Re: [PATCH] vsprintf: Do not break early boot with probing addresses

On Tue 2019-05-14 14:37:51, Steven Rostedt wrote:

[ Purple is a nice shade on the bike shed. ;-) ]

On Tue, 14 May 2019 11:02:17 +0200
Geert Uytterhoeven [off-list ref] wrote:

quoted

On Tue, May 14, 2019 at 10:29 AM David Laight [off-list ref] wrote:

quoted

quoted

And I like Steven's "(fault)" idea.
How about this:

      if ptr < PAGE_SIZE              -> "(null)"
      if IS_ERR_VALUE(ptr)            -> "(fault)"

      -ss  

Or:
        if (ptr < PAGE_SIZE)
                return ptr ? "(null+)" : "(null)";

Hmm, that is useful.

quoted

quoted

        if IS_ERR_VALUE(ptr)
                return "(errno)"  

I still prefer "(fault)" as is pretty much all I would expect from a
pointer dereference, even if it is just bad parsing of, say, a parsing
an MAC address. "fault" is generic enough. "errno" will be confusing,
because that's normally a variable not a output.

quoted

Do we care about the value? "(-E%u)"?

That too could be confusing. What would (-E22) be considered by a user
doing an sprintf() on some string. I know that would confuse me, or I
would think that it was what the %pX displayed, and wonder why it
displayed it that way. Whereas "(fault)" is quite obvious for any %p
use case.

This discussion clearly shows that it is hard to make anyone happy.

I considered switching to "(fault)" because there seems to be more
people in favor of this.

But there is used also "(einval)" when an unsupported pointer
modifier is passed. The idea is to show error codes that people
are familiar with.

It might have been better to use the uppercase "(EFAULT)" and
"(EINVAL)" to make it more obvious. But I wanted to follow
the existing style with the lowercase "(null)".

As of now, I think that we should keep it as is unless there is
some wider agreement on a change.

Best Regards,
Petr