Re: [PATCH] sunrpc: Fix trace events to store data in the struct

[Steven Rostedt <rostedt@goodmis.org>]

On Tue, 24 Feb 2015 14:19:07 +0000
James Hogan <james.hogan@imgtec.com> wrote:

> Hi Steven,
> 
> On 24/02/15 14:09, Steven Rostedt wrote:
> > On Tue, 24 Feb 2015 11:47:56 +0000
> > James Hogan <james.hogan@imgtec.com> wrote:
> > 
> > 
> > 
> >>  	TP_printk("xprt=0x%p addr=%pIScp pid=%d flags=%s", __entry->xprt,
> >> -		(struct sockaddr *)&__entry->xprt->xpt_remote,
> > 
> > There's actually nothing wrong with the above even if xprt is NULL.
> > It's not dereferencing the structure, it is just getting the address of
> > what would be dereference.
> 
> I think that corresponds to the %pIScp format which I presumed does
> dereference the pointer?

Ah, I missed the "__entry->xprt" part :-p


> 
> Looking at Documentation/printk-formats.txt I see:
> 
> > IPv4/IPv6 addresses (generic, with port, flowinfo, scope):
> > ...
> >         %pISpc  1.2.3.4:12345   or [1:2:3:4:5:6:7:8]:12345
> 
> Same applies below. Should these formats still be avoided?

No, we can still use them.

I assume that the %pISpc expects a "struct sockaddr" passed to it as
that is what is typecast in the print. We might as well make the ss into
that structure instead of a struct sockaddr_storage, as it looks like
the storage one is much larger, and we only care about the sockaddr
part. Let's not waste the ring buffer if we don't need to.

-- Steve