1085364 - Intel HD4000 not in the blacklist code?

Status: NEW

(Core :: Graphics, defect)

Description

[Nicolas Silva [:nical]]

the enumeration of intel device family doesn't contain IntelHD4000, here:
http://dxr.mozilla.org/mozilla-central/source/widget/xpwidgets/GfxDriverInfo.h#61

and in the .cpp, the code that associates device ids with families doesn't mention the device id 0x0166 (which is HD4000). I am new to the blacklisting business so maybe it isn't big deal, but in light of our current 33 release craziness I'd rather make sure we don't miss some faulty drivers.

Comment 1

[Nicolas Silva [:nical]]

Attachment: add hd4000 to the device families

Not sure if it's what we should do or not

Comment 2

[Nicolas Silva [:nical]]

Comment on attachment 8507882 [details] [diff] [review]
add hd4000 to the device families

Review of attachment 8507882 [details] [diff] [review]:
-----------------------------------------------------------------

::: widget/xpwidgets/GfxDriverInfo.cpp
@@ +168,5 @@
>      case IntelHD3000:
>        APPEND_DEVICE(0x0126);
>        break;
> +    case IntelHD4000:
> +      APPEND_DEVICE(0x0166);

Is this a place where we append all the devices we know of, or just the ones which we have had issues with? Looking on the web for other hd4000 device ids I found 0x0162 too.

Comment 3

[Nicolas Silva [:nical]]

Comment on attachment 8507882 [details] [diff] [review]
add hd4000 to the device families

Review of attachment 8507882 [details] [diff] [review]:
-----------------------------------------------------------------

Jeff you names shows up here and there in GfxInfo.cpp (hg blame) so you get to review this (unless you can find a better person to do the review)

Comment 4

[Nicolas Silva [:nical]]

Comment on attachment 8507882 [details] [diff] [review]
add hd4000 to the device families

Review of attachment 8507882 [details] [diff] [review]:
-----------------------------------------------------------------

Whoever can review this first or even if you want to pair-review it. Mind you that I don't know all of this and that if this patch happens to be correct we would consider adding it to the next chemspill release...

Comment 5

[Jeff Muizelaar [:jrmuizel]]

Comment on attachment 8507882 [details] [diff] [review]
add hd4000 to the device families

Review of attachment 8507882 [details] [diff] [review]:
-----------------------------------------------------------------

::: widget/xpwidgets/GfxDriverInfo.cpp
@@ +168,5 @@
>      case IntelHD3000:
>        APPEND_DEVICE(0x0126);
>        break;
> +    case IntelHD4000:
> +      APPEND_DEVICE(0x0166);

It's hard to say. This code has seen enough iteration and is somewhat ownerless so there's no longer really a vision for what it should be. It's worth noting that there's already duplication of device ID's between families. What is the intended use for this entry?

Comment 6

[Jeff Muizelaar [:jrmuizel]]

(In reply to Nicolas Silva [:nical] from comment #2)
> Comment on attachment 8507882 [details] [diff] [review]
> add hd4000 to the device families
> 
> Review of attachment 8507882 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> ::: widget/xpwidgets/GfxDriverInfo.cpp
> @@ +168,5 @@
> >      case IntelHD3000:
> >        APPEND_DEVICE(0x0126);
> >        break;
> > +    case IntelHD4000:
> > +      APPEND_DEVICE(0x0166);
> 
> Is this a place where we append all the devices we know of, or just the ones
> which we have had issues with? Looking on the web for other hd4000 device
> ids I found 0x0162 too.

I have a 0x0162 and the problem doesn't happen there. It might be better to block based on device id instead of family.

Comment 7

[Benoit Jacob [:bjacob] (mostly away)]

Comment on attachment 8507882 [details] [diff] [review]
add hd4000 to the device families

Cancelling review request until jeff's concerns are dealt with.

Comment 8

[Milan Sreckovic [:milan] (needinfo for best results)]

Seems like we should have a solution where we can downloadable blacklist a device family our software doesn't know about?  Separate bug, I guess, or part of bug 838845.