| Summary: | [RFE] Add support to ksu for k5login_directory | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 7 | Reporter: | Matthew Richardson <m.richardson> |
| Component: | krb5 | Assignee: | Robbie Harwood <rharwood> |
| Status: | CLOSED UPSTREAM | QA Contact: | Karel Srot <ksrot> |
| Severity: | unspecified | Docs Contact: | |
| Priority: | low | ||
| Version: | 7.2 | CC: | dpal, m.richardson, pkis |
| Target Milestone: | rc | Keywords: | FutureFeature |
| Target Release: | --- | ||
| Hardware: | Unspecified | ||
| OS: | Linux | ||
| URL: | http://krbdev.mit.edu/rt/Ticket/Display.html?id=8397 | ||
| Whiteboard: | |||
| Fixed In Version: | Doc Type: | Bug Fix | |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2016-11-17 19:13:06 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
|
Description
Matthew Richardson
2016-04-25 09:08:24 UTC
Can you please provide strace logs which include how strace was invoked? ksu is a setuid program and I would like to verify the interaction between it and strace. A little wary of posting a whole strace of a ksu process, in case any csecrets end up in it, but I'm running it as follows (using principal mattr@REALM):
sudo strace -fF -o /tmp/ksu.out sudo -u mattr ksu root -n mattr
Looking at all stats, reads etc in strace the only apparent ones of relevance are:
3872 stat("/root/.k5login", 0x7ffcf3955b40) = -1 ENOENT (No such file or directory)
3872 stat("/root/.k5users", 0x7ffcf3955b40) = -1 ENOENT (No such file or directory)
This is the same regardless of the state of k5login_directory option.
I've had a quick look at the source in the src rpm, and it does look like this is hard-coded for ksu.
lib/krb5/os/localauth_k5login.c provides a 'get_k5login_filename' method which does the 'try k5login_directory/look in homedir' logic.
However this doesn't seem to be called anywhere in clients/ksu/ with ksu.h having:
#define KRB5_LOGIN_NAME ".k5login"
#define KRB5_USERS_NAME ".k5users"
Which is then used through the rest of ksu's code.
I've had a quick look at the latest upstream code from MIT and ksu.h looks pretty identical, so I suspect this is really an upstream bug... don't know if that's something I need to open separately with the krb5 project, or if this bug report can be 'redirected' to them?
(In reply to Matthew Richardson from comment #4) > A little wary of posting a whole strace of a ksu process, in case any > csecrets end up in it, but I'm running it as follows (using principal > mattr@REALM): Sorry, assumed you had a realm set up for testing not something in production. The information you've provided should be (more than) sufficient; thanks! Created upstream ticket to track this. |