From: Stanislav K. <sta...@or...> - 2013-08-27 11:45:27
|
On 08/27/2013 03:22 PM, ch...@su... wrote: > Hi! >>>> But we can't rely on the fact that group ids of a user invoking the >>>> testcases are fit into __old_gid_t. >>>> >>>> And the purpose of current GETGROUPS definition is to handle such cases. >>>> >>>> A situation may happen where root user is a member of a group with a >>>> _big_ group id. >>>> >>>> So I'm not 100% sure about fixing getgroups part of the same header. >>>> >>>> Or did I miss something? >>> What I mean is that we should check if the compatibility getgroups >>> syscall exists and if not abort the test (as the setgroups will do). >> I understand, but should it be a part of syscalls/setgroups testcases? > Ideally we should move all the syscall magic into more general place, > some subdirectory of include/ I guess. > >> Maybe we should modify a little bit syscalls/getgroups tests and add >> this check there? > That would not help much, to test setgroups() you need to be able to > call getgroups() as well, thus I think that we should move them into > more general location. > > And moreover it looks like getgroups testcases currently are not able to > call the compat syscalls at all. Please, look at patches ([PATCH V2 1/2], [PATCH V2 2/2]) I sent today. I tried to fix this behaviour. |