>>> On 04.09.17 at 13:12, <roger....@citrix.com> wrote: > On Mon, Sep 04, 2017 at 06:59:24PM +0800, Lan Tianyu wrote: >> On 2017年09月01日 17:57, Roger Pau Monné wrote: >> > On Thu, Aug 31, 2017 at 01:01:48AM -0400, Lan Tianyu wrote: >> >> @@ -92,7 +92,12 @@ static struct acpi_20_madt *construct_madt(struct >> >> acpi_ctxt *ctxt, >> >> sz = sizeof(struct acpi_20_madt); >> >> sz += sizeof(struct acpi_20_madt_intsrcovr) * 16; >> >> sz += sizeof(struct acpi_20_madt_ioapic); >> >> - sz += sizeof(struct acpi_20_madt_lapic) * hvminfo->nr_vcpus; >> >> + >> >> + if (hvminfo->nr_vcpus < 128) >> > >> > This should be done based on APIC ID. >> >> There will be a problem how to get max apic id. Should we use the max >> vcpu index to get max APIC id? > > IMHO, this should become a loop that iterates over each vCPU getting > it's APIC ID, and add either a lapic or x2apic struct size to sz.
Indeed. Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xen.org https://lists.xen.org/xen-devel