Request for Intel 915 hw information

Jesse Barnes jbarnes at
Thu Sep 20 07:49:42 PDT 2007

On Tuesday, September 18, 2007 6:28:08 pm Krzysztof Halasa wrote:
> Hi,
> Simple questions for you Intel experts. I'm using interlaced mode
> on PIPE A:
> How do I read currently displayed scan line number, and/or field
> number? Is it the following?
> #define PIPEAFRAMEHIGH          0x70040
> #define PIPE_FRAME_HIGH_MASK    0x0000ffff
> #define PIPEAFRAMEPIXEL         0x70044
> #define PIPE_FRAME_LOW_MASK     0xff000000
> #define PIPE_FRAME_LOW_SHIFT    24
> (and 0x71040 + 0x71044 for PIPEB)?

Yeah, plus PIPE*_DSL

> On my 915 PIPEACONF, PIPEA_DSL and the above registers produce:
> 0x80C00000 288       0          0x1207A04E
> 0x80C00000 288       0          0x1307A03E
> 0x80C00000 288       0          0x1407A026
> 0x80C00000 288       0          0x1507A06E
> 0x80C00000 288       0          0x1607A06A
> 0x80C00000 288       0          0x1707A0CA
> 0x80C00000 288       0          0x1807A126
> 0x80C00000 288       0          0x1907A0EB
> 0x80C00000 288       0          0x1A07A0C2
> 0x80C00000 288       0          0x1B07A149
> 0x80C00000 288       0          0x1C07A076
> 0x80C00000 288       0          0x1D07A19C
> 0x80C00000 288       0          0x1E07A031
> Why is FRAME_HIGH always 0 for PIPE A (the active one)? Not implemented
> on 915?

No, this should work on i915.  FRAMEPIXEL's upper bits seem to be increasing 
slowly too (assuming you're printf'ing them while your app is running and not 
stuffing them into a buffer to dump out later).  Are you sure you should be 
looking at pipe A?  On a laptop, only pipe B will be active unless you have 
an external display plugged in.

> Should I expect the 8 MSBs of FRAME_PIXEL to always be the frame (field)
> count? On all chips?
> Can I assume that even/odd frame# in FRAME_PIXEL = even/odd field in
> output, or something like that?

In the case of an interlaced mode, I think the frame count registers will 
count whole fields, not individual interlaced frames.

> What is PIPEA_DSL? Is it implemented on 915 and/or other chips?

Yeah, i915 has it.  It gives you the actual scanline number currently being 
scanned out.  Again for interlaced modes it'll give you a field relative 
number, so you'd expect all even numbers for one frame, then all odd numbers 
the next.

> Is it possible to generate IRQ only on the second (final) vertical
> retrace?

Yeah, that's how things *should* work afaik.  Are you seeing something 

> Obviously I'm trying to reliably sync interlaced video to the full
> interlaced frame (in IRQ handler), is
> 	field_parity = (INREG(PIPEAFRAMEPIXEL) >> 24) & 1
> a reliable way to do it?

See above, I think you'll get field counts here, though I'll have to check 
some more recent docs to be sure.

> How about that Xvideo overlay, is it possible to have no vertical
> scaling etc. in interlaced mode? Current code causes the chip
> to deinterlace somehow, even if I have 576 visible lines and 576-line
> video stream.

Not sure about the video overlay, I'll have to look that up.


More information about the xorg mailing list