On Tue, Nov 28, 2006 at 11:27:46AM -0800, Andrew Morton wrote: > > The problem SEEMS to be in a read cycle. Since corruptions reproduce > > even with a dummy no-op cipher, Are you in a position to define 'corruption' more precisely yet? Bits getting flipped? Data coming back from a wrong place on disk (write a known pattern)? Random data being returned? What proportion of the read data is wrong? Is it affected by the read size? Or the raid5 stripe size? Is it only at particular offsets? etc. etc. Alasdair -- agk@xxxxxxxxxx --------------------------------------------------------------------- dm-crypt mailing list - http://www.saout.de/misc/dm-crypt/ To unsubscribe, e-mail: dm-crypt-unsubscribe@xxxxxxxx For additional commands, e-mail: dm-crypt-help@xxxxxxxx
