Stolen Memory <- i830M video chip

Stolen Memory <- i830M video chip

[Michael De Nil]

hey


working here on an Asus L1-laptop which contains the Intel 830M graphical
chip. when loading the agpgart module, it prints in syslog that there is
only 1 Meg 'stolen ram'. like that, it's not possable to run X @ 1024x768
with more then 256 colors.

i searched on the intel-website, which told me hat i should be able to
change this setting in my bios. *not*

can't i reserve any more ram myselve by selecting linux only to use 256 -
8 Meg or something @ boot-time ?


tnx !!
	michael


btw: sry for my n00b'ism ;)


-----------------------------------------------------------------------
                Michael De Nil -- michael@aerythmic.be
       Linux LiSa 2.4.18 #4 SMP ma apr 1 11:11:48 CEST 2002 i686
      22:44:01 up  4:46,  3 users,  load average: 1.07, 0.46, 0.21
-----------------------------------------------------------------------

Re: Stolen Memory <- i830M video chip

[Blue Lang]

On Thu, 11 Apr 2002, Michael De Nil wrote:

> i searched on the intel-website, which told me hat i should be able to
> change this setting in my bios. *not*
>
> can't i reserve any more ram myselve by selecting linux only to use 256 -
> 8 Meg or something @ boot-time ?

the dell c400 has this same problem. there is an excellent web page on one
person's experiences with it, but i can't find it right now.. anyways, his
result was that the bios was actually broken and that it would take a f/w
update from Dell to fix it. a little googling should yield the results you
want.

i assume it's probably the same situation with your laptop.

good luck,

-- 
Blue Lang                                      http://www.b-side.org/~blue
           editor, b-side.org                        http://www.b-side.org
           bug generation unit, alanthia mud             alanthia.org 1536
           integration engineer, veritas software   http://www.veritas.com

Re: i830M video chip (X driver deficient)

[Eric W. Biederman]

Blue Lang <blue@b-side.org> writes:

> On Thu, 11 Apr 2002, Michael De Nil wrote:
> 
> > i searched on the intel-website, which told me hat i should be able to
> > change this setting in my bios. *not*
> >
> > can't i reserve any more ram myselve by selecting linux only to use 256 -
> > 8 Meg or something @ boot-time ?
> 
> the dell c400 has this same problem. there is an excellent web page on one
> person's experiences with it, but i can't find it right now.. anyways, his
> result was that the bios was actually broken and that it would take a f/w
> update from Dell to fix it. a little googling should yield the results you
> want.
> 
> i assume it's probably the same situation with your laptop.

It isn't memory related at all.  The problem is that the X driver uses 
the video BIOS to set the display modes, instead of setting the
display mode by itself as it should.  I don't know if there are enough
docs available from intel about this but that is the problem.

I have a coworker with one of these and I tracked problem down that
far.  With agpgart the kernel allocates the rest of the memory for X
as needed.

Eric

Re: i830M video chip (X driver deficient)

[Blue Lang]

On 12 Apr 2002, Eric W. Biederman wrote:

> It isn't memory related at all.  The problem is that the X driver uses
> the video BIOS to set the display modes, instead of setting the
> display mode by itself as it should.  I don't know if there are enough
> docs available from intel about this but that is the problem.

erm.. I thought that's what I said. Anyways, here is that link I was
talking about:

http://www.cse.unsw.edu.au/~chak/linux/c400.html

Good luck,

-- 
Blue Lang                                      http://www.b-side.org/~blue
           editor, b-side.org                        http://www.b-side.org
           bug generation unit, alanthia mud             alanthia.org 1536
           integration engineer, veritas software   http://www.veritas.com

Re: i830M video chip (X driver deficient)

[Eric W. Biederman]

Blue Lang <blue@b-side.org> writes:

> On 12 Apr 2002, Eric W. Biederman wrote:
> 
> > It isn't memory related at all.  The problem is that the X driver uses
> > the video BIOS to set the display modes, instead of setting the
> > display mode by itself as it should.  I don't know if there are enough
> > docs available from intel about this but that is the problem.
> 
> erm.. I thought that's what I said. Anyways, here is that link I was
> talking about:
> 
> http://www.cse.unsw.edu.au/~chak/linux/c400.html
> 

My point the problem is not memory or the BIOS not allocating enough
memory.  The problem is (a) using the BIOS and (b) the BIOS not believing
it can to do the job when X has enough memory.

All I have seen in the X error messages was an error message that the BIOS
could not set the video mode.  My response, why the heck is the driver
doing BIOS calls.  I'm not a fan at all of using someone else's drivers
after the kernel loads.

Heck I'm not even a fan of using a closed source BIOS.

Eric

Re: Stolen Memory <- i830M video chip

[Abraham vd Merwe]

[-- Attachment #1: Type: text/plain, Size: 1452 bytes --]

Hi Michael!

It's a problem with your BIOS. There is no way to set the video modes unless
you use local memory or 8mb stolen since the setmode BIOS calls checks
whether the mode can fit in the stolen memory, whether you populated the GTT
with extra pages or not.

It is therefore a OEM BIOS problem. Dell was notified about this a long time
ago - I thought they fixed it in the mean time. Try moaning about this to
Asus as well...

> working here on an Asus L1-laptop which contains the Intel 830M graphical
> chip. when loading the agpgart module, it prints in syslog that there is
> only 1 Meg 'stolen ram'. like that, it's not possable to run X @ 1024x768
> with more then 256 colors.
> 
> i searched on the intel-website, which told me hat i should be able to
> change this setting in my bios. *not*
> 
> can't i reserve any more ram myselve by selecting linux only to use 256 -
> 8 Meg or something @ boot-time ?

-- 

Regards
 Abraham

Two wrights don't make a rong, they make an airplane.  Or bicycles.

__________________________________________________________
 Abraham vd Merwe - 2d3D, Inc.

 Device Driver Development, Outsourcing, Embedded Systems

  Cell: +27 82 565 4451         Snailmail:
   Tel: +27 21 761 7549            Block C, Aintree Park
   Fax: +27 21 761 7648            Doncaster Road
 Email: abraham@2d3d.co.za         Kenilworth, 7700
  Http: http://www.2d3d.com        South Africa


[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Stolen Memory <- i830M video chip

[Josh Fryman]

> It is therefore a OEM BIOS problem. Dell was notified about this a long time
> ago - I thought they fixed it in the mean time. Try moaning about this to
> Asus as well...

this isn't fixed in the new dell C400 laptops (still).  however, i argue that
BIOS problem or no, a solution _does_ exist.  there is a 3rd party that
sells commercial X server drop-ins for new chipsets.  one of the guys here
bought their X server for $40 or so, and now has full-color & -resolution 
without getting a BIOS update.

so if it's a BIOS problem that can only be fixed by Dell, how were these guys
able to do the fix?  and why can't the open source guys (XFree or Linux kernel) 
seem to do the same?

-josh

ps > company is xig.com : Dell C400 w/generic I830-M support:
http://www.xig.com/Pages/Atop/LaptopIndividualSupportSpecs/Dell-IndividualLaptops/LatitudeC400.html

Re: Stolen Memory <- i830M video chip

[Abraham vd Merwe]

[-- Attachment #1: Type: text/plain, Size: 1746 bytes --]

Hi Josh!

> > It is therefore a OEM BIOS problem. Dell was notified about this a long time
> > ago - I thought they fixed it in the mean time. Try moaning about this to
> > Asus as well...
> 
> this isn't fixed in the new dell C400 laptops (still).  however, i argue that
> BIOS problem or no, a solution _does_ exist.  there is a 3rd party that
> sells commercial X server drop-ins for new chipsets.  one of the guys here
> bought their X server for $40 or so, and now has full-color & -resolution 
> without getting a BIOS update.
> 
> so if it's a BIOS problem that can only be fixed by Dell, how were these guys
> able to do the fix?  and why can't the open source guys (XFree or Linux kernel) 
> seem to do the same?

*sigh* You're not listening to me. The _BIOS_ is the problem. If you don't
use the _BIOS_, you don't have a problem. We (2d3D, Inc. the people who
wrote the driver) are paid to use the BIOS. We have a non-bios setmode, but
were not using it because I'm not allowed to release that source.

I've told people before: If you want it to work, write a non-BIOS setmode.
I'll even give you tips if you try, but that is all I can do. Sorry, life is
tough.

-- 

Regards
 Abraham

Rules for Academic Deans:
	(1)  HIDE!!!!
	(2)  If they find you, LIE!!!!
		-- Father Damian C. Fandal

__________________________________________________________
 Abraham vd Merwe - 2d3D, Inc.

 Device Driver Development, Outsourcing, Embedded Systems

  Cell: +27 82 565 4451         Snailmail:
   Tel: +27 21 761 7549            Block C, Aintree Park
   Fax: +27 21 761 7648            Doncaster Road
 Email: abraham@2d3d.co.za         Kenilworth, 7700
  Http: http://www.2d3d.com        South Africa


[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Stolen Memory <- i830M video chip

[Abraham vd Merwe]

[-- Attachment #1: Type: text/plain, Size: 1973 bytes --]

Hi Josh!

> i've cut the key bits out for back-reference.  (btw, your .sig is kinda
> hefty :)
> 
> > > so if it's a BIOS problem that can only be fixed by Dell, how were these guys
> > > able to do the fix?  and why can't the open source guys (XFree or Linux kernel) 
> > > seem to do the same?
> > 
> > I've told people before: If you want it to work, write a non-BIOS setmode.
> > I'll even give you tips if you try, but that is all I can do. 
> 
> ok, let me see if i'm following what you're saying.  the XFree86 drivers don't
> use the BIOS or anything else - they rely on the kernel.  the kernel, however, 
> is relying on the BIOS to tell it what's going on.  when the I8x0 video interface
> activates, it asks the BIOS which lies through its teeth.

No, the X Server is using the BIOS to set video modes. Since the video
chipset doesn't have any onboard memory, it needs to "steal" some of your
system memory. The BIOS does that for you (hence the term stolen memory).
However, if your BIOS steals only 1mb memory, it isn't enough to support
high resolution modes.

This isn't a problem since we allocate more memory by stealing some memory
in the kernel and then populating the card's page table with those pages.
The problem is that the BIOS don't know about those extra pages and when you
do a setmode, it checks whether there is enough memory, sees only 1mb and
then refuses to set the mode...

-- 

Regards
 Abraham

I give you the man who -- the man who -- uh, I forgets the man who?
		-- Beauregard Bugleboy

__________________________________________________________
 Abraham vd Merwe - 2d3D, Inc.

 Device Driver Development, Outsourcing, Embedded Systems

  Cell: +27 82 565 4451         Snailmail:
   Tel: +27 21 761 7549            Block C, Aintree Park
   Fax: +27 21 761 7648            Doncaster Road
 Email: abraham@2d3d.co.za         Kenilworth, 7700
  Http: http://www.2d3d.com        South Africa


[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: Stolen Memory <- i830M video chip

[Michael De Nil]

On Fri, 12 Apr 2002, Abraham vd Merwe wrote:

> I've told people before: If you want it to work, write a non-BIOS setmode.
> I'll even give you tips if you try, but that is all I can do. Sorry, life is
> tough.

if anyone is inresting in doing this, they can access my system to test /
write.

plz contact me if you need my system.


tnx!
	michael


-----------------------------------------------------------------------
                Michael De Nil -- michael@aerythmic.be
       Linux LiSa 2.4.18 #4 SMP ma apr 1 11:11:48 CEST 2002 i686
     11:04:01 up 46 min,  2 users,  load average: 0.00, 0.06, 0.07
-----------------------------------------------------------------------

Re: Stolen Memory <- i830M video chip

[Alan Cox]

> BIOS problem or no, a solution _does_ exist.  there is a 3rd party that
> sells commercial X server drop-ins for new chipsets.  one of the guys here
> bought their X server for $40 or so, and now has full-color & -resolution 
> without getting a BIOS update.
> 
> so if it's a BIOS problem that can only be fixed by Dell, how were these guys
> able to do the fix?  and why can't the open source guys (XFree or Linux kernel) 
> seem to do the same?

Because nobody will give the open source guys a look at the top sekkrit elite
documentation the mere viewing of which would immediately put the vendor out
of business...

If you are going to buy a Linux laptop always check the linux-laptops web site
first and never forget which company(or companies) screwed you last time. Its
possibly worth talking to Dell but the last bad "Dell" laptop BIOS problems
Linux showed up got nowhere because Dell just rebadged that product. You
might in fact have to moan at Compal or similar depending on which Dell you
have.

Alan

Re: Stolen Memory <- i830M video chip

[Michael De Nil]

On Thu, 11 Apr 2002, Michael De Nil wrote:

> hey
>
>
> working here on an Asus L1-laptop which contains the Intel 830M graphical
> chip. when loading the agpgart module, it prints in syslog that there is
> only 1 Meg 'stolen ram'. like that, it's not possable to run X @ 1024x768
> with more then 256 colors.


Asus will take care of this, recieved following mail:



>Date: Mon, 15 Apr 2002 20:40:21 +0800
>From: tsd <tsd@asus.com.tw>
>To: michael@aerythmic.be
>Subject: Re:Asus L1-serie notebook -> i830M chipset
>
>
>Dear Sir
>We will try our best to release the new bios which can fix this bug , and
>you
>can install linux successfully...
>but you still need to wait for the latest bios release...
>Thank You


Michael


-----------------------------------------------------------------------
                Michael De Nil -- michael@aerythmic.be
       Linux LiSa 2.4.18 #4 SMP ma apr 1 11:11:48 CEST 2002 i686
      18:18:01 up 22:31,  1 user,  load average: 0.00, 0.00, 0.00
-----------------------------------------------------------------------