On Fri, Jun 08, 2018 at 09:24:49AM -0500, Eric Blake wrote: > On 06/07/2018 06:14 PM, Michael S. Tsirkin wrote: > > On Thu, Jun 07, 2018 at 04:31:09PM -0600, Ross Zwisler wrote: > > > This commit: > > > > > > commit aa78a16d8645 ("hw/i386: Rename 2.13 machine types to 3.0") > > > > > > updated the name used to create the q35 machine, which in turn changed the > > > SSDT table which is generated when we run "make check": > > > > > > acpi-test: Warning! SSDT mismatch. Actual [asl:/tmp/asl-QZDWJZ.dsl, > > > aml:/tmp/aml-T8JYJZ], Expected [asl:/tmp/asl-DTWVJZ.dsl, > > > aml:tests/acpi-test-data/q35/SSDT.dimmpxm]. > > > > > > Here's the only difference, aside from the checksum: > > > > > > < Name (MEMA, 0x07FFF000) > > > --- > > > > Name (MEMA, 0x07FFE000) > > > > Weird. How come the phys address changes just because of machine name? > > "2.13" is a different length than "3.0"; depending on whatever other > alignment coincidences or sharing of similar substrings are in place, this > obviously shuffled enough data that the one byte change then reflects into > an entire page boundary difference.
I don't think we expose the version number to guests though - do we? > -- > Eric Blake, Principal Software Engineer > Red Hat, Inc. +1-919-301-3266 > Virtualization: qemu.org | libvirt.org