On Thu, Jun 09, 2016 at 12:25:26 +0200, Jiri Denemark wrote:
On Thu, Jun 09, 2016 at 12:12:19 +0200, Peter Krempa wrote:
On Thu, Jun 09, 2016 at 11:56:17 +0200, Jiri Denemark wrote:
Our current detection code uses just the number of CPU features which need to be added/removed from the CPU model to fully describe the CPUID data. The smallest number wins. But this may sometimes generate wrong results as one can see from the fixed test cases. This patch modifies the algorithm to prefer the CPU model with matching signature even if this model results in a longer list of additional features.
Signed-off-by: Jiri Denemark <jdenemar@redhat.com> --- src/cpu/cpu_map.xml | 16 +++ src/cpu/cpu_x86.c | 148 +++++++++++++++++++-- .../cputestdata/x86-cpuid-Core-i7-5600U-guest.xml | 11 +- tests/cputestdata/x86-cpuid-Core2-E6850-guest.xml | 5 +- tests/cputestdata/x86-cpuid-Core2-E6850-host.xml | 5 +- tests/cputestdata/x86-cpuid-Xeon-5110-guest.xml | 5 +- tests/cputestdata/x86-cpuid-Xeon-5110-host.xml | 5 +- tests/cputestdata/x86-host+penryn-force-result.xml | 4 +- 8 files changed, 176 insertions(+), 23 deletions(-)
ACK
Thanks, I pushed all but 8/10 and 10/10 which need to wait until Eduardo's queue is pulled into QEMU master.
Since Skylake-Client was pushed to QEMU master yesterday I pushed the remaining two patches in this series. Jirka