Re: [PATCH] olpc.fth - grow the root filesystem partition on boot
On Fri, Mar 16, 2012 at 11:55:54PM -0400, Paul Fox wrote: i'm a little confused. i know i reported that filesystem resize was quite fast now, but then james reported that a resize took a really long time still, and i haven't retested. if a resize takes a really long time, then i don't think we can do it in initramfs, can we? that earlier test was on external card. retested today with six units using eMMC, and the resize time for 4GB to 8GB was around 14sec, as you had reported. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
New boot animation for 12.1.0
Hi, I've just pushed everything needed for a new boot animation for 12.1.0, it will arrive in the next build. The old olpc-bootanim system was barely working in F17, mostly due to the move to systemd. As the init sequence is no longer linear (it includes a lot of parallelization) it would be hard to say how to fix olpc-bootanim - the key issue is that calculating progress through boot is now far from trivial. The new animation is based on plymouth, which is the tool used by Fedora. This means that we'll no longer have to put up with olpc-bootanim breaking each cycle, we now use what upstream uses. plymouth does have a solution to the progress problem, but it is overcomplicated. It works something like this: it takes the amount of time needed for the last few system startups, averages it, and computes progress-of-the-moment based on the timing and the events that systemd sends during boot. It also applies functions to these calculations so that the progress movement is smooth, rather than jumpy. The new theme I've (quickly) implemented doesn't bother showing progress - it just shows a flashing dot, and then three dots when the bootup process is finishing. We do have the option of adding a progress display later, but I personally prefer the simplistic approach. Plymouth's progress calculation is pretty good but is not always reliable, and it the amount of calculation and smoothing needed seems wasteful - it is important that our boot animation is lightweight so that we don't drain system resources. However, maybe my simple blinking-dot design could be prettified a little, without losing simplicity. Please give it a spin and share your opinion here. I will now focus on some other tasks but can revisit this at a later date. Here's the code: http://dev.laptop.org/git/users/dsd/plymouth-theme-olpc/ http://dev.laptop.org/git/users/dsd/plymouth-plugin-olpc/ The last time I seriously looked at plymouth for the XO was in 2009, and I don't think anyone has done so since then. In 2009 we noted that it produced a significant boot delay on XO-1, so thats why rejected it. Since then, it has improved enormously in that it doesn't redraw the whole screen on every frame change (which was dragging us down a lot), and also, our theme is considerably less complex (less screen updates) than plymouth's normal ones. I have also studied the codebase and code structure and I have a good understanding of it, enough to say that the design is OK and that it should not increase boot time unacceptably like before (or if it does, the bug is embedded in the design). We can examine the performance impact in the coming weeks. Thanks, Daniel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [PATCH] olpc.fth - grow the root filesystem partition on boot
Hi James, On Fri, Mar 16, 2012 at 5:58 PM, James Cameron qu...@laptop.org wrote: Consensus is to implement a partition resize in OpenFirmware as the last step of ''fs-update''. This doesn't preclude the initramfs doing it, and the initramfs will do the filesystem resize. In OpenFirmware, we might either: * do the resize only if the .zd file requests it, (a new ''resize:'' line after ''zblocks-end:''), or * do the resize regardless, thus affecting previous builds once a firmware update occurs. I think I see a potential problem here. The problem is that during upgrades, we run fs-update before upgrading firmware, and a firmware upgrade is necessary for this new functionality. Lets say an XO-1.5 is running 11.3.0 with a firmware that doesn't support the resizing. The XO is upgraded to 12.1.0 final via fs-update, and the 12.1.0 image requests a resize (if that is the road chosen). However, as the fs-update is done with the old firmware, the resize does not happen. The system reboots, upgrades the firmware, and then boots into a system with no free disk space. Maybe I have just made a good argument as to why this needs to be done in the initramfs: to avoid tricky upgrade considerations. Daniel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [PATCH] olpc.fth - grow the root filesystem partition on boot
On Mon, Mar 19, 2012 at 4:30 PM, Daniel Drake d...@laptop.org wrote: I think I see a potential problem here. The problem is that during upgrades, we run fs-update before upgrading firmware, and a firmware upgrade is necessary for this new functionality. (...) Maybe I have just made a good argument as to why this needs to be done in the initramfs: to avoid tricky upgrade considerations. +10. More generally, having it all in the initrd seems a good self-contained approach. Splitting something across a boundary needs a sizable reward to outweigh the coordination/integration costs. m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
mmp-camera DMA crash fix
Hi Jon, Saadia, all
in my initial testing, this seems to be working _much_ better.
- I don't get garbled audio
- I don't get crashes
- dmesg shows lots of cma dma_alloc_from_contiguous() and
vb2_common_vm_{open,close} spam
- the captured picture is always greenish, with the top band in a
different luminance, this does not happen if I just take a picture,
without involving audio
This is of course minor :-)
cheers,
martin
On Sat, Mar 17, 2012 at 10:53 AM, Martin Langhoff mar...@laptop.org wrote:
We have what looks like a fix for the Record crash! Yum update your
kernel and rsync the kernel+inird into place, and help us test :-)
m
On Sat, Mar 17, 2012 at 10:15 AM, Saadia Husain Baloch
saa...@laptop.org wrote:
Jon C.
These are definitely a breakthrough! I have not been able to crash Record
(and I'm good at managing crashes), and there are no green screens of death
either. Chris Ball has pushed them to our code tree so more people can test
them as well.
I have noticed one thing in the short testing I did this morning: the audio
is garbled/unintelligible on these recorded segments. Let me see if anyone
else is experiencing this, but I didn't have it before.
Thank you for all your help
-Saadia
On Fri, Mar 16, 2012 at 6:51 PM, Jonathan Corbet cor...@lwn.net wrote:
On Fri, 16 Mar 2012 14:53:58 -0600
Jonathan Corbet cor...@lwn.net wrote:
I want to pound on it for a bit longer, then I'll have a set of patches
to
send in. Stay tuned...
And now they are ready. There's a whole set of stuff in my repo,
including the other bug fixes I've sent around, a couple you haven't seen
yet (including one for #11644), and, at the end of the series, the
demotion of the Release printk spam to debug level.
With these patches in, I cannot make my machine crash no matter how hard I
try; I think the problems are truly stomped. My apologies for creating
them in the first place and causing you to lose a lot of time.
jon
The following changes since commit
943cd42c06440bbc139b775a73fbf55786e44087
are available in the git repository at:
git://git.lwn.net/linux-2.6.git olpc-camera-fixes
for you to fetch changes up to 3b944fd17f650c9f589924f0f26de7c36c444516:
marvell-cam: Demote the release print to debug level (2012-03-16
16:29:50 -0600)
Jonathan Corbet (7):
marvell-cam: ensure that the camera stops when requested
marvell-cam: Remove broken owner logic
marvell-cam: Increase the DMA shutdown timeout
marvell-cam: fix the green screen of death
marvell-cam: Don't signal multiple frame completions in
scatter/gather mode
mmp-camera: Don't power up the sensor on resume
marvell-cam: Demote the release print to debug level
drivers/media/video/marvell-ccic/mcam-core.c | 35
drivers/media/video/marvell-ccic/mcam-core.h | 1 -
drivers/media/video/marvell-ccic/mmp-driver.c | 13 ++---
3 files changed, 32 insertions(+), 17 deletions(-)
--
mar...@laptop.org -- Software Architect - OLPC
- ask interesting questions
- don't get distracted with shiny stuff - working code first
- http://wiki.laptop.org/go/User:Martinlanghoff
--
mar...@laptop.org -- Software Architect - OLPC
- ask interesting questions
- don't get distracted with shiny stuff - working code first
- http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel
Re: [PATCH] olpc.fth - grow the root filesystem partition on boot
On Mon, Mar 19, 2012 at 02:30:22PM -0600, Daniel Drake wrote: Maybe I have just made a good argument as to why this needs to be done in the initramfs: to avoid tricky upgrade considerations. Yes, you must do it in initramfs for that reason. However the initramfs implementation is not yet ready. I've reduced the OpenFirmware change, removing the .zd file resize: verb support, leaving just the new utility command fs-resize. There's no change to fs-update, and no need to change olpc-os-builder. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Installing gstreamer media codecs on XO-1
I have tried to install media codecs on an XO-1 running 11.3.0 so that we can play FLV video files directly (from a flashdrive, school server or network location)., in either Sugar or GNOME. (I have not been able to find a way to do that using Sugar). I had trouble getting it work using the instructions at http://wiki.laptop.org/go/GStreamer#Totem_plugin - I think it is outdated. After some research and trial and error (lots!!) I have it working, some using XO-1 may find useful The below works on an XO-1 running 11.3.0 freshly installed. It works in Gnome using Totem/Movie Player and in Sugar using Jukebox Activity. It plays mp3 audio and FLV video OK (such as the Khan Academy collection, which we have loaded on the school servers in project schools) We also add the Flash plug in for the browser and disable click to view. This allows embedded flash animations and FLV videos accessed with Browse, Youtube, etc to play with good performance. Note that Flash version 11 is much better than v10. So bringing it all together, I reproduced the above using the following. - XO-1 running 11.3.0, fresh install - In Gnome view in a terminal, as su - wireless Internet connection yum localinstall --nogpgcheck http://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-stable.noa rch.rpm http://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-stab le.noarch.rpm yum install -y gstreamer-plugins-ugly yum install -y gstreamer-ffmpeg rm /home/olpc/Activities/Browse.activity/agent-stylesheet.css rm /home/olpc/Activities/Browse.activity/clickToView.xml navigate to the flash rpm and run rpm -Uhv flash.rpm where flash.rpm is the Flash verion 11 RPM for Linux 32 bit from Adobe. Questions 1. Is the above the right way to do it, if someone with more Fedora experience than I can verify.. 2. In our narrowband countries it takes an hour and downloads a lot per laptop, this is unworkable in PNG with large numbers of XOs and where the bandwidth is so expensive and unreliable. Isn't there a way to do this offline with a download, something we can run on a flashdrive? 3. is the update step above necessary (it requires downloading 33MB) 4. In doing the above am I violating a ton of licenses? David Leeming Solomon Islands Rural Link www.rurallink.com.sb ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Installing gstreamer media codecs on XO-1
I have tried to install media codecs on an XO-1 running 11.3.0 so that we can play FLV video files directly (from a flashdrive, school server or network location)., in either Sugar or GNOME. I had trouble getting it work using the instructions at http://wiki.laptop.org/go/GStreamer#Totem_plugin - I think it is outdated. After some research and trial and error (lots!!) I have it working, some using XO-1 may find useful The below works on an XO-1 running 11.3.0 freshly installed. It works in Gnome using Totem/Movie Player and in Sugar using Jukebox Activity. It plays mp3 audio and FLV video OK (such as the Khan Academy collection, which we have loaded on the school servers in project schools) We also add the Flash plug in for the browser and disable click to view. This allows embedded flash animations and FLV videos accessed with Browse, Youtube, etc to play with good performance. Note that Flash version 11 is much better than v10. So bringing it all together, I reproduced the above using the following. - XO-1 running 11.3.0, fresh install - In Gnome view in a terminal, as su - wireless Internet connection yum localinstall --nogpgcheck http://download1.rpmfusion.org/free/fedora/rpmfusion-free-release-stable.noa rch.rpm http://download1.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-stab le.noarch.rpm yum install -y gstreamer-plugins-ugly yum install -y gstreamer-ffmpeg rm /home/olpc/Activities/Browse.activity/agent-stylesheet.css rm /home/olpc/Activities/Browse.activity/clickToView.xml navigate to the flash rpm and run rpm -Uhv flash.rpm where flash.rpm is the Flash verion 11 RPM for Linux 32 bit from Adobe. Questions 1. Is the above the right way to do it, if someone with more Fedora experience than I can verify.. 2. In our narrowband countries it takes an hour and downloads a lot per laptop, this is unworkable in PNG with large numbers of XOs and where the bandwidth is so expensive and unreliable. Isn't there a way to do this offline with a download, something we can run on a flashdrive? 3. is the update step above necessary (it requires downloading 33MB) 4. In doing the above am I violating a ton of licenses? David Leeming Solomon Islands Rural Link www.rurallink.com.sb ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: New boot animation for 12.1.0
If the indication can track fraction_completed - so much the better - but I do not expect the user to leave the room if the boot process takes longer than he expects. Seeing only 25% to go is frosting; the chocolate_cake is knowing no problems thus far. Watching the number of files in /var/run/*pid and/or similar places might give a good indication of how much progress has been made. (as compared to just we can update the display so we aren't locked up yet) I don't know how to cleanly translate that into a percent. The total might depend on what is installed and/or what extra I/O gear is plugged in this time. It might be reasonable to collect a list of normal files for this distribution (say in a file) and only count them. That would let you convert to %, but initialization might finish early if a package is not installed or broken such that it crashes. -- These are my opinions, not necessarily my employer's. I hate spam. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: New boot animation for 12.1.0
On Mon, Mar 19, 2012 at 4:21 PM, Daniel Drake d...@laptop.org wrote: I've just pushed everything needed for a new boot animation for 12.1.0, it will arrive in the next build. Yay! Eager to see it! The new animation is based on plymouth, Good to hear Plymouth is a bit better than it used to be. With the move to systemd, a pretty spinner will be the best we can do; the closest cousin to systemd is launchd on OSX, and programmers there also threw their hands in the air, and made a spinner. The simple times when you could guess what got stuck based on the last dot shown are long gone. Questions - How easy is it to implement something nice? I see Manuel's already been asking you about it, and it'd be great to have something cool ;-) - Something easily customizable? -- thinking of OOB splicing in images cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Installing gstreamer media codecs on XO-1
On Mon, Mar 19, 2012 at 8:55 PM, David Leeming da...@leeming-consulting.com wrote: 1. Is the above the right way to do it, if someone with more Fedora experience than I can verify.. That's exactly how I'd do it. 2. In our narrowband countries it takes an hour and downloads a lot per laptop, this is unworkable in PNG with large numbers of XOs and where the bandwidth is so expensive and unreliable. Isn't there a way to do this offline with a download, something we can run on a flashdrive? Install yum-utils, use yumdownloader --resolve target packages , you end up with all the rpms, so then rpm -Uvh *rpm. 3. is the update step above necessary (it requires downloading 33MB) Yep 4. In doing the above am I violating a ton of licenses? That's complicated. I've studied software licensing, but _I am not your lawyer_, this isn't legal advice, etc, etc. From my time under books, Adobe Flash - Not violating licenses if you do it as an end-user, if you do it as a distributor, you're expected to fill out a form, and get approval. Gstreamer codecs - Not violating licenses - Some (most!) codecs are heavily patented with software patents, depending on the jurisdiction, you're in the clear, in the gray, or in the red. It is safe to say it's impossible to write a useful program without walking over a dozen sw patents, so it's a sadly nonsensical space. m -- martin.langh...@gmail.com mar...@laptop.org -- Software Architect - OLPC - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Sugar 0.94.1
So Sugar 0.94 is changing to 0.94.1 http://git.sugarlabs.org/sugar-toolkit/mainline/commit/9010760a107e7661e75bec91f4d13da85d384052 ### Checking POT for Sugar Toolkit (0.94) ## From git.sugarlabs.org:sugar-toolkit/mainline bab9c63..9010760 sucrose-0.94 - origin/sucrose-0.94 * [new tag] v0.94.1- v0.94.1 Updating bab9c63..9010760 Fast-forward configure.ac |2 +- src/sugar/graphics/icon.py | 24 2 files changed, 13 insertions(+), 13 deletions(-) I presume I should update the Glucose 0.94 language projects, but I want to check in with the developers first. cjl ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: How to run sugar on ArchLinux
On Tue, Mar 20, 2012 at 01:21:17PM +0800, lite li wrote: Hi, After I install sugar-unstable on ArchLinux by: yaourt sugar-unstable But I don't know how to run sugar on ArchLinux. Usually one of two methods: 1. run sugar-emulator, or 2. Log out of your desktop, then on the login screen look for a desktop option for Sugar. I use sugar-install-bundle installed the EToys activity. But when I launched EToys by sugar-launch org.vpri.EtoysActivity, I got this error: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.ServiceUnknown: The name org.laptop.Shell was not prvided by any .service files. Expected. sugar-launch is only useful when Sugar is running. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
