Quoting Collin Walling (2018-06-27 19:03:46)
On 06/22/2018 12:42 AM, Chris Venteicher wrote:
> Baseline cpu model using QEMU/QMP query-cpu-model-baseline
>
> query-cpu-model-baseline only compares two CPUModels so multiple
> exchanges are needed to evaluate more than two CPUModels.
> ---
> src/qemu/qemu_capabilities.c | 89 ++++++++++++++++++++++++++++++++++++
> src/qemu/qemu_capabilities.h | 4 ++
> 2 files changed, 93 insertions(+)
>
> diff --git a/src/qemu/qemu_capabilities.c b/src/qemu/qemu_capabilities.c
> index 658f88b327..1646284588 100644
> --- a/src/qemu/qemu_capabilities.c
> +++ b/src/qemu/qemu_capabilities.c
> @@ -5394,3 +5394,92 @@ virQEMUCapsStripMachineAliases(virQEMUCapsPtr qemuCaps)
> for (i = 0; i < qemuCaps->nmachineTypes; i++)
> VIR_FREE(qemuCaps->machineTypes[i].alias);
> }
> +
> +
> +/* in:
> + * cpus[0]->model = "z13-base";
> + * cpus[0]->features[0].name = "xxx";
> + * cpus[0]->features[1].name = "yyy";
> + * ***
> + * cpus[n]->model = "s390x";
> + * cpus[n]->features[0].name = "xxx";
> + * cpus[n]->features[1].name = "yyy";
> + *
> + * out:
> + * *baseline->model = "s390x";
> + * *baseline->features[0].name = "yyy";
> + *
> + * (ret==0) && (*baseline == NULL) if QMP baseline not supported by QEMU
> + */
> +int
> +virQEMUCapsQMPBaselineCPUModel(virQEMUCapsInitQMPCommandPtr cmd,
> + virCPUDefPtr *cpus,
> + virCPUDefPtr *baseline)
> +{
> + qemuMonitorCPUModelInfoPtr model_baseline = NULL;
> + qemuMonitorCPUModelInfoPtr new_model_baseline = NULL;
> + qemuMonitorCPUModelInfoPtr model_b = NULL;
> + bool migratable_only = true;
> + int ret = -1;
> + size_t i;
> +
> + int ncpus = virCPUDefListLength(cpus);
> +
> + VIR_DEBUG("ncpus = %i", ncpus);
> +
> + *baseline = NULL;
> +
> + if (!cpus || !cpus[0] || !cpus[1]) {
> + virReportError(VIR_ERR_INVALID_ARG, "%s", _("less than 2
cpus"));
> + goto cleanup;
> + }
> +
> + VIR_DEBUG("cpu[0]->model = %s", NULLSTR(cpus[0]->model));
> +
> + if (!(model_baseline = virQEMUCapsCPUModelInfoFromCPUDef(cpus[0])))
> + goto cleanup;
>
This check duplicates check in for loop.
> +
> + for (i = 1; i < ncpus; i++) {
> + virCPUDefPtr cpu = cpus[i];
The list of cpus is null-terminated, so you should be able to loop while cpus[i] != NULL
and
axe the virCPUDefListLength function.
Yep. Can probably start with i=0 and init model_baseline in for loop too.
> +
> + VIR_DEBUG("cpu[%lu]->model = %s", i, NULLSTR(cpu->model));
> +
> + if (!(model_b = virQEMUCapsCPUModelInfoFromCPUDef(cpu)))
> + goto cleanup;
> +
This check duplicates check below.
> + if (!model_baseline || !model_b) {
> + virReportError(VIR_ERR_INVALID_ARG, "%s", _("cpu without
content"));
> + goto cleanup;
> + }
> +
> + if (qemuMonitorGetCPUModelBaseline(cmd->mon, model_baseline,
> + model_b, &new_model_baseline) <
0)
> + goto cleanup;
> +
> + if (!new_model_baseline) {
> + VIR_DEBUG("QEMU didn't support baseline");
> + ret = 0;
> + goto cleanup;
> + }
> +
> + qemuMonitorCPUModelInfoFree(model_baseline);
> + qemuMonitorCPUModelInfoFree(model_b);
> +
> + model_b = NULL;
> +
> + model_baseline = new_model_baseline;
> + }
> +
> + if (!(*baseline = virQEMUCapsCPUModelInfoToCPUDef(migratable_only,
model_baseline)))
> + goto cleanup;
> +
> + VIR_DEBUG("baseline->model = %s",
NULLSTR((*baseline)->model));
> +
> + ret = 0;
> +
> + cleanup:
> + qemuMonitorCPUModelInfoFree(model_baseline);
> + qemuMonitorCPUModelInfoFree(model_b);
> +
> + return ret;
> +}
> diff --git a/src/qemu/qemu_capabilities.h b/src/qemu/qemu_capabilities.h
> index 9a62b014ac..6098378f6c 100644
> --- a/src/qemu/qemu_capabilities.h
> +++ b/src/qemu/qemu_capabilities.h
> @@ -591,6 +591,10 @@ void virQEMUCapsFilterByMachineType(virQEMUCapsPtr qemuCaps,
> qemuMonitorCPUModelInfoPtr virQEMUCapsCPUModelInfoFromCPUDef(const virCPUDef
*cpuDef);
> virCPUDefPtr virQEMUCapsCPUModelInfoToCPUDef(bool migratable_only,
qemuMonitorCPUModelInfoPtr model);
>
> +int virQEMUCapsQMPBaselineCPUModel(virQEMUCapsInitQMPCommandPtr cmd,
> + virCPUDefPtr *cpus,
> + virCPUDefPtr *baseline);
> +
> virFileCachePtr virQEMUCapsCacheNew(const char *libDir,
> const char *cacheDir,
> uid_t uid,
>
--
Respectfully,
- Collin Walling