Core Rendering with Glamor
I ve hacked up the intel driver to bypass all of the UXA paths
when Glamor is enabled so I m currently running an X server that uses
only Glamor for all rendering. There are still too many fall backs, and
performance for some operations is not what I d like, but it s
entirely usable. It supports DRI3, so I even have GL applications
running.
Core Rendering Status
I ve continued to focus on getting the core X protocol rendering
operations complete and correct; those remain a critical part of many
X applications and are a poor match for GL. At this point, I ve got
accelerated versions of the basic spans functions, filled rectangles,
text and copies.
GL and Scrolling
OpenGL has been on a many-year vendetta against one of the most common
2D accelerated operations copying data within the same object, even
when that operation overlaps itself. This used to be the most
performance-critical operation in X; it was used for scrolling your
terminal windows and when moving windows around on the screen.
Reviewing the OpenGL 3.x spec, Eric and I both read the glCopyPixels
specification as clearly requiring correct semantics for overlapping
copy operations it says that the operation must be equivalent to
reading the pixels and then writing the pixels. My CopyArea
acceleration thus uses this path for the self-copy case. However, the
ARB decided that having a well defined blt operation was too nice to
the users, so the current 4.4 specification adds explicit language to
assert that this is not well defined anymore (even in the face of the
existing language which is pretty darn unambiguous).
I suspect we ll end up creating an extension that offers what we need
here; applications are unlikely to stop scrolling stuff around, and
GPUs (at least Intel) will continue to do what we want. This is the
kind of thing that makes GL maddening for 2D graphics the GPU does
what we want, and the GL doesn t let us get at it.
For implementations not capable of supporting the required semantic,
someone will presumably need to write code that creates a temporary
copy of the data.
PBOs for fall backs
For operations which Glamor can t manage, we need to fall back to
using a software solution. Direct-to-hardware acceleration
architectures do this by simply mapping the underlying GPU object to
the CPU. GL doesn t provide this access, and it s probably a good
thing as such access needs to be carefully synchronized with GPU
access, and attempting to access tiled GPU objects with the CPU
require either piles of CPU code to de-tile accesses (ala wfb), or
special hardware detilers (like the Intel GTT).
However, GL does provide a fairly nice abstraction called pixel buffer
objects (PBOs) which work to speed access to GPU data from the CPU.
The fallback code allocates a PBO for each relevant X drawable, asks
GL to copy pixels in, and then calls fb, with the drawable now
referencing the temporary buffer. On the way back out, any potentially
modified pixels are copied back through GL and the PBOs are freed.
This turns out to be dramatically faster than malloc ing temporary
buffers as it allows the GL to allocate memory that it likes, and for
it to manage the data upload and buffer destruction asynchronously.
Because X pixmaps can contain many X windows (the root pixmap being
the most obvious example), they are often significantly larger than
the actual rendering target area. As an optimization, the code only
copies data from the relevant area of the pixmap, saving considerable
time as a result. There s even an interface which further restricts
that to a subset of the target drawable which the Composite
function uses.
Using Scissoring for Clipping
The GL scissor operation provides a single clipping rectangle. X
provides a list of rectangles to clip to. There are two obvious ways
to perform clipping here either perform all clipping in software,
or hand each X clipping rectangle in turn to GL and re-execute the
entire rendering operation for each rectangle.
You d think that the former plan would be the obvious choice; clearly
re-executing the entire rendering operation potentially many times is
going to take a lot of time in the GPU.
However, the reality is that most X drawing occurs under a single
clipping rectangle. Accelerating this common case by using the
hardware clipper provides enough benefit that we definitely want to
use it when it works. We could duplicate all of the rendering paths
and perform CPU-based clipping when the number of rectangles was above
some threshold, but the additional code complexity isn t obviously
worth the effort, given how infrequently it will be used. So I haven t
bothered. Most operations look like this:
Allocate VBO space for data
Fill VBO with X primitives
loop over clip rects
glScissor()
glDrawArrays()
This obviously out-sources as much of the problem as possible to the
GL library, reducing the CPU time spent in glamor to a minimum.
A Peek at Some Code
With all of these changes in place, drawing something like a list of
rectangles becomes a fairly simple piece of code:
First, make sure the program we want to use is available and can be
used with our GC configuration:
prog = glamor_use_program_fill(pixmap, gc,
&glamor_priv->poly_fill_rect_program,
&glamor_facet_polyfillrect);
if (!prog)
goto bail_ctx;
Next, allocate the VBO space and copy all of the X data into it. Note
that the data transfer is simply memcpy here that s because we
break the X objects apart in the vertex shader using instancing,
avoiding the CPU overhead of computing four corner coordinates.
/* Set up the vertex buffers for the points */
v = glamor_get_vbo_space(drawable->pScreen, nrect * (4 * sizeof (GLshort)), &vbo_offset);
glEnableVertexAttribArray(GLAMOR_VERTEX_POS);
glVertexAttribDivisor(GLAMOR_VERTEX_POS, 1);
glVertexAttribPointer(GLAMOR_VERTEX_POS, 4, GL_SHORT, GL_FALSE,
4 * sizeof (GLshort), vbo_offset);
memcpy(v, prect, nrect * sizeof (xRectangle));
glamor_put_vbo_space(screen);
Finally, loop over the pixmap tile fragments, and then over the clip
list, selecting the drawing target and painting the rectangles:
glEnable(GL_SCISSOR_TEST);
glamor_pixmap_loop(pixmap_priv, box_x, box_y)
int nbox = RegionNumRects(gc->pCompositeClip);
BoxPtr box = RegionRects(gc->pCompositeClip);
glamor_set_destination_drawable(drawable, box_x, box_y, TRUE, FALSE, prog->matrix_uniform, &off_x, &off_y);
while (nbox--)
glScissor(box->x1 + off_x,
box->y1 + off_y,
box->x2 - box->x1,
box->y2 - box->y1);
box++;
glDrawArraysInstanced(GL_TRIANGLE_STRIP, 0, 4, nrect);
GL texture size limits
X pixmaps use 16 bit dimensions for width and height, allowing them to
be up to 65536 x 65536 pixels. Because the X coordinate space is
signed, only a quarter of this space is actually useful, which makes
the useful size of X pixmaps only 32767 x 32767. This is still larger
than most GL implementations offer as a maximum texture size though,
and while it would be nice to just say we don t allow pixmaps larger
than GL textures , the reality is that many applications expect to be
able to allocate such pixmaps today, largely to hold the ever
increasing size of digital photographs.
Glamor has always supported large X pixmaps; it does this by splitting
them up into tiles, each of which is no larger than the largest
texture supported by the driver. What I ve added to Glamor is some
simple macros that walk over the array of tiles, making it easy for
the rendering code to support large pixmaps without needing any
special case code.
Glamor also had some simple testing support you can compile the
code to ignore the system-provided maximum texture size and supply
your own value. This code had gone stale, and couldn t work as there
were parts of the code for which tiling support just doesn t make
sense, like the glyph cache, or the X scanout buffer. I fixed things
so that you could leave those special cases as solitary large tiles
while breaking up all other pixmaps into tiles no larger than 32
pixels square.
I hope to remove the single-tile case and leave the code supporting
only the multiple-tile case; we have to have the latter code, and so
having the single-tile code around simply increases our code size for
not obvious benefit.
Getting accelerated copies between tiled pixmaps added a new
coordinate system to the mix and took a few hours of fussing until it
was working.
Rebasing Many (many) Times
I m sure most of us remember the days before git; changes were often
monolithic, and the notion of changing
how the changes were made for
the sake of clarity never occurred to anyone. It used to be that the
final code was the only interesting artifact; how you got there didn t
matter to anyone. Things are different today; I probably spend a third
of my development time communicating
how the code should change with
other developers by changing the sequence of patches that are to be
applied.
In the case of Glamor, I ve now got a set of 28 patches. The first few
are fixes outside of the glamor tree that make the rest of the server
work better. Then there are a few general glamor infrastructure
additions. After that, each core operation is replaced, one a at a
time. Finally, a bit of stale code is removed. By sequencing things in
a logical fashion, I hope to make review of the code easier, which
should mean that people will spend less time trying to figure out what
I did and be able to spend more time figuring out if what I did is
reasonable and correct.
Supporting Older Versions of GL
All of the new code uses vertex instancing to move coordinate
computation from the CPU to the GPU. I m also pulling textures apart
using integer operations. Right now, we should correctly fall back to
software for older hardware, but it would probably be nicer to just
fall back to simpler GL instead. Unless everyone decides to go buy
hardware with new enough GL driver support, someone is going to need
to write simplified code paths for glamor.
If you ve got such hardware, and are interested in making it work
well, please take this as an opportunity to help yourself and others.
Near-term Glamor Goals
I m pretty sure we ll have the code in good enough shape to merge
before the window closes for X server 1.16. Eric is in charge of the
glamor tree, so it s up to him when stuff is pulled in. He and Markus
Wick have also been generating code and reviewing stuff, but we could
always use additional testing and review to make the code as good as
possible before the merge window closes.
Markus has expressed an interest in working on Glamor as a part of the
X.org summer of code this year; there s clearly
plenty of work to do
here, Eric and I haven t touched the render acceleration stuff at all,
and that code could definitely use some updating to use more modern GL
features.
If that works as well as the core rendering code changes, then we can
look forward to a Glamor which offers GPU-limited performance for
classic X applications, without requiring GPU-specific drivers for
every generation of every chip.
In a project I recently worked on we wanted to provide a jenkins-debian-glue based setup on Amazon s EC2 for building Debian and Ubuntu packages. The idea is to keep a not-so-strong powered Jenkins master up and running 24 7, while stronger machines serving as Jenkins slaves should be launched only as needed. The project setup in question is fully open sourced (more on that in a separate blog post), hereby I am documenting the EC2 setup in usage.
Jenkins master vs. slave
Debian source packages are generated on Jenkins master where a checkout of the Git repository resides. The Jenkins slaves do the actual workload by building the binary packages and executing piuparts (.deb package installation, upgrading, and removal testing tool) on the resulting binary packages. The Debian packages (source+binaries) are then provided back to Jenkins master and put into a reprepro powered Debian repository for public usage.
Preparation
The starting point was one of the official Debian AMIs (x86_64, paravirtual on EBS). We automatically deployed jenkins-debian-glue on the system which is used as Jenkins master (we chose a m1.small instance for our needs).
We started another instance, slightly adjusted it to already include jenkins-debian-glue related stuff out-of-the-box (more details in section Reduce build time below) and created an AMI out of it. This new AMI ID can be configured for usage inside Jenkins by using the Amazon EC2 Plugin (see screenshot below).
IAM policy
Before configuring EC2 in Jenkins though start by adding a new user (or group) in AWS s IAM (Identity and Access Management) with a custom policy. This ensures that your EC2 user in Jenkins doesn t have more permissions than really needed. The following policy should give you a starting point (we restrict the account to allow actions only in the EC2 region eu-west-1, YMMV):
As you can see the configuration also includes a launch script. This script ensures that slaves are set up as needed (provide all the packages and scripts that are required for building) and always get the latest configuration and scripts before starting to serve as Jenkins slave.
Now your setup should be ready for launching Jenkins slaves as needed:
NOTE: you can use the Instance Cap configuration inside the advanced Amazon EC2 Jenkins configuration section to place an upward limit to the number of EC2 instances that Jenkins may launch. This can be useful for avoiding surprises in your AWS invoices. 
Notice though that the cap numbers are calculated for all your running EC2 instances, so be aware if you have further machines running under your account, you might want to e.g. further restrict your IAM policy then.
Reduce build time
Using a plain Debian AMI and automatically installing jenkins-debian-glue and further jenkins-debian-glue-buildenv* packages on each slave startup would work but it takes time. That s why we created our own AMI which is nothing else than an official Debian AMI with the bootstrap.sh script (which is referred to in the screenshot above) already executed. All the necessary packages are pre-installed and also all the cowbuilder environments are already present then. From time to time we start the instance again to apply (security) updates and execute the bootstrap script with its &dash&dashupdate option to also have all the cowbuilder systems up2date. Creating a new AMI is a no-brainer and we can then use the up2date system for our Jenkins slaves, if something should break for whatever reason we can still fall back to an older known-to-be-good AMI.
Final words
How to set up your Jenkins jobs for optimal master/slave usage, multi-distribution support (Debian/Ubuntu) and further details about this setup are part of another blog post.
Thanks to Andreas Granig, Victor Seva and Bernhard Miklautz for reading drafts of this.
Until this time next month I'll be posting code-based discussions only.
Recently I've been wanting to explore creating clustered services, because clusters are definitely things I use professionally.
My initial attempt was to write an auto-clustering version of memcached, because that's a useful tool. Writing the core of the service took an hour or so:
since people keep talking to me about my RC bug fixing activities, I thought
it might be time again for a short report. to be honest, I mostly stopped my
almost daily work at some point in december, partly because the overall
number of RC bugs affecting both 
I woke up at 2am, and then spent over an hour trying to get back to sleep. That
was phenomenally annoying. Then at 5am, Zoe's white noise app on her Nexus 7
decided to go a bit wonky (as it does from time to time), which caused her to
wake up. I couldn't convince her tablet to behave, so I just let her get in bed
with me, but neither of us went back to sleep. I think at about 6am I told her
she could turn on the TV so I could try and squeeze in a bit more rest.
I managed to forget to put the oats in the Thermomix when I made porridge this
morning. It took me a while to figure out why the consistency and colour didn't
look quite right when I dished it up. Yep, it was one of those mornings.
Zoe had her trial ballet class this morning, at 9:45am in Coorparoo. I managed
to get Zoe organised a bit early, so we were the first ones there. It's in a
church hall, and has a nice floor, with a mirrored wall and barres on each
long wall.
Eventually the other girls arrived, there were 5 regulars in total. A good
sized class.
I had mixed feelings when I discovered that parents weren't normally able to be
present in the dance studio to watch the classes. I was today, because it was a
trial, but once Zoe starts regular classes, I'll have to wait outside. I'm
disappointed I won't be able to watch, but it'll be nice to have 45 minutes to
myself.
Zoe did really well in the class today. She picked everything up very quickly,
and I don't think she'll be negatively impacted by having entered in Term 2.
She seemed very comfortable and looked like she enjoyed herself.
Not long after we arrived, the heavens opened in a massive downpour, which
continued when we had to leave. It was a fair dash back to the car. I'd had
plans of a picnic lunch in New Farm Park, but scrapped them on account of the
weather.
We got home, and had a late morning tea, and then Zoe wanted to play with her
Marble Run. After that, we gave Smudge a good brushing and then Zoe wanted to
do some craft and painting.
I let Zoe set herself up, and it was fun to see what bits and pieces she chose
to dig out of the arts and craft bins. She wanted to make another crown for
Mummy's boss, but we don't have any crown making stuff at home (yet), so she
ended up starting to decorate a wand and gluing a pom pom on a shoe box, before
moving on to painting her cardboard box cubby house and rocket. She wanted to
start doing some hand painting right around lunch time, but I managed to
convince her to do that after lunch.
We had some lunch, and then she did a little bit more painting and stickering
before her nap. She passed out for her nap very quickly.
After her nap, I asked her if she wanted to paint or do science, and without
hesitation, she chose science!
I delved into the "Weather" chapter of 365 Science Experiments and
we did a couple with ice cubes and salt and then she saw one where you stick a
balloon over the neck of a glass bottle and put it in hot water and the balloon
"inflates", so we did that one as well. We also popped a balloon to simulate
thunder. It was all nice and easy, without a lot of preparation required.
After that, we did a bit of a clean up and then Zoe watched some of her DVDs
from the library while she waited for Sarah to arrive to pick her up.
It was a good day today. No meltdowns.
To provide geolocation-aware answers with 
Moreover, caching makes the problem worse.
To solve this problem, a new
Not many authoritative servers support this extension (
(ARMv7 A15 dual core)
I ve 
A new minor release 0.11.1 of
I switched to 
I wrote about