From 2d7709867147de9cf702213b844a50d238ca90b9 Mon Sep 17 00:00:00 2001 From: Thomas Schwinge Date: Fri, 30 Jul 2010 16:32:52 +0200 Subject: open_issues/dir-lookup_authority: New. --- open_issues/dir-lookup_authority.mdwn | 68 +++++++++++++++++++++++++++++++++++ 1 file changed, 68 insertions(+) create mode 100644 open_issues/dir-lookup_authority.mdwn diff --git a/open_issues/dir-lookup_authority.mdwn b/open_issues/dir-lookup_authority.mdwn new file mode 100644 index 00000000..64866eb5 --- /dev/null +++ b/open_issues/dir-lookup_authority.mdwn @@ -0,0 +1,68 @@ +[[!meta copyright="Copyright © 2010 Free Software Foundation, Inc."]] + +[[!meta license="""[[!toggle id="license" text="GFDL 1.2+"]][[!toggleable +id="license" text="Permission is granted to copy, distribute and/or modify this +document under the terms of the GNU Free Documentation License, Version 1.2 or +any later version published by the Free Software Foundation; with no Invariant +Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license +is included in the section entitled [[GNU Free Documentation +License|/fdl]]."]]"""]] + +[[!tag open_issue_hurd]] + +IRC, unknown channel, unknown date. + + I have discovered a bug in the dir-lookup protocol though + Currently, I'm investigating the bug a bit further + when doing dir-lookups with several path components, the look-up is done with the authority of the user who opened the directory, as opposed to the user doing the lookup + e.g, consider foo/bar/baz, where bar can only be used by its owner and foo and baz are world readable + if foo is opened, then transferred to another user, he can open baz, which he shouldn't be able to + this is possible where foo/bar/baz is within a single translator, and the lookup is done in a single dir-lookup + cfhammar: I'm not sure this is a bug + I have a test case that triggers the bug, and another that doesn't which currently confuses me + cfhammar: it's probably not very usual to pass around open directory ports; but if somebody does it, it's probably actually desired that it keeps the authority + it's kinda consistent with passing normal FDs + antrik: it should only allow accesses to entries not sub-entries + antrik: it isn't allowed in Linux atleast, and I'm guessing it's mandated by posix + also note that a more common scenario is a process that opens a directory and then drops authority + probably more common, that is + cfhammar: I'm not really familiar with directory access functions... I wasn't even aware that it's possible to pass around directory FDs + but if it is, it would indeed be good to know what POSIX says about this + cfhammar: I don't see how this is related?... + antrik: after the process has dropped authority it can still make lookups in directories that it should no longer be able to + cfhammar: interesting point... + cfhammar: do you think this is fixable? + antrik: Not without (defacto) changing the interface + e.g only looking up a singe path component at a time + or doing the auth check lazily on io_reauthenticate + cfhammar: yeah, obviously it's not possible without an API change. I just wonder whether it's possible without throwing the current auth/lookup mechanism overboard alltogether... + antrik: both my solutions are only minor changes to the API, but fairly major in the sense that we need to change all callers :-( + diskfs_S_dir_lookup is a very large function, for example + cfhammar: OK + cfhammar: I wonder whether there is a possible transition path without breaking all existing installations... + we could provide a new RPC while supporting the old one + note that changing fs.defs only affects glibc and the Hurd, normal apps should be fine + cfhammar: have you posted your findings to the ML yet? + No, I'm still investigating why my second test-case doesn't trigger the bug + Intrestingly it's the one using all POSIX functions... + Perhaps its a bug that maskes the lookup bug ;-) + I guess there is some quirk which you do not fully understand yet :-) + Oh, there's always a new quirk to find in the Hurd :-) + antrik: seems that dir_lookup isn't buggy after all + antrik: as all FDs are reauthenticated on setauth + ah + antrik: and (presumably) ports are unauthenticated and reauthenticated when transfered + yeah, that's the idea behind the auth protocol... + users obtain specific capabilities by authenticating generic ports against their own ID + I didn't really have a coherent view on how open flags are handled on reauth + it seems open flags always win, so that a O_READ port that is unauthed is still readable + not sure what you mean + if I open a file to read it, then reauth it with a user that isn't permitted to read it, I can still read from it + (as it should be) + by contrast permission to do lookups in a directory is determined by who authed it + so I won't be able to do lookups after a reauth, if it's not permitted by the file bits + Mmm, openat should however be able to + since you've first opened the directory with the auth + it isn't since open FDs are reauthed on setauth + not sure whether it should though, Linux behaves the same way atleast + though it could be done with POSIX.2008's O_SEARCH open flag -- cgit v1.2.3