+++ This bug was initially created as a clone of Bug #120168 +++

From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i586; en-US; rv:1.4.2)
Gecko/20040308

Description of problem:
During code review I could a number of serious programming bugs:

If there was a failure reading config data, the memory buffer was
taken to be valid.

On a read failure of /dev/urandom, buffer was taken to be valid. The
function could not indicate that a failure occurred or even retry if
errno == EINTR.

Upon a read failure, salt may be mostly 0's.

During unlock, a structure was being passed by value to a variadic
function.

length returned by read was being caught in an unsigned variable which
would later cause write to output 4294967295 bytes of data.

flcose was being passed a NULL pointer in many places.



Version-Release number of selected component (if applicable):
libuser-0.51.7-7.1.1

How reproducible:
Always

Steps to Reproduce:
Found during code review
    

Additional info:

I will attach a patch that fixes all of these issues. I did not review
prior versions to see if they are susceptible to all these problems.

-- Additional comment from linux_4ever@yahoo.com on 2004-04-06 12:10 EST --
Created an attachment (id=99149)
Patch to fix all the problems listed above.

Please apply before fedora core 2 final.




This issue should also affect RHEL2.1

Please see the parent bug for more information.