On 19 July 2018 at 16:24, Richard Henderson <richard.hender...@linaro.org> wrote: > On 07/19/2018 06:05 AM, Alex Bennée wrote: >> I'm confused by the lock_user_struct/unlock_user_struct which AFAICT are >> basically access checks. Is there an implied locking I'm missing? > > I have no idea why those have that name. There's no lock, > just a validation of the underlying pages and a transform to > the corresponding host address.
Mmm, it's unclear. It came in via 53a5960aadd542dd, and the impression I get is that the idea was to try to decouple from knowing that the guest memory layout is flat, as a preparation for possible support for softmmu-in-usermode... thanks -- PMM
