Re: [PATCH] proc: Avoid mixing integer types in mem_rw()

[Marcelo Henrique Cerri <marcelo.cerri@canonical.com>]

[-- Attachment #1: Type: text/plain, Size: 1312 bytes --]

On Wed, May 12, 2021 at 04:19:26PM +0300, Alexey Dobriyan wrote:
> On Wed, May 12, 2021 at 09:52:12AM -0300, Marcelo Henrique Cerri wrote:
> > Use size_t when capping the count argument received by mem_rw(). Since
> > count is size_t, using min_t(int, ...) can lead to a negative value
> > that will later be passed to access_remote_vm(), which can cause
> > unexpected behavior.
> > 
> > Since we are capping the value to at maximum PAGE_SIZE, the conversion
> > from size_t to int when passing it to access_remote_vm() as "len"
> > shouldn't be a problem.
> 
> > --- a/fs/proc/base.c
> > +++ b/fs/proc/base.c
> > @@ -854,7 +854,7 @@ static ssize_t mem_rw(struct file *file, char __user *buf,
> >  	flags = FOLL_FORCE | (write ? FOLL_WRITE : 0);
> >  
> >  	while (count > 0) {
> > -		int this_len = min_t(int, count, PAGE_SIZE);
> > +		size_t this_len = min_t(size_t, count, PAGE_SIZE);
> 
> As much as I don't like signed integers, VFS caps read/write lengths
> at INT_MAX & PAGE_MASK, so casting doesn't change values.

Although that should always be true we had a recent example of a
caller that wasn't properly capping it and since count is unsigned
anyway it makes sense to keep this value positive which might mitigate
similar issues in the future.

-- 
Regards,
Marcelo


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 659 bytes --]