SCIENTIFIC-LINUX-USERS Archives

April 2008

SCIENTIFIC-LINUX-USERS@LISTSERV.FNAL.GOV

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Christopher Hunter <[log in to unmask]>
Reply To:
Christopher Hunter <[log in to unmask]>
Date:
Tue, 29 Apr 2008 12:27:41 -0400
Content-Type:
text/plain
Parts/Attachments:
text/plain (116 lines)
You might try 915resolution.
http://www.geocities.com/stomljen/

This used to be in the sl-contrib yum repository, not sure if it's still 
around.

That won't solve vesa driver problems.

The flashing screen problem indicates your monitor cannot support the 
current screen resolution.

> On Mon, 28 Apr 2008, John Summerfield wrote:
> 
>> John Summerfield wrote:
>>>  John Summerfield wrote:
>>>
>>>   I have installed 64-bit SL5.1 on this machine
>>> >  HP DC7700, Intel(R) Pentium(R) Dual  CPU  E2140  @ 1.60GHz, 5 Gbytes, >  2x80 Gb (Windows XP on the first).
>>> > >  _That_ is (almost) the worst.
>>> > >  Additionally, I can't change consoles, either with the usual keyboard >  magic or, having logged in remotely so as to look around, using the chvt >  command.
> 
> We have found that with the Intel chipsets an XFree86/Xorg you seem to have two options:
> 
>   use the intel driver and suffer problems with changing VT or restarting
>   X causing problems and lock-ups
> 
>   use the vesa driver and have no/poor acceleration and a lack of some
>   other features (like DPMS support).
> 
> we try to avoid (where possible) using any (usually built-in) intel graphics chipset with X by adding a graphics card to all desktop machines.
> 
> Of course the cards supported properly by Xorg in SL5 isn't a huge list we have had good experiences with the AIT X300/X600 cards but those are no longer commonly available.
> 
> Any of the newer cards (X1300 etc) arn't supported by the ati/radeon driver so one is back to using vesa there.
> 
> With the vesa driver these seems to be no simple way to tell if things like DPMS, DDC probing will actually work (reliably).  I'm told that on some cards they do work but it doesn't seem to do so on the cards I've tested here...
> 
>>> >  In fact, the keyboard (an HP USB keyboard no less) seems to be locked >  up.
>>>
>>>  I installed the non-Xen kernel and X seemed okay with it, but I didn't
>>>  test much. I rebooted to Xen and, again, X _looked_ okay.
>>>
>>>  I could switch to tty1, but when i switched back X didn't reinitialise
>>>  properly. Instead, it flashed between grey and black.
>>>
>>>  control-alt-BS kills it, but on restarting the flashing resumes. Nor does
>>>  this do better:
>>>
>>>  [root@bobtail ~]# telinit 3
>>>  [root@bobtail ~]# telinit 5
>>>
>>>  I will now return to the non-Xen kernel and see what's what.
>>>  <reboots>
>>>  That problem does not occur. However, the login screen is displayed at
>>>  800x600. I logged in, and a couple of "xvidtune -next" command had it at
>>>  1280x1024 (the screen's resolution), but the desktop was an 800x600 window
>>>  the viewer's left, top and partly replicated (it extends off-screen)
>>>  right, top.
> 
> It probably isn't picking up the monitor data (DDC) properly so is running with the safest of all resolutions.  You may be able to manually tell it the right data by specifying the exact monitor model in the advanced settings of system-config-display or manually editing the Xorg.conf
> 
> Does it list anything interesting about the monitor or DDC in the xorg log?
> 
>>>  From there, control-alt-BS killed X and it restarted at the correct
>>>  resolution!
>>>
>>>  <reboots>
>>>  <shrug>
>>>  Didn't happen this time.
>>>  OTOH it's back to its flashing behaviour.
>>>
>>>  I did this:
>>>  telinit 3;sleep 20s;telinit 5
>>>  and now the bottom of the login screen's at the top, the area from (about)
>>>  username and down is light grey and the whole thing's flashing.
>>>
>>>
>>>  I should also mention that every time, when X first starts there's a area
>>>  of video corruption at the bottom ot the screen, 3 cm or so, briefly and
>>>  then the screen (maybe) comes good.
> 
> That is quite common.  The video chipset gets re-programmed to the new video mode before the memory is cleared so you see junk based on what was displayed previously.  e.g. on our Dell machines we see a corrupted version of the Dell boot logo flash past...
> 
>>>  I have an IBM Thinkcentre with this graphics card:
>>>  00:02.0 VGA compatible controller: Intel Corporation 82865G Integrated
>>>  Graphics Controller (rev 02)
>>>
>>>  as compared with HP's choice
>>>  [root@bobtail ~]# lspci -s 00:02.0
>>>  00:02.0 VGA compatible controller: Intel Corporation 82Q963/Q965
>>>  Integrated Graphics Controller (rev 02)
>>>  [root@bobtail ~]#
>>>
>>>  I copied the Thinkcentre's configuration across, the result's the same.
>>>
>>>  The vesa driver does not work at all. It gives lots of promising messages,
>>>  a few warnings and no errors (EE messages).
>>>
>>>  the intel driver doesn't support it.
> 
> So if neither the intel nor vesa driver works what driver are you using to get anything displayed?
> 
>> Does nobody have an idea? The system's completely unusable at present.
> 
> Add a different video card.  I like ATI cards but I know that others prever nvidia models (which have the advantage of still being available though last time I checked we didn't get *good* acceleration with the 'nv' driver)...
> 
> As a rule of thumb the older/cheaper the video card the more likely that Xorg actually supports it since it will probably have been out for longer and had many people trying to make it work.
> 
> On laptops (or some desktops) where one can't alter the video hardware you can't add a new card so have to live with what is provided.
> 
> On my (test) HP 6720s laptop which has an intel graphics chipset the Xorg in both ubuntu feisty (710) and hardy (804) seem to drive it ok with the intel driver.  Of course I'm not doing anything complex there.  Perhaps the chipsets in common models of laptop get more widely tested than those in desktop systems...

-- 
Chris Hunter
Systems Programmer, Astronomy Yale University

ATOM RSS1 RSS2