Description of problem:
The crypt(3) man page states:
       In the SHA implementation the entire key is significant
       (instead of only the first 8 bytes in MD5).
This is incorrect.  Both MD5 and SHA consider the entire key; it's DES that only uses the first 8 bytes.

This was fixed upstream in man-pages-3.25; see the attached patch.

Version-Release number of selected component (if applicable):
man-pages-3.22-12.el6

Additional info:
From the man-pages-3.25/Changes file:
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
crypt.3
    Petr Baudis
        Correct note on key portion significance
            As Marcel Moreaux notes:

                The Linux manpage for crypt()[1] contains the following
                statement as the last sentence of the NOTES section:

                    In the SHA implementation the entire key is significant
                    (instead of only the first 8 bytes in MD5).

                It should probably say "DES" where it says "MD5" (and maybe
                "MD5/SHA" where it says "SHA"), because in MD5 password hashing,
                the entire key is significant, not just the first 8 bytes.

            This patch fixes the wording.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~