| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
| |
Locking and unlocking a non-recursive mutex is a simple memory
operation and should not fail on any reasonable platform with correct
usage. A pthread mutex can return EDEADLK on lock or EPERM on unlock,
or EINVAL if the mutex is uninitialized, but all of these conditions
would reflect serious bugs in the calling code.
Change the k5_mutex_lock and k5_mutex_unlock wrappers to return void
and adjust all call sites. Propagate this change through
k5_cc_mutex_lock and k5_cc_mutex_unlock as well.
|
|
|
|
|
|
|
|
|
|
|
| |
Calls to gssint_mecherrmap_map_errcode would result in calling
mecherror_copy with a zero-length mech OID, which would result in an
OID with 0 for length and malloc(0) for elements. On platforms which
return non-null from malloc(0), gssint_mecherrmap_destroy() wouldn't
free the elements pointer.
Avoid calling malloc(0) and don't use the length field to decide
whether to free an elements pointer.
|
|
|
|
|
|
|
| |
Signed-off-by: Kevin Wasserman <kevin.wasserman@painless-security.com>
Signed-off-by: Sam Hartman <hartmans@debian.org>
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@25200 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
| |
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@23457 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
|
|
|
| |
egregious formatting quirks. Add emacs mode settings to flag
untabified source files.
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@20876 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add several new gcc warning flags, used in the KfM build process.
Put declarations before code.
Fix a bunch of signed/unsigned type mixes, mostly by changing variable
types to unsigned int.
Fix constness in handling name of default ccache name.
Make sure functions get declared with prototypes:
krb5int_pthread_loaded krb5int_gmt_mktime krb5int_aes_encrypt
krb5int_aes_decrypt gssint_mecherrmap_init gssint_mecherramp_get.
Don't shadow global names: stat accept index open encrypt. Fix
variable shadowing in LDAP ASN.1 support.
Don't define unused krb5int_local_addresses.
Don't export internal krb5_change_set_password.
Fix error return indications from gssint_oid_to_mech.
Create and use k5-gmt_mktime.h to provide one global declaration of
krb5int_gmt_mktime, needed before we've generated krb5.h on some
platforms.
Not incorporated from initial patch: const changes in function
signatures.
ticket: 6096
status: open
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@20697 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
| |
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@20351 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
by only one mechanism. Revert RPC code to relying on this.
Build error-mapping code on a bidirectional map instead of a simple
array. When a status code is returned but has been seen returned from
a different mechanism already, generate a new number, starting at
100,000.
Use gssrpcint_printf for some more debugging code.
ticket: 5654
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19919 dc483132-0cff-0310-8789-dd5450dbe970
|
|
|
|
|
|
|
|
| |
do not include unistd.h on Windows because it doesn't exist
ticket: new
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19851 dc483132-0cff-0310-8789-dd5450dbe970
|
|
This patch creates a mapping in the mechglue/spnego code to modify
mechanism status codes when passing them back to the application, so
that mechglue's display_status dispatcher can determine the correct
mechanism to dispatch to.
This is part of the "get enhanced error messages from gssapi
applications" project; ticket 5590 has updates to the Kerberos 5
mechanism to extract enhanced error messages (when there are any) from
the Kerberos library.
util/gen.pl, util/t_*.pm: New code generation script and templates.
lib/gssapi/generic: Add a new, global mapping that enumerates the
{mechOID,status} pairs as they're seen, allowing a magic mechOID value
to indicate com_err error codes from mechglue and spnego, and
reserving status code 0 for unknown errors. Preload the Kerberos
"wrong principal" error code once for each mechanism OID used for
Kerberos, so the entries get fixed positions (1-3) in the table.
lib/gssapi/gss_libinit.c: Call the initializer and destructor
functions.
lib/gssapi/mechglue, lib/gssapi/spnego: Enter all mechanism-generated
or locally-generated status codes into the mapping table, and return
the table index to the application. Do the reverse in display_status,
to get the messages from the mechanism..
lib/rpc: Define new function gssrpcint_printf to use for debugging
instead of printf, to redirect output away from dejagnu; add a couple
more debugging calls. Check for minor status codes 1-3 now instead of
KRB5KRB_AP_WRONG_PRINC.
tests/dejagnu/krb-standalone/gssftp.exp: Test getting more detailed
error messages back, by having the ftp client attempt to authenticate
to a non-existent service, and examining the error message for the
service principal name.
ticket: new
git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19831 dc483132-0cff-0310-8789-dd5450dbe970
|