OK. Thx Andreas！ 在2013年12月20日 19:18， Andreas Färber写道： Am 19.12.2013 15:20,
schrieb Paolo Bonzini: > Il 18/12/2013 19:03, Andreas Färber ha scritto: >>>>
@@ -61,7 +69,6 @@ static void ioapic_common_realize(DeviceState *dev, Error
**errp) >>>> { >>>> IOAPICCommonState *s = IOAPIC_COMMON(dev); >>>>
IOAPICCommonClass *info; >>>> - static int ioapic_no; >>>> >>>> if
(ioapic_no >= MAX_IOAPICS) { >>>> error_setg(errp, "Only %d ioapics
allowed", MAX_IOAPICS); >> ... while the check for max. IOAPICs still happens
in common code. >> >> Do we need to count KVM IOAPICs as well? Or can we
consolidate this into >> the non-KVM version and keep it static there? > > KVM
only supports one IOAPIC. Creating a second fails with EEXIST. As it turns
out, MAX_IOAPICS is 1, so covers both cases. No KVM ioctl actually happens on
realize for using said EEXIST, so I moved the increment back to where it was,
minimizing changes. If we want to redesign it, that can still be done in a
follow-up patch. Thanks, Andreas -- SUSE LINUX Products GmbH, Maxfeldstr. 5,
90409 Nürnberg, Germany GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB
16746 AG Nürnberg
