vm86 - DCC probes don't work

vm86 - DCC probes don't work

[Rik van Riel]

After digging around a bit on a system that hangs when trying to
start X, I tried another system and was luckier, when running
system-config-display:

do_wp_page: bogus page at address 00000449
VM: killing process ddcprobe
PCI: Obtained IRQ 16 for device 0000:00:02.0
[drm] Initialized i915 1.1.0 20040405 on minor 0:
do_wp_page: bogus page at address 00000449
VM: killing process ddcprobe

I guess some of our low physical memory mappings in domain 0
aren't complete ?

X did start, btw ...

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

Re: vm86 - DCC probes don't work

[Jon Smirl]

ddcprobe is trying to read hardware I2C buses. On a radeon card these
are part of the mmio address space. Not sure where they appear on the
i915 but it should be easy to find in the source.

On Tue, 22 Feb 2005 13:45:31 -0500 (EST), Rik van Riel <riel@redhat.com> wrote:
> After digging around a bit on a system that hangs when trying to
> start X, I tried another system and was luckier, when running
> system-config-display:
> 
> do_wp_page: bogus page at address 00000449
> VM: killing process ddcprobe
> PCI: Obtained IRQ 16 for device 0000:00:02.0
> [drm] Initialized i915 1.1.0 20040405 on minor 0:
> do_wp_page: bogus page at address 00000449
> VM: killing process ddcprobe
> 
> I guess some of our low physical memory mappings in domain 0
> aren't complete ?
> 
> X did start, btw ...
> 
> --
> "Debugging is twice as hard as writing the code in the first place.
> Therefore, if you write the code as cleverly as possible, you are,
> by definition, not smart enough to debug it." - Brian W. Kernighan
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/xen-devel
> 


-- 
Jon Smirl
jonsmirl@gmail.com


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

RE: vm86 - DCC probes don't work

[Ian Pratt]

 
> ddcprobe is trying to read hardware I2C buses. On a radeon card these
> are part of the mmio address space. Not sure where they appear on the
> i915 but it should be easy to find in the source.

I suspect the i915 driver is going at the address using __va rather than
using isa_bus_to_virt. Please can someone investigate and send a .patch
(and feed upstream too).

Thanks,
Ian

> On Tue, 22 Feb 2005 13:45:31 -0500 (EST), Rik van Riel 
> <riel@redhat.com> wrote:
> > After digging around a bit on a system that hangs when trying to
> > start X, I tried another system and was luckier, when running
> > system-config-display:
> > 
> > do_wp_page: bogus page at address 00000449
> > VM: killing process ddcprobe
> > PCI: Obtained IRQ 16 for device 0000:00:02.0
> > [drm] Initialized i915 1.1.0 20040405 on minor 0:
> > do_wp_page: bogus page at address 00000449
> > VM: killing process ddcprobe



-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_ide95&alloc_id\x14396&op=click

RE: vm86 - DCC probes don't work

[Rik van Riel]

On Tue, 22 Feb 2005, Ian Pratt wrote:

> > ddcprobe is trying to read hardware I2C buses. On a radeon card these
> > are part of the mmio address space. Not sure where they appear on the
> > i915 but it should be easy to find in the source.
> 
> I suspect the i915 driver is going at the address using __va rather than
> using isa_bus_to_virt. Please can someone investigate and send a .patch
> (and feed upstream too).

Dccprobe is trying to do BIOS calls from vm86 mode.

I wonder if we're just not mapping certain physical addresses
into vm86 mode (in domain 0) that the program expects there to
be ...

-- 
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

Re: vm86 - DCC probes don't work

[Jon Smirl]

On Tue, 22 Feb 2005 17:31:36 -0500 (EST), Rik van Riel <riel@redhat.com> wrote:
> On Tue, 22 Feb 2005, Ian Pratt wrote:
> 
> > > ddcprobe is trying to read hardware I2C buses. On a radeon card these
> > > are part of the mmio address space. Not sure where they appear on the
> > > i915 but it should be easy to find in the source.
> >
> > I suspect the i915 driver is going at the address using __va rather than
> > using isa_bus_to_virt. Please can someone investigate and send a .patch
> > (and feed upstream too).
> 
> Dccprobe is trying to do BIOS calls from vm86 mode.
> 
> I wonder if we're just not mapping certain physical addresses
> into vm86 mode (in domain 0) that the program expects there to
> be ...

Intel has not released the source to the i915 VBIOS. You'll have to
debug it the hard way with breakpoints on the emulation hooks in vm86.
It might be going for ioports instead of pci memory. lspci -v should
tell you all of the hardware on the i915.

I also believe the only way to set a video mode on the i915 is with a
call to the VBIOS. That will probably be broken too.

Radeon hardware would be easier to get running. All of the source for
it is in the X tree.

> 
> --
> "Debugging is twice as hard as writing the code in the first place.
> Therefore, if you write the code as cleverly as possible, you are,
> by definition, not smart enough to debug it." - Brian W. Kernighan
> 


-- 
Jon Smirl
jonsmirl@gmail.com


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

RE: vm86 - DCC probes don't work

[Ian Pratt]

> I wonder if we're just not mapping certain physical addresses
> into vm86 mode (in domain 0) that the program expects there to
> be ...

If ddcprobe is using vm86 mode shouldn't it be mapping the appropriate
section of /dev/mem that it expects the 16bit code to access? Its not
attempting BIOS calls is it? 

BTW, where are the 'moving parts' of ddcprobe? I can only find the
python wrapper.  

Ian


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_ide95&alloc_id\x14396&op=click

Re: vm86 - DCC probes don't work

[Jon Smirl]

vbetool uses the same lrmi library as ddcprobe it might be easier to debug.

http://www.srcf.ucam.org/~mjg59/vbetool/


On Tue, 22 Feb 2005 22:59:19 -0000, Ian Pratt <m+Ian.Pratt@cl.cam.ac.uk> wrote:
> 
> > I wonder if we're just not mapping certain physical addresses
> > into vm86 mode (in domain 0) that the program expects there to
> > be ...
> 
> If ddcprobe is using vm86 mode shouldn't it be mapping the appropriate
> section of /dev/mem that it expects the 16bit code to access? Its not
> attempting BIOS calls is it?
> 
> BTW, where are the 'moving parts' of ddcprobe? I can only find the
> python wrapper.
> 
> Ian
> 


-- 
Jon Smirl
jonsmirl@gmail.com


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click