On Fri, 2008-02-15 at 22:27 +0000, Alan Cox wrote:
quoted text
> > The dbdma start is mostly harmless (things don't get posted for -that-
> > long), though I suppose it's worth fixing. Would reading back dmactl do
> > in that case or do you foresee any kind of side effect ? (Maybe only
> > doing it for MMIO ?)
> 
> The dmactl read back should be just fine, or any other DMA register (eg
> status).
> 
> > As for SRST, I'm not totally confident how safe it is to read back
> > there while doing the reset sequence, so I'm tempted to really only
> > do it for MMIO and use altstat rather than ctl/stat (the later tends
> > to have side effects which we don't want here).
> 
> Agreed - we know some controllers crap themselves spectacularly on
> anything which causes a SATA data transfer to be needed during a reset so
> the status is probably safest. The fact its not fixed is because nobody
> has sat down to figure out what is safe.
> 
> > The main problem from here is that I don't know whether we are using
> > MMIO or PIO from libata-core. Maybe I can add a host flag indicate
> > that such flushing is needed ?
> 
> Easier to add that to the ioxxxx ops I suspect (ioispio/ioismmio say) ?

Maybe, though that will involve fixing all the arch versions which do
different things. In fact, I'm not even sure I can tell you 100% after
the fact on ppc64, I have to double check.

I'd rather stick a flag in there to be safe, also since altstatus isn't
always there (which is why I used nsect in the test patch I sent to
Guennadi). I'm pretty sure I can rely on all MMIO controllers having an
altstatus but I'd rather still make that explicit with a host flag to
avoid unintended consequences to others.

Ben.


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