Re: [libvirt PATCH v2 01/20] cpu_x86: Simplify x86ParseCPUID

On 11/4/21 5:27 PM, Tim Wiederhake wrote: > ... by using virXMLProp*() helpers. These only require a xmlNodePtr > and > do not need a xmlXPathContextPtr. Reflect that in the function > signature. > > Signed-off-by: Tim Wiederhake <twiederh(a)redhat.com&gt; > --- >  src/cpu/cpu_x86.c | 43 +++++++++++++++++-------------------------- >  1 file changed, 17 insertions(+), 26 deletions(-) > > diff --git a/src/cpu/cpu_x86.c b/src/cpu/cpu_x86.c > index 1b829e5658..b0f6fa8f34 100644 > --- a/src/cpu/cpu_x86.c > +++ b/src/cpu/cpu_x86.c > @@ -1021,37 +1021,28 @@ x86FeatureNames(virCPUx86Map *map, >   >   >  static int > -x86ParseCPUID(xmlXPathContextPtr ctxt, > +x86ParseCPUID(xmlNodePtr node, >                virCPUx86DataItem *item) >  { > -    virCPUx86CPUID *cpuid; > -    unsigned long eax_in, ecx_in; > -    unsigned long eax, ebx, ecx, edx; > -    int ret_eax_in, ret_ecx_in, ret_eax, ret_ebx, ret_ecx, > ret_edx; > - > -    memset(item, 0, sizeof(*item)); > +    virCPUx86CPUID cpuid; >   > -    eax_in = ecx_in = 0; > -    eax = ebx = ecx = edx = 0; > -    ret_eax_in = virXPathULongHex("string(@eax_in)", ctxt, > &eax_in); > -    ret_ecx_in = virXPathULongHex("string(@ecx_in)", ctxt, > &ecx_in); > -    ret_eax = virXPathULongHex("string(@eax)", ctxt, &eax); > -    ret_ebx = virXPathULongHex("string(@ebx)", ctxt, &ebx); > -    ret_ecx = virXPathULongHex("string(@ecx)", ctxt, &ecx); > -    ret_edx = virXPathULongHex("string(@edx)", ctxt, &edx); > +    memset(&cpuid, 0, sizeof(cpuid)); Alternatively, let compiler clear the variable out in variable declaration:     virCPUx86CPUID cpuid = { 0 };

> > -    if (ret_eax_in < 0 || ret_ecx_in == -2 || > -        ret_eax == -2 || ret_ebx == -2 || ret_ecx == -2 || ret_edx > == -2) > +    if (virXMLPropUInt(node, "eax_in", 0, VIR_XML_PROP_REQUIRED, > &cpuid.eax_in) < 0) > +        return -1; > +    if (virXMLPropUInt(node, "ecx_in", 0, VIR_XML_PROP_NONE, > &cpuid.ecx_in) < 0) > +        return -1; > +    if (virXMLPropUInt(node, "eax", 0, VIR_XML_PROP_NONE, > &cpuid.eax) < 0) > +        return -1; > +    if (virXMLPropUInt(node, "ebx", 0, VIR_XML_PROP_NONE, > &cpuid.ebx) < 0) > +        return -1; > +    if (virXMLPropUInt(node, "ecx", 0, VIR_XML_PROP_NONE, > &cpuid.ecx) < 0) > +        return -1; > +    if (virXMLPropUInt(node, "edx", 0, VIR_XML_PROP_NONE, > &cpuid.edx) < 0) >          return -1; >   >      item->type = VIR_CPU_X86_DATA_CPUID; > -    cpuid = &item->data.cpuid; > -    cpuid->eax_in = eax_in; > -    cpuid->ecx_in = ecx_in; > -    cpuid->eax = eax; > -    cpuid->ebx = ebx; > -    cpuid->ecx = ecx; > -    cpuid->edx = edx; > +    item->data.cpuid = cpuid; >      return 0; >  } >   > @@ -1122,7 +1113,7 @@ x86FeatureParse(xmlXPathContextPtr ctxt, >      for (i = 0; i < n; i++) { >          ctxt->node = nodes[i]; >          if (virXMLNodeNameEqual(nodes[i], "cpuid")) { > -            if (x86ParseCPUID(ctxt, &item) < 0) { > +            if (x86ParseCPUID(ctxt->node, &item) < 0) { I believe nodes[i] can be passed directly. Because in the next patch you rework the else branch and if that used nodes[i] too then 'ctxt- >node = nodes[i]' line could be dropped. >                  virReportError(VIR_ERR_INTERNAL_ERROR, >                                 _("Invalid cpuid[%zu] in %s > feature"), >                                 i, feature->name); > @@ -1812,7 +1803,7 @@ virCPUx86DataParse(xmlXPathContextPtr ctxt) >      for (i = 0; i < n; i++) { >          ctxt->node = nodes[i]; >          if (virXMLNodeNameEqual(nodes[i], "cpuid")) { > -            if (x86ParseCPUID(ctxt, &item) < 0) { > +            if (x86ParseCPUID(ctxt->node, &item) < 0) { >                  virReportError(VIR_ERR_INTERNAL_ERROR, >                                 _("failed to parse cpuid[%zu]"), > i); >                  return NULL; > Same here. But looking more into the future, doesn't matter really, because in pach 03/20 you replace all of this with a function and use xmlNode directly. Michal