Re: Kernel for 12.1.0
On Tue, Apr 10, 2012 at 5:25 PM, Chris Ball c...@laptop.org wrote: Hi Dan, We have arm-3.3 and arm-3.4 branches available for use in 12.1.0. Great! Are these ready for build inclusion now, or are there still some big parts broken? 3.4 final will probably release in early June, and is at -rc2 now. Are we sure we want to go with 3.3 rather than 3.4 for 12.1.0? I think 3.3 is the best option. 3.4 lands a bit too late for comfort in our release cycle, and we've already got a decent amount of testing with 3.3. Plus it lines us up nicely with F17. Thanks, Daniel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Kernel for 12.1.0
On Wed, Apr 11, 2012 at 2:39 PM, Daniel Drake d...@laptop.org wrote: On Tue, Apr 10, 2012 at 5:25 PM, Chris Ball c...@laptop.org wrote: Hi Dan, We have arm-3.3 and arm-3.4 branches available for use in 12.1.0. Great! Are these ready for build inclusion now, or are there still some big parts broken? 3.4 final will probably release in early June, and is at -rc2 now. Are we sure we want to go with 3.3 rather than 3.4 for 12.1.0? I think 3.3 is the best option. 3.4 lands a bit too late for comfort in our release cycle, and we've already got a decent amount of testing with 3.3. Plus it lines us up nicely with F17. F17 plans to ship with 3.3.x at this stage and will rebase to 3.4 likely post F17 GA. Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Kernel for 12.1.0
Hi, On Wed, Apr 11 2012, Daniel Drake wrote: We have arm-3.3 and arm-3.4 branches available for use in 12.1.0. Great! Are these ready for build inclusion now, or are there still some big parts broken? Still waiting for CMA+galcore and suspend/resume, so not ready yet -- probably next week. 3.4 final will probably release in early June, and is at -rc2 now. Are we sure we want to go with 3.3 rather than 3.4 for 12.1.0? I think 3.3 is the best option. 3.4 lands a bit too late for comfort in our release cycle, and we've already got a decent amount of testing with 3.3. Plus it lines us up nicely with F17. Okay! Sounds good, thanks. - Chris. -- Chris Ball c...@laptop.org http://printf.net/ One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Porting alsa-info to sysfs -- for ARM platforms
At Mon, 9 Apr 2012 17:35:52 -0400, Martin Langhoff wrote: Hi Jaroslav, Takashi, I am addressing you, because you seem to be the alsa-info maintainers. I am part of OLPC's dev team, and our latest XO laptop model is based on a very power-efficient ARM CPU. One of the oddities of the ARM platform is that it doesn't have a PCI bus, so alsa works fine, but the PCI bus isn't there. More practically, lspci isn't there so alsa-info doesn't want to play ball. The same info can be obtained from sysfs on modern linuxen. Are there any plans to port (or fall back) to reading sysfs? Or maybe a better way? Better to fix alsa-info.sh to be runnable without lspci, yes. Does the patch below suffice? thanks, Takashi --- diff --git a/utils/alsa-info.sh b/utils/alsa-info.sh index fd7df96..30c55ca 100755 --- a/utils/alsa-info.sh +++ b/utils/alsa-info.sh @@ -392,12 +392,6 @@ trap cleanup 0 if [ $PROCEED = yes ]; then -if [[ -z $LSPCI ]] -then - echo This script requires lspci. Please install it, and re-run this script. - exit 0 -fi - #Fetch the info and store in temp files/variables DISTRO=`grep -ihs buntu\|SUSE\|Fedora\|PCLinuxOS\|MEPIS\|Mandriva\|Debian\|Damn\|Sabayon\|Slackware\|KNOPPIX\|Gentoo\|Zenwalk\|Mint\|Kubuntu\|FreeBSD\|Puppy\|Freespire\|Vector\|Dreamlinux\|CentOS\|Arch\|Xandros\|Elive\|SLAX\|Red\|BSD\|KANOTIX\|Nexenta\|Foresight\|GeeXboX\|Frugalware\|64\|SystemRescue\|Novell\|Solaris\|BackTrack\|KateOS\|Pardus /etc/{issue,*release,*version}` KERNEL_VERSION=`uname -r` @@ -408,8 +402,6 @@ KERNEL_OS=`uname -o` ALSA_DRIVER_VERSION=`cat /proc/asound/version |head -n1|awk {'print $7'} |sed 's/\.$//'` get_alsa_library_version ALSA_UTILS_VERSION=`amixer -v |awk {'print $3'}` -VENDOR_ID=`lspci -vn |grep 040[1-3] | awk -F':' '{print $3}'|awk {'print substr($0, 2);}' $TEMPDIR/vendor_id.tmp` -DEVICE_ID=`lspci -vn |grep 040[1-3] | awk -F':' '{print $4}'|awk {'print $1'} $TEMPDIR/device_id.tmp` LAST_CARD=$((`grep ]: /proc/asound/cards | wc -l` - 1 )) ESDINST=$(which esd 2/dev/null| sed 's|^[^/]*||' 2/dev/null) @@ -432,7 +424,9 @@ fi cat /proc/asound/modules 2/dev/null|awk {'print $2'}$TEMPDIR/alsamodules.tmp cat /proc/asound/cards $TEMPDIR/alsacards.tmp -lspci |grep -i multi\|audio$TEMPDIR/lspci.tmp +if [[ -n $LSPCI ]]; then +lspci |grep -i multi\|audio$TEMPDIR/lspci.tmp +fi #Check for HDA-Intel cards codec#* cat /proc/asound/card*/codec\#* $TEMPDIR/alsa-hda-intel.tmp 2 /dev/null @@ -537,6 +531,7 @@ echo $FILE cat $TEMPDIR/alsacards.tmp $FILE echo $FILE echo $FILE +if [[ -n $LSPCI ]]; then echo !!PCI Soundcards installed in the system $FILE echo !!-- $FILE echo $FILE @@ -549,6 +544,7 @@ echo $FILE lspci -vvn |grep -A1 040[1-3] $FILE echo $FILE echo $FILE +endif if [ $SNDOPTIONS ] then ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Announcing Q3C05 for XO-1.5
On 9 April 2012 13:25, James Cameron qu...@laptop.org wrote: - add new fs-save command [1] for preparing an image copy of internal storage, [1] http://wiki.laptop.org/go/OLPC_Firmware_q4d09/fs-save How useful is this for a layperson to clone an XO's setup across a school/classroom? Is this a replacement for http://wiki.laptop.org/go/Imaging_for_XO-1.5 ? Do we still need to manually mitigate the side effects? http://wiki.laptop.org/go/Imaging/Side_effects Thanks, Sridhar Sridhar Dhanapalan Engineering Manager One Laptop per Child Australia ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
XO still bootable with incomplete fs-update
Perhaps I misunderstood, but I thought that fs-update had been modified to make the XO unbootable unless the process was allowed to complete. I think this was achieved by blanking the first block and only writing it properly at the end. I am finding that XOs that have received an incomplete fs-update (e.g. if power was cut in the middle) still proceed to the boot process. Given that the OS hasn't been completely written, the behaviour after that is unpredictable. This can result in countless problems in the field. Is there a transparent and foolproof way to ensure that the XO will only boot if the OS writing is allowed to complete? This applies to a NANDblaster receive as well. Thanks, Sridhar Sridhar Dhanapalan Engineering Manager One Laptop per Child Australia ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Announcing Q3C05 for XO-1.5
On Thu, Apr 12, 2012 at 03:10:16PM +1000, Sridhar Dhanapalan wrote: On 9 April 2012 13:25, James Cameron qu...@laptop.org wrote: - add new fs-save command [1] for preparing an image copy of internal ?storage, [1] ?http://wiki.laptop.org/go/OLPC_Firmware_q4d09/fs-save How useful is this for a layperson to clone an XO's setup across a school/classroom? Not useful. The layperson would require training in http://wiki.laptop.org/go/Ok and use of USB drives, and then once the image is obtained they would need to be shown how to run zhashfs in olpc-os-builder. We do not have a program available at the moment for the reverse of fs-copy, that is, reading an image from USB drive and writing it to internal storage. It would take about 15 minutes to run, as opposed to the usual 5 minutes for fs-update. This would not be a difficult program to write, estimated four hours effort. But I don't think we have a need for it at the moment, and our deployment support team does not wish to encourage imaging. I'd have to get some buy-in. Both steps could be wrapped in a boot script to avoid typing by the layperson. Is this a replacement for http://wiki.laptop.org/go/Imaging_for_XO-1.5 ? No, not entirely. It is a slow replacement for the Capture Internal microSD step http://wiki.laptop.org/go/Imaging_for_XO-1.5#Capture_Internal_microSD Do we still need to manually mitigate the side effects? http://wiki.laptop.org/go/Imaging/Side_effects Yes. Jerry could also write a startup script that does these actions if the serial number changes since last boot. But that would have an unpleasant side-effect of losing all saved data if either (a) the serial number can't be read, or (b) the internal microSD card is moved to another unit as part of a repair. Making /home/olpc a symlink to /home/$SERIAL_NUMBER might be useful. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: XO still bootable with incomplete fs-update
No, fs-update was not modified for this. zhashfs in olpc-os-builder was modified to place the zero block last in the .zd file. This would have worked in your favour. We found that erase commands sent to certain microSD cards would fail. We removed the erase commands from the start of fs-update. This would have worked against you. You might modify zhashfs further to write an empty zero block first, then write the remaining blocks, then write the real zero block. What you are seeing is the re-use of the previous installation's partition table against the new installation. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Server-devel] Looking for new low power server hardware candidate
When I took the picture in the following url, I was focusing on what it would take to run off of 12V deep cycle battery: http://schoolserver.wordpress.com/xs-installation/xs-0-7-running-on-xo-1-5/ I'm concerned with packaging, and physical robustness in a real school setting. Maybe we could get someone with ME skills to dream up a cheap package for all the accessory items. We probably don't need the DC to DC inverter and the usb hub. But then we don't have an extra port for sneaker-net, or an adult sized usb keyboard. At the fall 2011 summit, there was a general call for a turnkey XS that just worked. If we could solve the form factor problem, the XO1.75 might be a good solution. I think it was Sameer who was telling me that in Australia, they are thinking about one XS per classroom. In that setting, seems to me that XO1.75 (even with only 512MB memory) would be more than adequate. George On Tue, Apr 10, 2012 at 10:55 PM, John Watlington w...@laptop.org wrote: Why not an XO-1.75 ? On Apr 10, 2012, at 8:04 PM, Martin Langhoff wrote: On Mon, Apr 2, 2012 at 2:03 AM, George Hunt georgejh...@gmail.com wrote: I left my fitpc2 and msi servers in the Philippines, hoping they would be pressed into service in a classroom situation. So now I'm in the market for another toy. If you have time towork with us through some hitches, I'd recommend an ARM server. At this stage I'd say one of the Marvell/Globalscale Plug servers (dreamplug for example), or a trimslice. Either option will need a combination of the OS on internal SD/eMMC and the storage on an ext HDD (via USB probably). 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
G'day George, The trouble with a full mechanical engineering treatment of this is that there's no telling what size each of the parts will be. For something similar I used a kitchen vegetable rack. This is a plastic shelf, with feet, with a square grid pattern. It is often used for potatos and onions in food storage here, since these two vegetables are best stored without light. The grid pattern allows equipment to be anchored using cable ties. It also has good airflow. Equipment can be anchored above and below the floor of the rack. Cables can be threaded through widened holes. You may be able to simplify your design a bit: 1. move from a USB HDD to a fast high spec SD card, saves one port, enough saving to avoid the USB hub and DC/DC inverter, 2. locate a USB hub that takes a 12V input, 3. decide on only one Ethernet interface. A minor thing; I would not have the input USB cable beyond the baseboard. It may be hit. -- James Cameron http://quozl.linux.org.au/ ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On 04/11/2012 04:52 PM, James Cameron wrote: G'day George, The trouble with a full mechanical engineering treatment of this is that there's no telling what size each of the parts will be. For something similar I used a kitchen vegetable rack. This is a plastic shelf, with feet, with a square grid pattern. It is often used for potatos and onions in food storage here, since these two vegetables are best stored without light. The grid pattern allows equipment to be anchored using cable ties. It also has good airflow. Equipment can be anchored above and below the floor of the rack. Cables can be threaded through widened holes. You may be able to simplify your design a bit: 1. move from a USB HDD to a fast high spec SD card, saves one port, enough saving to avoid the USB hub and DC/DC inverter, 2. locate a USB hub that takes a 12V input, 3. decide on only one Ethernet interface. A minor thing; I would not have the input USB cable beyond the baseboard. It may be hit. We have been using MSI Windbox DE-220 at Nepal and it is performing good at deployment sites. -- Abhishek Singh System Engineer Open Learning Exchange (OLE) Nepal साझा शिक्षा ई-पाटी http://www.olenepal.org Tel: +977-1-551 ext. 102 signature.asc Description: OpenPGP digital signature ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On Tue, Apr 10, 2012 at 10:55 PM, John Watlington w...@laptop.org wrote: Why not an XO-1.75 ? Good point. And XO + AP + HDD would work fantastic. George, how many users per server? If 100, an XO-1.75 will do ok. Want to sign up for the Contributors Programme (search in the wiki for the URL). XO-1.75, Plug or Trimslice will do fine with a recent Fedora for ARM (from the upcoming F17 series) -- we just need to recompile the XS specific packages. Most of them will just work. AFAIK, ejabberd and xs-config will need some work, and I can help you with those. 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On Wed, Apr 11, 2012 at 1:53 PM, Martin Langhoff martin.langh...@gmail.com wrote: On Tue, Apr 10, 2012 at 10:55 PM, John Watlington w...@laptop.org wrote: Why not an XO-1.75 ? Good point. And XO + AP + HDD would work fantastic. George, how many users per server? If 100, an XO-1.75 will do ok. Want to sign up for the Contributors Programme (search in the wiki for the URL). XO-1.75, Plug or Trimslice will do fine with a recent Fedora for ARM (from the upcoming F17 series) -- we just need to recompile the XS specific packages. Most of them will just work. AFAIK, ejabberd and xs-config will need some work, and I can help you with those. It might be worthwhile seeing what we can get into mainline Fedora for the XS releases so that as the new RHEL releases come along it will just work as well as be usable on ARM platforms. Peter ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On Wed, Apr 11, 2012 at 5:29 AM, George Hunt georgejh...@gmail.com wrote: When I took the picture in the following url, I was focusing on what it would take to run off of 12V deep cycle battery: http://schoolserver.wordpress.com/xs-installation/xs-0-7-running-on-xo-1-5/ That's very cool! I'm concerned with packaging, and physical robustness in a real school setting. Maybe we could get someone with ME skills to dream up a cheap package for all the accessory items. I can't speak much about ME, but I can suggest looking at some TP-LINK and Sapido branded small APs that take USB power. You can have a USB-powered HDD as well, and you still have a free USB port on the XO. At the fall 2011 summit, there was a general call for a turnkey XS that just worked. If we could solve the form factor problem, the XO1.75 might be a good solution. I'm exploring that path with a variant of the Dreamplug, but that won't happen overnight. I think it was Sameer who was telling me that in Australia, they are thinking about one XS per classroom. In that setting, seems to me that XO1.75 (even with only 512MB memory) would be more than adequate. One XS per classroom is a _bad_ idea for other reasons. One AP per classroom is a good idea, OTOH, and an XO-1.75 can probably handle a mid-sized school OK. 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On Apr 10, 2012, at 5:04 PM, Martin Langhoff wrote: On Mon, Apr 2, 2012 at 2:03 AM, George Hunt georgejh...@gmail.com wrote: I left my fitpc2 and msi servers in the Philippines, hoping they would be pressed into service in a classroom situation. So now I'm in the market for another toy. If you have time towork with us through some hitches, I'd recommend an ARM server. At this stage I'd say one of the Marvell/Globalscale Plug servers (dreamplug for example), or a trimslice. Either option will need a combination of the OS on internal SD/eMMC and the storage on an ext HDD (via USB probably). The Kirkwood based systems such plug computers can boot both the kernel and OS from a hard drive. I have been talking with one of he Fedora ARM team about this and am going to send them an email about what is involved and some other things about the kirkwood based devices. 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
[Server-devel] XS on ARM update.
Martin, I have been meaning to email you an update about XS on ARM for a while. I had to take a hiatus from what I was doing with it, but started back on it a while back. I had been experimenting with getting XS components running on ARM. That was with Fedora 13 for ARM. With Fedora 13 for ARM ejabberd behaved nicely. I was in the process of moving to Fedora 15 but then Fedora 15 stayed in perpetual alpha and I wanted to be on a release that was well supported. Then Fedora 17 for ARM koji is building much faster than expected so I am in the process of moving over to Fedora 17. Note that currently Puppet relies on some hard coded intel specific code. I would like to chat with you some time about this as I want to discuss what is essential for the XS and which is the correct git branches etc to pull from for some olpc specific bits. . Robert Howard ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] XS on ARM update.
On Wed, Apr 11, 2012 at 12:13 PM, rihowa...@gmail.com rihowa...@gmail.com wrote: I have been meaning to email you an update about XS on ARM for a while. I had to take a hiatus from what I was doing with it, but started back on it a while back. Great that you're back on track. True, Fedora branches were unpredictable, but the F17 push is solid. That's where it's at. Note that currently Puppet relies on some hard coded intel specific code. Oh really? I'm assuming it'd be in the facter code, using lspci and dmidecode to get the facts about hardware? Short term it only needs to handle a missing lspci and dmidecode without panicking -- XO hardware is pretty fixed for a given model anyway -- long term, Puppet will want to datamine sysfs where available. Not sure if it'll replace lspci in all cases, but on HW that doesn't have a PCI bus, well... :-) I would like to chat with you some time about this as I want to discuss what is essential for the XS and which is the correct git branches etc to pull from for some olpc specific bits. Let's flesh it out here. What are your doubts? AFAIK, most things we have for the XS should work on latest F17 ARM, some things will need a rebuild from SRPM. Exceptions: - There is a tricky situation with ejabberd which will need careful handling. See the mailing list archive. For XS 0.7 we found the problem late and IIRC the fix we applied there won't work for F17. So some work is needed here. - xs-configuration handling of network-- note that this is _optional_. - All the services will be using old style init scripts. We need to invest time in either testing that they work with systemd (compat isn't perfect), or porting them to systemd (and testing that they work too ;-) ). 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On 11 April 2012 22:59, Martin Langhoff martin.langh...@gmail.com wrote: On Wed, Apr 11, 2012 at 5:29 AM, George Hunt georgejh...@gmail.com wrote: I think it was Sameer who was telling me that in Australia, they are thinking about one XS per classroom. In that setting, seems to me that XO1.75 (even with only 512MB memory) would be more than adequate. It's just an idea for us. We haven't actioned anything. One XS per classroom is a _bad_ idea for other reasons. One AP per classroom is a good idea, OTOH, and an XO-1.75 can probably handle a mid-sized school OK. Why is it such a bad idea? The thought was to do away with registration, moodle and other unnecessary services and focus only on the XMPP server. Cheers, Sridhar ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Looking for new low power server hardware candidate
On Wed, Apr 11, 2012 at 11:57 PM, Sridhar Dhanapalan srid...@laptop.org.au wrote: Why is it such a bad idea? The thought was to do away with registration, moodle and other unnecessary services and focus only on the XMPP server. You want to run a network of federated XMPP servers? It's madness. Rather, it's not madness, but until demonstrated/automated otherwise it's a high-maintenance-per-classroom setup. And the federated XMPP stuff isn't widely used == widely tested. We get obvious and clear bugs in parts of the XMPP implementation that are used (or should be used) _everywhere_. And this is on what is reportedly the best XMPP implementation available. My appetite for putting an exotic feature into use in the _middle_ of a deployment plan is... just not there. In any case, what's the upside of one-XS-per-classroom? Cost, administration, reliance on federated-XMPP all seem downsides/risks to me. 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 ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel