Re: [PATCH 2/2] ocfs2: Replace strlcpy with strscpy

From: Kees Cook
Date: Thu Aug 31 2023 - 17:04:28 EST


On Thu, Aug 31, 2023 at 03:28:32PM -0400, Azeem Shaikh wrote:
> On Wed, Aug 30, 2023 at 7:06 PM Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> >
> > On Wed, Aug 30, 2023 at 09:54:26PM +0000, Azeem Shaikh wrote:
> > > strlcpy() reads the entire source buffer first.
> > > This read may exceed the destination size limit.
> > > This is both inefficient and can lead to linear read
> > > overflows if a source string is not NUL-terminated [1].
> > > In an effort to remove strlcpy() completely [2], replace
> > > strlcpy() here with strscpy().
> > >
> > > Direct replacement is assumed to be safe here since
> > > it's ok for `kernel_param_ops.get()` to return -errno [3].
> > > This changes the behavior such that instead of silently ignoring the
> > > case when sizeof(@buffer) < DLMFS_CAPABILITIES, we now return error.
> > >
> > > [1] https://www.kernel.org/doc/html/latest/process/deprecated.html#strlcpy
> > > [2] https://github.com/KSPP/linux/issues/89
> > > [3] https://elixir.bootlin.com/linux/v6.5/source/include/linux/moduleparam.h#L52
> > >
> > > Signed-off-by: Azeem Shaikh <azeemshaikh38@xxxxxxxxx>
> > > ---
> > > fs/ocfs2/dlmfs/dlmfs.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/fs/ocfs2/dlmfs/dlmfs.c b/fs/ocfs2/dlmfs/dlmfs.c
> > > index 33e529de93b2..b001eccdd2f3 100644
> > > --- a/fs/ocfs2/dlmfs/dlmfs.c
> > > +++ b/fs/ocfs2/dlmfs/dlmfs.c
> > > @@ -80,7 +80,7 @@ static int param_set_dlmfs_capabilities(const char *val,
> > > static int param_get_dlmfs_capabilities(char *buffer,
> > > const struct kernel_param *kp)
> > > {
> > > - return strlcpy(buffer, DLMFS_CAPABILITIES,
> > > + return strscpy(buffer, DLMFS_CAPABILITIES,
> > > strlen(DLMFS_CAPABILITIES) + 1);
> > > }
> >
> > This is another case of "accidentally correct".
> >
> >
> > param->get() is hooked here, in the sysfs "show" callback:
> >
> > static ssize_t param_attr_show(struct module_attribute *mattr,
> > struct module_kobject *mk, char *buf)
> > {
> > int count;
> > struct param_attribute *attribute = to_param_attr(mattr);
> >
> > if (!attribute->param->ops->get)
> > return -EPERM;
> >
> > kernel_param_lock(mk->mod);
> > count = attribute->param->ops->get(buf, attribute->param);
> > kernel_param_unlock(mk->mod);
> > return count;
> > }
> >
> > Meaning ultimately this will show up here, if I'm reading names right:
> > /sys/module/ocfs/parameters/dlmfs_capabilities
> >
> > Anyway, the "count" being returned would be quite bad if
> > DLMFS_CAPABILITIES were dynamic and larger than PAGE_SIZE (the size of
> > the sysfs buffer).
> >
> > For this case, I would say replace strlcpy with sysfs_emit:
> >
> > return sysfs_emit(buffer, DLMFS_CAPABILITIES);
> >
>
> Thanks, sending out a v2 for this. Out of curiosity - why sysfs_emit?
> Is it because DLMFS_CAPABILITIES is a hard-coded string?

It's basically that sysfs_emit() Does The Right Thing for these
callbacks: it tracks the offset of the buffer and checks for buffer
overflow. (All the sysfs callback work on a page-aligned page-sized
buffer, and the API lacks any length info -- it's "understood" to be
page-sized.) So, since this is doing a string copy into what is
ultimately a sysfs buffer, it's best to use sysfs_emit().

--
Kees Cook