Ever since mid-December 2008, none of the newer XO-1 builds (beyond 8.2.x) were able to properly show any "moving pictures" for me.
I ended up substituting another video driver in place of xv. That gave me a smaller picture (sometimes showing only "snapshots" rather than a "movie").
Having applied the ticket #9590 fix (adding 'Virtual 1200 900' to the "Screen" section of xorg.conf), I changed my os11 systems back to using the xv video driver. I tested under identical conditions both with and without the #9590 fix. That single change made the difference -- now on my XO_1 systems, "moving pictures" work properly in F11-on-XO1 builds.
Prior to the application of the #9590 fix I could not honestly recommend any builds beyond 8.2.x to an XO-1 owner. Now I can.
mikus
p.s. My current Mplayer command line:
mplayer -framedrop -vo xv -vfm ffmpeg -lavdopts skipframe=nonref:skiploopfilter=nonkey:fast=1
Mikus,
F11 is almost out of the oven!!
Thanks Mikus & everybody for all of your hard testing! Special kudos to SMParish for doing the builds!
Here at .PY we have been able to setup the os image build machinery (thanks Bernie!) so we plan to roll out a signed image soon.. Stay tuned!
Any other major blocker that deserves attention?
Raúl
On Wed, 2010-01-27 at 01:25 -0600, Mikus Grinbergs wrote:
Ever since mid-December 2008, none of the newer XO-1 builds (beyond 8.2.x) were able to properly show any "moving pictures" for me.
I ended up substituting another video driver in place of xv. That gave me a smaller picture (sometimes showing only "snapshots" rather than a "movie").
Having applied the ticket #9590 fix (adding 'Virtual 1200 900' to the "Screen" section of xorg.conf), I changed my os11 systems back to using the xv video driver. I tested under identical conditions both with and without the #9590 fix. That single change made the difference -- now on my XO_1 systems, "moving pictures" work properly in F11-on-XO1 builds.
Prior to the application of the #9590 fix I could not honestly recommend any builds beyond 8.2.x to an XO-1 owner. Now I can.
mikus
p.s. My current Mplayer command line:
mplayer -framedrop -vo xv -vfm ffmpeg -lavdopts skipframe=nonref:skiploopfilter=nonkey:fast=1
olpc mailing list olpc@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/olpc
On Jan 27, 2010, at 2:29 AM, Raul Gutierrez Segales wrote:
Any other major blocker that deserves attention?
Have the items here been resolved?
https://www.redhat.com/archives/fedora-olpc-list/2010-January/msg00020.html
Reuben K. Caron asked:
Have the items here been resolved? https://www.redhat.com/archives/fedora-olpc-list/2010-January/msg00020.html
My own personal view is that the items cited are not in the 'blocker' category. Rather, my question is: "Will 'fixing' be frozen once F11-on-XO1 is made generally available, or can some items be addressed after the initial release ?" [I do not know the answer.]
mikus
On Wed, 2010-01-27 at 04:29 -0300, Raul Gutierrez Segales wrote:
Here at .PY we have been able to setup the os image build machinery (thanks Bernie!) so we plan to roll out a signed image soon.. Stay tuned!
Today I've been testing activities and collaboration. Everything feels snappy and works decently.
Any other major blocker that deserves attention?
The only oddity I've seen is that libertas still has some reliability problems when automatic power management is switched on.
Libertas' PM always has been a real PITA. I heard plenty of good kernel hackers swear between their teeth because of it. Marcelo would often swear aloud too :-)
Today I promised to work on a cleaner patch for the main outstanding PM issue (http://dev.laptop.org/ticket/9967), but there is certainly something else going on. For instance, we seem to loose association for no good reason. After a few suspend/wakeup cycles, I even managed to make the Libertas disappear from the USB bus!
In fact, this might not even be a 2.6.31 regression. To my (somewhat fuzzy) memory, automatic power management always has been disabled in OLPC's official builds due to similar reliability issues.
At some point we may choose to give up trying to make libertas behave and keep the power management feature disabled for good.
On Wed, 2010-01-27 at 20:57 -0300, Bernie Innocenti wrote:
In fact, this might not even be a 2.6.31 regression. To my (somewhat fuzzy) memory, automatic power management always has been disabled in OLPC's official builds due to similar reliability issues.
Yesterday I played a little with automatic power management on an XO-1 with build 8.2.1.
While I could not make libertas fail the same way of 2.6.31, suspend was often making me loose association. NetworkManager did not seem to reassociate automatically on wakeup, I had to do it manually.