[libvirt] [PATCH v3 1/2] Add VIR_TYPED_PARAM_STRING

[Eric Blake]

On 09/23/2011 12:40 AM, Hu Tao wrote:
> This makes string can be transported between client and server.
> For compatibility,
>
>      o new server should not send strings to old client if it
>        doesn't see the flag VIR_DOMAIN_TYPED_STRING_OKAY.
>      o new client that wants to be able to send/receive strings
>        should always set the flag VIR_DOMAIN_TYPED_STRING_OKAY;
>        if it is rejected by a old server that doesn't understand
>        VIR_DOMAIN_TYPED_STRING_OKAY, then the client should have
>        a second try with out flag VIR_DOMAIN_TYPED_STRING_OKAY
>        to cope with an old server.

These points should probably be documented in libvirt.h.in, as well.

I'm also thinking that libvirt.c should do some argument validation - 
for every API that uses a virTypedParameter array, it should validate 
that no VIR_TYPED_PARAM_STRING occurs if the flag is not present.

> +++ b/daemon/remote.c
> @@ -672,6 +672,15 @@ remoteSerializeTypedParameters(virTypedParameterPtr params,
>           case VIR_TYPED_PARAM_BOOLEAN:
>               val[i].value.remote_typed_param_value_u.b = params[i].value.b;
>               break;
> +        case VIR_TYPED_PARAM_STRING:
> +            if (params[i].value.s) {
> +                val[i].value.remote_typed_param_value_u.s = strdup(params[i].value.s);
> +                if (val[i].value.remote_typed_param_value_u.s == NULL) {
> +                    virReportOOMError();
> +                    goto cleanup;
> +                }
> +            }
> +            break;
>           default:

Memory leak.  The cleanup: label must free any successfully allocated 
remote_typed_param_value_u.s contents prior to the failure point.

>               virNetError(VIR_ERR_RPC, _("unknown parameter type: %d"),
>                           params[i].type);
> @@ -750,6 +759,14 @@ remoteDeserializeTypedParameters(remote_typed_param *args_params_val,
>               params[i].value.b =
>                   args_params_val[i].value.remote_typed_param_value_u.b;
>               break;
> +        case VIR_TYPED_PARAM_STRING:
> +            params[i].value.s =
> +                strdup(args_params_val[i].value.remote_typed_param_value_u.s);
> +            if (params[i].value.s == NULL) {
> +                virReportOOMError();
> +                goto cleanup;
> +            }
> +            break;
>           default:
>               virNetError(VIR_ERR_INTERNAL_ERROR, _("unknown parameter type: %d"),
>                           params[i].type);

Memory leak.  The cleanup: label must now iterate over params, and free 
any successfully allocated params[i].value.s allocated prior to a failure.

> diff --git a/include/libvirt/libvirt.h.in b/include/libvirt/libvirt.h.in
> index 39155a6..448a0e7 100644
> --- a/include/libvirt/libvirt.h.in
> +++ b/include/libvirt/libvirt.h.in
> @@ -205,6 +205,7 @@ typedef enum {
>       VIR_DOMAIN_AFFECT_CURRENT = 0,      /* Affect current domain state.  */
>       VIR_DOMAIN_AFFECT_LIVE    = 1<<  0, /* Affect running domain state.  */
>       VIR_DOMAIN_AFFECT_CONFIG  = 1<<  1, /* Affect persistent domain state.  */
> +    VIR_DOMAIN_TYPED_STRING_OKAY = 1<<  2,
>   } virDomainModificationImpact;

I'm not sure if this is the best place to stick this enum value; it 
might fit better if it is listed (and documented!) closer to the rest of 
virTypedParameterType stuff.  But I agree with setting it to the value 
of 1<<2, since all current clients of virTypedParameterType seem to be 
using VIR_DOMAIN_AFFECT_* and nothing further.

>
>   /**
> @@ -489,7 +490,8 @@ typedef enum {
>       VIR_TYPED_PARAM_LLONG   = 3, /* long long case */
>       VIR_TYPED_PARAM_ULLONG  = 4, /* unsigned long long case */
>       VIR_TYPED_PARAM_DOUBLE  = 5, /* double case */
> -    VIR_TYPED_PARAM_BOOLEAN = 6  /* boolean(character) case */
> +    VIR_TYPED_PARAM_BOOLEAN = 6, /* boolean(character) case */
> +    VIR_TYPED_PARAM_STRING  = 7  /* string case */

Put a trailing comma after 7, so that the next addition (if any) doesn't 
have to touch two lines like this addition did.

>   } virTypedParameterType;
>
>   /**
> @@ -520,6 +522,7 @@ struct _virTypedParameter {
>           unsigned long long int ul;  /* type is ULLONG */
>           double d;                   /* type is DOUBLE */
>           char b;                     /* type is BOOLEAN */
> +        char *s;                    /* type is STRING */

Here, I'd use:

char *s; /* type is STRING, see also VIR_DOMAIN_TYPED_STRING_OKAY */

>       } value; /* parameter value */
>   };
>
> diff --git a/src/remote/remote_driver.c b/src/remote/remote_driver.c
> index 1217d94..85eaeea 100644
> --- a/src/remote/remote_driver.c
> +++ b/src/remote/remote_driver.c
> @@ -1276,6 +1276,15 @@ remoteSerializeTypedParameters(virTypedParameterPtr params,
>           case VIR_TYPED_PARAM_BOOLEAN:
>               val[i].value.remote_typed_param_value_u.b = params[i].value.b;
>               break;
> +        case VIR_TYPED_PARAM_STRING:
> +            if (params[i].value.s) {
> +                val[i].value.remote_typed_param_value_u.s = strdup(params[i].value.s);
> +                if (val[i].value.remote_typed_param_value_u.s == NULL) {
> +                    virReportOOMError();
> +                    goto cleanup;
> +                }
> +            }
> +            break;

Counterpart memory leak.

>           default:
>               remoteError(VIR_ERR_RPC, _("unknown parameter type: %d"),
>                   params[i].type);
> @@ -1347,6 +1356,14 @@ remoteDeserializeTypedParameters(remote_typed_param *ret_params_val,
>               params[i].value.b =
>                   ret_params_val[i].value.remote_typed_param_value_u.b;
>               break;
> +        case VIR_TYPED_PARAM_STRING:
> +            params[i].value.s =
> +                strdup(ret_params_val[i].value.remote_typed_param_value_u.s);
> +            if (params[i].value.s == NULL) {
> +                virReportOOMError();
> +                goto cleanup;
> +            }
> +            break;

And another memory leak.

-- 
Eric Blake   eblake at redhat.com    +1-801-349-2682
Libvirt virtualization library http://libvirt.org