Re: [PATCH v2] proc: fix /proc/*/map_files lookup some more

From: Alexey Dobriyan
Date: Wed Feb 21 2018 - 16:02:21 EST


On Wed, Feb 21, 2018 at 12:04:03PM -0800, Andrew Morton wrote:
> On Wed, 21 Feb 2018 22:53:40 +0300 Alexey Dobriyan <adobriyan@xxxxxxxxx> wrote:
>
> > I totally forgot that _parse_integer() accepts arbitrary amount of
> > leading zeroes leading to the following:
> >
> > OK
> > # readlink /proc/1/map_files/56427ecba000-56427eddc000
> > /lib/systemd/systemd
> >
> > bogus
> > # readlink /proc/1/map_files/00000000000056427ecba000-56427eddc000
> > /lib/systemd/systemd
> > # readlink /proc/1/map_files/56427ecba000-00000000000056427eddc000
> > /lib/systemd/systemd
> >
> > ...
> >
> > --- a/fs/proc/base.c
> > +++ b/fs/proc/base.c
> > @@ -1916,6 +1916,8 @@ static int dname_to_vma_addr(struct dentry *dentry,
> > unsigned long long sval, eval;
> > unsigned int len;
> >
> > + if (str[0] == '0' && str[1])
> > + return -EINVAL;
> > len = _parse_integer(str, 16, &sval);
> > if (len & KSTRTOX_OVERFLOW)
> > return -EINVAL;
> > @@ -1927,6 +1929,8 @@ static int dname_to_vma_addr(struct dentry *dentry,
> > return -EINVAL;
> > str++;
> >
> > + if (str[0] == '0' && str[1])
> > + return -EINVAL;
> > len = _parse_integer(str, 16, &eval);
> > if (len & KSTRTOX_OVERFLOW)
> > return -EINVAL;
>
> I don't know this code and I'm all confused.
>
> - why is the code designed to accept addresses of "0"?

Now I'm confused.
Code rejects, say ,'07ff...-...' because printing with %lx-%lx would never
produce leading zero.

> - how do we know that the first digit of a VMA address will never be 0?

Except when address is exactly 0 but this case is handled by looking at
the second character.