Re: Writing to a const pointer: is this supposed to happen?

[Greg Kroah-Hartman <gregkh@linuxfoundation.org>]

On Wed, Jun 24, 2020 at 02:34:45PM +0200, Kars Mulder wrote:
> On Tuesday, June 23, 2020 21:55 CEST, Pavel Machek wrote:
> > Odd, indeed... but not likely to cause immediate problems.
> >
> > You may want to cc relevant maintainers, or even run git
> > blame and contact author.
> 
> Thank you for your response.
> 
> The code was written by Kai-Heng Feng, whom I shall CC. The code is
> part of the usbcore module, which does not have a maintainer listed in
> MAINTAINERS, but the patch and most other recent patches to usbcore
> were signed off exclusively by Greg Kroah-Hartman, so I guess that
> makes him the de facto maintainer? I'll CC him as well.
> 
> I'm not sure whether it is easy to read the previous messages of this
> thread if you got CC'ed just now, so I'll repeat/paraphrase the
> important part of my initial mail for your convenience:
> 
> > In the file drivers/usb/core/quirks.c, I noticed that the function
> > quirks_param_set writes to a const pointer, and would like to check
> > whether this is ok with the kernel programming practices. Here are
> > the relevant lines from the function (several lines omitted):
> >
> > 	static int quirks_param_set(const char *val, const struct kernel_param *kp) {
> > 		char *p, *field;
> > 		for (i = 0, p = (char *)val; p && *p;) {
> > 			field = strsep(&p, ":");
> >
> > In here a const pointer *val is cast into a non-const pointer and
> > then written to by the function strsep, which replaces the first
> > occurrence of the ':' token with a null-byte. Is this allowed?
> 
> CC: Kai-Heng Feng <kai.heng.feng@canonical.com>
> CC: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

It's not the nicest thing, but as you have noticed, it all works just
fine, right?

Have you hit any runtime issues with this code doing this?  These
strings should be held in writable memory, right?  Or do you see a
codepath where that is not the case?  If so, please feel free to submit
a patch to fix this up (and probably fix up a number of other "set"
functions that deal with struct kernel_param as well.)

thanks,

greg k-h