On Mon, Jul 25, 2022 at 07:55:23AM +0000, xuyang2018.jy@xxxxxxxxxxx wrote: > > on 2022/05/21 0:04, Yang Xu wrote: > > The current_umask() is stripped from the mode directly in the vfs if the > > filesystem either doesn't support acls or the filesystem has been > > mounted without posic acl support. > > > > If the filesystem does support acls then current_umask() stripping is > > deferred to posix_acl_create(). So when the filesystem calls > > posix_acl_create() and there are no acls set or not supported then > > current_umask() will be stripped. > > > > This patch is also designed to test kernel patchset behaviour > > "move S_ISGID stripping into the vfs" > > https://patchwork.kernel.org/project/linux-fsdevel/list/?series=635692 > > The kernel patch has been merged into linux-next branch[1]. > > Does anyone review this fstests patch or give some comment? > > I plan to remove setgid_create_umask_idmapped_in_userns and > setgid_create_umask_idmapped cases because they doesn't trigger bug. > > Just treat it as a kernel regression test ie 31c01ce18 ("generic: add > test for tmpfs POSIX ACLs") format. > > CC Christian to confirm that whether need to test idmapped and > idmapped_in_userbs situation . You've also sent a patch to LTP I saw which is good. I'm not completely sure why you want to remove setgid_create_umask_idmapped{_in_userns}. But as long as we have tests for setgid stripping on idmapped mounts I'm fine with this. Christian
