On Wed, Jun 23, 2010 at 5:39 PM, James Cameron <qu...@laptop.org> wrote: > My calculations were based on an invalid assumption so I'll have to > retest.
So we don't have a test that shows the problem. Drat. Here are my notes anyway - - Happy to see that the commit below made it into 2.6.31 and is in our kernel. Wonder if tweaking MMC_SPI_BLOCKSATONCE to findthe best outcome or just to match whatever OFW is doing helps - (update: I see Richard's post too - maybe line up ext3/4 block size too). http://dev.laptop.org/git/olpc-2.6/commit/?h=olpc-2.6.31&id=5cf20aa557e8f9dd5af302b8f33972082479753a - Test disabling ext3/4 journalling. We have a hotspot in ext3/4 metadata ops, and it's probably doing "short writes" in a way that is likely to drive the FTL bananas. cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - 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