Re: [FOR 2.6.18 FIX][PATCH] drm: radeon flush TCL VAP for vertexprogram enable/disable

[Dave Airlie]

> It looks ok to me, although you have to look into the caller to see why it
> does what it does.
>
> It would be "prettier" if it changed the size and data of the incoming
> packet instead, but the code as is isn't actually set up to be able to do
> that (the size setup and verification stuff is done before the fixup).
>
> That said, I'd have expected that the VAP state flush is really something
> that the _client_ should do when it generates the commands, not the kernel
> after the fact. Although maybe the kernel could keep track of whether the
> flush is needed at all, and since we apparently allow untrusted generation
> of packets, maybe this is the right approach..

The problem is of course if the userspace side does it, the lockup is 
simple to trigger, we'd like if we can to stop them triggering it, we 
don't stop every lockup, but it would be nice to stop the ones we can...

Dave.

--
David Airlie, Software Engineer
http://www.skynet.ie/~airlied / airlied at skynet.ie
Linux kernel - DRI, VAX / pam_smb / ILUG

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/