On Sat, Jun 24, 2017 at 01:43:14PM +0800, JingPiao Chen wrote:
> On Sat, Jun 24, 2017 at 04:36:34AM +0300, Dmitry V. Levin wrote:
> > On Sat, Jun 24, 2017 at 09:06:06AM +0800, JingPiao Chen wrote:
> > > static int
> > > print_nlmsghdr(struct tcb *tcp,
> > > const int fd,
> > > int family,
> > > const struct nlmsghdr *const nlmsghdr)
> > > {
> > > ...
> > > const int hdr_family = (nlmsghdr->nlmsg_type < NLMSG_MIN_TYPE)
> > > ? NL_FAMILY_DEFAULT
> > > : (family != NL_FAMILY_DEFAULT
> > > ? family : get_fd_nl_family(tcp, fd));
> >
> > Feel free to change this if NLMSG_DONE needs a family specific decoding.
>
> I have send a patch. I've also thought about rewriting NLMSG_DONE,
> because only three family send useful data: drivers/connector/connector.c,
> drivers/infiniband/core/iwpm_util.c and net/netfilter/nfnetlink_log.c.
> In the end, I think it's good enough, but I want to ask your opinion.
>
> $ git grep "NLMSG_DONE"
>
> drivers/connector/connector.c: nlh = nlmsg_put(skb, 0, msg->seq,
> NLMSG_DONE, size, 0);
> drivers/infiniband/core/iwpm_util.c: nlh->nlmsg_type = NLMSG_DONE;
>
> These two family I do not understand what they send.
drivers/connector/connector.c:cn_netlink_send_mult() sends struct cn_msg
followed by optional data.
drivers/infiniband/core/iwpm_util.c:send_nlmsg_done() sends no data.
> net/netfilter/nfnetlink_log.c:
> static void
> __nfulnl_send(struct nfulnl_instance *inst)
> {
> if (inst->qlen > 1) {
> struct nlmsghdr *nlh = nlmsg_put(inst->skb, 0, 0,
> NLMSG_DONE,
> sizeof(struct nfgenmsg),
> 0);
> if (WARN_ONCE(!nlh, "bad nlskb size: %u, tailroom %d\n",
> inst->skb->len, skb_tailroom(inst->skb))) {
> kfree_skb(inst->skb);
> goto out;
> }
> }
> nfnetlink_unicast(inst->skb, inst->net, inst->peer_portid,
> MSG_DONTWAIT);
> out:
> inst->qlen = 0;
> inst->skb = NULL;
> }
>
> It allocate struct nfgenmsg, but do not initialize it.
Yes, that looks odd.
> Following family data length is zero. Default decoder can handle them.
>
> drivers/net/team/team.c: nlh = nlmsg_put(skb, portid, seq, NLMSG_DONE, 0,
> flags | NLM_F_MULTI);
>
> drivers/scsi/scsi_transport_iscsi.c: int t = done ? NLMSG_DONE : type;
> static void
> iscsi_if_rx(struct sk_buff *skb)
> {
> ...
> err = iscsi_if_send_reply(group, nlh->nlmsg_seq,
> nlh->nlmsg_type, 0, 0, ev, sizeof(*ev));
> ...
> }
>
> kernel/audit.c: int t = done ? NLMSG_DONE : type;
> kernel/auditfilter.c:
> static void audit_list_rules(int seq, struct sk_buff_head *q)
> {
> ...
> skb = audit_make_reply(seq, AUDIT_LIST_RULES, 1, 1, NULL, 0);
> ...
> }
>
> net/core/devlink.c: NLMSG_DONE, 0, flags | NLM_F_MULTI);
> net/core/devlink.c: NLMSG_DONE, 0, flags | NLM_F_MULTI);
> net/core/devlink.c: NLMSG_DONE, 0, flags | NLM_F_MULTI);
Yes, besides drivers/connector/connector.c, looks like the current parser
handles NLMSG_DONE well.
--
ldv
signature.asc
Description: PGP signature
------------------------------------------------------------------------------ Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot
_______________________________________________ Strace-devel mailing list Strace-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/strace-devel
