[PATCH] i915: make vbl interrupts work properly on i965g/gm

Previous thread: iwl4965 and driver merging policy by Benjamin Herrenschmidt on Thursday, September 27, 2007 - 9:39 pm. (7 messages)

Next thread: 2.6.23-rc8 network problem. Mem leak? ip1000a? by linux on Thursday, September 27, 2007 - 10:06 pm. (5 messages)
To: <torvalds@...>, Andrew Morton <akpm@...>
Cc: <linux-kernel@...>, <jbarnes@...>
Date: Thursday, September 27, 2007 - 10:05 pm

Hi Linus,

The attached patch is to fix a bug reported on 965gm chipsets (lots of new
laptops), I think distros will all have to patch this in to fix it, so can
we get it into the 2.6.23 final?

(Otherwise I'll wait until stable..)

Dave.

To: Dave Airlie <airlied@...>
Cc: <torvalds@...>, Andrew Morton <akpm@...>, <linux-kernel@...>
Date: Thursday, September 27, 2007 - 10:09 pm

Without this patch, my 965GM drops vblank interrupts, so I'd really like to
see it upstream ASAP too.

Acked-by: Jesse Barnes <jesse.barnes@intel.com>
-

To: Jesse Barnes <jbarnes@...>
Cc: Dave Airlie <airlied@...>, <linux-kernel@...>
Date: Friday, September 28, 2007 - 4:02 pm

Sorry to hijack this thread, but I have a question about i915
interrupts (945GM specifically) that I haven't gotten an answer to
yet, and this thread reminded me of it so I thought I'd ask again:
I hacked the kernel drm stuff on my laptop with 945GM graphics to do
pci_enable_msi() on the graphics device, and I don't see any
interrupts occur.

So can someone with access to details of the 945GM tell me if MSI
interrupts should work, and if so if there's some internal registers
that need to be poked to enable it (beyond what pci_enable_msi()
does).

It's not super-urgent-- I just like using shiny things like MSI, and I
already have ahci, iwl3945 and e1000 working with MSI on my laptop, so
the graphics are the only thing left with an MSI capability that's not
working.

Thanks!
Roland
-

To: Roland Dreier <rdreier@...>
Cc: Dave Airlie <airlied@...>, <linux-kernel@...>
Date: Friday, September 28, 2007 - 4:07 pm

I don't have a 945 to test with, but Dave might...

Jesse
-

To: Jesse Barnes <jbarnes@...>
Cc: Dave Airlie <airlied@...>, <linux-kernel@...>
Date: Friday, September 28, 2007 - 6:23 pm

> I don't have a 945 to test with, but Dave might...

Actually I'm not looking for testing (I can test fine on my own laptop :)
I was just hoping someone with docs could tell me "MSI is documented to be
broken on 945GM" or "Oh yeah, 945GM requires you to set SECRET_MSI_ENABLE_BIT
in register FOO before it generates MSIs."

As I said, it's not a big deal but I'd sort of like to know whether my
MSI enable hack at least has a chance at working before I try too hard
to get it working.

Thanks,
Roland
-

To: Roland Dreier <rdreier@...>
Cc: Dave Airlie <airlied@...>, <linux-kernel@...>
Date: Friday, September 28, 2007 - 7:14 pm

I don't see anything in the docs (either public or private) that would
indicate that MSI is broken on those chips, so I'd expect it to work...

Jesse

-

Previous thread: iwl4965 and driver merging policy by Benjamin Herrenschmidt on Thursday, September 27, 2007 - 9:39 pm. (7 messages)

Next thread: 2.6.23-rc8 network problem. Mem leak? ip1000a? by linux on Thursday, September 27, 2007 - 10:06 pm. (5 messages)