Re: [PATCH] mm/hmm: fix unused variable warnings

[Arnd Bergmann <arnd@arndb.de>]

On Tue, Mar 5, 2019 at 1:18 PM Anshuman Khandual
<anshuman.khandual@arm.com> wrote:
> On 03/05/2019 01:30 AM, Arnd Bergmann wrote:
> > When CONFIG_HUGETLB_PAGE is disabled, the only use of the variable 'h'
> > is compiled out, and the compiler thinks it is unnecessary:
> >
> > mm/hmm.c: In function 'hmm_range_snapshot':
> > mm/hmm.c:1015:19: error: unused variable 'h' [-Werror=unused-variable]
> >     struct hstate *h = hstate_vma(vma);
>
> After doing some Kconfig hacks like (ARCH_WANT_GENERAL_HUGETLB = n) on an
> X86 system I got (HUGETLB_PAGE = n and HMM = y) config. But was unable to
> hit the build error. Helper is_vm_hugetlb_page() seems to always return
> false when HUGETLB_PAGE = n. Would not the compiler remove the entire code
> block including the declaration for 'h' ?
>
> #ifdef CONFIG_HUGETLB_PAGE
> #include <linux/mm.h>
> static inline bool is_vm_hugetlb_page(struct vm_area_struct *vma)
> {
>         return !!(vma->vm_flags & VM_HUGETLB);
> }
> #else
> static inline bool is_vm_hugetlb_page(struct vm_area_struct *vma)
> {
>         return false;
> }
> #endif

The is_vm_hugetlb_page() check is unrelated to the warning here,
the problem is that huge_page_shift() is defined as

#define huge_page_shift(h) PAGE_SHIFT

when CONFIG_HUGETLB_PAGE is disabled, so after preprocessing,
the only reference to the variable is removed.

     Arnd