[PATCHv2] nvme: avoid bogus CRTO values

[Keith Busch]

On Thu, Sep 14, 2023 at 10:48:55PM +0300, Felix Yan wrote:
> 
> Thanks, verified that it works well here.

Thanks, okay if I append your Tested-by: in the patch?
 
> I noticed only one very small issue: dev_warn_once seems to only print once
> when multiple devices are affected. It may be more ideal if it prints once
> for each device, but I don't know how to really achieve that...

There's no good way to do that, unfortunately. We'd have to create a
custom "print once" based on some driver specific flag for this path,
but that's overkill for this issue, IMO. I feel it should be sufficient
just to know that the fallback is happening, and doesn't really matter
for an admin scanning the logs to see it appear for each device. My main
concern was printing it on every reset; that level of repitition would
definitely cause alarm for some people.