Re: [PATCH] rapidio: Avoid bogus __alloc_size warning

[Kees Cook]

On Thu, Sep 09, 2021 at 01:27:52PM -0700, Andrew Morton wrote:
> On Thu,  9 Sep 2021 09:14:09 -0700 Kees Cook <keescook@xxxxxxxxxxxx> wrote:
> 
> > GCC 9.3 (but not later) incorrectly evaluates the arguments to
> > check_copy_size(), getting seemingly confused by the size being returned
> > from array_size(). Instead, perform the calculation once, which both
> > makes the code more readable and avoids the bug in GCC.
> > 
> >    In file included from arch/x86/include/asm/preempt.h:7,
> >                     from include/linux/preempt.h:78,
> >                     from include/linux/spinlock.h:55,
> >                     from include/linux/mm_types.h:9,
> >                     from include/linux/buildid.h:5,
> >                     from include/linux/module.h:14,
> >                     from drivers/rapidio/devices/rio_mport_cdev.c:13:
> >    In function 'check_copy_size',
> >        inlined from 'copy_from_user' at include/linux/uaccess.h:191:6,
> >        inlined from 'rio_mport_transfer_ioctl' at drivers/rapidio/devices/rio_mport_cdev.c:983:6:
> >    include/linux/thread_info.h:213:4: error: call to '__bad_copy_to' declared with attribute error: copy destination size is too small
> >      213 |    __bad_copy_to();
> >          |    ^~~~~~~~~~~~~~~
> > 
> > But the allocation size and the copy size are identical:
> > 
> > 	transfer = vmalloc(array_size(sizeof(*transfer), transaction.count));
> > 	if (!transfer)
> > 		return -ENOMEM;
> > 
> > 	if (unlikely(copy_from_user(transfer,
> > 				    (void __user *)(uintptr_t)transaction.block,
> > 				    array_size(sizeof(*transfer), transaction.count)))) {
> 
> That's an "error", not a warning.  Or is this thanks to the new -Werror?

This is a "regular" error (__bad_copy_to() uses __compiletime_error()).

> Either way, I'm inclined to cc:stable on this, because use of gcc-9 on
> older kernels will be a common thing down the ages.
> 
> If it's really an "error" on non-Werror kernels then definitely cc:stable.

I would expect that as only being needed if __alloc_size was backported
to -stable, which seems unlikely.

-- 
Kees Cook