krb5.git
17 years agoTest for existence of identity/obj before try to delete files
Kevin Koch [Fri, 2 Mar 2007 20:42:26 +0000 (20:42 +0000)]
Test for existence of identity/obj before try to delete files

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19202 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoOn unload, free up g_mechSet and g_mechList
Ken Raeburn [Fri, 2 Mar 2007 05:30:30 +0000 (05:30 +0000)]
On unload, free up g_mechSet and g_mechList

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19201 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoupdate dependencies
Ken Raeburn [Fri, 2 Mar 2007 04:17:12 +0000 (04:17 +0000)]
update dependencies

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19199 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDelete more stuff for 'clean' and 'distclean' targets
Ken Raeburn [Fri, 2 Mar 2007 03:48:59 +0000 (03:48 +0000)]
Delete more stuff for 'clean' and 'distclean' targets

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19198 dc483132-0cff-0310-8789-dd5450dbe970

17 years agovalgrind detects uninitialized (but really unused) bytes in 'queue'
Ken Raeburn [Thu, 1 Mar 2007 02:19:41 +0000 (02:19 +0000)]
valgrind detects uninitialized (but really unused) bytes in 'queue'

The gsstest program exports a GSSAPI security context to a blob in
memory, writes that memory to a file, and reads it back to use it.
Under valgrind, the writing phase triggers a warning about
uninitialized storage.

The "queue" structure as implemented in generic/util_ordering.c holds
an array of values, some of which may never be initialized.  As far as
I can tell, those uninitialized values are never used before being
initialized, either, but valgrind doesn't know that.

This patch zaps the structure contents (including the array) before
using the queue object.

ticket: new
target_version: 1.6.1
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19196 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokrb5_get_cred_from_kdc fails to null terminate the tgt list
Jeffrey Altman [Thu, 1 Mar 2007 01:49:11 +0000 (01:49 +0000)]
krb5_get_cred_from_kdc fails to null terminate the tgt list

if the next tgt in a cross-realm traversal cannot be
obtained find_nxt_kdc() was calling krb5_free_creds()
on the last tgt in the list but was failing to nullify
the pointer to the cred that was just freed.

if there were no additional tgts obtained,
krb5_get_cred_from_kdc() would return a non-NULL terminated
cred list to the caller.  This would result in a crash
when attempting to manipulate the non-existent cred past
the end of the list.

This commit nullifies the credential pointer in
find_nxt_kdc() after the call to krb5_free_creds()

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19195 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago Add intalert.h which should have been added as part of
Jeffrey Altman [Wed, 28 Feb 2007 18:00:43 +0000 (18:00 +0000)]
Add intalert.h which should have been added as part of
revision 19189

ticket: 5452

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19191 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoWindows - some apps define ssize_t as a preprocessor symbol
Jeffrey Altman [Wed, 28 Feb 2007 07:10:33 +0000 (07:10 +0000)]
Windows - some apps define ssize_t as a preprocessor symbol

ticket: new
component: krb5-libs
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19190 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM Improved Alert Management
Jeffrey Altman [Wed, 28 Feb 2007 07:01:21 +0000 (07:01 +0000)]
NIM Improved Alert Management

This patch implements the new Alert Management functionality.

Many improvements to avoid race conditions and improve resource
  tracking.

ticket: new
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19189 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoUnused macro NEG removed
Ken Raeburn [Mon, 26 Feb 2007 23:36:46 +0000 (23:36 +0000)]
Unused macro NEG removed

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19183 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoWhen importing a name with a valid (non-null) handle but a zero
Ken Raeburn [Sat, 24 Feb 2007 00:58:05 +0000 (00:58 +0000)]
When importing a name with a valid (non-null) handle but a zero
length, set GSS_S_BAD_NAME but not GSS_S_CALL_INACCESSIBLE_READ.

ticket: 5445
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19172 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoIf a reflection is detected, zap the message buffer pointer output
Ken Raeburn [Sat, 24 Feb 2007 00:56:23 +0000 (00:56 +0000)]
If a reflection is detected, zap the message buffer pointer output
argument as well as actually freeing the buffer.  (Found while using
the gsstest option to exercise error conditions.)

ticket: 5445
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19171 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoInitialize "loopback" field in newly allocated name structure
Ken Raeburn [Thu, 22 Feb 2007 22:57:34 +0000 (22:57 +0000)]
Initialize "loopback" field in newly allocated name structure

ticket: 5445
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19170 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoEnsure consistancy between prototypes and functions
Ezra Peisach [Mon, 19 Feb 2007 02:37:10 +0000 (02:37 +0000)]
Ensure consistancy between prototypes and functions

I am using an older compiler that is complaining that prototypes
do not match the functions they reference.  The issue is that a number
of prototypes are using "const int foo" while the function is "int foo".
From a caller sense it makes no difference - but the compiler is correct
they are different.

All is now consistant.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19169 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoConditionalize some ipv6 bits
Ken Raeburn [Fri, 16 Feb 2007 21:43:28 +0000 (21:43 +0000)]
Conditionalize some ipv6 bits

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19168 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoConditionalize more bits that might not be defined by the OS
Ken Raeburn [Fri, 16 Feb 2007 21:40:44 +0000 (21:40 +0000)]
Conditionalize more bits that might not be defined by the OS

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19167 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't use struct in6_addr if we don't have IPv6 support
Ken Raeburn [Fri, 16 Feb 2007 21:12:40 +0000 (21:12 +0000)]
Don't use struct in6_addr if we don't have IPv6 support

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19165 dc483132-0cff-0310-8789-dd5450dbe970

17 years agorename krb5_server_decrypt_ticket_keyblock() to
Tom Yu [Tue, 13 Feb 2007 00:35:48 +0000 (00:35 +0000)]
rename krb5_server_decrypt_ticket_keyblock() to
krb5int_server_decrypt_ticket_keyblock()

ticket: 5349

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19159 dc483132-0cff-0310-8789-dd5450dbe970

17 years agohack to permit GetEnvironmentVariable usage without requiring getenv() conversion
Jeffrey Altman [Mon, 12 Feb 2007 14:54:28 +0000 (14:54 +0000)]
hack to permit GetEnvironmentVariable usage without requiring getenv() conversion

Windows has a major flaw when it comes to the use of getenv/putenv.
getenv/putenv do not modify the actual environment of the process.
Instead, they modify a copy of the environment block at the time the
C Runtime Library was initialized for the current module.  In other
words, the C Runtime Library environment block for the executable
is not the same as the C Runtime Library environment block for the
krb5_32.dll library, etc.

This results in problems when a process wants to set the default
ccache name outside the krb5_context.  The krb5_context default ccname
disappears when the context is destroyed.  gss_acquire_cred() suffers
from the creation and destruction of krb5_contexts and therefore the
krb5_context default ccname cannot be used to set a default ccname.
Instead, the process environment must be used.

In order to modify the process environment, SetEnvironmentVariable()
must be used.  However, this does not result in the C Runtime Library
environment blocks being updated.  putenv() does not see the definition
of "KRB5CCNAME".

This patch modifies get_os_ccname() for Windows to check
GetEnvironmentVariable() before checking the registry.  This hack will
work as long as there is no "KRB5CCNAME" variable in the C Runtime
Library environment block.

The long term solution is to replace all calls to getenv and putenv
with GetEnvironmentVariable/SetEnvironmentVariable for Windows.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19154 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoIn the resolve function - ensure on memory allocation error -
Ezra Peisach [Mon, 12 Feb 2007 00:19:47 +0000 (00:19 +0000)]
In the resolve function - ensure on memory allocation error -
deref of null pointer does not happen.

Also include kt-int.h for prototypes.

ticket: 5411

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19153 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokrb5_kt_default_name should take an unsized length
Ezra Peisach [Mon, 5 Feb 2007 20:39:59 +0000 (20:39 +0000)]
krb5_kt_default_name should take an unsized length

The internal code to krb5_kt_default_name was casting the length to size_t.
Change prototype to take an unsigned - which makes sense in how the code uses
it.

ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19144 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokrb5_kt_get_type should return const char *.
Ezra Peisach [Mon, 5 Feb 2007 15:32:13 +0000 (15:32 +0000)]
krb5_kt_get_type should return const char *.

The code returns a pointer to static structures. Just to enforce the
assumption that users can not change the returned data.

Change prototype of krb5_kt_get_type to return const char *. The other
changes are to clean up warnings - no change in code - usage assumed const.

ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19143 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoAdd a new program to perform various tests on the WRFILE: and MEMORY: keytabs
Ezra Peisach [Sun, 4 Feb 2007 03:08:47 +0000 (03:08 +0000)]
Add a new program to perform various tests on the WRFILE: and MEMORY: keytabs

I developed this program to test functionality of the MEMORY keytab - which
resulted in the numerous fixes that have been committed recently.

Tests all functioanlity of keytabs except for krb5_kt_default() and
krb5_kt_read_service_key() - although essential functionality tested.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19142 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMEMORY keytab krb5_kt_get_entry sets enctype to 0 if unspecified
Ezra Peisach [Sun, 4 Feb 2007 02:49:54 +0000 (02:49 +0000)]
MEMORY keytab krb5_kt_get_entry sets enctype to 0 if unspecified

If one invokes krb5_kt_get_entry() for a memory keytab w/ an unspecified
enctype (i.e. 0) - the returned keytab has it's enctype set to 0 as well.

Also - in copying out the found keytab_entry - when a kvno is unspecified,
the last match found is used  - not the one with the highest vno.
This was caused by copying out the variable "entry" and not "match".

Ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19141 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMEMORY keytab does not copy keytab_entry keyblock contents
Ezra Peisach [Sun, 4 Feb 2007 02:03:48 +0000 (02:03 +0000)]
MEMORY keytab does not copy keytab_entry keyblock contents

In krb5_kt_add_entry: The MEMORY keytab does not make a copy of the
keytab_entry keyblock contents - but instead retains a pointer to the
incomming one.

In krb5_kt_get_entry and krb5_kt_get_next - a pointer to internal
keyblock contents memory is returned to the caller - which is subsequently
freed when tht caller invokes krb5_free_keytab_entry_contents.

Solution is to use krb5_copy_keyblock_contents() instead of simply copying
the structure.

Ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19140 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoThe default K4 compatibility mode is now none, not preauth. Also document
Russ Allbery [Sat, 3 Feb 2007 18:33:23 +0000 (18:33 +0000)]
The default K4 compatibility mode is now none, not preauth.  Also document
that the valid values for v4_mode are the valid arguments to the -4 flag
to krb5kdc.

Ticket: 2724
Component: krb5-doc
Target_Version: 1.6.1
Tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19139 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMemory leak in new memory keytab
Ezra Peisach [Sat, 3 Feb 2007 18:09:56 +0000 (18:09 +0000)]
Memory leak in new memory keytab

The memory keytab introduced in ticket 5411 suffers from a memory leak
in the Heimdal compatible kt_close() handling and thread destructor in which
the cursor->entry is not freed.

ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19138 dc483132-0cff-0310-8789-dd5450dbe970

17 years agobuffer overflow in krb5_kt_get_name
Ezra Peisach [Thu, 1 Feb 2007 19:17:31 +0000 (19:17 +0000)]
buffer overflow in krb5_kt_get_name

krb5_kt_get_name() allows the called to specify the size of the buffer to copy
the name into. The size must be big enough for the tailing nul character.

If one specified a buffer length that is precisely the strlen w/o allowing for
the nul - the functions would copy one past the end of the buffer.

No code in our tree would be subject this problem - as buffers in use are 1024
or BUFSIZ....

The logic failure was:

strlen(p+1) vs. strlen(p)+1

The code is essentially duplicated in the three changed files.

Ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19137 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoUse krb5.h now for kdb errors, not kdb5_err.h
Ken Raeburn [Thu, 1 Feb 2007 01:05:41 +0000 (01:05 +0000)]
Use krb5.h now for kdb errors, not kdb5_err.h

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19136 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't need to initialize error tables
Ken Raeburn [Thu, 1 Feb 2007 00:09:07 +0000 (00:09 +0000)]
Don't need to initialize error tables

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19133 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't need to register WRFILE: keytab type, it's already in the compiled-in
Ken Raeburn [Thu, 1 Feb 2007 00:02:36 +0000 (00:02 +0000)]
Don't need to register WRFILE: keytab type, it's already in the compiled-in
list.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19132 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoFix merge botches:
Tom Yu [Wed, 31 Jan 2007 00:53:11 +0000 (00:53 +0000)]
Fix merge botches:

restore krb5_get_init_creds_opt_set_change_password_prompt()

undo accidental reversion of preauth_tryagain change to
krb5_get_init_creds()

ticket: 5420
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19131 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoremove krb5_get_init_creds_opt_set_pkinit() for now
Tom Yu [Wed, 31 Jan 2007 00:52:59 +0000 (00:52 +0000)]
remove krb5_get_init_creds_opt_set_pkinit() for now

ticket: 5420
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19130 dc483132-0cff-0310-8789-dd5450dbe970

17 years agosort, remove duplicates
Ken Raeburn [Tue, 30 Jan 2007 23:37:36 +0000 (23:37 +0000)]
sort, remove duplicates

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19129 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoupdate def-check to look at preauth_plugin.h
Tom Yu [Tue, 30 Jan 2007 21:40:20 +0000 (21:40 +0000)]
update def-check to look at preauth_plugin.h

ticket: 5420
status: open

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19128 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoget_init_creds_opt extensibility
Tom Yu [Tue, 30 Jan 2007 21:38:47 +0000 (21:38 +0000)]
get_init_creds_opt extensibility

 r18922@cathode-dark-space:  coffman | 2006-12-04 18:30:15 -0500
 First cut at making the get_init_creds_opt structure extendable
 and adding library functions to set options for preauthentication
 plugins.

 This does *not* include a compatibility function to work like
 Heimdal's krb5_get_init_creds_opt_set_pkinit() function.

 Hopefully, the test code that doesn't belong in kinit.c is
 obvious.

 r18929@cathode-dark-space:  coffman | 2006-12-07 10:01:20 -0500
 Remove extra "user_id" parameter.

 Add function which duplicates the Heimdal interface (if we can agree on
 what the matching attribute names should be).

 r18934@cathode-dark-space:  coffman | 2006-12-08 15:28:03 -0500
 Update to use the simplified interface for krb5_get_init_creds_opt_set_pa()

 Add code in kinit to process "-X" options as preauth options and pass
 them along.

 r18936@cathode-dark-space:  coffman | 2006-12-11 12:04:26 -0500
 Move prototypes for get_init_creds_opt_get_pa() and
 krb5_get_init_creds_opt_free_pa() into the
 preauth_plugin.h header rather than krb5.hin.

ticket: new
status: open
component: krb5-libs

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19127 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago This commit adds a thread safe MEMORY keytab implementation
Jeffrey Altman [Tue, 30 Jan 2007 17:21:56 +0000 (17:21 +0000)]
This commit adds a thread safe MEMORY keytab implementation
that is compatible with Heimdal 0.7.  Each successful resolve
returns a handle to a keytab and increases the internal
reference count.  Each close invalidates the handle and
decreases the reference count.  When the reference count hits
zero, the keytab is destroyed.

When a kt_destroy function is added, the automatic destruction
after close behavior will be removed.

ticket: 5411

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19126 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago Conditionalize ssize_t declaration
Jeffrey Altman [Tue, 30 Jan 2007 11:23:59 +0000 (11:23 +0000)]
Conditionalize ssize_t declaration

ticket: 5419

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19125 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago The fix for the command line race conditions
Jeffrey Altman [Tue, 30 Jan 2007 11:22:45 +0000 (11:22 +0000)]
The fix for the command line race conditions
broke the ability to cancel and restart the
Obtain New Credentials dialog

ticket: 5414

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19124 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMicrosoft Windows Visual Studio does not define ssize_t
Jeffrey Altman [Tue, 30 Jan 2007 11:17:35 +0000 (11:17 +0000)]
Microsoft Windows Visual Studio does not define ssize_t

Microsoft defines size_t but not ssize_t.
Provide a definition in win-mac.h that follows the style
used by Microsoft providing a preprocessor macro
SSIZE_T_DEFINED which can be used for conditional
definition and testing in case Microsoft decides to add
it later.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19123 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoKFW: 32-bit builds use the pismere krbv4w32.dll library
Jeffrey Altman [Tue, 30 Jan 2007 11:02:32 +0000 (11:02 +0000)]
KFW: 32-bit builds use the pismere krbv4w32.dll library

Only the 32-bit builds should use the pismere krbv4w32.dll
library.   Pismere does not provide 64-bit support.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19122 dc483132-0cff-0310-8789-dd5450dbe970

17 years agosort
Ken Raeburn [Sat, 27 Jan 2007 08:46:09 +0000 (08:46 +0000)]
sort

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19119 dc483132-0cff-0310-8789-dd5450dbe970

17 years agosort
Ken Raeburn [Sat, 27 Jan 2007 02:41:29 +0000 (02:41 +0000)]
sort

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19114 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM Bug Fixes
Jeffrey Altman [Wed, 24 Jan 2007 00:09:13 +0000 (00:09 +0000)]
NIM Bug Fixes

Document User Interface Callbacks

Fix a race condition when performing renewal actions
triggered by command line parameters.

When importing credentials, kickoff a renewal after
the credentials after the API: ccache is created.

Another fix for identity expiration states.  This one
fixes the behavior of the system tray icon.

ticket: new
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19110 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoideas for passing error msg info across kadm5 api
Ken Raeburn [Tue, 23 Jan 2007 23:15:36 +0000 (23:15 +0000)]
ideas for passing error msg info across kadm5 api

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19109 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoAdd const to variable to cleanup warnings
Ezra Peisach [Tue, 23 Jan 2007 20:44:06 +0000 (20:44 +0000)]
Add const to variable to cleanup warnings

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19108 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoInclude k5-int.h for krb5_cc_register() prototype. Remove defintion of
Ezra Peisach [Tue, 23 Jan 2007 20:40:58 +0000 (20:40 +0000)]
Include k5-int.h for krb5_cc_register() prototype. Remove defintion of
krb5_cc_dfl_ops - as k5-int.h defines.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19107 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoChange second argument to krb5_cc_register to take a const krb5_cc_ops *. Makes
Ezra Peisach [Tue, 23 Jan 2007 20:17:08 +0000 (20:17 +0000)]
Change second argument to krb5_cc_register to take a const krb5_cc_ops *. Makes
sense - when krb5_Xcc_ops currently defined as such - and assignment of value
assumes same.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19106 dc483132-0cff-0310-8789-dd5450dbe970

17 years agogss_krb5int_unseal_token_v3() - change bodysize argument to unsigned -
Ezra Peisach [Tue, 23 Jan 2007 02:57:32 +0000 (02:57 +0000)]
gss_krb5int_unseal_token_v3() - change bodysize argument to unsigned -
which is consistent with how it is invoked and used.  Cleans up two
signed/unsigned warnings.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19105 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokt_file.c memory leak on error in krb5_kt_resolve / krb5_kt_wresolve
Jeffrey Altman [Tue, 23 Jan 2007 01:18:17 +0000 (01:18 +0000)]
kt_file.c memory leak on error in krb5_kt_resolve / krb5_kt_wresolve

The krb5_kt_resolve and krb5_kt_wresolve functions leak the 'data'
memory allocation if the lock cannot be initialized.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19104 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoInclude fake-addrinfo.h before krb5.h to get proper structure declarations
Ezra Peisach [Mon, 22 Jan 2007 14:32:15 +0000 (14:32 +0000)]
Include fake-addrinfo.h before krb5.h to get proper structure declarations

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19103 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoOnly declare cc_fcc_entry if USE_KEYRING_CCACHE is not defined
Ezra Peisach [Sun, 21 Jan 2007 19:15:02 +0000 (19:15 +0000)]
Only declare cc_fcc_entry if USE_KEYRING_CCACHE is not defined

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19102 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago small fixes
Jeffrey Altman [Sat, 20 Jan 2007 22:21:13 +0000 (22:21 +0000)]
small fixes

ticket: 5408

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19101 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago Move files to their correct location
Jeffrey Altman [Sat, 20 Jan 2007 21:10:10 +0000 (21:10 +0000)]
Move files to their correct location

ticket: 5408

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19100 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoUse intermediate target to prevent parallel invocations of 'make includes' in lib...
Ken Raeburn [Sat, 20 Jan 2007 20:30:36 +0000 (20:30 +0000)]
Use intermediate target to prevent parallel invocations of 'make includes' in lib/krb4

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19099 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoUse move-if-changed and a timestamp file to avoid re-running commands and
Ken Raeburn [Sat, 20 Jan 2007 19:11:34 +0000 (19:11 +0000)]
Use move-if-changed and a timestamp file to avoid re-running commands and
updating krb5.h when there are no textual changes.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19098 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't copy headers to .../include
Ken Raeburn [Sat, 20 Jan 2007 18:47:32 +0000 (18:47 +0000)]
Don't copy headers to .../include

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19097 dc483132-0cff-0310-8789-dd5450dbe970

17 years agomake depend
Ken Raeburn [Sat, 20 Jan 2007 18:44:55 +0000 (18:44 +0000)]
make depend

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19096 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoBuild krb5.h using the _err.h files from the error_tables directory, not
Ken Raeburn [Sat, 20 Jan 2007 18:42:06 +0000 (18:42 +0000)]
Build krb5.h using the _err.h files from the error_tables directory, not
additional copies of them.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19095 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't explicitly include error table headers that are also in krb5.h
Ken Raeburn [Sat, 20 Jan 2007 18:32:03 +0000 (18:32 +0000)]
Don't explicitly include error table headers that are also in krb5.h

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19094 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMark as C source for Emacs mode selection
Ken Raeburn [Sat, 20 Jan 2007 18:06:27 +0000 (18:06 +0000)]
Mark as C source for Emacs mode selection

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19093 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMove osconf input file to include directory instead of keeping it in its own
Ken Raeburn [Sat, 20 Jan 2007 17:23:36 +0000 (17:23 +0000)]
Move osconf input file to include directory instead of keeping it in its own
subdirectory.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19092 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* Makefile.in (osconf.h): Use move-if-changed script
Ken Raeburn [Sat, 20 Jan 2007 17:05:02 +0000 (17:05 +0000)]
* Makefile.in (osconf.h): Use move-if-changed script

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19091 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* pre.in (MOVEIFCHANGED): New variable
Ken Raeburn [Sat, 20 Jan 2007 17:04:25 +0000 (17:04 +0000)]
* pre.in (MOVEIFCHANGED): New variable

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19090 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoRemove unused file
Ken Raeburn [Sat, 20 Jan 2007 16:56:32 +0000 (16:56 +0000)]
Remove unused file

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19089 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoRemove varargs.h paths, always use stdarg.h
Ken Raeburn [Sat, 20 Jan 2007 16:54:37 +0000 (16:54 +0000)]
Remove varargs.h paths, always use stdarg.h

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19088 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoInclude stdarg.h, and don't conditionalize declarations of functions with
Ken Raeburn [Sat, 20 Jan 2007 12:52:34 +0000 (12:52 +0000)]
Include stdarg.h, and don't conditionalize declarations of functions with
va_list arguments.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19087 dc483132-0cff-0310-8789-dd5450dbe970

17 years agomake depend
Ken Raeburn [Sat, 20 Jan 2007 12:33:00 +0000 (12:33 +0000)]
make depend

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19086 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoHandle function names immediately preceded by "*", like "*strdup"
Ken Raeburn [Sat, 20 Jan 2007 12:13:15 +0000 (12:13 +0000)]
Handle function names immediately preceded by "*", like "*strdup"
in k5-int.h+krb5.h.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19085 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMark as C source for Emacs mode selection
Ken Raeburn [Sat, 20 Jan 2007 12:07:41 +0000 (12:07 +0000)]
Mark as C source for Emacs mode selection

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19084 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDon't define KRB5_PRIVATE, since it's not tested any more
Ken Raeburn [Sat, 20 Jan 2007 12:05:22 +0000 (12:05 +0000)]
Don't define KRB5_PRIVATE, since it's not tested any more

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19083 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMove all KRB5_PRIVATE declarations from krb5.h(in) to k5-int.h, and drop the
Ken Raeburn [Sat, 20 Jan 2007 11:33:36 +0000 (11:33 +0000)]
Move all KRB5_PRIVATE declarations from krb5.h(in) to k5-int.h, and drop the
KRB5_PRIVATE preprocessor test.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19082 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* Makefile.in (private-and-public-decls): New target file, created from
Ken Raeburn [Sat, 20 Jan 2007 11:29:07 +0000 (11:29 +0000)]
* Makefile.in (private-and-public-decls): New target file, created from
k5-int.h and krb5.h.
(verify-calling-conventions-krb5): Use it instead of krb5.h.
(clean-unix): Get rid of it.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19081 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoInclude k5-int.h instead of krb5.h when 'private' functions are needed
Ken Raeburn [Sat, 20 Jan 2007 11:02:58 +0000 (11:02 +0000)]
Include k5-int.h instead of krb5.h when 'private' functions are needed

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19080 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* pre.in (.et.h, .et.c): Use temporary files in the current directory with
Ken Raeburn [Sat, 20 Jan 2007 10:59:19 +0000 (10:59 +0000)]
* pre.in (.et.h, .et.c): Use temporary files in the current directory with
fixed names, instead of incorporating the shell pid.
* post.in (clean-unix): Delete the temporary files.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19079 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoKFW MSI installer corrections
Jeffrey Altman [Sat, 20 Jan 2007 05:50:41 +0000 (05:50 +0000)]
KFW MSI installer corrections

Add missing registry keys and values for
Network Identity Manager plug-ins.

Correct short names.

ticket: new
tags: pullup
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19078 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM - Context sensitive system tray menu and more
Jeffrey Altman [Sat, 20 Jan 2007 05:42:30 +0000 (05:42 +0000)]
NIM - Context sensitive system tray menu and more

  Adds context sensitive menus for renew and destroy
to the system tray menu.  Select either all identities
    or one of the identities with credentials.

Increases the API to 1.1.9 and adds a new interactive
callback mechanism.

ticket: new
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19077 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDelete t_cccursor on cleanup. Add t_cc and t_cccursor to dependencies
Ken Raeburn [Sat, 20 Jan 2007 04:49:02 +0000 (04:49 +0000)]
Delete t_cccursor on cleanup.  Add t_cc and t_cccursor to dependencies

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19076 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoAdd t_cccursor to UNIX tests
Ken Raeburn [Sat, 20 Jan 2007 04:45:20 +0000 (04:45 +0000)]
Add t_cccursor to UNIX tests

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19075 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoAdded KDC timesyncing support to the CCAPI ccache backend for CCAPI v5 and
Alexandra Ellwood [Thu, 18 Jan 2007 23:24:13 +0000 (23:24 +0000)]
Added KDC timesyncing support to the CCAPI ccache backend for CCAPI v5 and
later.  v5 is the first version of the CCAPI to support the kdc timesyncing
API.

ticket: 5403

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19071 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago In addition to setting the kpasswd port after
Jeffrey Altman [Thu, 18 Jan 2007 15:28:07 +0000 (15:28 +0000)]
In addition to setting the kpasswd port after
searching for kadmind host addresses we must also
set the socket type according to the request.

ticket: 5393
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19070 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM Kerberos 5 Provider corrections
Jeffrey Altman [Thu, 18 Jan 2007 12:43:58 +0000 (12:43 +0000)]
NIM Kerberos 5 Provider corrections

When validating a Kerberos 5 principal name, the request
to the KDC should not request forwardable, renewable, or
proxiable options as these may be blocked by policy and
will result in the return of an error.

Always treat the Kerberos 5 principal name as valid
unless the KDC returns an error that clearly indicates that
the principal name does not exist.

Use a MEMORY: ccache for temporary storage instead of an
API: ccache.

Initialize pointer values with NULL instead of 0.

ticket: new
tags: pullup
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19069 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM Correct Visual Identity Expiration Status
Jeffrey Altman [Thu, 18 Jan 2007 12:39:06 +0000 (12:39 +0000)]
NIM Correct Visual Identity Expiration Status

The visual status for the identities in NIM 1.1 was based
upon the highest alert status of any credential that was
associated with the identity.  The correct behavior that is
now implemented is that the identity status should be based
solely upon the expiration state of the credentials obtained
by the identity provider.

For example, the Kerberos v5 identity provider will based the
identity expiration status on the initial TGT.  Service tickets
with short lifetimes that expire do not prevent the acquisition
of additional service tickets.  Therefore, the identity should
not be listed as expired.

ticket: new
tags: pullup
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19068 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM Kerberos v4 configuration dialog
Jeffrey Altman [Thu, 18 Jan 2007 12:33:47 +0000 (12:33 +0000)]
NIM Kerberos v4 configuration dialog

The Kerberos v4 options for individual identities
was never wired.  The controls were visible but they
did not do anything.  Implement them now for NIM 1.2.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19067 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoNIM string tables
Jeffrey Altman [Thu, 18 Jan 2007 12:31:23 +0000 (12:31 +0000)]
NIM string tables

Update the string tables for NIM so that they are consistent.
Always use "Kerberos v5" or "Kerberos v4".  Refer to credentials
instead of tickets.  Do not abbreviate "Network Identity Manager".
Etc.

ticket: new
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19066 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago sendto_kdc.c: use of a variable index into a dynamically
Jeffrey Altman [Thu, 18 Jan 2007 11:35:33 +0000 (11:35 +0000)]
   sendto_kdc.c: use of a variable index into a dynamically
allocated array to determine the sizeof() an object makes
it unclear what type of object is involved.  It also requires
a runtime check instead of a compile time replacement.
Not to mention that it could lead to the evaluation of an
uninitialized variable as was done in this case.  Replace
sizeof(array index variable) with sizeof(type).

memset() the correct data structure.

ticket: 5394
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19065 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* t_cccursor.c (do_chk): Reverse sense of test of do_chk_one return value
Ken Raeburn [Wed, 17 Jan 2007 00:01:56 +0000 (00:01 +0000)]
* t_cccursor.c (do_chk): Reverse sense of test of do_chk_one return value

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19064 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoFix typo in checked-in version
Ken Raeburn [Tue, 16 Jan 2007 23:29:46 +0000 (23:29 +0000)]
Fix typo in checked-in version

ticket: 5349

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19063 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago This commit adds two new functions, krb5_server_decrypt_ticket_keyblock
Jeffrey Altman [Tue, 16 Jan 2007 04:18:02 +0000 (04:18 +0000)]
  This commit adds two new functions, krb5_server_decrypt_ticket_keyblock
  (private) and krb5_server_decrypt_ticket_keytab (public).  These
  functions take a krb5_ticket as input and decrypt it using the provided
  key data.  The public function is useful for higher level application
  protocols such a TLS-KRB5 and AFS RX-KRB5 which exchange a service
  but do not use the AP-REQ/AP-REP messages.

  This commit also adds new functionality to kvno which permits kvno
  when provided a keytab as input to verify whether or not the keytab
  contains a key that can successfully decrypt the obtains service ticket.

ticket: 5349
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19062 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokfw wix installer - memory overwrite error
Jeffrey Altman [Sun, 14 Jan 2007 08:04:46 +0000 (08:04 +0000)]
kfw wix installer - memory overwrite error

   The custom handler allocates a buffer that is smaller
than is required to hold the input.  Allocate the correct
  sized buffer.

ticket: new
tags: pullup
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19061 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoUpdate to KFW NSIS installer
Jeffrey Altman [Sat, 13 Jan 2007 17:08:28 +0000 (17:08 +0000)]
Update to KFW NSIS installer

Update copyright date, samples directory, and shortcut
to NetIDMgr documentation

ticket: new
component: windows
tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19060 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoupdated Windows README
Jeffrey Altman [Sat, 13 Jan 2007 17:04:44 +0000 (17:04 +0000)]
updated Windows README

Revise the readme text for Windows.  Remove references
to old beta SDKs.

ticket: new
tags: pullup
component: windows

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19059 dc483132-0cff-0310-8789-dd5450dbe970

17 years agokrb5-1.7-prerelease
Tom Yu [Thu, 11 Jan 2007 00:13:46 +0000 (00:13 +0000)]
krb5-1.7-prerelease

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19056 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMITKRB5-SA-2006-003: mechglue argument handling too lax
Tom Yu [Tue, 9 Jan 2007 19:45:25 +0000 (19:45 +0000)]
MITKRB5-SA-2006-003: mechglue argument handling too lax

Fix mechglue argument checks so that output pointers are always
initialized regardless of whether the other arguments fail to validate
for some reason.  This avoids freeing of uninitialized pointers.

Initialize the gss_buffer_descs in ovsec_kadmd.c.

ticket: new
target_version: 1.6
tags: pullup
component: krb5-libs

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19043 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoMITKRB5-SA-2006-002: svctcp_destroy() can call uninitialized function pointer
Tom Yu [Tue, 9 Jan 2007 19:45:10 +0000 (19:45 +0000)]
MITKRB5-SA-2006-002: svctcp_destroy() can call uninitialized function pointer

Explicitly null out xprt->xp_auth when AUTH_GSSAPI is being used, so
that svctcp_destroy() will not call through an uninitialized function
pointer after code in svc_auth_gssapi.c has destroyed expired state
structures.  We can't unconditionally null it because the RPCSEC_GSS
implementation needs it to retrieve state.

ticket: new
target_version: 1.6
tags: pullup
component: krb5-libs

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19042 dc483132-0cff-0310-8789-dd5450dbe970

17 years agocreate KDC database directory
Ken Raeburn [Mon, 8 Jan 2007 22:51:23 +0000 (22:51 +0000)]
create KDC database directory

The default directory for the KDC database files doesn't exist after a
normal "make install".  Reported by Jeff Blaine.

* Makefile.in (INSTALLMKDIRS): Add var (really localstatedir) and var/krb5kdc
to directories to be created at "make install" time, even though it'll be
empty on most machines (since most probably aren't KDCs).

ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19039 dc483132-0cff-0310-8789-dd5450dbe970

17 years agocrash creating db2 database in non-existent directory
Ken Raeburn [Mon, 8 Jan 2007 22:41:01 +0000 (22:41 +0000)]
crash creating db2 database in non-existent directory

* kdb_db2.c (krb5_db2_db_create): If the creation of the first database file
fails, return the error, instead of attempting to create the second (and using
a null pointer as an input string in formatting a filename).

Reported by Jeff Blaine.

ticket: new

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19038 dc483132-0cff-0310-8789-dd5450dbe970

17 years ago* lib.in (clean-libs): Delete darwin.exports and hpux10.exports.
Ken Raeburn [Sat, 6 Jan 2007 08:30:46 +0000 (08:30 +0000)]
* lib.in (clean-libs): Delete darwin.exports and hpux10.exports.
* libnover.in (clean-libs): Use $(DYNOBJEXT) instead of $(SHLIBEXT) for
extension on object to delete.  Also delete darwin.exports and hpux10.exports.

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19037 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDocument what the kadmind ACL is for
Russ Allbery [Fri, 5 Jan 2007 21:42:38 +0000 (21:42 +0000)]
Document what the kadmind ACL is for

Add a sentence documenting the purpose of the kadmind ACL to the node
explaining how to create it.

Ticket: 5279
Component: krb5-doc
Version_Reported: 1.5.1
Target_Version: 1.6
Tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19036 dc483132-0cff-0310-8789-dd5450dbe970

17 years agoDocument KDC behavior without stash file
Russ Allbery [Thu, 4 Jan 2007 22:32:41 +0000 (22:32 +0000)]
Document KDC behavior without stash file

After the discussion of the optional stash file, document the effects of
not creating a stash file.

ticket: new
Component: krb5-doc
Version_Reported: 1.5.1
Target_Version: 1.6
Tags: pullup

git-svn-id: svn://anonsvn.mit.edu/krb5/trunk@19035 dc483132-0cff-0310-8789-dd5450dbe970