Re: [OLPC-AU] WLAN failure in 20x XO-1.5s [WAS: Re: NANDblaster failing]
On Tue, Feb 15, 2011 at 03:48:24PM +1100, Sridhar Dhanapalan wrote: > After upgrading the Q3A50 one to Q3A62, I get slightly different > output from the WLAN test: > > SDHCI: Error: ISR = 8000 ESR = 1 Command Timeout, > Command reg: 50a Mode reg: 0 Arg reg: 0 > Recent commands (decimal): 0 5 > Stopping > Selftest failed due to abort Thanks. The wireless card did not respond to commands. Normally a card will respond to commands. Speculation as to cause: (a) open circuit, (b) short circuit, (c) damaged card, (d) damaged host. It would require further analysis by Martin. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: NANDblaster failing
On Mon, Feb 14, 2011 at 09:39:33PM -0800, Hal Murray wrote: > 7500 isn't very many packets. Good point. It isn't the packet loss that I'm looking for, (there will always be some, and even more when flooding), but rather complaints of SDHCI problems from the kernel, to match against the complaints from OpenFirmware. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Developer locking an unlocked XO
On Tue, Feb 15, 2011 at 04:11:30PM +1100, Sridhar Dhanapalan wrote: > On 15 February 2011 13:27, Chris Ball wrote: > > Hi, > > > > ? > # You can reverse the disable-security command by entering > > ? > enable-security at the 'ok' prompt. Security will then be > > ? > permanently enabled until disabled again. > > > > Yes, but that'll use OLPC's keys (if they were installed in > > manufacturing). ?You might want to use your own keys, which > > would involve a different procedure. > > Is that the developer key mentioned at > http://wiki.laptop.org/go/Firmware_security#Deployment_Key_Manufacturing_Data_Tags > ? Yes, that is one of the OLPC master keys. See the list of five public keys? Just after it is the text: "An OLPC "master" version of _each_ of those public keys is stored within the Open Firmware image, so that it will be rewritten upon a firmware update." I've emphasised "each". So if you have not injected your deployment keys into a laptop that is in the field, and then you go and enable-security on it, then you will require an OLPC developer key to unlock it again. I can't imagine you wanting that. Inject your deployment keys first. The OLPC "master" keys will always be there if you use OLPC firmware. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Developer locking an unlocked XO
Hi, > Is that the developer key mentioned at > http://wiki.laptop.org/go/Firmware_security#Deployment_Key_Manufacturing_Data_Tags > ? Yes, it's that set of keys. I'd suggest discussing the rest off-list with Martin/OLPCA. - Chris. -- Chris Ball One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: NANDblaster failing
> I don't see any problems in Linux. Following James' advice, I associated > with an AP and pinged the gateway for a few hours. Over ~7500 pings, I had > 1% packet loss and no errors in /ver/log/dmesg. 7500 isn't very many packets. Try ping -f. (f for flood, need to be root) It takes me 13 seconds to get to 7500 packets. If that doesn't find a problem, try longer packets. In case you don't have a man page handy... -c xxx (c for count, stops after xxx packets) -s xxx (s for size, sends packets of xxx bytes) -f puts an "interesting" load on things. You might try XO to XO rather than going through the AP and/or loading the router. (but the lights on the AP and/or router might be helpful) Or XO to PC rather than router... -- 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: Developer locking an unlocked XO
On 15 February 2011 13:27, Chris Ball wrote: > Hi, > > > # You can reverse the disable-security command by entering > > enable-security at the 'ok' prompt. Security will then be > > permanently enabled until disabled again. > > Yes, but that'll use OLPC's keys (if they were installed in > manufacturing). You might want to use your own keys, which > would involve a different procedure. Is that the developer key mentioned at http://wiki.laptop.org/go/Firmware_security#Deployment_Key_Manufacturing_Data_Tags ? We are currently trying to get a handle on this process so that we can get our keys applied in the factory. Sridhar ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [OLPC-AU] WLAN failure in 20x XO-1.5s [WAS: Re: NANDblaster failing]
On 15 February 2011 15:25, James Cameron wrote: > On Tue, Feb 15, 2011 at 02:37:21PM +1100, Sridhar Dhanapalan wrote: >> Possibly related to this, we have at least 20x XO-1.5s that have >> completely dead WLAN. The OpenFirmware wireless test fails >> immediately. > > In what way does the test fail? > > What version of OpenFirmware was used to test with? Trying with two that I have here, I get the same error: SHDCI: Error: ISR = 8000 ESR = 1 Command Timeout Stopping Selftest failed due to abort The firmware was: Q3A50 A3A36 After upgrading the Q3A50 one to Q3A62, I get slightly different output from the WLAN test: SDHCI: Error: ISR = 8000 ESR = 1 Command Timeout, Command reg: 50a Mode reg: 0 Arg reg: 0 Recent commands (decimal): 0 5 Stopping Selftest failed due to abort >> The serial numbers are: > > Make sure you talk to Martin about this. devel@lists.laptop.org tends > to be focused on development discussions rather than deployment support. Sure, will do. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [OLPC-AU] WLAN failure in 20x XO-1.5s [WAS: Re: NANDblaster failing]
On Tue, Feb 15, 2011 at 02:37:21PM +1100, Sridhar Dhanapalan wrote: > Possibly related to this, we have at least 20x XO-1.5s that have > completely dead WLAN. The OpenFirmware wireless test fails > immediately. In what way does the test fail? What version of OpenFirmware was used to test with? > The serial numbers are: Make sure you talk to Martin about this. devel@lists.laptop.org tends to be focused on development discussions rather than deployment support. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
WLAN failure in 20x XO-1.5s [WAS: Re: NANDblaster failing]
On 15 February 2011 14:27, Sridhar Dhanapalan wrote: > On 12 February 2011 01:39, Martin Langhoff wrote: >> On Fri, Feb 11, 2011 at 2:54 AM, James Cameron wrote: >>> I'm quite familiar with that part of the SDHCI implementation in the >>> firmware. >> >> And practical experience in the field backs James' and Chris' notes >> and recommendations :-) >> >> The wlan card is acting up. Maybe I can add a suggestion to test it >> under OFW as a NANDBlaster receiver, as well as testing it under >> Linux. > > It seems to work fine as a NANDblaster receiver. It also passes the > firmware tests . The only thing it can't do is act as a NANDblaster > sender. By that I mean it runs for 30-60 minutes and then starts > spitting out the errors I mentioned in my other message. > > I don't see any problems in Linux. Following James' advice, I > associated with an AP and pinged the gateway for a few hours. Over > ~7500 pings, I had 1% packet loss and no errors in /ver/log/dmesg. > > The serial number for that XO is SHC01400234. > Possibly related to this, we have at least 20x XO-1.5s that have completely dead WLAN. The OpenFirmware wireless test fails immediately. The serial numbers are: SHC01600DFB SHC01901C07 SHC01901776 SCH01901C7B SHC01901D75 SHC019011CA SHC019011B9 SHC01901144 SHC01901164 SHC0190104A SHC01901670 SHC01901755 SHC01901708 SHC0190113F SHC01901357 SHC0190136C SHC019010FC SHC01901D53 SHC01901D7A SHC01901D69 ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: NANDblaster failing
On 12 February 2011 01:39, Martin Langhoff wrote: > On Fri, Feb 11, 2011 at 2:54 AM, James Cameron wrote: >> I'm quite familiar with that part of the SDHCI implementation in the >> firmware. > > And practical experience in the field backs James' and Chris' notes > and recommendations :-) > > The wlan card is acting up. Maybe I can add a suggestion to test it > under OFW as a NANDBlaster receiver, as well as testing it under > Linux. It seems to work fine as a NANDblaster receiver. It also passes the firmware tests . The only thing it can't do is act as a NANDblaster sender. By that I mean it runs for 30-60 minutes and then starts spitting out the errors I mentioned in my other message. I don't see any problems in Linux. Following James' advice, I associated with an AP and pinged the gateway for a few hours. Over ~7500 pings, I had 1% packet loss and no errors in /ver/log/dmesg. The serial number for that XO is SHC01400234. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Developer locking an unlocked XO
Hi, > # You can reverse the disable-security command by entering > enable-security at the 'ok' prompt. Security will then be > permanently enabled until disabled again. Yes, but that'll use OLPC's keys (if they were installed in manufacturing). You might want to use your own keys, which would involve a different procedure. - Chris. -- Chris Ball One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Re: Developer locking an unlocked XO
# You can reverse the disable-security command by entering enable-security at the 'ok' prompt. Security will then be permanently enabled until disabled again. http://wiki.laptop.org/go/Activation_and_developer_keys (sorry sent blank reply just before) > Hi, > >> Is it possible to lock an XO that came from the factory unlocked? >> >> All of our XOs are currently unlocked and we are comfortable with >> that for the moment. However, we may want to change this in the >> future. > > Yes, entirely possible; Martin can help. > > - Chris. > -- > Chris Ball > One Laptop Per Child > ___ > Devel mailing list > Devel@lists.laptop.org > http://lists.laptop.org/listinfo/devel > > _ > This mail has been virus scanned by Australia On Line > see http://www.australiaonline.net.au/mailscanning ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Re: Developer locking an unlocked XO
> Hi, > >> Is it possible to lock an XO that came from the factory unlocked? >> >> All of our XOs are currently unlocked and we are comfortable with >> that for the moment. However, we may want to change this in the >> future. > > Yes, entirely possible; Martin can help. > > - Chris. > -- > Chris Ball > One Laptop Per Child > ___ > Devel mailing list > Devel@lists.laptop.org > http://lists.laptop.org/listinfo/devel > > _ > This mail has been virus scanned by Australia On Line > see http://www.australiaonline.net.au/mailscanning ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Developer locking an unlocked XO
Hi, > Is it possible to lock an XO that came from the factory unlocked? > > All of our XOs are currently unlocked and we are comfortable with > that for the moment. However, we may want to change this in the > future. Yes, entirely possible; Martin can help. - Chris. -- Chris Ball One Laptop Per Child ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Developer locking an unlocked XO
Is it possible to lock an XO that came from the factory unlocked? All of our XOs are currently unlocked and we are comfortable with that for the moment. However, we may want to change this in the future. Thanks, Sridhar Sridhar Dhanapalan Technical Manager One Laptop per Child Australia M: +61 425 239 701 E: srid...@laptop.org.au A: G.P.O. Box 731 Sydney, NSW 2001 W: www.laptop.org.au ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Planning to package our 'audited' libtommath and libtomcrypt
I am approaching bios-crypto again -- with the intention to split off - libtommath -- it's trivial to reuse the spec from the official pkg, and the delta is small - libtomcrypt -- same - bios-crypto -- the binaries and some low-level scripts - bios-crypto-utils -- will move to a separate repo & rpm the shell & python utils we use in the XS Here is some background on the "audited" libs http://lists.fedoraproject.org/pipermail/devel/2009-December/043679.html There will probably be limited interest from Fedora proper, but at least there'll be a pacakge in case anyone is crazy enough to push more on this track :-) 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: Bug Report
On Mon, Feb 14, 2011 at 3:50 PM, Walter Bender wrote: > Using an XO-1.5 with an external USB->VGA, the image projected fine, > but it did not display on the XO screen. This is on 10.1.3. (It used > to show the image mirrored on both displays. It is inconvenient to > have the display only on the projector, since you need to turn your > back to the audience.) Expected, not-a-bug. Unfortunately, mirroring involved a lot of complex maneuvering that led to several other bugs, and very slow performance. I took the executive decision to have it only displayed on the external display, which can be awkward, but works faster and more consistently. You can see for yourself all the glitches and bugs discussed and battled in https://dev.laptop.org/ticket/10210 when we used Xephyr for mirroring. cheers, m -- 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
Created '/git/packages' for fedpkg-style repos, reorganized XS source repos
In dev.laptop.org, I created /git/packages -- to store git repos that follow the fedpkg style. Very useful for any packages that aren't in Fedora, or where we patch or frob the pkg slightly. And converted almost all the XS-related packages to use this infra. Changes - Trimmed trailing '.git' from the repo path. You will have to edit your .git/config file to match. I fixed all references to the affected repos in wiki.laptop.org . - Split off the spec files to their own fedpkg-style repos - The fedpkg-style repos have their funny branch setup. I removed any strange branch setup from the actual source repos. - There were 2 "invisible" pam_sotp repos tracking spec and patches under /git/users/martin/ -- merged the good bits in them, fedpkg'd the results. Now lives under /git/packages as it should. This affects pam_sotp xs-activation xs-activity-server xs-logos xs-otp xs-release xs-rsync xs-tools moodle and xs-config will see a major revamp, so I haven't attacked them yet. 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
Bug Report
Couldn't get Trac to cooperate, so I am filing this bug report by email: Using an XO-1.5 with an external USB->VGA, the image projected fine, but it did not display on the XO screen. This is on 10.1.3. (It used to show the image mirrored on both displays. It is inconvenient to have the display only on the projector, since you need to turn your back to the audience.) -walter -- Walter Bender Sugar Labs http://www.sugarlabs.org ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
F14 losing keyboard on S/R XO-1.5 -- #10650
Hi folks, we are finding that this hits us very often, and makes working on the XO a damn pain. We had seen it but thought it wasn't frequent. Now that we're working full steam ahead on F14, the complaints mark the passing of each hour. Daniel's promised to look at it, but the whole situation might benefit from extra eyes, specially those trained on input devices and s/r codepaths ;-) https://dev.laptop.org/ticket/10650 thanks! 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: F15 glibc again fails on AMD Geode LX
Hi All, On Fri, Feb 4, 2011 at 5:37 PM, John Gilmore wrote: > FYI: Early Fedora 15 builds don't run on the Geode, again. This time, > people seem to be on the issue, and may resolve it without much work > from OLPC. But I think it would be worth spending some testing time > to make sure it's really resolved, so the final F15 can be used as > a basis for an OLPC release for the XO-1. This is now (or at least should be) fixed in F-15 and rawhide (note we've now branched in preparation for the Alpha release) so it would be great if someone could confirm that its working as expected. Unfortunately I'm travelling at the moment without access to a XO-1 to be able to test. Regards, Peter ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: localpkg helper script for fedpkg -- take two
On 2/14/11 8:51 AM, Martin Langhoff wrote: > On Tue, Jan 25, 2011 at 4:11 PM, Martin Langhoff > wrote: >> If you are using or experimenting with fedpkg, I am putting some >> simple helper bits in an accessory "localpkg" python script (at >> http://dev.laptop.org/git/users/martin/localpkg > > Updated and added a README - > > lpkg - some extra commands to complement fedpkg > > Getting started > > - You have your fedpkg-style checkouts tracking spec files > and patches. We'll put them in ~/fpkg for this tour. > > - You have the actual src code checkouts. They use git. > We'll have them in ~/src for this tour. > > Install > > - clone localpkg into your ~/src > - $ ln -s ~/src/localpkg/lpkg ~/bin/lpkg > > Use > > - $ cd ~/src/sugar # hack, commit. lpkg will help you build > (with fedpkg) from the tip of HEAD of this checkout. > Use git describe to see the "version" that will be set > in the resulting RPMs and SRPMs. > > - $ cd ~/fpkg/sugar # make sure you are in the right branch. > fedpkg "guesses" the build target based on the name of the > remote branch you are tracking. > > - $ git config --set lpkg.checkout ~/src/sugar > alternatively, you can pass --git-checkout to lpkg > > - $ lpkg prep # prepares spec and source tarball > $ fedpkg local # builds it locally > $ lpkg lint # run rpmlint :-) > ... # test it! does it work? > $ lpkg publish # scp to dev.laptop.org:public_rpms// > > cheers, > > > > m This looks like good stuff. There is another similar project that even predated fedpkg, called Tito. https://github.com/dgoodwin/tito There might be some things of interest to you there. -- Jesse Keating Fedora -- Freedom² is a feature! identi.ca: http://identi.ca/jkeating ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
localpkg helper script for fedpkg -- take two
On Tue, Jan 25, 2011 at 4:11 PM, Martin Langhoff wrote: > If you are using or experimenting with fedpkg, I am putting some > simple helper bits in an accessory "localpkg" python script (at > http://dev.laptop.org/git/users/martin/localpkg Updated and added a README - lpkg - some extra commands to complement fedpkg Getting started - You have your fedpkg-style checkouts tracking spec files and patches. We'll put them in ~/fpkg for this tour. - You have the actual src code checkouts. They use git. We'll have them in ~/src for this tour. Install - clone localpkg into your ~/src - $ ln -s ~/src/localpkg/lpkg ~/bin/lpkg Use - $ cd ~/src/sugar # hack, commit. lpkg will help you build (with fedpkg) from the tip of HEAD of this checkout. Use git describe to see the "version" that will be set in the resulting RPMs and SRPMs. - $ cd ~/fpkg/sugar # make sure you are in the right branch. fedpkg "guesses" the build target based on the name of the remote branch you are tracking. - $ git config --set lpkg.checkout ~/src/sugar alternatively, you can pass --git-checkout to lpkg - $ lpkg prep # prepares spec and source tarball $ fedpkg local # builds it locally $ lpkg lint # run rpmlint :-) ... # test it! does it work? $ lpkg publish # scp to dev.laptop.org:public_rpms// cheers, m ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Help with signing messages
On Sun, Feb 13, 2011 at 7:38 PM, Michael Stone wrote: > a) Online signing w/ HTTPS: That's a very good idea. I had mentioned using OLPC BIOS Crypto to Esteban, but the main issue is ensuring the msg comes from a trusted network host, so HTTPS does the job perfectly fine, and is the simplest to implement. 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: Help with signing messages
Hi Michael (sorry for confuse) , > > Thanks for your answers. This will be very useful to me. > > Regarding your doubt > > > I'm confused here because, above, you said that a typical message might be > >> "Your laptop will be blocked tomorrow, please update your blacklist." >> >> Is this kind of message supposed to be: >> >> * received and displayed by all laptops? >>* received by all laptops and displayed by one or more "addressed" >> laptops? >>* received and displayed only by addressed laptops? >> >> Maybe this example was not the best. "Classes start tomorrow" can be more > representative. > > The other kind of message "Your laptop will be blocked tomorrow, please > update your blacklist." can be displayed by internal dbus message (offline). > > Regards, > Esteban. > ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: Help with signing messages
Hi Daniel, Thanks for your answers. This will be very useful to me. Regarding your doubt I'm confused here because, above, you said that a typical message might be > "Your laptop will be blocked tomorrow, please update your blacklist." > > Is this kind of message supposed to be: > > * received and displayed by all laptops? >* received by all laptops and displayed by one or more "addressed" > laptops? >* received and displayed only by addressed laptops? > > Maybe this example was not the best. "Classes start tomorrow" can be more representative. The other kind of message "Your laptop will be blocked tomorrow, please update your blacklist." can be displayed by internal dbus message (offline). Regards, Esteban. ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: OFW and journal recovery warning
G'day Yioryos, The "ext3 journal needs recovery" is emitted by OpenFirmware when the ext3 filesystem contains a bit flag set in the metadata. Linux sets this bit in various situations. fsck.ext3 does not *report* this bit. When the filesystem is mounted, even read-only, the kernel mount will cause the ext3 journal to be recovered. It has very little to do with fsck. The "ext3 journal needs recovery" is emitted only when a file is opened. You'll get one for each open while the feature flag is set. OpenFirmware does not clear the feature flag. OpenFirmware blocks writes to the filesystem while the feature flag is set. The flag is EXT3_FEATURE_INCOMPAT_RECOVER defined in linux/ext2_fs.h, value 0x0004, and is read from field s_feature_incompat in struct ext2_super_block. This is done in line 53 of ext2fs/sb.fth http://tracker.coreboot.org/trac/openfirmware/browser/ofw/fs/ext2fs/sb.fth The filesystem is not actually clean. It has this bit set. However, there's different opinions on what "clean" is. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
OFW and journal recovery warning
When booting the XO-1 (q2e45) or XO-1.5 (q3a61) from an ext3 formatted USB stick or external SDcard, the OFW reports that the "ext3 journal needs recovery". However, stopping the boot process at this exact point (power button), rebooting from the internal flash/nand (running F14 builds) and fsck.ext3'ing the external device reports that the file system is clean. Furthermore, rebooting immediately after that from the external device there is no OFW warning. I tried to see how OFW determines the journaling status but is beyond me I'm afraid. How does it? and why is it reporting journal issues when the fs is actually clean? I would appreciate any suggestions Thx ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
