On 2012-02-02 20:13, Steven Lang wrote: > The only thing I can think is maybe there was some sort of other issue > where perhaps a DMA IO was happening before the memory write, > unrelated to any of the code there, but the write barrier just > happened to fix it. But even that seems like it would be a bit of a > stretch to imagine it actually being a problem. I was thinking that could've been the issue as well, but in that case it would be a kernel (or driver) bug in that it didn't flush. -- Jens Axboe -- To unsubscribe from this list: send the line "unsubscribe fio" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html