XO-1.75 - Kernel news
martin at laptop.org
Sun Mar 11 12:04:31 EDT 2012
So I guess it's a risky kernel to try :-)
I've had a couple more runs with be20def and I've had it hang on the
2nd or 3rd attempt at recording. On the next boot, it didn't hang
after >20 recordings. Not sure what the noticeable variation
correlates to (if anything).
On Sat, Mar 10, 2012 at 11:28 PM, Saadia Husain Baloch
<saadia at laptop.org> wrote:
> Our emails crossed, and I would like to warn people that I experienced some
> questionable results with just be20def applied, after my machine got a
> corruption of the mmc "EXT4-fs (mmcblk0p2): group descriptors corrupted!"
> which stops the machine from booting. I have never had this before, and the
> kernel I am using is the head of arm-3.0-ramp with just this one mcam-core.c
> patch applied.
> Other than the corruption the crash still happens for me on two machines.
> On Sat, Mar 10, 2012 at 10:46 PM, Martin Langhoff <martin at laptop.org> wrote:
>> The very latest kernel for XO-1.75 11.3.x builds, which you can get
>> with yum update (and then follow
>> http://wiki.laptop.org/go/Kernel#Installing_kernel_RPMs ) is very
>> Please update to it and let us know how it fares for you!
>> What's interesting?
>> = be20def marvell-cam: ensure that the camera stops when requested
>> Opening Record and recording many videos or many audio files does not
>> end up in a hard freeze (#11657). At least not for me, helps us
>> There are still some oddities when using Record, but those may be
>> different (lesser) bugs. For example, the camera will sometimes get a
>> green snapshot for the audio recording.
>> = 4ae137c libertas: drop commands if the libertas host is asleep
>> Using aggressive suspend/resume (Control Panel > Power > Automatic
>> power management) no longer leads to hangs in busy wireless networks
>> (#11658). At least not for me, help us corroborate.
>> On Friday, I left one XO with 4ae137c with powerd configured to sleep
>> after 2s of idle, and a second XO pinging it every 5. Checked them in
>> the office earlier, and both were still going at it.
>> Grepping the log of the ping sender, it had >18K successful pings, and
>> 127 failures. Not perfect, but not a bad ratio either.
>> martin at 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 at lists.laptop.org
martin at laptop.org -- Software Architect - OLPC
- ask interesting questions
- don't get distracted with shiny stuff - working code first
More information about the Devel