Message ID | 20230405103819.151246-1-steven.price@arm.com (mailing list archive) |
---|---|
State | Mainlined, archived |
Headers | show |
Series | [v2] smaps: Fix defined but not used smaps_shmem_walk_ops | expand |
On Wed, Apr 05, 2023 at 11:38:19AM +0100, Steven Price wrote: > When !CONFIG_SHMEM smaps_shmem_walk_ops is defined but not used, > triggering a compiler warning. To avoid the warning remove the #ifdef > around the usage. This has no effect because shmem_mapping() is a stub > returning false when !CONFIG_SHMEM so the code will be compiled out, > however we now need to also provide a stub for shmem_swap_usage(). > > Fixes: 7b86ac3371b7 ("pagewalk: separate function pointers from iterator data") > Reported-by: kernel test robot <lkp@intel.com> > Link: https://lore.kernel.org/oe-kbuild-all/202304031749.UiyJpxzF-lkp@intel.com/ > Signed-off-by: Steven Price <steven.price@arm.com> > --- > I've implemented Jason's suggestion of removing the #ifdef around the > usage and prodiving a stub for shmem_swap_usage() instead. This seems OK Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> > extern bool shmem_is_huge(struct inode *inode, pgoff_t index, bool shmem_huge_force, > struct mm_struct *mm, unsigned long vm_flags); > +#ifdef CONFIG_SHMEM > extern unsigned long shmem_swap_usage(struct vm_area_struct *vma); > +#else > +static inline unsigned long shmem_swap_usage(struct vm_area_struct *vma) > +{ > + return 0; > +} > +#endif Though it would be nice if this file didn't have so many ifdef blocks Jason
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c index 6a96e1713fd5..cb49479acd2e 100644 --- a/fs/proc/task_mmu.c +++ b/fs/proc/task_mmu.c @@ -782,7 +782,6 @@ static void smap_gather_stats(struct vm_area_struct *vma, if (start >= vma->vm_end) return; -#ifdef CONFIG_SHMEM if (vma->vm_file && shmem_mapping(vma->vm_file->f_mapping)) { /* * For shared or readonly shmem mappings we know that all @@ -803,7 +802,7 @@ static void smap_gather_stats(struct vm_area_struct *vma, ops = &smaps_shmem_walk_ops; } } -#endif + /* mmap_lock is held in m_start */ if (!start) walk_page_vma(vma, ops, mss); diff --git a/include/linux/shmem_fs.h b/include/linux/shmem_fs.h index 103d1000a5a2..762c37b32bd4 100644 --- a/include/linux/shmem_fs.h +++ b/include/linux/shmem_fs.h @@ -94,7 +94,14 @@ int shmem_unuse(unsigned int type); extern bool shmem_is_huge(struct inode *inode, pgoff_t index, bool shmem_huge_force, struct mm_struct *mm, unsigned long vm_flags); +#ifdef CONFIG_SHMEM extern unsigned long shmem_swap_usage(struct vm_area_struct *vma); +#else +static inline unsigned long shmem_swap_usage(struct vm_area_struct *vma) +{ + return 0; +} +#endif extern unsigned long shmem_partial_swap_usage(struct address_space *mapping, pgoff_t start, pgoff_t end);
When !CONFIG_SHMEM smaps_shmem_walk_ops is defined but not used, triggering a compiler warning. To avoid the warning remove the #ifdef around the usage. This has no effect because shmem_mapping() is a stub returning false when !CONFIG_SHMEM so the code will be compiled out, however we now need to also provide a stub for shmem_swap_usage(). Fixes: 7b86ac3371b7 ("pagewalk: separate function pointers from iterator data") Reported-by: kernel test robot <lkp@intel.com> Link: https://lore.kernel.org/oe-kbuild-all/202304031749.UiyJpxzF-lkp@intel.com/ Signed-off-by: Steven Price <steven.price@arm.com> --- I've implemented Jason's suggestion of removing the #ifdef around the usage and prodiving a stub for shmem_swap_usage() instead. fs/proc/task_mmu.c | 3 +-- include/linux/shmem_fs.h | 7 +++++++ 2 files changed, 8 insertions(+), 2 deletions(-)