MMIO and gcc re-ordering issue

[James Bottomley]

On Tue, 2008-05-27 at 10:38 -0700, Roland Dreier wrote:
> > Actually, this specifically should not be.  The need for mmiowb on altix
>  > is because it explicitly violates some of the PCI rules that would
>  > otherwise impede performance.   The compromise is that readX on altix
>  > contains the needed dma flush but there's a variant operator,
>  > readX_relaxed that doesn't (for drivers that know what they're doing).
>  > The altix critical drivers have all been converted to use the relaxed
>  > form for performance, and the unconverted ones should all operate just
>  > fine (albeit potentially more slowly).
> 
> Is this a recent change?  Because as of October 2007, 76d7cc03
> ("IB/mthca: Use mmiowb() to avoid firmware commands getting jumbled up")
> was needed.  But this was involving writel() (__raw_writel() actually,
> looking at the code), not readl().  But writel_relaxed() doesn't exist
> (and doesn't make sense).

Um, OK, you've said write twice now ... I was assuming you meant read.
Even on an x86, writes are posted, so there's no way a spin lock could
serialise a write without an intervening read to flush the posting
(that's why only reads have a relaxed version on altix).  Or is there
something else I'm missing?

James