You are right and some calling functions like krb5_copy_keyblock do allocate, but not krb5_get_credentials(_core) if I now read the code right.
Markus ----- Original Message ----- From: "Ken Raeburn" <[EMAIL PROTECTED]> To: "Markus Moeller" <[EMAIL PROTECTED]> Cc: <[email protected]> Sent: Monday, October 15, 2007 11:23 PM Subject: Re: Bug in krb5_keyblock_data function on Solaris 10/Opensolaris > On Oct 15, 2007, at 17:44, Markus Moeller wrote: >> If other people come across this issue(noticable when >> krb5_get_credentials >> is used in cross-realm situations) it had been identified in Opensolaris >> snv_57 (http://bugs.opensolaris.org/view_bug.do?bug_id=6565115) but >> still >> hasn't been fixed in source >> http://cvs.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/uts/ >> common/gssapi/mechs/krb5/krb5/krb/copy_key.c > > Your suggested fix is to have krb5_copy_keyblock_data allocate the new > storage, but the comment documenting the function says it assumes it's > already been allocated, and the body of krb5_copy_keyblock further down > does allocate the new storage before calling krb5_copy_keyblock. So if > I'm following this right, you'd be adding a memory leak to > krb5_copy_keyblock, wouldn't you? > > ________________________________________________ Kerberos mailing list [email protected] https://mailman.mit.edu/mailman/listinfo/kerberos
