Re: [RFC PATCH v6 4/6] tracing/probe: Support user-space dereference

[Masami Hiramatsu <mhiramat@kernel.org>]

On Wed, 8 May 2019 11:22:37 -0400
Steven Rostedt <rostedt@goodmis.org> wrote:

> > > >  Per-Probe Event Filtering
> > > >  -------------------------
> > > > diff --git a/Documentation/trace/uprobetracer.rst b/Documentation/trace/uprobetracer.rst
> > > > index 4346e23e3ae7..de8812c932bc 100644
> > > > --- a/Documentation/trace/uprobetracer.rst
> > > > +++ b/Documentation/trace/uprobetracer.rst
> > > > @@ -42,16 +42,17 @@ Synopsis of uprobe_tracer
> > > >     @+OFFSET	: Fetch memory at OFFSET (OFFSET from same file as PATH)
> > > >     $stackN	: Fetch Nth entry of stack (N >= 0)
> > > >     $stack	: Fetch stack address.
> > > > -   $retval	: Fetch return value.(*)
> > > > +   $retval	: Fetch return value.(\*1)
> > > >     $comm	: Fetch current task comm.
> > > > -   +|-offs(FETCHARG) : Fetch memory at FETCHARG +|- offs address.(**)
> > > > +   +|-[u]OFFS(FETCHARG) : Fetch memory at FETCHARG +|- OFFS address.(\*2)(\*3)
> > > >     NAME=FETCHARG     : Set NAME as the argument name of FETCHARG.
> > > >     FETCHARG:TYPE     : Set TYPE as the type of FETCHARG. Currently, basic types
> > > >  		       (u8/u16/u32/u64/s8/s16/s32/s64), hexadecimal types
> > > >  		       (x8/x16/x32/x64), "string" and bitfield are supported.  
> > > 
> > > Hmm, shouldn't uprobes default to userspace. Isn't the purpose mostly
> > > to find out what's going on in userspace. Perhaps we should add a 'k'
> > > annotation to uprobes to denote that it's for kernel space, as that
> > > should be the exception and not the norm.  
> > 
> > No, uprobe can not access kernel space, because it doesn't have the
> > current kernel context. Note that all registers, stacks which
> > can be accessed from uprobe handler are user-space. We can not access
> > kernel context from that. See below
> > 
> > > > -  (*) only for return probe.
> > > > -  (**) this is useful for fetching a field of data structures.
> > > > +  (\*1) only for return probe.
> > > > +  (\*2) this is useful for fetching a field of data structures.
> > > > +  (\*3) Unlike kprobe event, "u" prefix will just be ignored.  
> > 
> > Thus the 'u' is just ignored on uprobe event.
> 
> I totally missed the footnote here. Can we stress this point more up in
> the "User Memory Access" section. Specifically state something like:
> "Uprobes only access userspace memory, thus the 'u' is not required,
> and if it is added to a uprobe, it will simply be ignored".

Sorry, I missed this mail. 

Since the "User Memory Access" section is only in kprobetrace.rst, I think
mentioning uprobe-events in kprobetrace.rst is meaningless. Uprobe user
might read uprobetracer.rst instead of kprobetrace.rst.
So I think it is enough to mention it as a footnote in uprobetracer.rst.

Thank you,

-- 
Masami Hiramatsu <mhiramat@kernel.org>