On Tue, Apr 26, 2022 at 03:03:33PM +0200, Geert Uytterhoeven wrote:
> On Tue, Apr 19, 2022 at 9:20 PM Kees Cook <keescook@chromium.org> wrote:
> > On Sun, Apr 17, 2022 at 07:26:33AM +0800, kernel test robot wrote:
> > > tree:   https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git for-next/lkdtm
> > > head:   2e53b877dc1258d4ac3de98f496bb88ec3bf5e25
> > > commit: 2e53b877dc1258d4ac3de98f496bb88ec3bf5e25 [6/6] lkdtm: Add CFI_BACKWARD to test ROP mitigations
> > > config: sparc-randconfig-r026-20220417 (https://download.01.org/0day-ci/archive/20220417/202204170710.LwCV5Ec0-lkp@intel.com/config)
> 
> Just received a failure for m68k-allmodconfig on next-20220426:
> 
> drivers/misc/lkdtm/cfi.c:62:37: error: cast from pointer to integer of
> different size [-Werror=pointer-to-int-cast]
> drivers/misc/lkdtm/cfi.c:62:10: error: cast to pointer from integer of
> different size [-Werror=int-to-pointer-cast]
> drivers/misc/lkdtm/cfi.c:62:37: error: cast from pointer to integer of
> different size [-Werror=pointer-to-int-cast]
> drivers/misc/lkdtm/cfi.c:62:10: error: cast to pointer from integer of
> different size [-Werror=int-to-pointer-cast]
> 
> > > compiler: sparc-linux-gcc (GCC) 11.2.0
> > > reproduce (this is a W=1 build):
> > >         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
> > >         chmod +x ~/bin/make.cross
> > >         # https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git/commit/?id=2e53b877dc1258d4ac3de98f496bb88ec3bf5e25
> > >         git remote add kees https://git.kernel.org/pub/scm/linux/kernel/git/kees/linux.git
> > >         git fetch --no-tags kees for-next/lkdtm
> > >         git checkout 2e53b877dc1258d4ac3de98f496bb88ec3bf5e25
> > >         # save the config file to linux build tree
> > >         mkdir build_dir
> > >         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross O=build_dir ARCH=sparc SHELL=/bin/bash drivers/misc/lkdtm/
> > >
> > > If you fix the issue, kindly add following tag as appropriate
> > > Reported-by: kernel test robot <lkp@intel.com>
> > >
> > > All warnings (new ones prefixed by >>):
> > >
> > >    drivers/misc/lkdtm/cfi.c: In function 'set_return_addr_unchecked':
> > > >> drivers/misc/lkdtm/cfi.c:62:37: warning: cast from pointer to integer of different size [-Wpointer-to-int-cast]
> > >       62 |         ((__force __typeof__(addr))((__force u64)(addr) | PAGE_OFFSET))
> > >          |                                     ^
> > >    drivers/misc/lkdtm/cfi.c:72:13: note: in expansion of macro 'no_pac_addr'
> > >       72 |         if (no_pac_addr(*ret_addr) == expected)
> > >          |             ^~~~~~~~~~~
> >
> > Hmm. I don't see why this is warning. The macro looks like untagged_addr():
> >
> > #define untagged_addr(addr)     ({ \
> >         u64 __addr = (__force u64)(addr); \
> >         __addr &= __untagged_addr(__addr); \
> >         (__force __typeof__(addr))__addr; \
> > })
> 
> The offending macro is:
> 
> #define no_pac_addr(addr)      \
>           ((__force __typeof__(addr))((__force u64)(addr) | PAGE_OFFSET))
> 
> On 32-bit, addr is a 32-bit pointer, which is cast to a 64-bit integer first,
> then to a 32-bit pointer again.
> All of these need intermediate casts to uintptr_t.

As in this?

#define no_pac_addr(addr)      \
	((__force __typeof__(addr))((uintptr_t)(addr) | PAGE_OFFSET))



-- 
Kees Cook