On Thu, Nov 22, 2007 at 03:28:57AM -0800, Richard L. Hamilton wrote: > "ephemeral" UIDs/GIDs resulting from non-persistent Windows SID mappings > shouldn't be stored persistently (so ZFS will store the SIDs themselves, or > at least a compacted representation of them?). > > I got that. But what about tmpfs? It's not persistent across a reboot, > so does it (or what would be the risk vs benefit of allowing it if it doesn't) > store them? Curious...
There's not a conceptual issue or risk with storing ephemeral IDs inside of tmpfs. However, the risk/issue with doing so is that in general we want to have the kernel programming model that if you are not FUID-aware (VFS_XID) then we never pass ephemeral IDs across the VFS boundary, and if you are VFS_XID-aware, then you're supposed to follow a known programming interface to handle conversion. We should avoid having more special-cases where that programming model is bypassed if we can help it, because that model should be what ties consumers to appropriate asserts, debugging support, etc. -Mike -- Mike Shapiro, Solaris Kernel Development. blogs.sun.com/mws/ _______________________________________________ opensolaris-code mailing list opensolaris-code@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/opensolaris-code
