Sync cifs-utils 2:6.7-1 (main) from Debian unstable (main)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
cifs-utils (Ubuntu) |
Fix Released
|
Wishlist
|
Unassigned |
Bug Description
Please sync cifs-utils 2:6.7-1 (main) from Debian unstable (main)
Explanation of the Ubuntu delta and why it can be dropped:
* Merge from Debian unstable (LP: #1660372). Remaining changes:
- debian/
/
TL;DR the patch in our delta is to prevent a harmless error message from
popen() leaking through to the console in the case systemd-
does not exist in /bin, in which case the code falls back to using getpass():
ubuntu@61-b1:~$ mount /ds216/downloads
sh: 1: systemd-
Password for andreas@
ubuntu@61-b1:~$
It just doesn't seem worth carrying a delta for that. It's best to file
an upstream bug to have it properly fixed, which I did:
https:/
Longer story:
When that patch was added by Ubuntu, the upstream code was like this:
"""
is_
&& (lstat(
&& (a.st_dev != b.st_dev);
if (is_systemd_
(...)
if (asprintf(&cmd, "/bin/systemd-
free (cmd);
}
"""
If popen() can't find the binary, the code falls back to getpass(), but popen
prints an "ugly" error message before:
ubuntu@61-b1:~$ mount /ds216/downloads
sh: 1: systemd-
Password for andreas@
ubuntu@61-b1:~$
The ubuntu patch had the effect of not even trying to popen the missing binary,
thus avoiding the error message.
Upstream later changed the code slightly (see bug
https:/
"""
is_
&& (lstat(
&& (a.st_dev != b.st_dev);
(...)
if (asprintf(&cmd, "systemd-
free (cmd);
}
"""
That now searches for systemd-
the error message is still printed, and getpass() is again used as a fallback.
The Ubuntu patch is still useful, but not complete now: it should search for
the binary in $PATH to match what popen does.
Perhaps the easiest solution is to just hide the popen error, since it's handled anyway?
- if (asprintf(&cmd, "systemd-
+ if (asprintf(&cmd, "systemd-
I filed an upstream bug with that suggestion:
https:/
Changelog entries since current artful version 2:6.6-5ubuntu1:
cifs-utils (2:6.7-1) unstable; urgency=medium
* New upstream release
- a6662ed: cifs.upcall: switch group IDs when handling an upcall
- 109c189: cifs.upcall: drop capabilities early in program
- ed97e4e: cifs.upcall: allow scraping of KRB5CCNAME out of initiating
task's /proc/<pid>/environ file (Closes: #854607)
- ec387af4: cifs.upcall: trim even more capabilities
- 2dcecd2: cifs.upcall: unset $KRB5CCNAME when creating new credcache from
keytab
- 9eaa21e: cifs.upcall: don't do env scraping when uid is 0
- 69949ba: cifs.upcall: use a MEMORY: ccache when instantiating from a keytab
- and more commits...
-- Mathieu Parent <email address hidden> Wed, 08 Mar 2017 22:21:54 +0100
Changed in cifs-utils (Ubuntu): | |
importance: | Undecided → Wishlist |
This bug was fixed in the package cifs-utils - 2:6.7-1
---------------
cifs-utils (2:6.7-1) unstable; urgency=medium
* New upstream release
- a6662ed: cifs.upcall: switch group IDs when handling an upcall
- 109c189: cifs.upcall: drop capabilities early in program
- ed97e4e: cifs.upcall: allow scraping of KRB5CCNAME out of initiating
task's /proc/<pid>/environ file (Closes: #854607)
- ec387af4: cifs.upcall: trim even more capabilities
- 2dcecd2: cifs.upcall: unset $KRB5CCNAME when creating new credcache from
keytab
- 9eaa21e: cifs.upcall: don't do env scraping when uid is 0
- 69949ba: cifs.upcall: use a MEMORY: ccache when instantiating from a keytab
- and more commits...
-- Mathieu Parent <email address hidden> Wed, 08 Mar 2017 22:21:54 +0100