the client, and just to handle the rendering side. This work isn't as simple as his vga_switcheroo implementation, but it required changes to the Linux kernel and the Graphics Execution Manager (GEM), the DRI2 protocol, the X Server and DRI2 modules, and then the actual Linux hardware drivers.
All of this code has already been published as a proof-of-concept, but David shares on his blog that he's unlikely to personally take this work further by upstreaming the code. He has been successful though in using this code to offload the rendering work from an Intel IGP that's driving a display to a discrete ATI graphics processor.
Right now Intel and ATI hardware is supported, but NVIDIA GPUs could be supported too. This work depends upon a system using DRI2 (albeit with these out-of-tree patches) and a compositing manager must be running. David also shares, "To make this as good as Windows we need to seriously re-architect the X server + drivers. At the moment you can't load an X driver without having a screen to attach it to, I don't really want a screen for the slave driver, however I still have to have one all setup and doing nothing and hopefully not getting in the way. We'd need to separate screen + drivers a lot better. Having some sort of dynamic screens would probably fall out of this work if someone decides to actually do it."
It would be wonderful if this work on Prime could be continued and it works its way upstream or that someone takes the reigns from David to continue on with this GPU offloading work for open-source drivers. First though it may make more sense to focus on getting decent performance out of a single GPU before dealing with multi-GPU excitement.
| 26
| 1,760,738,427.451302
|
https://www.phoronix.com/news/ODA0Ng
|
X.Org SoC: Gallium3D H.264, OpenGL 3.2, GNU/Hurd
|
Michael Larabel
|
There's a few months left until it's summertime in the northern hemisphere, but Google is already preparing for their annual Summer of Code (SoC) project as are their projects involved. X.Org will once again be part of the Summer of Code program where Google pays various student developers to work on different free software projects. While nothing is yet officially determined for the X.Org SoC work, there are some ideas expressed by the X.Org developers for any interested students.
So far on the 2010 idea list is Gallium H.264 decoding, an OpenGL 3.2 State Tracker, an OpenGL fixed function "super shader" implementation, automated testing, VT switching for GNU/Hurd, and porting DRM to GNU/Hurd. The H.264 video decoding for Tungsten's 3D driver architecture would involve writing a VDPAU state tracker that implements this popular (and quite good) NVIDIA API. The GNU/Hurd ideas involve bringing the Direct Rendering Manager code to this GNU operating system, which consists of memory management (GEM/TTM), kernel mode-setting, and then the 3D code too. What makes the GNU/Hurd work special is the interesting implications presented by its micro-kernel design.
The "automated testing" SoC idea mentions: "Ideas: IRC/mail notification of new regressions, graphing of the performance data cairobench and x11perf give us (with automatic marking of regressions), more functional tests." Any semi-technical student could easily accomplish this in a weekend with the tools that we provide... The Phoronix Test Suite already fully supports the automated testing of x11perf and Cairobench will be soon (still waiting on Intel to hand over the test profile), Phoromatic can then be used for aggregating the results from the runs that are scheduled daily or triggered on an action (say a new Git commit), then Phoromatic Tracker provides the public interface and the other features like automatic regression detecting, etc. Other graphics tests too could be easily strapped on that are already supported by our GPLv3 software. We're already doing this for the entire Linux kernel on a daily basis at kernel-tracker.phoromatic.com and it basically took a day to setup this tracker. We also have other trackers on the way too, but we certainly would love to see X.Org get involved with better testing. About the only actual work required would be setting up a small shim to allow jhautobuild that the X.Org project uses for automated builds that allows it to connect-in and send triggers to the Phoromatic.
The full list of Google Summer of Code ideas for the X.Org project can be found on the project Wiki.
| 24
| 1,760,738,428.157547
|
https://www.phoronix.com/news/ODAzNA
|
X Server 1.7.6 Restores The RECORD Extension
|
Michael Larabel
|
While X Server 1.8 should be out later this month, enough bug-fixes have come along since X Server 1.7.5 (the last scheduled maintenance release) to warrant a new version. Peter Hutterer pushed out the first release candidate for X Server 1.7.6. With X Server 1.7.6 there will be many bug-fixes atop 1.7.5, including this being the first 1.7.x release where the server's RECORD extension is actually working. There's also about a dozen other fixes to the DIX, xselinux, and other areas of the server stack.
The xorg-server 1.7.6 RC1 release announcement can be read on the mailing list while the final release should be out this month.
| 2
| 1,760,738,428.654238
|
https://www.phoronix.com/news/ODAwNQ
|
The State of The X.Org Foundation 2010
|
Michael Larabel
|
Along with announcing the X.Org Foundation board of director results, Bart Massey also issued the 2010 State of the X.Org Foundation report. This lengthy report on the state of the foundation for this year can be read on the mailing list. It really doesn't detail the financial situation well like many are after, but just an overview of the X.Org happenings now and going forward. Below are a few take-aways from the report.
- The X.Org work with Google Summer of Code 2009 was successful but X.Org Endless Vacation of Code was not. The X.Org Endless Vacation of Code project was spawned from Google's Summer of Code, but it was really not promoted and there didn't end up being any student developers contributing.
- XDS 2010 is in France this September, but the foundation is already looking at XDC 2011. Next year's X Developers' Conference may be in Brazil.
- Funding options for the X.Org Foundation will be again explored in 2010 after they stopped actively collecting money from sponsors a few years back.
- The report's conclusion: "The state of the X.Org Foundation is strong. The X Window System continues to succeed. Each year that I have prepared this report I have confidently predicted that the following year will be the Year of the Linux Desktop. However, I do not believe that 2010 will be the Year of the Linux Desktop, due to decreasing interest in desktops and laptops. I am quite sure that 2010 will be the Year of Mobile Linux."
| 1
| 1,760,738,428.668664
|
https://www.phoronix.com/news/ODAxMQ
|
XGI Working On Big Linux Patch For Big Feature
|
Michael Larabel
|
XGI's Linux driver was written off as dead back in the summer of 2008 when Ian Romanick who had been working on the open-source XGI Linux driver through his work at IBM had left the company to go join Intel where he now works on Mesa and Intel's open-source 3D stack. XGI Technology really hasn't put out new hardware lately and has been flying under the radar, but this past January a long-time XGI employee began working on this code.
Jong Lin is the XGI engineer who has been working for the company for sometime and is someone that we communicated with years ago when XGI was in its prime. Last month we found out that for the xf86-video-xgi driver stack he was working on 2D EXA acceleration, EDID with I2C, custom modes of mode-line, and support for the ARM CPU architecture. Jong isn't too familiar with the X.Org development process and this morning he has more questions about it.
Hi Gaetan,
It would be a problem to divide the big patch into small ones because of big features such as EXA support. The individual feature of EXA support is still a big patch.
As I know, Ian was the maintainer of XGI driver by 5/2008 but he hasn't been in charge of it for a long time. This is what XGI's Jong Lin wrote today to the xorg-devel mailing list. It looks like Jonh is getting ready to send a batch of patches to the list, but apparently the patches are big because of "big features" (hopefully more than just EXA) so Jong is concerned about the length. We'll see what patches end up getting submitted.
| 1
| 1,760,738,429.201257
|
https://www.phoronix.com/news/ODAyNQ
|
Luc Creates X.Org Board of Director Summaries
|
Michael Larabel
|
Last month we reported on issues within the X.Org Foundation that were raised by Luc Verhaegen over the lack of transparency and issues within this crucial organization that backs the development of the X.Org Server. One of the problems expressed was over the lack of minutes that have not been readily available for the X.Org Foundation Board of Director meetings nor any IRC logs. Well, that is now being changed with the meeting logs being published and the IRC information also being published for any interested individuals to tune in to the chat.
As another step, Luc has now published X.Org Board of Director meeting summaries going back to last year. While some meeting summaries are lacking some information due to not all of the IRC logs being published, those interested in these text summaries can view this directory.
| 0
| 1,760,738,429.209383
|
https://www.phoronix.com/news/Nzk5NA
|
Issues Within The X.Org Foundation?
|
Michael Larabel
|
For the past two weeks elections have been going on by X.Org members to elect five people to serve as board of directors for the X.Org Foundation, the formal 501(c)(3) organization that backs the development of the X.Org project. The elections for the board of directors takes place annually replacing four of the eight members each time around, but this year the elections have been particularly interesting. The X.Org Foundation itself isn't in the public spotlight too much and there really isn't much in the way of public communication and involvement outside of this gang of eight. Since the elections started there has been a rather explosive mailing list discussion started by Luc Verhaegen and it has revealed new details about this foundation.
It all started out by Luc asking about company affiliations of the candidates seeking a position on the board. The X.Org Foundation bylaws states that only two people from a given company are allowed a seat on the board, but this year there are three candidates from Intel: Eric Anholt, Keith Packard, and Carl Worth. After that, Luc dived in with another question asking about minutes from the X.Org Foundation meetings and the inability for others to see them. Additionally, as an X.Org member, he inquired about the financial position of the X.Org Foundation since that information isn't readily available to its members or the general public. In response to Luc's question, Daniel Stone admitted that they do "an abysmal job at keeping minutes" and that its hard for the board members to even always meet on IRC due to time zone differences. It was then debated why a summary of the X.Org Foundation is not published on the Wiki after the state of the X.Org Foundation is supposed to be discussed at every XDC/XDS conference, but not every X.Org member is present at these events. The lack of time was basically that answer. The X.Org board meetings that take place on IRC are also supposed to be open to the public, but the details surrounding those meetings aren't readily available, up until now.
Next the discussion turned to the financial position of the X.Org Foundation as even some board members are uncertain about the cash levels and haven't seen statements in years. Keith Packard them stated that the foundation currently has around $125,000 USD on hand and then detailed the $10,380 worth of costs to the X.Org Foundation in the past year, which included XDS 2008, XDC 2009, and web hosting for the project. Once that was said, however, the discussion turned to the price of web-hosting. The X.Org Foundation is paying $3,000 USD annually for hosting at MIT across three servers even though they aren't all utilized. For a few days the discussion then turned to these servers and what should be done with them.
Yesterday, Egbert Eich of Novell mentioned on the list that at the start of 2007 the X.Org Foundation had $222,816 within their accounts. If this is the case, their burn rate for the past three years is approximately $33,000 annually, which is a far cry from the $10,380 that was stated for the past year.
So far the last of these questions have gone unanswered, but as always we are watching the mailing list for updates. Hopefully this ongoing discussion will result in greater communication and transparency within the X.Org Foundation as it seems the status quo causes grief for many on both sides of the table.
| 11
| 1,760,738,429.712162
|
https://www.phoronix.com/news/ODAwNA
|
AMD Gets A Seat With The X.Org Foundation
|
Michael Larabel
|
While concerns have been raised about the state of the X.Org Foundation, the 2010 elections have ended and were not extended though it was requested by some X.Org members. The new X.Org Foundation board members include Alex Deucher, Keith Packard, Matthieu Herrb, Matthias Hopf, and Eric Anholt. Alex Deucher had the most votes to be seated and this his first time sitting on the board and now provides some AMD representation where he works on their open-source driver stack and documentation. With Matthias Hopf on the board, Novell GmbH has a say as Egbert Eich of the same Novell X team is no longer serving.
The X.Org Foundation 2010 election results can be found on the X.Org mailing list.
| 6
| 1,760,738,429.724
|
https://www.phoronix.com/news/Nzk4OQ
|
Talking Radeon GLSL From X@FOSDEM 2010
|
Michael Larabel
|
Last week we provided an update as to our X@FOSDEM 2010 video status. Well, it's still taken some time to get these recordings out there that were taken earlier this month in Brussels, Belgium. While I am back to my normal working schedule, the uploading has been hampered by the file size of these 720p HD recordings. The videos are too long to be hosted on YouTube, we don't have the bandwidth to allocate for these massive video files for non-Premium users, and most of the other video sites out there don't allow video recordings of an hour in length or file sizes greater than 1GB or have a decent Flash Player.
However, with many of the presentation slides being out there, the video is not nearly as important as the audio. After all, the videos are just capturing a portion of the slides that were projected due to the projection screen's proximity to the camera. Most of our video files are 3~4GB in size and once compressing them the video quality is shot. The audio quality this year though is quite good. As a result, we have started uploading MP3 files of the FOSDEM talks.
If you want to listen to Nicolai Hähnle's talk about the R300 GLSL (GL Shading Language) status and road-map, that 40 minute recording is now available. We have hosted the MP3 at this TooFiles location that allows streaming from its Flash Player or download in MP3 format. Here are his slides (ODP format) if you wish to follow along.
More of the recordings are on the way.
| 49
| 1,760,738,430.23595
|
https://www.phoronix.com/news/Nzk5MA
|
Listen Now: Luc's Heated Talk From X@FOSDEM
|
Michael Larabel
|
The most heated talk this year during FOSDEM in the X.Org development room was certainly the talk by Luc Verhaegen with his ambitions to clean up the Linux graphics driver stack. Building the entire X.Org stack can be a mess and there is certainly areas to improve upon in the development process and making it easier for end-users and others to test out this latest code. Luc's goal for this is to create unified trees for each driver that contain all of the driver-specific code rather than having various bits scattered all over the place.
Luc's proposal involves reorganizing the kernel DRM, separating out the various parts of the DRM, moving out driver-specific code from libdrm, splitting out the mode-setting / memory management components, and to modularize Mesa and provide a formal API for this graphics stack. Luc has done a proof of concept for this new unified tree design using his VIA xf86-video-unichrome driver and encourages other X.Org developers to do the same. Luc didn't expect this talk to be received positively by all developers that were there in Brussels, and that certainly was not the case.
Click here to listen to the hour long MP3 of Luc's unified graphics tree talk. It's complete with the heated Q&A from Intel's Eric Anholt and others. Afterwards the polarized debate and discussion spilled into our forums. Luc's slides that go along with the audio can be found in this Phoronix article.
| 17
| 1,760,738,430.245159
|
https://www.phoronix.com/news/Nzk4Ng
|
Fifth Stable Update Ends Out X Server 1.7 Series
|
Michael Larabel
|
X Server 1.7 was released back in October with support for X Input 2.0, Multi-Pointer X, and other improvements after having arrived several months late. Since that initial 1.7 release some five months ago there have been new stable updates / point releases made available quite regularly. These updates incorporate bug and security fixes -- some of which work has been back-ported from the current X Server 1.8 code that should be released towards the end of March. X Server 1.7.5, however, was just released and this will likely be the last official X.Org update to the xorg-server 1.7 series.
X Server 1.7.5 doesn't have much to offer beyond the 1.7.5 release candidates from weeks ago, but mostly smaller changes scattered throughout the X Server code-base. Pater Hutterer in announcing the X Server 1.7.5 release had mentioned, "This one is the last scheduled update. The patch flow is slow enough now to
stop the fortnightly tarballs and do releases on a need-per-need basis." In other words, an X Server 1.7.6 release could come about, but only if the need really arises with an unforeseen situation.
The X Server 1.7.5 release announcement can be read on the X.Org mailing list. X Server 1.8 should be released in just over a month, but some distributions releasing in the coming months (such as Ubuntu 10.04 LTS) are sticking with the X Server 1.7 until H2'2010.
| 9
| 1,760,738,430.795361
|
https://www.phoronix.com/news/Nzk4MA
|
X Server 1.8 Release Candidate Is Here
|
Michael Larabel
|
Keith Packard has just made available the first release candidate of X Server 1.8 and confirms that its release schedule is still on track. Snapshots and the Git code for X Server 1.8 go back to last year, but with a planned release by the end of March, Keith has now started working on release candidates.
X Server 1.8 will feature a move away from HAL with udev input handling and xorg.conf.d support. This major X Server update also has new DMX2 code and DRI2 2.2 support, but it is yet another X.Org Server update that goes without XKB2.
X Server 1.8 Release Candidate 1 contains dozens of changes since the earlier snapshots and then also what is available in the X Server 1.7 series that has been around since last October. NVIDIA's newest beta already supports X Server 1.8, but it will be many months before AMD is likely to support this updated X Server in their Catalyst Linux driver (they still lack support for X Server 1.7).
Other minor work in this release includes updates to EXA, KDrive, Cygwin/X, and XQuartz. The X Server 1.8 Release Candidate 1 announcement can be read on the X.Org mailing list. It's great to see that the new X.Org release process may actually be working for putting out timed, consistent releases.
| 17
| 1,760,738,431.482383
|
https://www.phoronix.com/news/Nzk3MQ
|
X@FOSDEM 2010 Video Status Update
|
Michael Larabel
|
For those waiting on our X@FOSDEM 2010 videos that we recorded, they still need to be uploaded. Each of the talks, which were less than an hour in length, are about 3GB in size with the original HD files. The Internet connection at FOSDEM in Brussels was better than last year, but still was slow and too unreliable for uploading these large video files. T-Mobile's WLAN connections in northern Germany seem to be not nearly as fast and reliable as they have been in the past in Bavaria, so this too is taking longer than anticipated and may mean that it's a few days until the uploads are in place and the videos embedded at Phoronix.
Until then, there is the following written coverage: How To Reverse Engineer A Motherboard BIOS, Cleaning Up The Linux Graphics Driver Stack, Notes From X@FOSDEM 2010: GLSL, X, Etc, and Jerome's Radeon KMS Short-Term TODO List.
There are also slides and presenter notes from some of the talks, including Eric Anholt's Cairo-GL talk, Jerome Glisse's KMS and Radeon work, and Nicolai Hähnle's R300 GLSL slides.
Besides the X@FOSDEM coverage, we also have videos from the CoreBoot talks the previous day. Those too still need to be uploaded.
| 9
| 1,760,738,431.984964
|
https://www.phoronix.com/news/Nzk2Nw
|
Notes From X@FOSDEM 2010: GLSL, X, Etc
|
Michael Larabel
|
X@FOSDEM is taking place at FOSDEM in Brussels, Belgium. Nicolai Hahnle and Daniel Stone provided talks on R300 GLSL compilation and X11 and its problems, respectively. Intel's Eric Anholt also ended up giving a very brief talk on the Cairo-GL project. Luc Verhaegen is now starting to talk on cleaning/integrating the Linux graphics stack.
Videos are forthcoming, but Nicolai's technical talk did produce a TODO list for the R300 GLSL compiler. While most GLSL shaders should work today, as proclaimed Nicolai, there still is missing flow control in vertex programs, loop support in fragment and vertex programs, support for additional instructions, and there is an immense amount of optimization work that can be done to Mesa and the OpenGL drivers.
Daniel's talk was on how users expect "every frame must be perfect" and some of the current problems include issues with RandR reconfiguration, video display programs, server implementations are awful, video tearing is common, and window reconfiguration is brutal. Daniel also briefly commented on the Wayland Display Server, but as he said, "X is the best since everything else doesn't work or doesn't exist." Daniel thinks someday Wayland might function according to him.
| 6
| 1,760,738,432.010618
|
https://www.phoronix.com/news/Nzk2NQ
|
X.Org Server 1.7.5 Is Just About Done
|
Michael Larabel
|
Peter Hutterer has put out a new release candidate for X Server 1.7.5, which also marks this point release as being just about complete. There are still two weeks left before the 1.7.5 release is expected to be made and then after that we still may see X.Org Server 1.7.6. In the just-made xorg-server 1.7.4.902 release are fixes for potential segmentation faults and other small changes.
The release announcement for the new X.Org Server 1.7.5 release candidate can be read on xorg-announce.
| 0
| 1,760,738,432.516318
|
https://www.phoronix.com/news/Nzk2Mg
|
XDS 2010 Is In Toulouse, France
|
Michael Larabel
|
While X@FOSDEM has turned into a mess due to a lack of participation and interest among the X.Org development community, plans are underway for the 2010 X Developers' Summit. This year's XDS is back in Europe and is taking place in Toulouse, France. The talks for the three day event have not yet been determined, but will come about leading up to September.
Details regarding XDS 2010 in Toulouse can be found on the X.Org Wiki. There should be coverage on Phoronix from this European event.
| 2
| 1,760,738,432.531264
|
https://www.phoronix.com/news/Nzk0MQ
|
X@FOSDEM 2010 Is Now A Half-Day Event
|
Michael Larabel
|
Well, it was bad enough when X@FOSDEM became a one day event (where for the past several years it has been a highly-populated two-day conference) at the upcoming Free Open Source Developers' European Meeting (FOSDEM) taking place in Brussels next weekend, but now it's not even a one day event. X@FOSDEM has just been sliced down to now just be a half-day conference... Well, five hours.
We already reported last week that X@FOSDEM 2010 will very likely be its last after FOSDEM organizers are not impressed at all over missing talks from the schedule, but now at the last minute they have stripped the X.Org project of its development room in the morning and handed it over to OpenMoko for some morning meetings.
X@FOSDEM 2010 will now start at noon on Sunday and then there are five, one-hour planned talks. After that, FOSDEM is over. That means there is no more room for any last minute talks, including no talk by Intel's Keith Packard, which has been somewhat of an annual tradition at FOSDEM. There's also no annual Nouveau status update nor any Gallium3D update from Tungsten Graphics / VMware. The speakers for this final X@FOSDEM include Nicolai Hähnle, Daniel Stone, Luc Verhaegen, Jerome Glisse, and Mikhail Gusarov. The topics covered include getting GLSL support in the ATI R300 Gallium3D driver, polishing X.Org, Luc's new, simple-to-maintain-and-update graphics driver stack, Radeon kernel mode-setting, and lastly is a talk about X on e-Paper. Details can be found on the X.Org Wiki.
This is rather unfortunate that X@FOSDEM is now just going to occupy a quarter of the time that it did in year's past at FOSDEM... Especially if you are like I and going from the US to Europe for providing this X.Org coverage, but Phoronix will have videos and live coverage for what is left of this event. There is also a rumor about the X Developers' Conference (XDC) making a return sometime this year in France.
| 7
| 1,760,738,433.034127
|
https://www.phoronix.com/news/NzkzMg
|
XGI Works On Their Linux Driver... From Windows
|
Michael Larabel
|
Two days ago we reported on XGI submitting open-source driver patches after they've basically been written off as dead for years and their Linux driver has been unmaintained. These patches actually were notable in that they provided EXA acceleration support, improved EDID, support for custom display modes, and compatibility with ARM-based systems. However, they didn't apply cleanly.
It turns out XGI actually works on their Linux driver code from Windows systems, as was said by Jong Lin. This actually begs the question whether XGI Technology actually tests out the xf86-video-xgi Linux driver or to what extent. However, as sad as it may be, soon enough they could actually have more notable code contributions than what VIA is doing with their open-source work and outpace their "efforts" from years past.
For what it's worth, XGI Technology primarily advertises their open-source graphics hardware support now and not their binary Linux drivers. In fact, it takes some navigating to find their Linux blobs, but the Volari Linux driver was last updated in January of 2006... This was their update that added Linux 2.6 kernel compatibility and 3D support. Don't even think about running this binary driver with a recent kernel or X Server. Though you can find their open-source drivers for the V3XE, XP10, Z7, and Z9 on this page. No open-source 3D stack has yet to be published. XGI Technology has also abandoned their NDA-ladened developer area that we had access to and was more interesting.
| 6
| 1,760,738,433.050527
|
https://www.phoronix.com/news/NzkxMw
|
X Server 1.7.5 Release Candidate 1
|
Michael Larabel
|
Peter Hutterer has announced the first release candidate for X Server 1.7.5. This new test build is coming less than a month after the release of X Server 1.7.4 and its change-log isn't particularly exciting, especially considering the fact that most of the X.Org developers are currently down in New Zealand for Linux.Conf.Au. X Server 1.7.5 RC1 (a.k.a. xorg-server 1.7.4.901) contains a couple XQuartz fixes and a few other bits, but Peter promises X Server 1.7.5 RC2 will be more lively.
The X Server 1.7.5 RC1 release announcement can be found on the X.Org mailing list. Traditionally there haven't been so many X Server point releases for a given branch, but X Server 1.8 isn't due out until the end of March and as a result a number of distributions shipping major updates this quarter or next are opting to use the stable X Server 1.7 series.
| 0
| 1,760,738,433.529252
|
https://www.phoronix.com/news/NzkwNA
|
Farewell To X@FOSDEM. 2010 Is The Last.
|
Michael Larabel
|
X@FOSDEM 2011? Forget about it. It is to much disappointment that we have to report this evening that the X@FOSDEM DevRoom will be ending this year, after the X.Org project has consistently held a development room at Europe's largest free software event for years. Two days ago we reported on the sad state of this year's X@FOSDEM schedule. Only half the schedule is filled and there is just two weeks left until the Free Open-Source Developers' European Meeting.
This half-filled schedule is even taking into account that the X.Org development room is now just a one-day event compared to two days in past years. X.Org developers are just not stepping up to speak. As of right now there are just four talks planned. Last year's X@FOSDEM schedule wasn't completely booked either, but at least there were enough sessions to cover most of the weekend.
Hopefully three more X.Org sessions end up coming about so at least this final X@FOSDEM will go out with a full day worth of content. FOSDEM organizers do not intend to grant X.Org another development room in forthcoming years due to this lackluster showing and there being many other free software projects that would like to host their own development rooms, but end up getting declined for the lack of space. This was confirmed in a blog post by Luc Verhaegen who has long organized X@FOSDEM. This is especially sad considering that there is no longer X Developers' Summit (XDS) and X Developers' Conference (XDC) each year but only one of the events, which now will be the sole X.Org event each year to discuss various X.Org projects, road-maps, state of various components, and a general gathering amongst most of the core X.Org development community.
Phoronix will be providing live notes from X@FOSDEM 2010 and also will be producing audio/video recordings from the talks as well. If the one-day schedule does not end up getting filled, I have proposed to Luc a possible Q&A of a few X.Org developers on stage and then any end-users in the audience wishing to ask general (non-tech support) questions about popular X.Org/graphics topics, their thoughts on different matters, the future of different drivers, etc. If this goes through, questions could possibly come in remotely with the session being made live via IRC or on Phoronix. If a Q&A like this would be of interest to you, please comment on this via clicking the "Comments & Discussion" link below.
| 16
| 1,760,738,433.538115
|
https://www.phoronix.com/news/Nzg5Ng
|
What Do You Want From Linux GPU Drivers In 2010?
|
Michael Larabel
|
Where there are the 2009 Linux Graphics Survey results from a few months back, we are interested in hearing what you would like from Linux graphics drivers in 2010. Whether it be specific features, overall improvements, a change in policies, or anything else that influences your Linux graphics experience.
In terms of 3D, on the open-source side there is Gallium3D coming and other core Mesa improvements that we have been talking about as of late, but what about with ATI/AMD and NVIDIA with their binary drivers? Both companies will be introducing new features in their binary Linux drivers in 2010, but what do you hope they provide?
Let us know.
| 92
| 1,760,738,434.049674
|
https://www.phoronix.com/news/Nzg5Mg
|
X.Org Talks For FOSDEM 2010 Still Missing
|
Michael Larabel
|
X@FOSDEM, the X.Org development room that's held at FOSDEM each year, has traditionally been a two day event that runs the duration of this major European free software meeting. This year though it turned into a one day affair, even though there is only one other X.Org conference held each year (which used to be two -- XDS and XDC -- until that fell apart in 2008). Even with this trimmed down schedule and not many opportunities to talk about X.Org each year at such conferences, the speaking schedule remains only half filled.
There's just a little over two weeks left until FOSDEM 2010 in Brussels, Belgium yet only four of the seven sessions are scheduled. Only one more talk has been added since earlier this month and that will be a session by Luc Verhaegen. Luc's description of his planned talk is below.
4 years after the modular X tree was released, we can clearly see that we did not fully satisfy all expectations and that we are really holding the free software desktop back. In this talk, the current situation gets analysed, and the next step, providing more integrated graphics driver stacks, a change that will make life easier for all involved, is introduced and demonstrated. Luc will be demonstrating a new "project" he has been working on since being unemployed and that is a more integrated Linux graphics driver stack. Wouldn't it be wonderful to be able to update your complete graphics driver stack in one swoop rather than having to build numerous separate packages, etc...? Luc thinks so and he'll provide his answer, but you'll need to be at this talk (or view our recordings and news postings afterwards) when all of the details are shared.
The entire morning of X@FOSDEM on the 7th of February remains unfilled. Intel's Keith Packard is expected to be at FOSDEM once again, so he will likely be giving a talk on X.Org in general or the Intel driver stack as he has been doing this for some years now, but nothing has yet been announced. Also missing from this schedule is any talk on the status of the Nouveau driver, which has also been a staple of X@FOSDEM for the past few years. This year's Nouveau status update would be particularly interesting with their DRM code going mainline in the Linux 2.6.33 kernel and likely a DDX driver release in the not too distant future.
Other sessions in the past at X@FOSDEM have included using LLVM for improved GPU performance, rebuilding the X.Org desktop, RandR 1.3, improving GPU power management, Gallium3D status updates, audio support for the X Server, and plenty of other interesting talks.
The current X@FOSDEM 2010 plans can be found on the X.Org Wiki.
| 11
| 1,760,738,434.743179
|
https://www.phoronix.com/news/Nzg3NA
|
VIA Releases New Documentation, But It's Old
|
Michael Larabel
|
A few weeks back we shared VIA's Linux TODO list, which was very disappointing to say the least. VIA Technologies has gone through several phases of trying to be "open-source friendly" and they certainly try to brand themselves as such, but it's still going to be a year (or longer) before they have a viable open-source driver stack and by that time the VIA hardware that's supported will be dated. Read the above-linked article for all of the details on the matter. This morning, however, VIA Technologies has let their puppets with the OpenChrome driver project release some new documentation -- but this documentation is not actually new.
OpenChrome's Xavier Bachleot wrote a mailing list message announcing the release of two new VIA documents. These documents cover the graphics core / 2D engine and 3D / video for VIA's Chrome 9 IGP, specifically the VIA VX855 and VX875 ASICs.
The VIA VX855 media system processor has been on the market now for almost a year, but its largely derived from the VX800 that has been around for two years. It hasn't taken VIA Technologies all this time to assemble the documentation and then release it, but it's clearly stated on both documents "Preliminary Revision 1.0; July 29, 2009." In other words, these documents have been around for at least six months before finally being uploaded for the public. It should have not taken a half-year either for these documents to clear VIA's legal review process. The OpenChrome developers have likely been sitting on these documents as well through their VIA NDAs for some amount of time too.
Both of these VIA documents are licensed under the Creative Commons and the core/2D documentation amounts of 110 pages while the 3D/video portion is 172 pages. These are the first public documents emerging from VIA Technologies since November of 2008 when they did their last documentation drop.
VIA's graphics processor documentation can be found on the X.Org server.
| 24
| 1,760,738,435.281212
|
https://www.phoronix.com/news/Nzg3Nw
|
Pulled: DRI 2.2 Protocol Requests, Swap Events
|
Michael Larabel
|
While X Server 1.8 left its primary development cycle and entered a period of bug-fixing at the end of last month, per its release schedule with planned readiness by the end of March, some late work does continue to get pulled into this next major X.Org Server release. On Monday, Intel's Jesse Barnes put in a pull request for one of his trees that adds in support for the DRI 2.2 protocol requests and new extensions for the X Server. This work, which bumps the DRI2 version to v2.2, has now been pulled and will be part of X Server 1.8.
Back in November we covered Clutter/Mutter work leading to a new GLX extension and this work delivers on this support. GLX_INTEL_swap_event is the new extension and it notifies the client when a buffer swap has been completed. The benefit of this extension is that glXSwapBuffers call can be made a-synchronous so that the client can continue on doing other work, but then will be notified when the swap has taken place so that it can return to preparing the next buffer.
The other part of this work is the new protocol requests for DRI2 2.2 and they include DRI2SwapBuffers, DRI2GetMSC, DRI2WaitMSC, DRI2WaitSBC and DRI2SwapInterval. These requests are used for supporting the SGI_video_sync, SGI_swap_interval, and OML_sync_control GLX extensions. It was back in October that we originally talked about these DRI2 sync and swap extensions.
This work requires new versions of dri2proto and glproto when building the X Server. The X.Org hardware drivers also need some changes to take advantage of these new X Server capabilities.
These new features can be found in X Server 1.8, which will be found in Fedora 13 and other distributions that aggressively update their X.Org packages, but Ubuntu users will need to wait for Ubuntu 10.10 before they see X Server 1.8/1.9. Other X Server 1.8 features include udev input handling (for Linux) to support moving away from HAL, xorg.conf.d configuration, and other improvements.
| 0
| 1,760,738,435.290181
|
https://www.phoronix.com/news/Nzg2NQ
|
X Server 1.7.4 Released
|
Michael Larabel
|
Version 1.7.4 of the X.Org Server has been released this morning. This point release continues to bring new bug-fixes to the X Server 1.7 series branch since its release last year. All major development work continues to be focused on X Server 1.8, which is expected for release in March.
X Server 1.7.4 just brings a collection of fixes, but none of the changes are too very exciting for Linux desktop end-users. Planned for release in five weeks is already X Server 1.7.5, which will continue bringing forth bug-fixes, some of which are back-ported from master / X Server 1.8. With some distributions like Ubuntu 10.04 LTS opting to stick with the X Server 1.7 series, these bug-fix releases do carry a bit more weight.
The release announcement for xorg-server 1.7.4 can be read on xorg-announce.
| 0
| 1,760,738,435.792389
|
https://www.phoronix.com/news/Nzg2OA
|
X@FOSDEM 2010 Talks Planned So Far
|
Michael Larabel
|
There's just one month left until the Free and Open-source Software Developers' European Meeting (FOSDEM) takes place once again in Brussels, Belgium. Like in past years, there will be an X.Org development room where various talks about X will be held, but this year it has turned into a one-day affair. Even with having half the time as past years to talk about X, the schedule is not even full at this point.
As is shown on the X.Org FOSDEM 2010 Wiki, as of this afternoon there are only three talks -- by Daniel Stone, Jerome Glisse, and Mikhail Gusarov. Daniel will be talking about polishing X.Org with areas like video APIs, RandR, Composite, and Composite and how basically X11 can not suck so badly, thereby eliminating the need for an X12 or Wayland.
Jerome will once again give another technical X.Org talk, this time on the status of Radeon kernel mode-setting and the GPU kernel interfaces. The last talk planned at this time is on "X on e-Paper", in regards to the requirements for getting the X.Org Server to work well with ePaper technology and its unique needs.
There still is room for an additional four talks, which at this point remains unfilled. Luc Verhaegen has shared that Keith Packard and Eric Anholt (Intel) will indeed be coming once again, so that is likely another talk about Intel and X.Org, but details have yet to be published.
I will be at X@FOSDEM again and will be providing live coverage along with audio/video feeds via Phoronix. Please subscribe to Phoronix Premium, use our shopping links, or even make a donation to help support this work. You can also follow this coverage on Facebook, Twitter, and Identi.ca.
| 1
| 1,760,738,435.80694
|
https://www.phoronix.com/news/Nzg1Ng
|
XKB2 Gets Cut Again, Maybe For X Server 1.9?
|
Michael Larabel
|
The end of 2009 marked the closure of the merge window for X Server 1.8, which means that HAL removal work got in, including xorg.conf.d and udev input handling. What didn't make it in time for X Server 1.8 though was XKB2. XKB2 has been talked about for quite a while now and going back a number of X Server releases, but now we will not see this revised version of the X Keyboard Extension until at least X Server 1.9. This though is not entirely surprising since XKB2 hasn't been discussed a lot as of late.
For those interested in learning more about the HAL input changes for X Server 1.8, Peter Hutterer just happened to publish a new blog post in which he goes over some of the changes at length. The final release of X Server 1.8 is expected in March.
| 7
| 1,760,738,436.330896
|
https://www.phoronix.com/news/Nzg2NA
|
DMX2 Bits For X.Org Server To Move Forward
|
Michael Larabel
|
Back in 2008, Novell's David Reveman published his own branch of the Distributed Multi-head X (DMX) server which he called dmx-2 as it was close to a complete rewrite of the original DMX implementation. David's DMX-2 branch was less complex but provided a greater set of features, including X-Video, RandR 1.2, and Composite support in a DMX environment, D-Bus configuration, and many other changes. This branch was never merged to master, but now Red Hat's Adam Jackson is looking at merging some of the DMX-2 to work into the mainline X Server.
The major items that Adam is looking at from DMX-2 include support for Composite when using Xinerama, exposing animated cursors to the DDX for acceleration, integrating RandR and Xinerama at the protocol layer, and various performance optimizations for for redirected windows and MIT-SHM.
Adam has re-based all of the old DMX-2 DIX (Device Independent X) code against the latest master code, which can currently be found in this branch. Nothing has yet been merged to the X.Org Server master branch, but there is currently an ongoing discussion.
X Server 1.8 just recently entered its bug-fixing state where no new features are supposed to be committed, but there's always room for exceptions. Otherwise this work will likely end up in X Server 1.9 to be released in late 2010.
| 2
| 1,760,738,436.338844
|
https://www.phoronix.com/news/Nzg0Mw
|
X Server 1.7.4 Is Just About Here
|
Michael Larabel
|
Work on the X Server 1.7.4 point release is finishing up and as a result Peter Hutterer has tagged the second release candidate, which may very well be the last test release before it goes gold. X Server 1.7.4 RC2 back-ports a number of fixes from X Server 1.8, switches the default keyboard layout to pc105 (from pc104), man-page fixes/updates, and miscellaneous other work.
The X Server 1.7.4 Release Candidate 2 announcement can be found on the X.Org mailing list. X Server 1.7.4 should be out within the next week or two. Meanwhile, the next major release, X Server 1.8, should be out by the end of March.
| 1
| 1,760,738,436.882565
|
https://www.phoronix.com/news/Nzg1NQ
|
TWM Updated For 2010: X Render Support, Etc
|
Michael Larabel
|
We are just a few days into 2010, but the standard (basic) window manager for X.Org, TWM, has finally received some updates. TWM has been around for 13 years and under the hands of many developers, but is now receiving some new development love from Eeri Kask. Eeri has been hacking away at TWM for a few years now and has made many improvements already, but this is the first time since September of 2008 that he is announcing some of his new work.
Eeri's latest TWM work includes making support for new options configurable through the autoconf build script (for those that prefer a traditional TWM), a number of new/updated keywords, X Render support has been added (including transparency support). new arguments added as a prerequisite to xcompmgr based composited transparency, and improvements to frame geometry / focus management.
More information on these new TWM window manager improvements can be found on the X.Org mailing list.
| 5
| 1,760,738,436.890626
|
https://www.phoronix.com/news/Nzg0NQ
|
Multi-Touch Gesture Recognition For Clutter
|
Michael Larabel
|
Multi-Pointer X has been fully supported by X.Org since the X Server 1.7 release earlier this year. The MPX support allows multiple input devices to work independently within a single running X Server. GTK has been hacked-up to allow for multi-touch events to some extent and Qt 4.6 was released with multi-touch support. The Intel developers working on Moblin and Clutter have also been working on multi-touch support for the Clutter tool-kit. In addition, they have also developed a gestures recognition framework for multi-touch (and for some gestures, single-touch) environments.
Clutter Gesture attempts to identify gestures generated from input events and then where applicable to send off these recognized gestures to the application that has focus. This framework is flexible to allow all sorts of gesture algorithms to be built-in and it can also be extended from the application side as well, but the currently supported set of input gestures include slide up, slide down, slide left, slide right, and touch&hold. Supported from the multi-touch side is the pinch/rotate gestures.
Clutter Gesture is not currently found in a released version of Clutter or Moblin, but over at Moblin.org are more details on this project along with how to experiment with this gestures framework using Moblin 2.1 and some extra RPM packages in a few simple steps.
| 0
| 1,760,738,437.38645
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.