Re: [PATCH] amd/iommu: fix logic bug in amd_iommu_report_page_fault()

From: David Rientjes
Date: Sun Aug 01 2021 - 23:52:48 EST

On Sat, 31 Jul 2021, Luigi Rizzo wrote:

> amd_iommu_report_page_fault() has two print paths, depending on whether or
> not it can find a pci device. But the code erroneously enters the second
> path if the rate limiter in the first path triggers:
> if (dev_data && ratelimit(A)) { A; } else if (ratelimit(B)) { B; }
> The correct code should be
> if (dev_data) { if (ratelimit(A)) { A;} } else if (ratelimit(B)) { B; }
> Signed-off-by: Luigi Rizzo <lrizzo@xxxxxxxxxx>

Acked-by: David Rientjes <rientjes@xxxxxxxxxx>

This would be very helpful so we don't erroneously classify these KERN_ERR
messages as dangerous.