Re: [PATCH] [v2] arm64/io: add constant-argument check

[Jason Gunthorpe]

On Tue, Jun 04, 2024 at 04:55:37PM +0100, Will Deacon wrote:
> On Tue, May 28, 2024 at 05:37:11PM +0200, Arnd Bergmann wrote:
> > From: Arnd Bergmann <arnd@xxxxxxxx>
> > 
> > In some configurations __const_iowrite32_copy() does not get inlined
> > and gcc runs into the BUILD_BUG():
> > 
> > In file included from <command-line>:
> > In function '__const_memcpy_toio_aligned32',
> >     inlined from '__const_iowrite32_copy' at arch/arm64/include/asm/io.h:203:3,
> >     inlined from '__const_iowrite32_copy' at arch/arm64/include/asm/io.h:199:20:
> > include/linux/compiler_types.h:487:45: error: call to '__compiletime_assert_538' declared with attribute error: BUILD_BUG failed
> >   487 |         _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> >       |                                             ^
> > include/linux/compiler_types.h:468:25: note: in definition of macro '__compiletime_assert'
> >   468 |                         prefix ## suffix();                             \
> >       |                         ^~~~~~
> > include/linux/compiler_types.h:487:9: note: in expansion of macro '_compiletime_assert'
> >   487 |         _compiletime_assert(condition, msg, __compiletime_assert_, __COUNTER__)
> >       |         ^~~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:39:37: note: in expansion of macro 'compiletime_assert'
> >    39 | #define BUILD_BUG_ON_MSG(cond, msg) compiletime_assert(!(cond), msg)
> >       |                                     ^~~~~~~~~~~~~~~~~~
> > include/linux/build_bug.h:59:21: note: in expansion of macro 'BUILD_BUG_ON_MSG'
> >    59 | #define BUILD_BUG() BUILD_BUG_ON_MSG(1, "BUILD_BUG failed")
> >       |                     ^~~~~~~~~~~~~~~~
> > arch/arm64/include/asm/io.h:193:17: note: in expansion of macro 'BUILD_BUG'
> >   193 |                 BUILD_BUG();
> >       |                 ^~~~~~~~~
> > 
> > Move the check for constant arguments into the inline function to ensure
> > it is still constant if the compiler decides against inlining it.
> > 
> > Fixes: ead79118dae6 ("arm64/io: Provide a WC friendly __iowriteXX_copy()")
> > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> > ---
> > v2:
> >  - fix both 32-bit and 64-bit copies
> >  - remove now-redundant macros
> > ---
> >  arch/arm64/include/asm/io.h | 24 +++++++++---------------
> >  1 file changed, 9 insertions(+), 15 deletions(-)
> 
> I think this is superseded by Mark's diff in reply to v1, right?
> 
> https://lore.kernel.org/r/ZlcODqVXTDh6n0h-@J2N7QTR9R3
>
> If so, Mark, please can you post that as a proper patch so that we can
> get this fixed?

I wouldn't say superseded, but I agree with Mark that we should have
the __always_inline added to the __iowrite64_copy() and
__const_memcpy_toio_aligned64() in addition to the stuff here.

When I originally wrote this I copied the fairly common pattern of
having the builtin_constant_p test inside a macro, but I see now it is
quite common to put that into an inline. Putting it in the inline is
definately much better, so I like Arnd's patch.

Arnd can you just make that addition and repost this?

Thanks,
Jason