Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > linux.debian.security > #5068 > unrolled thread

Re: Userdata stays in RAM after Logout and Relogin

Started byherve <herve@couvelard.com>
First post2016-11-04 17:40 +0100
Last post2016-11-06 10:40 +0100
Articles 6 — 5 participants

Back to article view | Back to linux.debian.security

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.


Contents

  Re: Userdata stays in RAM after Logout and Relogin herve <herve@couvelard.com> - 2016-11-04 17:40 +0100
    RE: Userdata stays in RAM after Logout and Relogin "Radzykewycz, T (Radzy)" <radzy@windriver.com> - 2016-11-04 20:30 +0100
      Re: Userdata stays in RAM after Logout and Relogin Felix Knecht <debian@felixknecht.de> - 2016-11-04 23:40 +0100
      Re: Userdata stays in RAM after Logout and Relogin "H. N." <deb_mailinglist@niemeczek.at> - 2016-11-05 01:20 +0100
        Re: Userdata stays in RAM after Logout and Relogin Gunnar Wolf <gwolf@debian.org> - 2016-11-06 01:00 +0100
          Re: Userdata stays in RAM after Logout and Relogin "H. N." <deb_mailinglist@niemeczek.at> - 2016-11-06 10:40 +0100

#5068 — Re: Userdata stays in RAM after Logout and Relogin

Fromherve <herve@couvelard.com>
Date2016-11-04 17:40 +0100
SubjectRe: Userdata stays in RAM after Logout and Relogin
Message-ID<szPH4-4eu-35@gated-at.bofh.it>
On 04/11/2016 17:01, H. N. wrote:
> Hello,
>
> Sometimes it occurs, that the background Image of my desktop of is
> somehow corrupted. Yesterday I noteced, that after logout and re-login
> with another account on my pc the backgroundimage got corrupted again
> (which is annoying, but occurs very rearely).
> At the backgroundimage I could see lots of pixels in all colours, but
> also parts of logos and Images used im Programs (GeoGebra, Darktable) as
> well as 3 nearly complete frames of at least 3 Videos I watched on
> Debians Chromium on Youtube, Fragments of IntelliJ's IDEA-Logo, some
> writing (several times 'Google', as well as other words and once the
> Time-display at the top of the Desktop) and other logos.
>
>
> So if this was been a public PC another user could find out easyly, what
> I had done at the PC, possibly even Passwords and really sensitive
> information.
> If another User knew about this problem he might be able to find out a
> lot more than, just that by reading whats written in the RAM.
>
> I wonder that the RAM used by the user isn't cleared after logout.
>
> My System:
> Debian 8, Gnome Desktop currently used.
> No major system modifications done
>
>
> I MIGHT NOT BE IN SECURITY MAILING LIST (I DON'T KNOW, HAD A FEW
> PROBLEMS HERE), SO PLEASE SEND RESPONSE-EMAILS DIRECTLY TO
> deb_mailinglist@niemeczek.at TOO.
>
>
>
>
Ram or graphic ram ?
could be a bug in the video driver ?

hervé

[toc] | [next] | [standalone]


#5070

From"Radzykewycz, T (Radzy)" <radzy@windriver.com>
Date2016-11-04 20:30 +0100
Message-ID<szSlA-5Uy-1@gated-at.bofh.it>
In reply to#5068
________________________________________
> From: herve [herve@couvelard.com]
> Sent: Friday, November 04, 2016 9:26 AM
> To: debian-security@lists.debian.org
> Subject: Re: Userdata stays in RAM after Logout and Relogin
> 
> On 04/11/2016 17:01, H. N. wrote:
> > Hello,
> >
> > Sometimes it occurs, that the background Image of my desktop of is
> > somehow corrupted.
<snip>
> > I wonder that the RAM used by the user isn't cleared after logout.
> >
> > My System:
> > Debian 8, Gnome Desktop currently used.
> > No major system modifications done
> >
> > I MIGHT NOT BE IN SECURITY MAILING LIST (I DON'T KNOW, HAD A FEW
> > PROBLEMS HERE), SO PLEASE SEND RESPONSE-EMAILS DIRECTLY TO
> > deb_mailinglist@niemeczek.at TOO.
> >
> Ram or graphic ram ?
> could be a bug in the video driver ?
> 
> hervé

That's what I assumed when I read it: video driver bug.
It might also be a hardware issue, where graphic RAM is not
cleared on reset.  Or given the nature of video devices, perhaps a
combination of the two: a driver bug that expects certain generic
behavior combined with a device that expects a custom, proprietary,
vendor-provided driver to work correctly.

I have no idea whether H.N.'s system uses it, but I guess that
the USB graphic devices (DisplayLink et.al.) are probably prone
to this kind of issue, both on the displays that they connect to
and to the other displays on the primary graphics device.

Enjoy!

-- radzy

[toc] | [prev] | [next] | [standalone]


#5071

FromFelix Knecht <debian@felixknecht.de>
Date2016-11-04 23:40 +0100
Message-ID<szVjr-7M5-13@gated-at.bofh.it>
In reply to#5070
On 11/04/2016 05:46 PM, Radzykewycz, T (Radzy) wrote:
> That's what I assumed when I read it: video driver bug.

Me too. I saw this even after a reboot once on a second monitor. With
the binary nvidia drivers and KDE.
And I remember reading about how you can read the video memory after a
reboot. Can't find any details though.

f

[toc] | [prev] | [next] | [standalone]


#5072

From"H. N." <deb_mailinglist@niemeczek.at>
Date2016-11-05 01:20 +0100
Message-ID<szWSd-sO-1@gated-at.bofh.it>
In reply to#5070

[Multipart message — attachments visible in raw view] — view raw

On 2016-11-04 17:46, Radzykewycz, T (Radzy) wrote:
> ________________________________________
>> From: herve [herve@couvelard.com]
>> Sent: Friday, November 04, 2016 9:26 AM
>> To: debian-security@lists.debian.org
>> Subject: Re: Userdata stays in RAM after Logout and Relogin
>>
>> On 04/11/2016 17:01, H. N. wrote:
>>> Hello,
>>>
>>> Sometimes it occurs, that the background Image of my desktop of is
>>> somehow corrupted.
> <snip>
>>> I wonder that the RAM used by the user isn't cleared after logout.
>>>
>>> My System:
>>> Debian 8, Gnome Desktop currently used.
>>> No major system modifications done
>>>
>>> I MIGHT NOT BE IN SECURITY MAILING LIST (I DON'T KNOW, HAD A FEW
>>> PROBLEMS HERE), SO PLEASE SEND RESPONSE-EMAILS DIRECTLY TO
>>> deb_mailinglist@niemeczek.at TOO.
>>>
>> Ram or graphic ram ?
>> could be a bug in the video driver ?
>>
>> hervé
> That's what I assumed when I read it: video driver bug.
> It might also be a hardware issue, where graphic RAM is not
> cleared on reset.  Or given the nature of video devices, perhaps a
> combination of the two: a driver bug that expects certain generic
> behavior combined with a device that expects a custom, proprietary,
> vendor-provided driver to work correctly.
>
> I have no idea whether H.N.'s system uses it, but I guess that
> the USB graphic devices (DisplayLink et.al.) are probably prone
> to this kind of issue, both on the displays that they connect to
> and to the other displays on the primary graphics device.
>
> Enjoy!
>
> -- radzy
>
Thanks for the help.

My setup, just for information:
Nvidia GTX790 (I think) with 4 screens connected.
Two of the screens are connected by Logilink Adapters DP to HDMI.
I am using the driver from Nvidias Website.

Can (and should I) report the problem to Nvidia?

(and recieving Messages from the Mailinglist luckily works)

[toc] | [prev] | [next] | [standalone]


#5073

FromGunnar Wolf <gwolf@debian.org>
Date2016-11-06 01:00 +0100
Message-ID<sAj2q-6aj-7@gated-at.bofh.it>
In reply to#5072

[Multipart message — attachments visible in raw view] — view raw

H. N. dijo [Sat, Nov 05, 2016 at 01:12:47AM +0100]:
> > That's what I assumed when I read it: video driver bug.
> > It might also be a hardware issue, where graphic RAM is not
> > cleared on reset.  Or given the nature of video devices, perhaps a
> > combination of the two: a driver bug that expects certain generic
> > behavior combined with a device that expects a custom, proprietary,
> > vendor-provided driver to work correctly.
> >(...)
>
> Thanks for the help.
> 
> My setup, just for information:
> Nvidia GTX790 (I think) with 4 screens connected.
> Two of the screens are connected by Logilink Adapters DP to HDMI.
> I am using the driver from Nvidias Website.
> 
> Can (and should I) report the problem to Nvidia?
> 
> (and recieving Messages from the Mailinglist luckily works)

Yes... Although I do not think they will care very much, but it falls
within their area of responsability; Debian cannot fix non-free
software.

Have you tried to reproduce this issue using the free nVidia drivers?
You might have some less capabilities (i.e. the ability to drive the
four screens) or speed, but it should help you determine whether it's
something caused by Debian or by their non-free drivers.

[toc] | [prev] | [next] | [standalone]


#5074

From"H. N." <deb_mailinglist@niemeczek.at>
Date2016-11-06 10:40 +0100
Message-ID<sAs5H-3Fm-1@gated-at.bofh.it>
In reply to#5073

[Multipart message — attachments visible in raw view] — view raw

On 2016-11-06 00:53, Gunnar Wolf wrote:

> H. N. dijo [Sat, Nov 05, 2016 at 01:12:47AM +0100]:
>>> That's what I assumed when I read it: video driver bug.
>>> It might also be a hardware issue, where graphic RAM is not
>>> cleared on reset.  Or given the nature of video devices, perhaps a
>>> combination of the two: a driver bug that expects certain generic
>>> behavior combined with a device that expects a custom, proprietary,
>>> vendor-provided driver to work correctly.
>>> (...)
>> Thanks for the help.
>>
>> My setup, just for information:
>> Nvidia GTX790 (I think) with 4 screens connected.
>> Two of the screens are connected by Logilink Adapters DP to HDMI.
>> I am using the driver from Nvidias Website.
>>
>> Can (and should I) report the problem to Nvidia?
>>
>> (and recieving Messages from the Mailinglist luckily works)
> Yes... Although I do not think they will care very much, but it falls
> within their area of responsability; Debian cannot fix non-free
> software.
>
> Have you tried to reproduce this issue using the free nVidia drivers?
> You might have some less capabilities (i.e. the ability to drive the
> four screens) or speed, but it should help you determine whether it's
> something caused by Debian or by their non-free drivers.
Since I am not an expert I fear corrupting my Debian (again) by messing
with the drivers.
So I am not going to try that. Sorry.




[toc] | [prev] | [standalone]


Back to top | Article view | linux.debian.security


csiph-web