On 10.06.2025 17:54, Roger Pau Monné wrote: > On Tue, Jun 10, 2025 at 04:13:40PM +0200, Jan Beulich wrote: >> On 10.06.2025 15:24, Roger Pau Monné wrote: >>> IMO the added complexity here is not worth the performance >>> improvement, not without a clear justification that it's needed. >> >> Well, okay, I'll simply consider the patch in this shape rejected then. >> I don't see much value in wasting further time on it. > > The code is already there, so I think there's value in this patch. > Did you see my suggestion in the email yo uare replying to about not > needing to add the is_iommu_enabled(d) || cache_flush_permitted(d) > checks?
I did, but it didn't look like we were able to come to an agreement. > With that dropped (if it's indeed OK), I would be fine with Acking the > patch. I guess I'll make one more try then, removing the instance where the setting of the "flush" local is made conditional, but retaining the -EPERM return path as was. Jan
