On 22/06/2020 08:10, Riz wrote: > Hi, > I am working with Dm-verity with FEC enabled. > I have defined the --fec-roots to be "4". > > So, My question is how many errors can actually be corrected in this case for example? Hi, the code used there is Reed-Solomon code, so see definition for these codes and it parameters. But there is some issue in cryptsetup so it does not work as expected in some situations. That said, I think this was you who reported the problem https://gitlab.com/cryptsetup/cryptsetup/-/issues/574 and actually it is related to https://gitlab.com/cryptsetup/cryptsetup/-/issues/462 We have some serious problem there, few weeks ago I spent some time to debug it but did not finished it. Anyway it is strange that kernel and userspace behaves differently (despite it is based on the same FEC library). I would say that there is a bug in cryptsetup when calculating redundant code (during format), but did not have time to study it further. Note that Google does no use cryptsetup but own utilities, so I guess it works as excected in Android. Unfortunately my student that worked with me on this already disappeared and I have currently urgent personal issues to solve so please be patient, we will debug it and fix it, but it will take some time. This is not trivial issue (well, maybe it is, but debugging is not trivial, at least for me ;-) Thanks, Milan _______________________________________________ dm-crypt mailing list dm-crypt@xxxxxxxx https://www.saout.de/mailman/listinfo/dm-crypt
