Bug 2118

Summary:	via: Invalid DMA header command.
Product:	DRI	Reporter:	Thomas Hellström <thomas>
Component:	DRM/other	Assignee:	Default DRI bug account <dri-devel>
Status:	RESOLVED FIXED	QA Contact:
Severity:	normal
Priority:	high
Version:	DRI git
Hardware:	x86 (IA32)
OS:	Linux (All)
Whiteboard:
i915 platform:		i915 features:

Description Thomas Hellström 2004-12-20 12:00:19 UTC

For some applications, The following message from the VIA command verifier appears:

*ERROR* Invalid / Unimplemented DMA HEADER command. 0x0

Which suggests either a bug in the verifier or in the Mesa driver. 
I'm entering this for reference and will try to track it down as soon as possible. 

/Thomas

Comment 1 Thomas Hellström 2004-12-20 15:05:44 UTC

This is caused by the verifier exiting to state_command after a single
HC_COMMAND_FIRE, followed by a 0x00 command. 

Seems like a HC_HEADER2 should be inserted inbetween. 

The command regulator actually accepts this, whereas the PCI path 
(taken from the original via dri driver) does not, and hangs the machine.
I suggest fixing up the Mesa driver not to issue this command sequence.

/Thomas

Comment 2 Thomas Hellström 2004-12-21 09:24:44 UTC

OK, previous message was a probably not correct. Apparently the command verifier
exits the state due to vertex data that is misinterpreted as a fire command. 

In the worst case this means that the command verifier needs to interprete and
count all the vertex data :(. I'm not modifying anything at the moment, since
both the command verifier and the PCI paths might need additions. This bug will
show itself as misrendering in a couple of applications.

Anybody with a detailed understanding of how the command regulator exits the
vertex data state?

/Thomas

Comment 3 Thomas Hellström 2005-11-24 05:19:51 UTC

Fixed some time ago.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.