On 18/05/17 15:31, Tim Deegan wrote: > Hi, > > At 15:02 +0100 on 18 May (1495119734), Andrew Cooper wrote: >> * When fabricating gl1e's from superpages, propagate the protection key as >> well, so the protection key logic sees the real key as opposed to 0. >> >> * Experimentally, the protection key checks are performed ahead of the other >> access rights. In particular, accesses which fail both protection key and >> regular permission checks yield PFEC_prot_key in the resulting pagefault. >> >> * Protection keys apply to all user mode data accesses, including accesses >> from supervisor code. > I think this would be clearer as "all data accesses to user-mode addresses".
Done. > >> PKRU WD applies to any data write, not just to >> mapping which are writable. However, a supervisor access without CR0.WP >> bypasses any protection from protection keys. >> >> Signed-off-by: Andrew Cooper <andrew.coop...@citrix.com> > Reviewed-by: Tim Deegan <t...@xen.org> Thanks. ~Andrew _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel
