(pls. delete prev. msg – “>” not visible here)

It always pains me to hear about another open-source project that is having trouble with SD cards.

OLPC XO-1, OpenMoko, SheevaPlug, NanoNote… the list goes on and on. All are small embedded systems running Linux. All have had trouble with SD. This has been going on for years, and I’m not sure any of these projects has really solved the problem. Even people running Linux on the EEE-PC have reported problems.

The SD card problem on OLPC XO-1 was noticed almost two years ago. They worked on this a long time, but the interim solution was to post a list of “believed to be compatible” cards on the wiki. I don’t know if they ever really got beyond this interim solution. In other words, I can pop any old SD card in my camera and it will Just Work, but will it work reliably if I put it in an XO-1 with the latest patches? I don’t know.

Here’s the first bug report I saw relating to this kind of problem, maybe it will give you some clues:

http://dev.laptop.org/ticket/6532

If you have any kind of suspend/resume functionality running on your system, that would be the first thing to look at — disable it and see if the problem goes away.

I hope this turns out to be a minor problem for you, easy to solve. Looking forward to the NanoNote! Good luck…

Even with using rootwait or rootdelay=60 – after the device is powered on for some time, next time it won’t recognize the SD-card anymore with DMA enabled.

You have to wait several hours – powered off – until the device will recognize the mmc correctly while booting again.

This may sound like the chip is overheating…

But the “funny” thing is: Once it’s bootet (from MMC or NAND) I’m able to perform I/O-operations on MMC for days – reading, writing, verifying – everything without problems (booting from NAND and mounting the MMC always works).