Looking through the archives and around the net, it seems that this is
a fairly common problem, but it isn't clear to me if this is an SD
card speed issue, or I've also seen reference to a kernel patch for
it.
Anyway, I was getting this intermittently on reboots, like 1 in 5 or
so. This morning I got 3 in a row and then the board came up. My
shop area is a little chilly, I'm wondering if it might be thermal,
also? Is there something I can/should do about this?
Unless “a little chilly” means below freezing, heat is not the problem.
It's acting like a race during init. I just updated the kernel to
maverick v2.6.37.2-x3, and it happens often, and many times in
a row now. It just took me 5 tries to get the board to boot.
What *does* cause this?
Found it: 6 month old kernel bug:
http://article.gmane.org/gmane.linux.ports.arm.omap/38260/match=sdhc+affected+preemption
sdhc affected by preemption.
<quote>
When going from PREEMPT_VOLUNTARY to PREEMPT_NONE the problem goes
away.
</quote>
-dave
That patch is already included in that version of the kernel he's using.:
http://bazaar.launchpad.net/~beagleboard-kernel/+junk/2.6-stable/view/head:/patches/beagle/0001-omap-mmc-Adjust-dto-to-eliminate-timeout-errors.patch
what's really weird, on my 2.6.38-rc image, that error still can
occur, BUT doesn't stop boot as it eventually finds the mmc card
anyways..
Regards,
this get's more interesting, since they have
"CONFIG_PREEMPT_VOLUNTARY=y" set for their 2.6.38-rc natty armel
config...
I have "CONFIG_PREEMPT" enabled in mine..
I'll try it on OFF on a couple mmc cards i have here..
Regards,