On 07/30/2013 03:59 PM, Eric Blake wrote:
On 07/30/2013 03:09 PM, Don Dugger wrote:
[V2 - per review comments change the flag name to be more descriptive and change errors from warnings to propogated errors.]
This aside belongs...
Currently the virConnectBaselineCPU API does not expose the CPU features that are part of the CPU's model. This patch adds a new flag, VIR_CONNECT_BASELINE_CPU_EXPAND_FEATURE, that causes the API to explictly list all features that are part of that model.
Signed-off-by: Don Dugger <donald.d.dugger@intel.com> ---
...here, after the '---', so that 'git am' won't turn it into a permanent part of libvirt.git history. Your patch didn't apply as is (are you properly rebasing to the latest libvirt.git?), but the conflict resolution in qemu_driver.c seemed pretty trivial. You also failed 'make syntax-check':
You also failed regular 'make': CC libvirt_cpu_la-cpu_x86.lo cpu/cpu_x86.c: In function 'x86DecodeCPUData': cpu/cpu_x86.c:1467:5: error: too few arguments to function 'x86Decode' return x86Decode(cpu, data->data.x86, models, nmodels, preferred); ^ cpu/cpu_x86.c:1356:1: note: declared here x86Decode(virCPUDefPtr cpu, ^ cpu/cpu_x86.c: At top level: cpu/cpu_x86.c:1948:5: error: initialization from incompatible pointer type [-Werror] .decode = x86DecodeCPUData, ^ cpu/cpu_x86.c:1948:5: error: (near initialization for 'cpuDriverX86.decode') [-Werror] cpu/cpu_x86.c: In function 'x86DecodeCPUData': cpu/cpu_x86.c:1468:1: error: control reaches end of non-void function [-Werror=return-type] } ^ -- Eric Blake eblake redhat com +1-919-301-3266 Libvirt virtualization library http://libvirt.org