Quoting Mika Kuoppala (2019-02-19 13:07:08) > Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> writes: > > > Introduce a new ABI method for detecting a wedged driver by reporting > > -EIO from DRM_IOCTL_I915_GEM_CONTEXT_CREATE. > > Need more on the 'why' department. What is lacking with > the method of submitting and noticing the wedged? This fits in with https://lists.freedesktop.org/archives/mesa-dev/2019-February/215469.html where we are recreating the contexting the context after receiving the -EIO from execbuf and so this allows us to actually break that loop if the driver is wedged. > Also detecting banned from wedged is not trivial. We are not detecting banned per-se, just saying that if the driver is wedged. > I am trying to figure out what is the userspace > need and flow wrt this. > > If we will have object parameters, should we > convey this type of information with status query? It's not an object property, it's the status of the driver. Currently the indication is throttle returning -EIO, but that would look mighty strange in the above recreate_context(). -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
