11:21 p.m.
On 2018-09-05 at 21:26, Erik Skultety wrote:
On Wed, Sep 05, 2018 at 04:36:27PM +0800, Shi Lei wrote:
the subject should read:
util: netlink: Introduce virNetlinkNewLink helper
> This patch introduces virNetlinkNewLink which wraps newlink code
> using libnl.
... virNetlinkNewLink helper which wraps the common libnl/netlink code to
create a new link.
Okay.
>
> Signed-off-by: Shi Lei <shi_lei(a)massclouds.com>
> ---
> src/libvirt_private.syms | 1 +
> src/util/virnetlink.c | 120 +++++++++++++++++++++++++++++++++++++--
> src/util/virnetlink.h | 13 +++++
> 3 files changed, 129 insertions(+), 5 deletions(-)
>
> diff --git a/src/libvirt_private.syms b/src/libvirt_private.syms
> index e2340d2..77c9b9e 100644
> --- a/src/libvirt_private.syms
> +++ b/src/libvirt_private.syms
> @@ -2447,6 +2447,7 @@ virNetlinkEventServiceStop;
> virNetlinkEventServiceStopAll;
> virNetlinkGetErrorCode;
> virNetlinkGetNeighbor;
> +virNetlinkNewLink;
> virNetlinkShutdown;
> virNetlinkStartup;
>
> diff --git a/src/util/virnetlink.c b/src/util/virnetlink.c
> index 8f06446..32aa62d 100644
> --- a/src/util/virnetlink.c
> +++ b/src/util/virnetlink.c
> @@ -420,10 +420,8 @@ virNetlinkDumpLink(const char *ifname, int ifindex,
> if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0)
> goto buffer_too_small;
>
> - if (ifname) {
> - if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0)
> - goto buffer_too_small;
> - }
> + if (ifname && nla_put_string(nl_msg, IFLA_IFNAME, ifname) < 0)
> + goto buffer_too_small;
unrelated to this patch...this kind of change should be a separate patch
somewhere at the beginning of the series...
Okay.
>
> # ifdef RTEXT_FILTER_VF
> /* if this filter exists in the kernel's netlink implementation,
> @@ -488,6 +486,118 @@ virNetlinkDumpLink(const char *ifname, int ifindex,
> }
>
>
> +/**
> + * virNetlinkNewLink:
> + *
> + * @ifname: Name of the link
> + * @type: The type of device, i.e., "bridge", "macvtap",
"macvlan"
s/Name/name
s/The/the
s/of device/of the device/
s/i.e.,/i.e.
Okay.
> + * @ifindex: The index for the 'link' device
I think ^this one should be part of the extra args, e.g. bridge doesn't use it
Okay.
> + * @data: The extra args for creating the netlink interface
actually extra_args is a better name for the argument :)
Yes. The extra_args is more clear.
> + * @error: for retrieving error code
s/for retrieving error code/netlink error code
Okay.
> + *
> + * Create a network "link" (aka interface aka device) with the given
> + * args. This works for many different types of network devices,
> + * including macvtap and bridges.
A generic wrapper to create a network link.
Okay.
> + *
> + * Returns 0 on success, -1 on fatal error.
> + */
> +int
> +virNetlinkNewLink(const char *ifname,
> + const char *type,
> + const int *ifindex,
> + virNetlinkNewLinkDataPtr data,
> + int *error)
> +{
> + struct nlmsgerr *err;
> + struct nlattr *linkinfo = NULL;
> + struct nlattr *infodata = NULL;
> + unsigned int buflen;
> + struct ifinfomsg ifinfo = { .ifi_family = AF_UNSPEC };
> + VIR_AUTOPTR(virNetlinkMsg) nl_msg = NULL;
> + VIR_AUTOFREE(struct nlmsghdr *) resp = NULL;
> +
> + *error = 0;
> +
> + if (!data)
> + return -1;
Extra args should not be required, BridgeCreate doesn't make use of them. That
said, I think it's reasonable to make @ifname and @type mandatory and fail
whenever we don't get either of them:
if (!ifname && !type)
virReportError(...);
Okay. I see.
> +
> + nl_msg = nlmsg_alloc_simple(RTM_NEWLINK,
> + NLM_F_REQUEST | NLM_F_CREATE | NLM_F_EXCL);
> + if (!nl_msg) {
> + virReportOOMError();
> + return -1;
> + }
> +
> + if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0)
> + goto buffer_too_small;
> +
> + if (ifindex && nla_put_u32(nl_msg, IFLA_LINK, *ifindex) < 0)
> + goto buffer_too_small;
> +
> + if (data->mac && nla_put(nl_msg, IFLA_ADDRESS, VIR_MAC_BUFLEN,
data->mac) < 0)
> + goto buffer_too_small;
The order of the data in the msg payload doesn't matter, right? If that's true,
we could move everything related to @extra_args and *not* related to nested
containers right before actually calling into virNetlinkCommand. If the order
does actually matter, then you can disregard this comment.
Right. I look through the rtnl_newlink function which uses nlmsg_parse which uses
nla_parse.
In nla_parse function, we can see that the order of the netlink attributes does *not*
matter.
So I could do as you said.
> +
> + if (ifname && nla_put_string(nl_msg, IFLA_IFNAME, ifname) < 0)
> + goto buffer_too_small;
> +
> + if (!(linkinfo = nla_nest_start(nl_msg, IFLA_LINKINFO)))
> + goto buffer_too_small;
> +
> + if (type && nla_put_string(nl_msg, IFLA_INFO_KIND, type) < 0)
> + goto buffer_too_small;
> +
> + if ((STREQ(type, "macvtap") || STREQ(type, "macvlan"))
&&
> + data->macvlan_mode && *data->macvlan_mode > 0) {
this check would get a bit uglier though:
if ((STREQ(type, "macvtap") || STREQ(type, "macvlan")) &&
extra_args &&
extra_args->macvlan_mode &&
*extra_args->macvlan_mode > 0)
Okay.
> + if (!(infodata = nla_nest_start(nl_msg, IFLA_INFO_DATA)))
> + goto buffer_too_small;
> +
> + if (nla_put_u32(nl_msg, IFLA_MACVLAN_MODE, *data->macvlan_mode) < 0)
> + goto buffer_too_small;
> +
> + nla_nest_end(nl_msg, infodata);
> + }
> +
> + nla_nest_end(nl_msg, linkinfo);
here you'd have:
if (extra_args) {
...
}
Okay. I see.
> +
> + if (virNetlinkCommand(nl_msg, &resp, &buflen, 0, 0, NETLINK_ROUTE, 0)
< 0)
> + return -1;
> +
> + if (buflen < NLMSG_LENGTH(0) || resp == NULL)
> + goto malformed_resp;
> +
> + switch (resp->nlmsg_type) {
> + case NLMSG_ERROR:
> + err = (struct nlmsgerr *)NLMSG_DATA(resp);
> + if (resp->nlmsg_len < NLMSG_LENGTH(sizeof(*err)))
> + goto malformed_resp;
> +
> + if (err->error < 0) {
> + *error = err->error;
> + return -1;
> + }
> + break;
> +
> + case NLMSG_DONE:
> + break;
> +
> + default:
> + goto malformed_resp;
> + }
> +
> + return 0;
> +
> + malformed_resp:
> + virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
> + _("malformed netlink response message"));
> + return -1;
> +
> + buffer_too_small:
> + virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
> + _("allocated netlink buffer is too small"));
> + return -1;
> +}
> +
> +
> /**
> * virNetlinkDelLink:
> *
> @@ -522,7 +632,7 @@ virNetlinkDelLink(const char *ifname, virNetlinkDelLinkFallback
fallback)
> if (nlmsg_append(nl_msg, &ifinfo, sizeof(ifinfo), NLMSG_ALIGNTO) < 0)
> goto buffer_too_small;
>
> - if (nla_put(nl_msg, IFLA_IFNAME, strlen(ifname)+1, ifname) < 0)
> + if (ifname && nla_put_string(nl_msg, IFLA_IFNAME, ifname) < 0)
> goto buffer_too_small;
...unrelated to this patch...
Okay.
>
> if (virNetlinkCommand(nl_msg, &resp, &recvbuflen, 0, 0,
> diff --git a/src/util/virnetlink.h b/src/util/virnetlink.h
> index 1e1e616..8163a81 100644
> --- a/src/util/virnetlink.h
> +++ b/src/util/virnetlink.h
> @@ -65,6 +65,19 @@ int virNetlinkDumpCommand(struct nl_msg *nl_msg,
> unsigned int protocol, unsigned int groups,
> void *opaque);
>
> +typedef struct _virNetlinkNewLinkData virNetlinkNewLinkData;
> +typedef virNetlinkNewLinkData *virNetlinkNewLinkDataPtr;
> +struct _virNetlinkNewLinkData {
> + const virMacAddr *mac; /* The MAC address of the device */
> + const uint32_t *macvlan_mode; /* The mode of macvlan */
this would contain const int *ifindex too...
Okay.
> +};
> +
> +int virNetlinkNewLink(const char *ifname,
> + const char *type,
> + const int *ifindex,
> + virNetlinkNewLinkDataPtr data,
> + int *error);
> +
> typedef int (*virNetlinkDelLinkFallback)(const char *ifname);
>
> int virNetlinkDelLink(const char *ifname, virNetlinkDelLinkFallback fallback);
The patch looks good, except for the few nitpicks I mentioned.
Erik
Thanks for your comments!
Shi Lei