[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: KDE changing system time



On Wed, 14 Aug 2002 23:06, Richard Ibbotson wrote:
> > Yes, the OxLug people are obviously on crack. This is the first I
> > have ever heard of it.
>
> Probably.  But, I've hardly ever seen a computer keep time when I used
> KDE on any distribution. So, I wouldn't understand what this might be
> about.
>
> I've built a lot of computers.  Also installed quite a few networks.
> Most of them GNU/Linux.

You want to have a flame-war on two lists instead of one?

I'll paste in one of my messages from the OxLug list below.

Also I ask again, which Debian developer told you that they were aware of 
such a problem?

Richard, your entire arguement is based on assertions that you are declining 
to prove.  You are quoting sources that you are refusing to identify.  You 
have quoted a professor as supporting your case without saying what he said 
or providing a reference.  You have refused to perform the tests I suggested 
to track down the problem, and ignored the evidence I have provided of 
systems running KDE without any attempt to change the time.

I am being forced to concluded that one of the following two explanations 
applies:
1)  KDE works fine but you broke your KDE systems through your own 
incompetance.
2)  You are making this all up entirely.

If you want anyone to believe otherwise then provide some evidence.



On Wed, 14 Aug 2002 23:06, Richard Ibbotson wrote:
> No idea.  Seen this repeated on Mandrake, Debian and SuSE over several
> releases.  The Debian KDE developers tell me that they are aware of
> it but didn't explain what it was about.

I did "su -" from a user domain account and then tried to use "date" to 
change the system clock, and received the following SE Linux messages in my 
log:
Aug 14 19:22:08 lyta kernel: avc:  denied  { sys_time } for  pid=29111 
comm=date capability=25 scontext=rjc:user_r:user_t tcontext=rjc:user_r:user_t 
tclass=capability
Aug 14 19:22:08 lyta kernel: avc:  denied  { sys_time } for  pid=29111 
comm=date capability=25 scontext=rjc:user_r:user_t tcontext=rjc:user_r:user_t 
tclass=capability

I then grepped for sys_time in my /var/log/kern.log* files and found no 
rejections (logs go from Jun 3).

Then I checked what domains are allowed sys_time capability, hwclock_t, 
apmd_t, ntpd_t, sysadm_t, init_t, and initrc_t.

These domains are only permitted in the system_r and sysadm_r roles, so 
barring a critical security flaw in the kdm code I wrote there is no way that 
buggy KDE code tried to change the time on my system in the last 6 weeks.  So 
if you know what part of KDE does it then it shouldn't be difficult to turn 
it off as I must have done.

Which Debian KDE developers told you about this?  I'd like to track this one 
down.

-- 
I do not get viruses because I do not use MS software.
If you use Outlook then please do not put my email address in your
address-book so that WHEN you get a virus it won't use my address in the
>From field.



Reply to: