On Fri, Sep 23, 2022 at 09:55:02PM +0530, K Prateek Nayak wrote:
Yes, I did see the patch, but Andreas replied to Dave's patch attached in v1 saying that he had seen this issue with AMD Athlon on a VIA chipset back in 2006. (https://lore.kernel.org/lkml/Yyy6l94G0O2B7Yh1@rhlx01.hs-esslingen.de/) Therefore, Dave's patch is not sufficient.
Well, AFAICT, that box is old and dead. And I'm being told that the AMD machines which we care for and which are still alive should not need the dummy read.
Which means, we could try the simple fix first and see who complains.