[PATCHv4 5/6] symbol lookup: introduce dereference_symbol_descriptor()
Sergey Senozhatsky
sergey.senozhatsky.work at gmail.com
Wed Dec 6 21:46:56 AEDT 2017
On (12/06/17 11:32), Petr Mladek wrote:
[..]
> > diff --git a/Documentation/printk-formats.txt b/Documentation/printk-formats.txt
> > index aa0a776c817a..02745028e909 100644
> > --- a/Documentation/printk-formats.txt
> > +++ b/Documentation/printk-formats.txt
> > @@ -61,41 +61,31 @@ Symbols/Function Pointers
> >
> > ::
> >
> > - %pF versatile_init+0x0/0x110
> > - %pf versatile_init
> > - %pS versatile_init+0x0/0x110
> > - %pSR versatile_init+0x9/0x110
> > + %pS versatile_init+0x0/0x110
> > + %ps versatile_init
> > + %pF versatile_init+0x0/0x110
> > + %pf versatile_init
> > + %pSR versatile_init+0x9/0x110
> > (with __builtin_extract_return_addr() translation)
> > - %ps versatile_init
> > - %pB prev_fn_of_versatile_init+0x88/0x88
> > + %pB prev_fn_of_versatile_init+0x88/0x88
>
> I was curious why so many lines were changed here. You converted
> the 2nd tab to spaces. I put back the tab. The result is:
ew... how did that happen. thanks for fixing up.
> > +static inline void *dereference_symbol_descriptor(void *ptr)
> > +{
> > +#ifdef HAVE_DEREFERENCE_FUNCTION_DESCRIPTOR
> > + struct module *mod;
> > +
> > + ptr = dereference_kernel_function_descriptor(ptr);
> > + if (is_ksym_addr((unsigned long)ptr))
> > + return ptr;
> > +
> > + preempt_disable();
> > + mod = __module_address((unsigned long)ptr);
> > + preempt_enable();
> > +
> > + if (mod)
> > + ptr = dereference_module_function_descriptor(mod, ptr);
> > +#endif
> > + return ptr;
> > +}
>
> It is a bit too long for an inline function but I did not find a
> better solution. It should always be defined and all suitable
> .c files are compiled only under certain configuration. Well,
> it is a nop on most architectures.
or we can move dereference_symbol_descriptor() to vsprintf.c,
since all the functions it depends on are now available either
as exported symbols or via kallsyms header file. not that it
annoys me, so we can keep it as it is.
-ss
More information about the Linuxppc-dev
mailing list