Hi Peter, On 9/16/19 5:23 AM, Peter Xu wrote: > On Fri, Sep 13, 2019 at 10:36:10AM +0200, Eric Auger wrote: >> Currently, when a notifier is attempted to be registered and its >> flags are not supported (especially the MAP one) by the IOMMU MR, >> we generally abruptly exit in the IOMMU code. The failure could be >> handled more nicely in the caller and especially in the VFIO code. >> >> So let's allow memory_region_register_iommu_notifier() to fail as >> well as notify_flag_changed() callback. >> >> All sites implementing the callback are updated. This patch does >> not yet remove the exit(1) in the intel_iommu and amd_iommu code. > > The idea looks sane to me, though how about using "Error **" instead > of returning int (or, both)? Just like the majority of rest of QEMU. None of the memory_region_register_iommu_notifier callsites would really use the Error object and pass it up. I am a bit reluctant to add it if not really used. There are still plenty of functions in the memory API that do not use any Error handle. Anyway I will follow recommendations if there is a consensus.
Thanks Eric > > Regards, >
