On Fri, 2019-08-16 at 15:47 -0500, Benjamin Marzinski wrote: > > Err.. This patch does do the remapping in code (in propsel.c) just as > you suggest.. right? I'm truly sorry. It was late when I looked at this patch, and obviously I got it all wrong. So, I almost ACK this patch; but I'd like to simplify the logic by using the delay_* parameters only if _none_ of the san_path_ parameters have been set by the user (and also, of course, if no marginal_path parameters have been set); similar to the logic of c8775cf "multipathd: marginal_path overrides san_path_err". Users should IMO not be able to mix these parameters. Cheers, Martin -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
