[02/11] x86: tboot: avoid Wstringop-overread-warning

From: Arnd Bergmann <arnd@arndb.de>

gcc-11 warns about using string operations on pointers that are
defined at compile time as offsets from a NULL pointer. Unfortunately
that also happens on the result of fix_to_virt(), which is a
compile-time constant for a constantn input:

arch/x86/kernel/tboot.c: In function 'tboot_probe':
arch/x86/kernel/tboot.c:70:13: error: '__builtin_memcmp_eq' specified bound 16 exceeds source size 0 [-Werror=stringop-overread]
   70 |         if (memcmp(&tboot_uuid, &tboot->uuid, sizeof(tboot->uuid))) {
      |             ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

I hope this can get addressed in gcc-11 before the release.

As a workaround, split up the tboot_probe() function in two halves
to separate the pointer generation from the usage. This is a bit
ugly, and hopefully gcc understands that the code is actually correct
before it learns to peek into the noinline function.

Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
---
 arch/x86/kernel/tboot.c | 44 ++++++++++++++++++++++++-----------------
 1 file changed, 26 insertions(+), 18 deletions(-)

On Mon, Mar 22, 2021 at 9:29 PM Ingo Molnar <mingo@kernel.org> wrote:
> * Arnd Bergmann <arnd@kernel.org> wrote:
> > From: Arnd Bergmann <arnd@arndb.de>

> This is indeed rather ugly - and the other patch that removes a debug
> check seems counterproductive as well.
>
> Do we know how many genuine bugs -Wstringop-overread-warning has
> caught or is about to catch?
>
> I.e. the real workaround might be to turn off the -Wstringop-overread-warning,
> until GCC-11 gets fixed?

See the [PATCH 0/11] message. The last two patches in the series are for
code that I suspect may be broken, the others are basically all false positives.

As gcc-11 is not released yet, I don't think we have to apply any of the
patches or disable the warning at the moment, but I posted all the patches
to get a better understanding on which of them should be addressed in
the kernel vs gcc.

       Arnd

* Martin Sebor <msebor@gmail.com> wrote:

> > I.e. the real workaround might be to turn off the -Wstringop-overread-warning,
> > until GCC-11 gets fixed?
> 
> In GCC 10 -Wstringop-overread is a subset of -Wstringop-overflow.
> GCC 11 breaks it out as a separate warning to make it easier to
> control.  Both warnings have caught some real bugs but they both
> have a nonzero rate of false positives.  Other than bug reports
> we don't have enough data to say what their S/N ratio might be
> but my sense is that it's fairly high in general.
> 
>   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=wstringop-overread
>   https://gcc.gnu.org/bugzilla/show_bug.cgi?id=wstringop-overflow
> 
> In GCC 11, all access warnings expect objects to be either declared
> or allocated.  Pointers with constant values are taken to point to
> nothing valid (as Arnd mentioned above, this is to detect invalid
> accesses to members of structs at address zero).
> 
> One possible solution to the known address problem is to extend GCC
> attributes address and io that pin an object to a hardwired address
> to all targets (at the moment they're supported on just one or two
> targets).  I'm not sure this can still happen before GCC 11 releases
> sometime in April or May.
> 
> Until then, another workaround is to convert the fixed address to
> a volatile pointer before using it for the access, along the lines
> below.  It should have only a negligible effect on efficiency.

Thank you for the detailed answer!

I think I'll go with Arnd's original patch - which makes the code a 
slightly bit cleaner by separating out the check_tboot_version() check 
into a standalone function.

The only ugly aspect is the global nature of the 'tboot' pointer - but 
that's a self-inflicted wound.

I'd also guess that the S/N ratio somewhat unfairly penalizes this 
warning right now, because the kernel had a decade of growing real 
fixes via other efforts such as static and dynamic instrumentation as 
well.

So the probability of false positive remaining is in fact higher, and 
going forward we should see a better S/N ratio of this warning. Most 
of which will never be seen by upstream maintainers, as the mishaps 
will stay at the individual developer level. :-)

Thanks,

	Ingo

From: Martin Sebor

> Sent: 22 March 2021 22:08

...
> In GCC 11, all access warnings expect objects to be either declared

> or allocated.  Pointers with constant values are taken to point to

> nothing valid (as Arnd mentioned above, this is to detect invalid

> accesses to members of structs at address zero).

> 

> One possible solution to the known address problem is to extend GCC

> attributes address and io that pin an object to a hardwired address

> to all targets (at the moment they're supported on just one or two

> targets).  I'm not sure this can still happen before GCC 11 releases

> sometime in April or May.


A different solution is to define a normal C external data item
and then assign a fixed address with an asm statement or in
the linker script.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)

From: David Laight

> Sent: 24 March 2021 09:12

> 

> From: Martin Sebor

> > Sent: 22 March 2021 22:08

> ...

> > In GCC 11, all access warnings expect objects to be either declared

> > or allocated.  Pointers with constant values are taken to point to

> > nothing valid (as Arnd mentioned above, this is to detect invalid

> > accesses to members of structs at address zero).

> >

> > One possible solution to the known address problem is to extend GCC

> > attributes address and io that pin an object to a hardwired address

> > to all targets (at the moment they're supported on just one or two

> > targets).  I'm not sure this can still happen before GCC 11 releases

> > sometime in April or May.

> 

> A different solution is to define a normal C external data item

> and then assign a fixed address with an asm statement or in

> the linker script.


Or stop gcc tracking the value by using:
	struct foo *foo = (void *)xxxxx;
	asm ("", "+r" (foo));

If the address is used more than once forcing it into
a register is also likely to generate better code.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)