[Cbe-oss-dev] elaboration of Axon driver?

[Arnd Bergmann]

On Thursday 18 December 2008, Marcus G. Daniels wrote:
> > I can't currently reproduce the problem here, but it would be interesting
> > to instrument the device driver to check how this happened. The specific
> > address should be used when you ask for a dma with 'remote status update'.
> >   
> Yeah, I found experimentally that the AXON_DMAFLAG_WRITE_REMOTE_STATUS 
> was the trigger.  The workaround I have now is to read(2) on the device 
> on the local end, with the device configured for 
> AXON_DMAFLAG_LOCAL_COMPLETION_SIGNAL.   Then I use the sma memory map to 
> copy over a signal to the remote (Opteron) side.  This doesn't happen 
> when the local side is the Opteron. 

Yeah, that's certainly not nice as a workaround. Are you able to install
patched kernels to try out possible fixes?

> And it seems that this workaround only works for a read(2) on the local end. 
> read on the remote just waits and waits.

Right, AFAICT there is no AXON_DMAFLAG_REMOTE_COMPLETION_SIGNAL which
you would need for this.

	Arnd <><