Re: [PATCH net 2/2 RESEND] ipmr: Add lockdep expression to ipmr_for_each_table macro

[Jakub Kicinski]

On Tue, 12 May 2020 22:47:10 +0530 Amol Grover wrote:
> > > This is a strange condition, IMHO. How can we be fine with either
> > > lock.. This is supposed to be the writer side lock, one can't have 
> > > two writer side locks..
> > > 
> > > I think what is happening is this:
> > > 
> > > ipmr_net_init() -> ipmr_rules_init() -> ipmr_new_table()
> > > 
> > > ipmr_new_table() returns an existing table if there is one, but
> > > obviously none can exist at init.  So a better fix would be:
> > > 
> > > #define ipmr_for_each_table(mrt, net)					\
> > > 	list_for_each_entry_rcu(mrt, &net->ipv4.mr_tables, list,	\
> > > 				lockdep_rtnl_is_held() ||		\
> > > 				list_empty(&net->ipv4.mr_tables))
> > >  
> 
> Jakub, I agree, this condition looks better (and correct) than the one I
> proposed. I'll do the changes as necessary. Also, do you want me to add
> the full trace to the git commit body as well? I omitted it on purpose
> to not make it messy.

In this case we can leave it at the depth of IPMR code + the caller, so:

[    1.534758]  ? ipmr_get_table+0x3c/0x70
[    1.535430]  ? ipmr_new_table+0x1c/0x60
[    1.536173]  ? ipmr_net_init+0x7b/0x170
[    1.536923]  ? register_pernet_subsys+0xd/0x30

This makes it clear that the problem happens at net namespace init.

Thanks!