On Fri, Sep 21, 2018 at 02:25:42PM +0000, Singh, Brijesh wrote:
[...]
> +static int amdvi_int_remap_ga(AMDVIState *iommu,
> + MSIMessage *origin,
> + MSIMessage *translated,
> + uint64_t *dte,
> + X86IOMMUIrq *irq,
> + uint16_t sid)
> +{
> + int ret;
> + struct irte_ga irte;
> +
> + /* get interrupt remapping table */
> + ret = amdvi_get_irte_ga(iommu, origin, dte, &irte, sid);
> + if (ret < 0) {
> + return ret;
> + }
> +
> + if (!irte.lo.fields_remap.valid) {
> + trace_amdvi_ir_target_abort("RemapEn is disabled");
> + return -AMDVI_IR_TARGET_ABORT;
> + }
> +
> + if (irte.lo.fields_remap.guest_mode) {
> + trace_amdvi_ir_err("guest mode is not zero");
Just to let you know that you mentioned about replacing some of the
traces with error_report_once() but I didn't see any in this latest
version. Again that's optional so traces would be fine, but I'm still
pointing that out in case you wanted that and you're posting a wrong
tree.
Otherwise it looks good to me.
Thanks,
--
Peter Xu
