Bug 638 – Hang due to invalid timestamp entry when using "Defaults timestamp_timeout=0"

Bug 638 - Hang due to invalid timestamp entry when using "Defaults timestamp_timeout=0"


Summary:	Hang due to invalid timestamp entry when using "Defaults timestamp_timeout=0"

Status:	RESOLVED FIXED

Product:	Sudo
Classification:	Unclassified
Component:	Sudoers
Version:	1.8.10
Hardware:	PC Linux

Importance:	low high
Assigned To:	Todd C. Miller

URL:

Depends on:
Blocks:
	Show dependency tree / graph

Reported:	2014-03-17 01:02 MDT by Evangelos Foutras
Modified:	2014-03-17 10:30 MDT (History)
CC List:	0 users

See Also:

Attachments
Avoid writing empty timestamp records when timestamp_timeout is 0 (1.02 KB, patch) 2014-03-17 07:14 MDT, Todd C. Miller	Details \| Diff
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.

Description Evangelos Foutras 2014-03-17 01:02:12 MDT

With "Defaults timestamp_timeout=0" specified in /etc/sudoers, sudo will create what appears to be an invalid timestamp entry. On subsequent invocations it will hang (after the correct password is typed in) and the following debug message is repeatedly logged (when debugging is configured in /etc/sudo.conf):

"wrong sized record, got 0, expected 40 @ ts_find_record() ./timestamp.c:125"

Tested this using sudo 1.8.10p1; didn't check 1.8.10 but I'm guessing it has the same problem. According to the user that originally reported this on the Arch Linux bug tracker, 1.8.9.p5 worked fine with the same settings.¹

Please let me know if you can reproduce the issue or if I need to provide more information.

¹ https://bugs.archlinux.org/task/39472

Comment 1 Todd C. Miller 2014-03-17 07:14:55 MDT

Created attachment 399 [details]
Avoid writing empty timestamp records when timestamp_timeout is 0

The attached patch is what I'm going to ship as part of 1.8.10p2

Comment 2 Evangelos Foutras 2014-03-17 10:22:47 MDT

(In reply to Todd C Miller from comment #1)
> Created attachment 399 [details]
> Avoid writing empty timestamp records when timestamp_timeout is 0
> 
> The attached patch is what I'm going to ship as part of 1.8.10p2

Tested 1.8.10p2 and can confirm that it fixes this issue.

Thanks!