On Sat 2018-04-07 17:26:40, Andy Shevchenko wrote:
> On Wed, 2018-04-04 at 10:58 +0200, Petr Mladek wrote:
> > There are few printk formats that make sense only with two or more
> > specifiers. Also some specifiers make sense only when a kernel feature
> > is enabled.
> > 
> > The handling of unknown specifiers is strange, inconsistent, and
> > even leaking the address. For example, netdev_bits() prints the
> > non-hashed pointer value or clock() prints "(null)".
> > 
> > The best solution seems to be in flags_string(). It does not print any
> > misleading value. Instead it calls WARN_ONCE() describing the unknown
> > specifier. Therefore it clearly shows the problem and helps to find
> > it.
> > 
> > Note that WARN_ONCE() used to cause recursive printk(). But it is safe
> > now because vscnprintf() is called in printk_safe context from
> > vprintk_emit().
> > 
> 
> > -   if (!IS_ENABLED(CONFIG_HAVE_CLK) || !clk)
> > +   if (!IS_ENABLED(CONFIG_HAVE_CLK)) {
> > +           WARN_ONCE(1, "Unsupported pointer format specifier:
> > %%pC\n");
> > +           return buf;
> > +   }
> > +
> > +   if (!clk)
> >             return string(buf, end, NULL, spec);
> 
> This change collides with my patch series. Can you elaborate what your
> thoughts are about my patches? Are you going incorporate them to your
> series? Should I send them independently?

Good question. I think that the best solution will be that I go
over your patchset and just add all valid ones into printk.git
for-4.18. Then I will base v5 of this patchset on top of it.

I should have done this earlier. But I did not expect that long
way for the access-check stuff. We originally planned to
do the access check first, see
https://lkml.kernel.org/r/1520000254.10722.389.ca...@linux.intel.com
But the access check patchset still need some love, so it makes
sense to switch the order.

Best Regards,
Petr

Reply via email to