Redhat maintain libpwdb. They release it in .rpm form and I guess you'll need to submit your bug report to them. [Redhat folk do read the pam-list, so I'm Cc:ing your email there.] Thanks for the suggested patch! Cheers Andrew Koni wrote: > > Dear Morgan, > > Sorry, I guess this got sent to old email address. I found this one on > the Linux PAM page a kernel.org. The attached is about memory leaks in > linux libpwdb-0.61. I did not find a new version yet, or a web page for > the library. If you can spare a moment, please tell me where I can find > the development site for libpwdb and any comments you have on the > solutions I used below, or, point me toward the active developers for > libpwdb. > > Cheers, > Koni > > -- > mhw6@cornell.edu > Koni (Mark Wright) > Solanaceae Genome Network 250 Emerson Hall - Cornell University > Strategic Forecasting 242 Langmuir Laboratory > Lightlink Internet http://www.lightlink.com/ > > "If I'm right 90% of the time, why quibble about the other 3%?" > > ------------------------------------------------------------------------ > > Subject: failure notice > Date: 5 Apr 2001 00:13:29 -0000 > From: MAILER-DAEMON@askatu.strategicforecasting.com > To: mhw6@cornell.edu (Mark Hamilton Wright) > > Hi. This is the qmail-send program at askatu.strategicforecasting.com. > I'm afraid I wasn't able to deliver your message to the following addresses. > This is a permanent error; I've given up. Sorry it didn't work out. > > <morgan@parc.power.net>: > Sorry, I wasn't able to establish an SMTP connection. (#4.4.1) > I'm not going to try again; this message has been in the queue too long. > > --- Below this line is a copy of the message. > > Return-Path: <mhw6@cornell.edu> > Received: (qmail 13321 invoked from network); 28 Mar 2001 23:13:21 -0000 > Received: from askatu.corsim.bogus.net (HELO txakoli) (root@192.168.2.6) > by askatu.corsim.bogus.net with SMTP; 28 Mar 2001 23:13:21 -0000 > Subject: Memory leaks in libpwdb-0.61 > From: Koni <mhw6@cornell.edu> > To: morgan@parc.power.net > Cc: gafton@sorosis.ro > Content-Type: text/plain > X-Mailer: Evolution (0.9/+cvs.2001.03.16.09.00 - Preview Release) > Date: 28 Mar 2001 18:12:44 -0500 > Mime-Version: 1.0 > > Hello, > > My name is Koni. In the process of developing some soon to be open > source software for a local ISP here in Ithaca, NY, which uses PAM to > authenticate users, I've tracked down some memory leaks in my program > which led to the PAM libraries, and ultimately to "libpwdb". At the end > of this email are diffs for pwdb_module.c and pwdb_start.c > > This is for pwdb version 0.61, as distrubted in RedHat-6.2. I didn't > find the project homepage for possibly more recent versions, so I > apologize if this is already known and fixed. > > In pwdb_module.c, two dynamically allocated/reallocated arrays were not > being freed. They are arrays of pointers, and the pointers are being > freed but not the array itself. I added the appropriate free() calls. > > In pwdb_start.c, something was calling _pwdb_delete_string(entry->name) > to free memory, but entry->name was set to NULL explicitly before this > call, which means _pwdb_delete_string will not free the string (nor will > it zero the memory before freeing!). I deleted the line that set > entry->name to NULL. > > I found no further memory leaks in PAM or pwdb and the core size of my > server is no longer growing with each authenticated connection. I'm > using shadow passwords on my system at the moment. Thus, I conclude > that, for the path of execution for authenticating a user using shadow > passwords, libpwdb with these fixes has no futher memory leaks. > > Thank you very very much for developing this software. PAM and pwdb have > greatly simplified authentication for us, as well as allowed an > impressive amount of options in handling user sessions and > authentication methods. > > Cheers, > Koni > > Diffs: > > for file: libpwdb/pwdb/pwdb_module.c > 309a310 > > if (pwdb_policy) free(pwdb_policy); > 315a317 > > if (pwdb_group_policy) free(pwdb_group_policy); > > for file: libpwdb/pwdb/pwdb_start.c > 77d76 > < entry->name = NULL; > > -- > mhw6@cornell.edu > Koni (Mark Wright) > Solanaceae Genome Network 250 Emerson Hall - Cornell University > Strategic Forecasting 242 Langmuir Laboratory > Lightlink Internet http://www.lightlink.com/ > > "If I'm right 90% of the time, why quibble about the other 3%?"
