Re: [PATCH v4 6/9] vsprintf: Factor out %pV handler as va_format()

[Petr Mladek]

On Fri 2018-04-06 07:19:15, Joe Perches wrote:
> On Fri, 2018-04-06 at 15:12 +0200, Petr Mladek wrote:
> > On Wed 2018-04-04 07:26:07, Joe Perches wrote:
> > > On Wed, 2018-04-04 at 10:58 +0200, Petr Mladek wrote:
> > > > Move the code from the long pointer() function. We are going to add a check
> > > > for the access to the address that will make it even more complicated.
> > > > 
> > > > This patch does not change the existing behavior.
> > > 
> > > But it might increase stack consumption.
> > > 
> > > As the %pV is recursive, this is may not be a good thing.
> > 
> > It seems to be safe to pass just a pointer to struct printf_spec.
> > In fact, it would make sense to use this also in string() and
> > __string() calls. Copying 64 bytes many times look useless.
> 
> huh?
> 
> struct printf_spec is 64 bits, the same size as a
> pointer on 64 bit systems.

Yes, it was a nonsense. It was late Friday here when I wrote it.


> I'm dubious about this entire patch series.

I would appreciate some constructive feedback.
What exactly is wrong? What you would suggest instead?

Best Regards,
Petr