Re: Distributing Patches in Email

From: Dirk Nimmich (nimmich@UNI-MUENSTER.DE)
Date: Fri Mar 03 2000 - 09:22:56 PST

  • Next message: Przemyslaw Frasunek: "Re: [ Hackerslab bug_paper ] Linux dump buffer overflow"

    Scott Blake wrote:
    > An exception the rule Marc mentions should be non-executable,
    > strongly signed updates. Concerned users can easily verify the
    > signature manually (the software does so automatically) to be
    > certain of the file's provenance and integrity.
    [...]
    > Btw, if anyone sees a flaw in our approach, I'd love to hear it.
    
    You didn't say anything about the verification of signed files and
    how those patches are applied, so the "generic" answer to this is:
    Replay attack with signed files known to have security bugs. Can be
    avoided if dates (of the signature, not of the message) and file
    versions are checked, too.
    



    This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 15:39:01 PDT