mm: establish const-correctness for pointer parameters
Subsystems (15)
v5 cover mm: establish const-correctness for pointer parameters
For improved const-correctness.
This patch series systematically adds const qualifiers to pointer parameters throughout the memory management subsystem, establishing a foundation for improved const-correctness across the entire Linux kernel.
Const-correctness provides multiple benefits:
1. Type Safety: The compiler enforces that functions marked as taking const parameters cannot accidentally modify the data, catching potential bugs at compile time rather than runtime.
2. Compiler Optimizations: When the compiler knows data won't be modified, it can generate more efficient code through better register allocation, code motion, and aliasing analysis.
3. API Documentation: Const qualifiers serve as self-documenting code, making it immediately clear to developers which functions are read-only operations versus those that modify state.
4. Maintenance Safety: Future modifications to const-correct code are less likely to introduce subtle bugs, as the compiler will reject attempts to modify data that should remain unchanged.
The memory management subsystem is a fundamental building block of the kernel. Most higher-level kernel subsystems (filesystems, drivers, networking) depend on mm interfaces. By establishing const-correctness at this foundational level:
1. Enables Propagation: Higher-level subsystems can adopt const-correctness in their own interfaces. Without const-correct mm functions, filesystems cannot mark their own parameters const when they need to call mm functions.
2. Maximum Impact: Changes to core mm APIs benefit the entire kernel, as these functions are called from virtually every subsystem.
3. Prevents Impedance Mismatch: Without const-correctness in mm, other subsystems must either cast away const (dangerous) or avoid using const altogether (missing optimization opportunities).
Each patch focuses on a specific header or subsystem component to ease review and bisection.
This work was initially posted as a single large patch:
https://lore.kernel.org/lkml/[email protected]/
Following feedback from Lorenzo Stoakes and David Hildenbrand, it has been split into focused, reviewable chunks. The approach was validated with a smaller patch that received agreement:
https://lore.kernel.org/lkml/[email protected]/
---
v1 -> v2:
- made several parameter values const (i.e. the pointer address, not just the pointed-to memory), as suggested by Andrew Morton and Yuanchu Xie
- drop existing+obsolete "extern" keywords on lines modified by these patches (suggested by Vishal Moola)
- add missing parameter names on lines modified by these patches (suggested by Vishal Moola)
- more "const" pointers (e.g. the task_struct passed to process_shares_mm())
- add missing "const" to s390, fixing s390 build failure
- moved the mmap_is_legacy() change in arch/s390/mm/mmap.c from 08/12 to 06/12 (suggested by Vishal Moola)
v2 -> v3:
- remove garbage from 06/12
- changed tags on subject line (suggested by Matthew Wilcox)
v3 -> v4:
- more verbose commit messages including a listing of function names (suggested by David Hildenbrand and Lorenzo Stoakes)
v4 -> v5:
- back to shorter commit messages after an agreement between David Hildenbrand and Lorenzo Stoakes was found
Max Kellermann (12):
mm: constify shmem related test functions for improved
const-correctness
mm: constify pagemap related test functions for improved
const-correctness
mm: constify zone related test functions for improved
const-correctness
fs: constify mapping related test functions for improved
const-correctness
mm: constify process_shares_mm() for improved const-correctness mm, s390: constify mapping related test functions for improved
const-correctness
parisc: constify mmap_upper_limit() parameter for improved
const-correctness
mm: constify arch_pick_mmap_layout() for improved const-correctness
mm: constify ptdesc_pmd_pts_count() and folio_get_private()
mm: constify various inline test functions for improved
const-correctness
mm: constify assert/test functions in mm.h
mm: constify highmem related functions for improved const-correctness
arch/arm/include/asm/highmem.h | 6 +-- arch/parisc/include/asm/processor.h | 2 +- arch/parisc/kernel/sys_parisc.c | 2 +- arch/s390/mm/mmap.c | 7 ++-- arch/sparc/kernel/sys_sparc_64.c | 3 +- arch/x86/mm/mmap.c | 7 ++-- arch/xtensa/include/asm/highmem.h | 2 +- include/linux/fs.h | 7 ++-- include/linux/highmem-internal.h | 44 +++++++++++---------- include/linux/highmem.h | 8 ++-- include/linux/mm.h | 56 +++++++++++++-------------- include/linux/mm_inline.h | 26 +++++++------ include/linux/mm_types.h | 4 +- include/linux/mmzone.h | 42 ++++++++++---------- include/linux/pagemap.h | 59 +++++++++++++++-------------- include/linux/sched/mm.h | 4 +- include/linux/shmem_fs.h | 4 +- mm/highmem.c | 10 ++--- mm/oom_kill.c | 7 ++-- mm/shmem.c | 6 +-- mm/util.c | 20 ++++++---- 21 files changed, 171 insertions(+), 155 deletions(-)
Signature · 1 line
2.47.2
PATCH v5 1/12 mm: constify shmem related test functions for improved const-correctness
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
--- include/linux/mm.h | 8 ++++---- include/linux/shmem_fs.h | 4 ++-- mm/shmem.c | 6 +++--- 3 files changed, 9 insertions(+), 9 deletions(-)
include/linux/mm.h +4 -4
@@ -979,11 +979,11 @@ static inline void vma_iter_set(struct vma_iterator *vmi, unsigned long addr)* The vma_is_shmem is not inline because it is used only by slow * paths in userfault. */ -bool vma_is_shmem (struct vm_area_struct *vma); -bool vma_is_anon_shmem (struct vm_area_struct *vma); +bool vma_is_shmem (const struct vm_area_struct *vma); +bool vma_is_anon_shmem (const struct vm_area_struct *vma);#else -static inline bool vma_is_shmem (struct vm_area_struct *vma) {return false; } -static inline bool vma_is_anon_shmem (struct vm_area_struct *vma) {return false; } +static inline bool vma_is_shmem (const struct vm_area_struct *vma) {return false; } +static inline bool vma_is_anon_shmem (const struct vm_area_struct *vma) {return false; }#endif int vma_is_stack_for_current (struct vm_area_struct *vma);
include/linux/shmem_fs.h +2 -2
@@ -99,9 +99,9 @@ extern unsigned long shmem_get_unmapped_area(struct file *, unsigned long addr, unsigned long len, unsigned long pgoff, unsigned long flags);extern int shmem_lock (struct file *file,int lock,struct ucounts *ucounts);#ifdef CONFIG_SHMEM -bool shmem_mapping (struct address_space *mapping); +bool shmem_mapping (const struct address_space *mapping);#else -static inline bool shmem_mapping (struct address_space *mapping) +static inline bool shmem_mapping (const struct address_space *mapping) {return false; }
mm/shmem.c +3 -3
@@ -275,18 +275,18 @@ static const struct vm_operations_struct shmem_vm_ops;static const struct vm_operations_struct shmem_anon_vm_ops;static struct file_system_type shmem_fs_type; -bool shmem_mapping (struct address_space *mapping) +bool shmem_mapping (const struct address_space *const mapping) {return mapping->a_ops == &shmem_aops; }EXPORT_SYMBOL_GPL (shmem_mapping); -bool vma_is_anon_shmem (struct vm_area_struct *vma) +bool vma_is_anon_shmem (const struct vm_area_struct *const vma) {return vma->vm_ops == &shmem_anon_vm_ops; } -bool vma_is_shmem (struct vm_area_struct *vma) +bool vma_is_shmem (const struct vm_area_struct *const vma) {return vma_is_anon_shmem (vma) || vma->vm_ops == &shmem_vm_ops; }
Signature · 1 line
2.47.2
Re: [PATCH v5 01/12] mm: constify shmem related test functions for improved const-correctness
Quoted 10 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 01/12] mm: constify shmem related test functions for improved const-correctness
In reply to Max KellermannQuoted 6 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Thanks.
Quoted 3 lines
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
LGTM, so:
Quoted 71 lines
--- include/linux/mm.h | 8 ++++---- include/linux/shmem_fs.h | 4 ++-- mm/shmem.c | 6 +++--- 3 files changed, 9 insertions(+), 9 deletions(-)include/linux/mm.h +4
-4@@ -979,11 +979,11 @@ static inline void vma_iter_set(struct vma_iterator *vmi, unsigned long addr)* The vma_is_shmem is not inline because it is used only by slow * paths in userfault. */ -bool vma_is_shmem (struct vm_area_struct *vma); -bool vma_is_anon_shmem (struct vm_area_struct *vma); +bool vma_is_shmem (const struct vm_area_struct *vma); +bool vma_is_anon_shmem (const struct vm_area_struct *vma);#else -static inline bool vma_is_shmem (struct vm_area_struct *vma) {return false; } -static inline bool vma_is_anon_shmem (struct vm_area_struct *vma) {return false; } +static inline bool vma_is_shmem (const struct vm_area_struct *vma) {return false; } +static inline bool vma_is_anon_shmem (const struct vm_area_struct *vma) {return false; }#endif int vma_is_stack_for_current (struct vm_area_struct *vma);include/linux/shmem_fs.h +2
-2@@ -99,9 +99,9 @@ extern unsigned long shmem_get_unmapped_area(struct file *, unsigned long addr, unsigned long len, unsigned long pgoff, unsigned long flags);extern int shmem_lock (struct file *file,int lock,struct ucounts *ucounts);#ifdef CONFIG_SHMEM -bool shmem_mapping (struct address_space *mapping); +bool shmem_mapping (const struct address_space *mapping);#else -static inline bool shmem_mapping (struct address_space *mapping) +static inline bool shmem_mapping (const struct address_space *mapping) {return false; }mm/shmem.c +3
-4@@ -275,18 +275,18 @@ static const struct vm_operations_struct shmem_vm_ops;static const struct vm_operations_struct shmem_anon_vm_ops;static struct file_system_type shmem_fs_type; -bool shmem_mapping (struct address_space *mapping) +bool shmem_mapping (const struct address_space *const mapping) {return mapping->a_ops == &shmem_aops; }EXPORT_SYMBOL_GPL (shmem_mapping); -bool vma_is_anon_shmem (struct vm_area_struct *vma) +bool vma_is_anon_shmem (const struct vm_area_struct *const vma) {return vma->vm_ops == &shmem_anon_vm_ops; } -bool vma_is_shmem (struct vm_area_struct *vma) +bool vma_is_shmem (const struct vm_area_struct *const vma) {return vma_is_anon_shmem (vma) || vma->vm_ops == &shmem_vm_ops; } --2.47.2
PATCH v5 2/12 mm: constify pagemap related test functions for improved const-correctness
In reply to Max KellermannWe select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
--- include/linux/pagemap.h | 57 +++++++++++++++++++++-------------------- 1 file changed, 29 insertions(+), 28 deletions(-)
include/linux/pagemap.h +29 -28
@@ -140,7 +140,7 @@ static inline int inode_drain_writes(struct inode *inode)return filemap_write_and_wait (inode->i_mapping); } -static inline bool mapping_empty (struct address_space *mapping) +static inline bool mapping_empty (const struct address_space *const mapping) {return xa_empty (&mapping->i_pages); }
@@ -166,7 +166,7 @@ static inline bool mapping_empty(struct address_space *mapping)* refcount and the referenced bit, which will be elevated or set in * the process of adding new cache pages to an inode. */ -static inline bool mapping_shrinkable (struct address_space *mapping) +static inline bool mapping_shrinkable (const struct address_space *const mapping) {void *head;
@@ -267,7 +267,7 @@ static inline void mapping_clear_unevictable(struct address_space *mapping)clear_bit (AS_UNEVICTABLE , &mapping->flags); } -static inline bool mapping_unevictable (struct address_space *mapping) +static inline bool mapping_unevictable (const struct address_space *const mapping) {return mapping &&test_bit (AS_UNEVICTABLE , &mapping->flags); }
@@ -277,7 +277,7 @@ static inline void mapping_set_exiting(struct address_space *mapping)set_bit (AS_EXITING , &mapping->flags); } -static inline int mapping_exiting (struct address_space *mapping) +static inline int mapping_exiting (const struct address_space *const mapping) {return test_bit (AS_EXITING , &mapping->flags); }
@@ -287,7 +287,7 @@ static inline void mapping_set_no_writeback_tags(struct address_space *mapping)set_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); } -static inline int mapping_use_writeback_tags (struct address_space *mapping) +static inline int mapping_use_writeback_tags (const struct address_space *const mapping) {return !test_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); }
@@ -333,7 +333,7 @@ static inline void mapping_set_inaccessible(struct address_space *mapping)set_bit (AS_INACCESSIBLE , &mapping->flags); } -static inline bool mapping_inaccessible (struct address_space *mapping) +static inline bool mapping_inaccessible (const struct address_space *const mapping) {return test_bit (AS_INACCESSIBLE , &mapping->flags); }
@@ -343,18 +343,18 @@ static inline void mapping_set_writeback_may_deadlock_on_reclaim(struct address_set_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline bool mapping_writeback_may_deadlock_on_reclaim (struct address_space *mapping) +static inline bool mapping_writeback_may_deadlock_on_reclaim (const struct address_space *const mapping) {return test_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline gfp_t mapping_gfp_mask (struct address_space * mapping) +static inline gfp_t mapping_gfp_mask (const struct address_space *const mapping) {return mapping->gfp_mask; }/* Restricts the given gfp_mask to what the mapping allows. */ -static inline gfp_t mapping_gfp_constraint (struct address_space *mapping, +static inline gfp_t mapping_gfp_constraint (const struct address_space *mapping,gfp_t gfp_mask) {return mapping_gfp_mask (mapping) & gfp_mask;
@@ -477,13 +477,13 @@ mapping_min_folio_order(const struct address_space *mapping) }static inline unsigned long -mapping_min_folio_nrpages (struct address_space *mapping) +mapping_min_folio_nrpages (const struct address_space *const mapping) {return 1UL <<mapping_min_folio_order (mapping); }static inline unsigned long -mapping_min_folio_nrbytes (struct address_space *mapping) +mapping_min_folio_nrbytes (const struct address_space *const mapping) {return mapping_min_folio_nrpages (mapping) <<PAGE_SHIFT ; }
@@ -497,7 +497,7 @@ mapping_min_folio_nrbytes(struct address_space *mapping)* new folio to the page cache and need to know what index to give it, * call this function. */ -static inline pgoff_t mapping_align_index (struct address_space *mapping, +static inline pgoff_t mapping_align_index (const struct address_space *const mapping,pgoff_t index) {return round_down (index,mapping_min_folio_nrpages (mapping));
@@ -507,7 +507,7 @@ static inline pgoff_t mapping_align_index(struct address_space *mapping,* Large folio support currently depends on THP. These dependencies are * being worked on but are not yet fixed. */ -static inline bool mapping_large_folio_support (struct address_space *mapping) +static inline bool mapping_large_folio_support (const struct address_space *mapping) {/* AS_FOLIO_ORDER is only reasonable for pagecache folios */ VM_WARN_ONCE((unsigned long)mapping & FOLIO_MAPPING_ANON,
@@ -522,7 +522,7 @@ static inline size_t mapping_max_folio_size(const struct address_space *mapping)return PAGE_SIZE <<mapping_max_folio_order (mapping); } -static inline int filemap_nr_thps (struct address_space *mapping) +static inline int filemap_nr_thps (const struct address_space *const mapping) {#ifdef CONFIG_READ_ONLY_THP_FOR_FS return atomic_read (&mapping->nr_thps);
@@ -936,7 +936,7 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping,* * Return: The index of the folio which follows this folio in the file. */ -static inline pgoff_t folio_next_index (struct folio *folio) +static inline pgoff_t folio_next_index (const struct folio *const folio) {return folio->index +folio_nr_pages (folio); }
@@ -965,7 +965,7 @@ static inline struct page *folio_file_page(struct folio *folio, pgoff_t index)* e.g., shmem did not move this folio to the swap cache. * Return: true or false. */ -static inline bool folio_contains (struct folio *folio,pgoff_t index) +static inline bool folio_contains (const struct folio *const folio,pgoff_t index) {VM_WARN_ON_ONCE_FOLIO (folio_test_swapcache (folio), folio);return index - folio->index <folio_nr_pages (folio);
@@ -1042,13 +1042,13 @@ static inline loff_t page_offset(struct page *page)/* * Get the offset in PAGE_SIZE (even for hugetlb folios). */ -static inline pgoff_t folio_pgoff (struct folio *folio) +static inline pgoff_t folio_pgoff (const struct folio *const folio) {return folio->index; } -static inline pgoff_t linear_page_index (struct vm_area_struct *vma, -unsigned long address) +static inline pgoff_t linear_page_index (const struct vm_area_struct *const vma, +const unsigned long address) {pgoff_t pgoff; pgoff = (address - vma->vm_start) >>PAGE_SHIFT ;
@@ -1468,7 +1468,7 @@ static inline unsigned int __readahead_batch(struct readahead_control *rac,* readahead_pos - The byte offset into the file of this readahead request. * @rac: The readahead request. */ -static inline loff_t readahead_pos (struct readahead_control *rac) +static inline loff_t readahead_pos (const struct readahead_control *const rac) {return (loff_t )rac->_index *PAGE_SIZE ; }
@@ -1477,7 +1477,7 @@ static inline loff_t readahead_pos(struct readahead_control *rac)* readahead_length - The number of bytes in this readahead request. * @rac: The readahead request. */ -static inline size_t readahead_length (struct readahead_control *rac) +static inline size_t readahead_length (const struct readahead_control *const rac) {return rac->_nr_pages *PAGE_SIZE ; }
@@ -1486,7 +1486,7 @@ static inline size_t readahead_length(struct readahead_control *rac)* readahead_index - The index of the first page in this readahead request. * @rac: The readahead request. */ -static inline pgoff_t readahead_index (struct readahead_control *rac) +static inline pgoff_t readahead_index (const struct readahead_control *const rac) {return rac->_index; }
@@ -1495,7 +1495,7 @@ static inline pgoff_t readahead_index(struct readahead_control *rac)* readahead_count - The number of pages in this readahead request. * @rac: The readahead request. */ -static inline unsigned int readahead_count (struct readahead_control *rac) +static inline unsigned int readahead_count (const struct readahead_control *const rac) {return rac->_nr_pages; }
@@ -1504,12 +1504,12 @@ static inline unsigned int readahead_count(struct readahead_control *rac)* readahead_batch_length - The number of bytes in the current batch. * @rac: The readahead request. */ -static inline size_t readahead_batch_length (struct readahead_control *rac) +static inline size_t readahead_batch_length (const struct readahead_control *const rac) {return rac->_batch_count *PAGE_SIZE ; } -static inline unsigned long dir_pages (struct inode *inode) +static inline unsigned long dir_pages (const struct inode *const inode) {return (unsigned long )(inode->i_size +PAGE_SIZE -1 ) >>PAGE_SHIFT ;
@@ -1523,8 +1523,8 @@ static inline unsigned long dir_pages(struct inode *inode)* Return: the number of bytes in the folio up to EOF, * or -EFAULT if the folio was truncated. */ -static inline ssize_t folio_mkwrite_check_truncate (struct folio *folio, -struct inode *inode) +static inline ssize_t folio_mkwrite_check_truncate (const struct folio *const folio, +const struct inode *const inode) {loff_t size =i_size_read (inode);pgoff_t index = size >>PAGE_SHIFT ;
@@ -1555,7 +1555,8 @@ static inline ssize_t folio_mkwrite_check_truncate(struct folio *folio,* Return: The number of filesystem blocks covered by this folio. */ static inline -unsigned int i_blocks_per_folio (struct inode *inode,struct folio *folio) +unsigned int i_blocks_per_folio (const struct inode *const inode, +const struct folio *const folio) {return folio_size (folio) >> inode->i_blkbits; }
Signature · 1 line
2.47.2
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 10 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Not only test functions but also some other "getters" in there.
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 1 line · David Hildenbrand
Not only test functions but also some other "getters" in there.
Oh, I thought getter/test function means the same for you. I'll amend the text.
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
In reply to Max KellermannQuoted 9 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
LGTM, so:
Quoted 237 lines
--- include/linux/pagemap.h | 57 +++++++++++++++++++++-------------------- 1 file changed, 29 insertions(+), 28 deletions(-)include/linux/pagemap.h +29
-29@@ -140,7 +140,7 @@ static inline int inode_drain_writes(struct inode *inode)return filemap_write_and_wait (inode->i_mapping); } -static inline bool mapping_empty (struct address_space *mapping) +static inline bool mapping_empty (const struct address_space *const mapping) {return xa_empty (&mapping->i_pages); }@@ -166,7 +166,7 @@ static inline bool mapping_empty(struct address_space *mapping)* refcount and the referenced bit, which will be elevated or set in * the process of adding new cache pages to an inode. */ -static inline bool mapping_shrinkable (struct address_space *mapping) +static inline bool mapping_shrinkable (const struct address_space *const mapping) {void *head;@@ -267,7 +267,7 @@ static inline void mapping_clear_unevictable(struct address_space *mapping)clear_bit (AS_UNEVICTABLE , &mapping->flags); } -static inline bool mapping_unevictable (struct address_space *mapping) +static inline bool mapping_unevictable (const struct address_space *const mapping) {return mapping &&test_bit (AS_UNEVICTABLE , &mapping->flags); }@@ -277,7 +277,7 @@ static inline void mapping_set_exiting(struct address_space *mapping)set_bit (AS_EXITING , &mapping->flags); } -static inline int mapping_exiting (struct address_space *mapping) +static inline int mapping_exiting (const struct address_space *const mapping) {return test_bit (AS_EXITING , &mapping->flags); }@@ -287,7 +287,7 @@ static inline void mapping_set_no_writeback_tags(struct address_space *mapping)set_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); } -static inline int mapping_use_writeback_tags (struct address_space *mapping) +static inline int mapping_use_writeback_tags (const struct address_space *const mapping) {return !test_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); }@@ -333,7 +333,7 @@ static inline void mapping_set_inaccessible(struct address_space *mapping)set_bit (AS_INACCESSIBLE , &mapping->flags); } -static inline bool mapping_inaccessible (struct address_space *mapping) +static inline bool mapping_inaccessible (const struct address_space *const mapping) {return test_bit (AS_INACCESSIBLE , &mapping->flags); }@@ -343,18 +343,18 @@ static inline void mapping_set_writeback_may_deadlock_on_reclaim(struct address_set_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline bool mapping_writeback_may_deadlock_on_reclaim (struct address_space *mapping) +static inline bool mapping_writeback_may_deadlock_on_reclaim (const struct address_space *const mapping) {return test_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline gfp_t mapping_gfp_mask (struct address_space * mapping) +static inline gfp_t mapping_gfp_mask (const struct address_space *const mapping) {return mapping->gfp_mask; }/* Restricts the given gfp_mask to what the mapping allows. */ -static inline gfp_t mapping_gfp_constraint (struct address_space *mapping, +static inline gfp_t mapping_gfp_constraint (const struct address_space *mapping,gfp_t gfp_mask) {return mapping_gfp_mask (mapping) & gfp_mask;@@ -477,13 +477,13 @@ mapping_min_folio_order(const struct address_space *mapping) }static inline unsigned long -mapping_min_folio_nrpages (struct address_space *mapping) +mapping_min_folio_nrpages (const struct address_space *const mapping) {return 1UL <<mapping_min_folio_order (mapping); }static inline unsigned long -mapping_min_folio_nrbytes (struct address_space *mapping) +mapping_min_folio_nrbytes (const struct address_space *const mapping) {return mapping_min_folio_nrpages (mapping) <<PAGE_SHIFT ; }@@ -497,7 +497,7 @@ mapping_min_folio_nrbytes(struct address_space *mapping)* new folio to the page cache and need to know what index to give it, * call this function. */ -static inline pgoff_t mapping_align_index (struct address_space *mapping, +static inline pgoff_t mapping_align_index (const struct address_space *const mapping,pgoff_t index) {return round_down (index,mapping_min_folio_nrpages (mapping));@@ -507,7 +507,7 @@ static inline pgoff_t mapping_align_index(struct address_space *mapping,* Large folio support currently depends on THP. These dependencies are * being worked on but are not yet fixed. */ -static inline bool mapping_large_folio_support (struct address_space *mapping) +static inline bool mapping_large_folio_support (const struct address_space *mapping) {/* AS_FOLIO_ORDER is only reasonable for pagecache folios */ VM_WARN_ONCE((unsigned long)mapping & FOLIO_MAPPING_ANON,@@ -522,7 +522,7 @@ static inline size_t mapping_max_folio_size(const struct address_space *mapping)return PAGE_SIZE <<mapping_max_folio_order (mapping); } -static inline int filemap_nr_thps (struct address_space *mapping) +static inline int filemap_nr_thps (const struct address_space *const mapping) {#ifdef CONFIG_READ_ONLY_THP_FOR_FS return atomic_read (&mapping->nr_thps);@@ -936,7 +936,7 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping,* * Return: The index of the folio which follows this folio in the file. */ -static inline pgoff_t folio_next_index (struct folio *folio) +static inline pgoff_t folio_next_index (const struct folio *const folio) {return folio->index +folio_nr_pages (folio); }@@ -965,7 +965,7 @@ static inline struct page *folio_file_page(struct folio *folio, pgoff_t index)* e.g., shmem did not move this folio to the swap cache. * Return: true or false. */ -static inline bool folio_contains (struct folio *folio,pgoff_t index) +static inline bool folio_contains (const struct folio *const folio,pgoff_t index) {VM_WARN_ON_ONCE_FOLIO (folio_test_swapcache (folio), folio);return index - folio->index <folio_nr_pages (folio);@@ -1042,13 +1042,13 @@ static inline loff_t page_offset(struct page *page)/* * Get the offset in PAGE_SIZE (even for hugetlb folios). */ -static inline pgoff_t folio_pgoff (struct folio *folio) +static inline pgoff_t folio_pgoff (const struct folio *const folio) {return folio->index; } -static inline pgoff_t linear_page_index (struct vm_area_struct *vma, -unsigned long address) +static inline pgoff_t linear_page_index (const struct vm_area_struct *const vma, +const unsigned long address) {pgoff_t pgoff; pgoff = (address - vma->vm_start) >>PAGE_SHIFT ;@@ -1468,7 +1468,7 @@ static inline unsigned int __readahead_batch(struct readahead_control *rac,* readahead_pos - The byte offset into the file of this readahead request. * @rac: The readahead request. */ -static inline loff_t readahead_pos (struct readahead_control *rac) +static inline loff_t readahead_pos (const struct readahead_control *const rac) {return (loff_t )rac->_index *PAGE_SIZE ; }@@ -1477,7 +1477,7 @@ static inline loff_t readahead_pos(struct readahead_control *rac)* readahead_length - The number of bytes in this readahead request. * @rac: The readahead request. */ -static inline size_t readahead_length (struct readahead_control *rac) +static inline size_t readahead_length (const struct readahead_control *const rac) {return rac->_nr_pages *PAGE_SIZE ; }@@ -1486,7 +1486,7 @@ static inline size_t readahead_length(struct readahead_control *rac)* readahead_index - The index of the first page in this readahead request. * @rac: The readahead request. */ -static inline pgoff_t readahead_index (struct readahead_control *rac) +static inline pgoff_t readahead_index (const struct readahead_control *const rac) {return rac->_index; }@@ -1495,7 +1495,7 @@ static inline pgoff_t readahead_index(struct readahead_control *rac)* readahead_count - The number of pages in this readahead request. * @rac: The readahead request. */ -static inline unsigned int readahead_count (struct readahead_control *rac) +static inline unsigned int readahead_count (const struct readahead_control *const rac) {return rac->_nr_pages; }@@ -1504,12 +1504,12 @@ static inline unsigned int readahead_count(struct readahead_control *rac)* readahead_batch_length - The number of bytes in the current batch. * @rac: The readahead request. */ -static inline size_t readahead_batch_length (struct readahead_control *rac) +static inline size_t readahead_batch_length (const struct readahead_control *const rac) {return rac->_batch_count *PAGE_SIZE ; } -static inline unsigned long dir_pages (struct inode *inode) +static inline unsigned long dir_pages (const struct inode *const inode) {return (unsigned long )(inode->i_size +PAGE_SIZE -1 ) >>PAGE_SHIFT ;@@ -1523,8 +1523,8 @@ static inline unsigned long dir_pages(struct inode *inode)* Return: the number of bytes in the folio up to EOF, * or -EFAULT if the folio was truncated. */ -static inline ssize_t folio_mkwrite_check_truncate (struct folio *folio, -struct inode *inode) +static inline ssize_t folio_mkwrite_check_truncate (const struct folio *const folio, +const struct inode *const inode) {loff_t size =i_size_read (inode);pgoff_t index = size >>PAGE_SHIFT ;@@ -1555,7 +1555,8 @@ static inline ssize_t folio_mkwrite_check_truncate(struct folio *folio,* Return: The number of filesystem blocks covered by this folio. */ static inline -unsigned int i_blocks_per_folio (struct inode *inode,struct folio *folio) +unsigned int i_blocks_per_folio (const struct inode *const inode, +const struct folio *const folio) {return folio_size (folio) >> inode->i_blkbits; } --2.47.2
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
In reply to Max KellermannQuoted 18 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>--- include/linux/pagemap.h | 57 +++++++++++++++++++++-------------------- 1 file changed, 29 insertions(+), 28 deletions(-)include/linux/pagemap.h +0
-0
include/linux/pagemap.h static inline int inode_drain_writes(struct inode *inode) · line 143
Quoted 5 lines
include/linux/pagemap.h +1
-1@@ -140,7 +140,7 @@ static inline int inode_drain_writes(struct inode *inode)return filemap_write_and_wait (inode->i_mapping); } -static inline bool mapping_empty (struct address_space *mapping) +static inline bool mapping_empty (const struct address_space *const mapping)
Generally - I'm not sure how useful this 'double' const-ification is.
const struct <type> *const <val>
^ ^ | | | | 1 2
Means:
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
I mean it's kinda nice to guarantee that this won't happen but I'm not sure if we're getting much value for the noise?
We also never mention that we're doing this in any commit message or the cover letter.
Quoted 223 lines
include/linux/pagemap.h +28
-28{return xa_empty (&mapping->i_pages); }@@ -166,7 +166,7 @@ static inline bool mapping_empty(struct address_space *mapping)* refcount and the referenced bit, which will be elevated or set in * the process of adding new cache pages to an inode. */ -static inline bool mapping_shrinkable (struct address_space *mapping) +static inline bool mapping_shrinkable (const struct address_space *const mapping) {void *head;@@ -267,7 +267,7 @@ static inline void mapping_clear_unevictable(struct address_space *mapping)clear_bit (AS_UNEVICTABLE , &mapping->flags); } -static inline bool mapping_unevictable (struct address_space *mapping) +static inline bool mapping_unevictable (const struct address_space *const mapping) {return mapping &&test_bit (AS_UNEVICTABLE , &mapping->flags); }@@ -277,7 +277,7 @@ static inline void mapping_set_exiting(struct address_space *mapping)set_bit (AS_EXITING , &mapping->flags); } -static inline int mapping_exiting (struct address_space *mapping) +static inline int mapping_exiting (const struct address_space *const mapping) {return test_bit (AS_EXITING , &mapping->flags); }@@ -287,7 +287,7 @@ static inline void mapping_set_no_writeback_tags(struct address_space *mapping)set_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); } -static inline int mapping_use_writeback_tags (struct address_space *mapping) +static inline int mapping_use_writeback_tags (const struct address_space *const mapping) {return !test_bit (AS_NO_WRITEBACK_TAGS , &mapping->flags); }@@ -333,7 +333,7 @@ static inline void mapping_set_inaccessible(struct address_space *mapping)set_bit (AS_INACCESSIBLE , &mapping->flags); } -static inline bool mapping_inaccessible (struct address_space *mapping) +static inline bool mapping_inaccessible (const struct address_space *const mapping) {return test_bit (AS_INACCESSIBLE , &mapping->flags); }@@ -343,18 +343,18 @@ static inline void mapping_set_writeback_may_deadlock_on_reclaim(struct address_set_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline bool mapping_writeback_may_deadlock_on_reclaim (struct address_space *mapping) +static inline bool mapping_writeback_may_deadlock_on_reclaim (const struct address_space *const mapping) {return test_bit (AS_WRITEBACK_MAY_DEADLOCK_ON_RECLAIM , &mapping->flags); } -static inline gfp_t mapping_gfp_mask (struct address_space * mapping) +static inline gfp_t mapping_gfp_mask (const struct address_space *const mapping) {return mapping->gfp_mask; }/* Restricts the given gfp_mask to what the mapping allows. */ -static inline gfp_t mapping_gfp_constraint (struct address_space *mapping, +static inline gfp_t mapping_gfp_constraint (const struct address_space *mapping,gfp_t gfp_mask) {return mapping_gfp_mask (mapping) & gfp_mask;@@ -477,13 +477,13 @@ mapping_min_folio_order(const struct address_space *mapping) }static inline unsigned long -mapping_min_folio_nrpages (struct address_space *mapping) +mapping_min_folio_nrpages (const struct address_space *const mapping) {return 1UL <<mapping_min_folio_order (mapping); }static inline unsigned long -mapping_min_folio_nrbytes (struct address_space *mapping) +mapping_min_folio_nrbytes (const struct address_space *const mapping) {return mapping_min_folio_nrpages (mapping) <<PAGE_SHIFT ; }@@ -497,7 +497,7 @@ mapping_min_folio_nrbytes(struct address_space *mapping)* new folio to the page cache and need to know what index to give it, * call this function. */ -static inline pgoff_t mapping_align_index (struct address_space *mapping, +static inline pgoff_t mapping_align_index (const struct address_space *const mapping,pgoff_t index) {return round_down (index,mapping_min_folio_nrpages (mapping));@@ -507,7 +507,7 @@ static inline pgoff_t mapping_align_index(struct address_space *mapping,* Large folio support currently depends on THP. These dependencies are * being worked on but are not yet fixed. */ -static inline bool mapping_large_folio_support (struct address_space *mapping) +static inline bool mapping_large_folio_support (const struct address_space *mapping) {/* AS_FOLIO_ORDER is only reasonable for pagecache folios */ VM_WARN_ONCE((unsigned long)mapping & FOLIO_MAPPING_ANON,@@ -522,7 +522,7 @@ static inline size_t mapping_max_folio_size(const struct address_space *mapping)return PAGE_SIZE <<mapping_max_folio_order (mapping); } -static inline int filemap_nr_thps (struct address_space *mapping) +static inline int filemap_nr_thps (const struct address_space *const mapping) {#ifdef CONFIG_READ_ONLY_THP_FOR_FS return atomic_read (&mapping->nr_thps);@@ -936,7 +936,7 @@ static inline struct page *grab_cache_page_nowait(struct address_space *mapping,* * Return: The index of the folio which follows this folio in the file. */ -static inline pgoff_t folio_next_index (struct folio *folio) +static inline pgoff_t folio_next_index (const struct folio *const folio) {return folio->index +folio_nr_pages (folio); }@@ -965,7 +965,7 @@ static inline struct page *folio_file_page(struct folio *folio, pgoff_t index)* e.g., shmem did not move this folio to the swap cache. * Return: true or false. */ -static inline bool folio_contains (struct folio *folio,pgoff_t index) +static inline bool folio_contains (const struct folio *const folio,pgoff_t index) {VM_WARN_ON_ONCE_FOLIO (folio_test_swapcache (folio), folio);return index - folio->index <folio_nr_pages (folio);@@ -1042,13 +1042,13 @@ static inline loff_t page_offset(struct page *page)/* * Get the offset in PAGE_SIZE (even for hugetlb folios). */ -static inline pgoff_t folio_pgoff (struct folio *folio) +static inline pgoff_t folio_pgoff (const struct folio *const folio) {return folio->index; } -static inline pgoff_t linear_page_index (struct vm_area_struct *vma, -unsigned long address) +static inline pgoff_t linear_page_index (const struct vm_area_struct *const vma, +const unsigned long address) {pgoff_t pgoff; pgoff = (address - vma->vm_start) >>PAGE_SHIFT ;@@ -1468,7 +1468,7 @@ static inline unsigned int __readahead_batch(struct readahead_control *rac,* readahead_pos - The byte offset into the file of this readahead request. * @rac: The readahead request. */ -static inline loff_t readahead_pos (struct readahead_control *rac) +static inline loff_t readahead_pos (const struct readahead_control *const rac) {return (loff_t )rac->_index *PAGE_SIZE ; }@@ -1477,7 +1477,7 @@ static inline loff_t readahead_pos(struct readahead_control *rac)* readahead_length - The number of bytes in this readahead request. * @rac: The readahead request. */ -static inline size_t readahead_length (struct readahead_control *rac) +static inline size_t readahead_length (const struct readahead_control *const rac) {return rac->_nr_pages *PAGE_SIZE ; }@@ -1486,7 +1486,7 @@ static inline size_t readahead_length(struct readahead_control *rac)* readahead_index - The index of the first page in this readahead request. * @rac: The readahead request. */ -static inline pgoff_t readahead_index (struct readahead_control *rac) +static inline pgoff_t readahead_index (const struct readahead_control *const rac) {return rac->_index; }@@ -1495,7 +1495,7 @@ static inline pgoff_t readahead_index(struct readahead_control *rac)* readahead_count - The number of pages in this readahead request. * @rac: The readahead request. */ -static inline unsigned int readahead_count (struct readahead_control *rac) +static inline unsigned int readahead_count (const struct readahead_control *const rac) {return rac->_nr_pages; }@@ -1504,12 +1504,12 @@ static inline unsigned int readahead_count(struct readahead_control *rac)* readahead_batch_length - The number of bytes in the current batch. * @rac: The readahead request. */ -static inline size_t readahead_batch_length (struct readahead_control *rac) +static inline size_t readahead_batch_length (const struct readahead_control *const rac) {return rac->_batch_count *PAGE_SIZE ; } -static inline unsigned long dir_pages (struct inode *inode) +static inline unsigned long dir_pages (const struct inode *const inode) {return (unsigned long )(inode->i_size +PAGE_SIZE -1 ) >>PAGE_SHIFT ;@@ -1523,8 +1523,8 @@ static inline unsigned long dir_pages(struct inode *inode)* Return: the number of bytes in the folio up to EOF, * or -EFAULT if the folio was truncated. */ -static inline ssize_t folio_mkwrite_check_truncate (struct folio *folio, -struct inode *inode) +static inline ssize_t folio_mkwrite_check_truncate (const struct folio *const folio, +const struct inode *const inode) {loff_t size =i_size_read (inode);pgoff_t index = size >>PAGE_SHIFT ;@@ -1555,7 +1555,8 @@ static inline ssize_t folio_mkwrite_check_truncate(struct folio *folio,* Return: The number of filesystem blocks covered by this folio. */ static inline -unsigned int i_blocks_per_folio (struct inode *inode,struct folio *folio) +unsigned int i_blocks_per_folio (const struct inode *const inode, +const struct folio *const folio) {return folio_size (folio) >> inode->i_blkbits; } --2.47.2
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 2 lines · Lorenzo Stoakes
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.
Only this was what my initial patch was about.
Quoted 2 lines
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
This wasn't my idea, it was Andrew Morton's idea, supported by Yuanchu Xie:
https://lore.kernel.org/lkml/CAJj2-QHVC0QW_4X95LLAnM=1g6apH==-OXZu65SVeBj0tSUcBg@mail.gmail.com/
You know that because you participated in the discussion. In that thread, nobody objected, so I took the time and adjusted all of my patches.
There is some value, but of course it's very small.
Note that I added the value-level "const" only to the implementation, never to prototypes, because it would have no effect there.
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 6 lines · Max Kellermann
Quoted 2 lines · Lorenzo Stoakes
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.
Only this was what my initial patch was about.
Right agreed then.
Quoted 6 lines
Quoted 3 lines
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
This wasn't my idea, it was Andrew Morton's idea, supported by Yuanchu Xie:
https://lore.kernel.org/lkml/CAJj2-QHVC0QW_4X95LLAnM=1g6apH==-OXZu65SVeBj0tSUcBg@mail.gmail.com/
Andrew said:
"Not that I'm suggesting that someone go in and make this change."
And Yuanchu said:
"Longer function readability would benefit from that, but it's IMO infeasible to do so everywhere."
(he also mentions it'd be good if gcc could wran on it).
So this isn't quite true actually.
Let's please just drop this, sorry.
The noise for multiple const params is too much, we can do it on a case-by-case basis going forward for larger functions.
Quoted 1 line
You know that because you participated in the discussion. In that
A point of advice given how this series has gone so far - please don't say things like this :)
I have an extremely heavy review load, and often work 12 hour days. I don't necessarily see everything though I try to.
I had missed this.
Just add a little civility and empathy it really helps, thanks.
Quoted 3 lines
thread, nobody objected, so I took the time and adjusted all of my patches.
There is some value, but of course it's very small.
Yeah sorry that you had extra work as a result.
But let's please drop this.
Quoted 3 lines
Note that I added the value-level "const" only to the implementation, never to prototypes, because it would have no effect there.
Right.
Thanks, Lorenzo
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 21 lines · Lorenzo Stoakes
Quoted 6 lines · Max Kellermann
Quoted 2 lines · Lorenzo Stoakes
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.
Only this was what my initial patch was about.
Right agreed then.
Quoted 6 lines
Quoted 3 lines
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
This wasn't my idea, it was Andrew Morton's idea, supported by Yuanchu Xie:
https://lore.kernel.org/lkml/CAJj2-QHVC0QW_4X95LLAnM=1g6apH==-OXZu65SVeBj0tSUcBg@mail.gmail.com/Andrew said:
"Not that I'm suggesting that someone go in and make this change."
Quoted 8 lines
And Yuanchu said:
"Longer function readability would benefit from that, but it's IMO infeasible to do so everywhere."
(he also mentions it'd be good if gcc could wran on it).
So this isn't quite true actually.
I understood it the same, that it would be nice if gcc treated incoming params (i.e. pointers, not pointed-to values) as const and warn otherwise, but not suggesting we should start doing that manually.
I personally agree that adding those extra "const" is of little value and makes the function definition lines longer and harder to read and so would rather not add those.
I mean we could first collectively decide (and that's not a review half-suggesting we do them) that we want them, and document that, but AFAIK that's not the case yet. While there's already an agreement that const pointed-to values is a good thing and nobody objects that.
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 35 lines · Vlastimil Babka
Quoted 21 lines · Lorenzo Stoakes
Quoted 6 lines · Max Kellermann
Quoted 2 lines · Lorenzo Stoakes
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed
value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.Only this was what my initial patch was about.
Right agreed then.
Quoted 6 lines
Quoted 3 lines
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
This wasn't my idea, it was Andrew Morton's idea, supported by Yuanchu Xie:
https://lore.kernel.org/lkml/CAJj2-QHVC0QW_4X95LLAnM=1g6apH==-OXZu65SVeBj0tSUcBg@mail.gmail.com/Andrew said:
"Not that I'm suggesting that someone go in and make this change."
Quoted 8 lines
And Yuanchu said:
"Longer function readability would benefit from that, but it's IMO infeasible to do so everywhere."
(he also mentions it'd be good if gcc could wran on it).
So this isn't quite true actually.
I understood it the same, that it would be nice if gcc treated incoming params (i.e. pointers, not pointed-to values) as const and warn otherwise, but not suggesting we should start doing that manually.
Quoted 8 lines
I personally agree that adding those extra "const" is of little value and makes the function definition lines longer and harder to read and so would rather not add those.
I mean we could first collectively decide (and that's not a review half-suggesting we do them) that we want them, and document that, but AFAIK that's not the case yet. While there's already an agreement that const pointed-to values is a good thing and nobody objects that.
Yeah, and discussed elsewhere in this series, it would also have to be clarified how to deal with the *const" (or const values in general) with function declaration vs. definition. I don't think we really have written-down rule for that yet.
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 02/12] mm: constify pagemap related test functions for improved const-correctness
Quoted 45 lines · David Hildenbrand
Quoted 35 lines · Vlastimil Babka
Quoted 21 lines · Lorenzo Stoakes
Quoted 6 lines · Max Kellermann
Quoted 2 lines · Lorenzo Stoakes
1. (most useful) Const pointer (const <type> *<param>) means that the dereffed
value is const, so *<param> = <val> or <param>-><field> = <val> are prohibited.Only this was what my initial patch was about.
Right agreed then.
Quoted 6 lines
Quoted 3 lines
2. (less useful) We can't modify the actual pointer value either, so e.g. <param> = <new param> is prohibited.
This wasn't my idea, it was Andrew Morton's idea, supported by Yuanchu Xie:
https://lore.kernel.org/lkml/CAJj2-QHVC0QW_4X95LLAnM=1g6apH==-OXZu65SVeBj0tSUcBg@mail.gmail.com/Andrew said:
"Not that I'm suggesting that someone go in and make this change."
Quoted 8 lines
And Yuanchu said:
"Longer function readability would benefit from that, but it's IMO infeasible to do so everywhere."
(he also mentions it'd be good if gcc could wran on it).
So this isn't quite true actually.
I understood it the same, that it would be nice if gcc treated incoming params (i.e. pointers, not pointed-to values) as const and warn otherwise, but not suggesting we should start doing that manually.
Quoted 8 lines
I personally agree that adding those extra "const" is of little value and makes the function definition lines longer and harder to read and so would rather not add those.
I mean we could first collectively decide (and that's not a review half-suggesting we do them) that we want them, and document that, but AFAIK that's not the case yet. While there's already an agreement that const pointed-to values is a good thing and nobody objects that.
Quoted 4 lines
Yeah, and discussed elsewhere in this series, it would also have to be clarified how to deal with the *const" (or const values in general) with function declaration vs. definition. I don't think we really have written-down rule for that yet.
For this series the consensus is clear that we should eliminate these and revert to const pointed-to values (1) only.
We can determine how we do this in future re: const actual pointers (2), but this series isn't the place.
So Max - can you respin with the (2) const-ification removed please.
Thanks.
PATCH v5 3/12 mm: constify zone related test functions for improved const-correctness
In reply to Max KellermannWe select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
--- include/linux/mmzone.h | 42 +++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-)
include/linux/mmzone.h +21 -21
@@ -1104,7 +1104,7 @@ static inline unsigned long promo_wmark_pages(const struct zone *z)return wmark_pages (z,WMARK_PROMO ); } -static inline unsigned long zone_managed_pages (struct zone *zone) +static inline unsigned long zone_managed_pages (const struct zone *const zone) {return (unsigned long )atomic_long_read (&zone->managed_pages); }
@@ -1128,12 +1128,12 @@ static inline bool zone_spans_pfn(const struct zone *zone, unsigned long pfn)return zone->zone_start_pfn <= pfn && pfn <zone_end_pfn (zone); } -static inline bool zone_is_initialized (struct zone *zone) +static inline bool zone_is_initialized (const struct zone *const zone) {return zone->initialized; } -static inline bool zone_is_empty (struct zone *zone) +static inline bool zone_is_empty (const struct zone *const zone) {return zone->spanned_pages ==0 ; }
@@ -1273,7 +1273,7 @@ static inline bool folio_is_zone_movable(const struct folio *folio)* Return true if [start_pfn, start_pfn + nr_pages) range has a non-empty * intersection with the given zone */ -static inline bool zone_intersects (struct zone *zone, +static inline bool zone_intersects (const struct zone *const zone,unsigned long start_pfn,unsigned long nr_pages) {if (zone_is_empty (zone))
@@ -1581,12 +1581,12 @@ static inline int local_memory_node(int node_id) { return node_id; };#define zone_idx (zone) ((zone) - (zone)->zone_pgdat->node_zones)#ifdef CONFIG_ZONE_DEVICE -static inline bool zone_is_zone_device (struct zone *zone) +static inline bool zone_is_zone_device (const struct zone *const zone) {return zone_idx (zone) ==ZONE_DEVICE ; }#else -static inline bool zone_is_zone_device (struct zone *zone) +static inline bool zone_is_zone_device (const struct zone *const zone) {return false; }
@@ -1598,19 +1598,19 @@ static inline bool zone_is_zone_device(struct zone *zone)* populated_zone(). If the whole zone is reserved then we can easily * end up with populated_zone() && !managed_zone(). */ -static inline bool managed_zone (struct zone *zone) +static inline bool managed_zone (const struct zone *const zone) {return zone_managed_pages (zone); }/* Returns true if a zone has memory */ -static inline bool populated_zone (struct zone *zone) +static inline bool populated_zone (const struct zone *const zone) {return zone->present_pages; }#ifdef CONFIG_NUMA -static inline int zone_to_nid (struct zone *zone) +static inline int zone_to_nid (const struct zone *const zone) {return zone->node; }
@@ -1620,7 +1620,7 @@ static inline void zone_set_nid(struct zone *zone, int nid) zone->node = nid; }#else -static inline int zone_to_nid (struct zone *zone) +static inline int zone_to_nid (const struct zone *zone) {return 0 ; }
@@ -1647,7 +1647,7 @@ static inline int is_highmem_idx(enum zone_type idx)* @zone: pointer to struct zone variable * Return: 1 for a highmem zone, 0 otherwise */ -static inline int is_highmem (struct zone *zone) +static inline int is_highmem (const struct zone *const zone) {return is_highmem_idx (zone_idx (zone)); }
@@ -1713,12 +1713,12 @@ static inline struct zone *zonelist_zone(struct zoneref *zoneref)return zoneref->zone; } -static inline int zonelist_zone_idx (struct zoneref *zoneref) +static inline int zonelist_zone_idx (const struct zoneref *const zoneref) {return zoneref->zone_idx; } -static inline int zonelist_node_idx (struct zoneref *zoneref) +static inline int zonelist_node_idx (const struct zoneref *const zoneref) {return zone_to_nid (zoneref->zone); }
@@ -2021,7 +2021,7 @@ static inline struct page *__section_mem_map_addr(struct mem_section *section)return (struct page *)map; } -static inline int present_section (struct mem_section *section) +static inline int present_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_MARKED_PRESENT )); }
@@ -2031,12 +2031,12 @@ static inline int present_section_nr(unsigned long nr)return present_section (__nr_to_section (nr)); } -static inline int valid_section (struct mem_section *section) +static inline int valid_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_HAS_MEM_MAP )); } -static inline int early_section (struct mem_section *section) +static inline int early_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_EARLY )); }
@@ -2046,27 +2046,27 @@ static inline int valid_section_nr(unsigned long nr)return valid_section (__nr_to_section (nr)); } -static inline int online_section (struct mem_section *section) +static inline int online_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_ONLINE )); }#ifdef CONFIG_ZONE_DEVICE -static inline int online_device_section (struct mem_section *section) +static inline int online_device_section (const struct mem_section *const section) {unsigned long flags =SECTION_IS_ONLINE |SECTION_TAINT_ZONE_DEVICE ;return section && ((section->section_mem_map & flags) == flags); }#else -static inline int online_device_section (struct mem_section *section) +static inline int online_device_section (const struct mem_section *const section) {return 0 ; }#endif #ifdef CONFIG_SPARSEMEM_VMEMMAP_PREINIT -static inline int preinited_vmemmap_section (struct mem_section *section) +static inline int preinited_vmemmap_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_VMEMMAP_PREINIT ));
@@ -2076,7 +2076,7 @@ void sparse_vmemmap_init_nid_early(int nid);void sparse_vmemmap_init_nid_late (int nid);#else -static inline int preinited_vmemmap_section (struct mem_section *section) +static inline int preinited_vmemmap_section (const struct mem_section *const section) {return 0 ; }
Signature · 1 line
2.47.2
Re: [PATCH v5 03/12] mm: constify zone related test functions for improved const-correctness
Quoted 10 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Also some getters hiding in between the test functions.
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 03/12] mm: constify zone related test functions for improved const-correctness
In reply to Max KellermannQuoted 9 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
On basis that we figure out whether we want the "const <type> *const <param>" thing or not, this otherwise LGTM so:
Quoted 180 lines
--- include/linux/mmzone.h | 42 +++++++++++++++++++++--------------------- 1 file changed, 21 insertions(+), 21 deletions(-)include/linux/mmzone.h +21
-22@@ -1104,7 +1104,7 @@ static inline unsigned long promo_wmark_pages(const struct zone *z)return wmark_pages (z,WMARK_PROMO ); } -static inline unsigned long zone_managed_pages (struct zone *zone) +static inline unsigned long zone_managed_pages (const struct zone *const zone) {return (unsigned long )atomic_long_read (&zone->managed_pages); }@@ -1128,12 +1128,12 @@ static inline bool zone_spans_pfn(const struct zone *zone, unsigned long pfn)return zone->zone_start_pfn <= pfn && pfn <zone_end_pfn (zone); } -static inline bool zone_is_initialized (struct zone *zone) +static inline bool zone_is_initialized (const struct zone *const zone) {return zone->initialized; } -static inline bool zone_is_empty (struct zone *zone) +static inline bool zone_is_empty (const struct zone *const zone) {return zone->spanned_pages ==0 ; }@@ -1273,7 +1273,7 @@ static inline bool folio_is_zone_movable(const struct folio *folio)* Return true if [start_pfn, start_pfn + nr_pages) range has a non-empty * intersection with the given zone */ -static inline bool zone_intersects (struct zone *zone, +static inline bool zone_intersects (const struct zone *const zone,unsigned long start_pfn,unsigned long nr_pages) {if (zone_is_empty (zone))@@ -1581,12 +1581,12 @@ static inline int local_memory_node(int node_id) { return node_id; };#define zone_idx (zone) ((zone) - (zone)->zone_pgdat->node_zones)#ifdef CONFIG_ZONE_DEVICE -static inline bool zone_is_zone_device (struct zone *zone) +static inline bool zone_is_zone_device (const struct zone *const zone) {return zone_idx (zone) ==ZONE_DEVICE ; }#else -static inline bool zone_is_zone_device (struct zone *zone) +static inline bool zone_is_zone_device (const struct zone *const zone) {return false; }@@ -1598,19 +1598,19 @@ static inline bool zone_is_zone_device(struct zone *zone)* populated_zone(). If the whole zone is reserved then we can easily * end up with populated_zone() && !managed_zone(). */ -static inline bool managed_zone (struct zone *zone) +static inline bool managed_zone (const struct zone *const zone) {return zone_managed_pages (zone); }/* Returns true if a zone has memory */ -static inline bool populated_zone (struct zone *zone) +static inline bool populated_zone (const struct zone *const zone) {return zone->present_pages; }#ifdef CONFIG_NUMA -static inline int zone_to_nid (struct zone *zone) +static inline int zone_to_nid (const struct zone *const zone) {return zone->node; }@@ -1620,7 +1620,7 @@ static inline void zone_set_nid(struct zone *zone, int nid) zone->node = nid; }#else -static inline int zone_to_nid (struct zone *zone) +static inline int zone_to_nid (const struct zone *zone) {return 0 ; }@@ -1647,7 +1647,7 @@ static inline int is_highmem_idx(enum zone_type idx)* @zone: pointer to struct zone variable * Return: 1 for a highmem zone, 0 otherwise */ -static inline int is_highmem (struct zone *zone) +static inline int is_highmem (const struct zone *const zone) {return is_highmem_idx (zone_idx (zone)); }@@ -1713,12 +1713,12 @@ static inline struct zone *zonelist_zone(struct zoneref *zoneref)return zoneref->zone; } -static inline int zonelist_zone_idx (struct zoneref *zoneref) +static inline int zonelist_zone_idx (const struct zoneref *const zoneref) {return zoneref->zone_idx; } -static inline int zonelist_node_idx (struct zoneref *zoneref) +static inline int zonelist_node_idx (const struct zoneref *const zoneref) {return zone_to_nid (zoneref->zone); }@@ -2021,7 +2021,7 @@ static inline struct page *__section_mem_map_addr(struct mem_section *section)return (struct page *)map; } -static inline int present_section (struct mem_section *section) +static inline int present_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_MARKED_PRESENT )); }@@ -2031,12 +2031,12 @@ static inline int present_section_nr(unsigned long nr)return present_section (__nr_to_section (nr)); } -static inline int valid_section (struct mem_section *section) +static inline int valid_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_HAS_MEM_MAP )); } -static inline int early_section (struct mem_section *section) +static inline int early_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_EARLY )); }@@ -2046,27 +2046,27 @@ static inline int valid_section_nr(unsigned long nr)return valid_section (__nr_to_section (nr)); } -static inline int online_section (struct mem_section *section) +static inline int online_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_ONLINE )); }#ifdef CONFIG_ZONE_DEVICE -static inline int online_device_section (struct mem_section *section) +static inline int online_device_section (const struct mem_section *const section) {unsigned long flags =SECTION_IS_ONLINE |SECTION_TAINT_ZONE_DEVICE ;return section && ((section->section_mem_map & flags) == flags); }#else -static inline int online_device_section (struct mem_section *section) +static inline int online_device_section (const struct mem_section *const section) {return 0 ; }#endif #ifdef CONFIG_SPARSEMEM_VMEMMAP_PREINIT -static inline int preinited_vmemmap_section (struct mem_section *section) +static inline int preinited_vmemmap_section (const struct mem_section *const section) {return (section && (section->section_mem_map &SECTION_IS_VMEMMAP_PREINIT ));@@ -2076,7 +2076,7 @@ void sparse_vmemmap_init_nid_early(int nid);void sparse_vmemmap_init_nid_late (int nid);#else -static inline int preinited_vmemmap_section (struct mem_section *section) +static inline int preinited_vmemmap_section (const struct mem_section *const section) {return 0 ; } --2.47.2
PATCH v5 4/12 fs: constify mapping related test functions for improved const-correctness
In reply to Max KellermannWe select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
--- include/linux/fs.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)
include/linux/fs.h +4 -3
@@ -537,7 +537,8 @@ struct address_space {/* * Returns true if any of the pages in the mapping are marked with the tag. */ -static inline bool mapping_tagged (struct address_space *mapping,xa_mark_t tag) +static inline bool mapping_tagged (const struct address_space *const mapping, +const xa_mark_t tag) {return xa_marked (&mapping->i_pages, tag); }
@@ -585,7 +586,7 @@ static inline void i_mmap_assert_write_locked(struct address_space *mapping)/* * Might pages of this file be mapped into userspace? */ -static inline int mapping_mapped (struct address_space *mapping) +static inline int mapping_mapped (const struct address_space *const mapping) {return !RB_EMPTY_ROOT (&mapping->i_mmap.rb_root); }
@@ -599,7 +600,7 @@ static inline int mapping_mapped(struct address_space *mapping)* If i_mmap_writable is negative, no new writable mappings are allowed. You * can only deny writable mappings, if none exists right now. */ -static inline int mapping_writably_mapped (struct address_space *mapping) +static inline int mapping_writably_mapped (const struct address_space *const mapping) {return atomic_read (&mapping->i_mmap_writable) >0 ; }
Signature · 1 line
2.47.2
Re: [PATCH v5 04/12] fs: constify mapping related test functions for improved const-correctness
Quoted 10 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 04/12] fs: constify mapping related test functions for improved const-correctness
In reply to Max KellermannQuoted 10 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 04/12] fs: constify mapping related test functions for improved const-correctness
In reply to Max KellermannQuoted 9 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
(Again, on basis of us figuring out whether we want the double-const)
LGTM, so:
Quoted 39 lines
--- include/linux/fs.h | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-)include/linux/fs.h +4
-4@@ -537,7 +537,8 @@ struct address_space {/* * Returns true if any of the pages in the mapping are marked with the tag. */ -static inline bool mapping_tagged (struct address_space *mapping,xa_mark_t tag) +static inline bool mapping_tagged (const struct address_space *const mapping, +const xa_mark_t tag) {return xa_marked (&mapping->i_pages, tag); }@@ -585,7 +586,7 @@ static inline void i_mmap_assert_write_locked(struct address_space *mapping)/* * Might pages of this file be mapped into userspace? */ -static inline int mapping_mapped (struct address_space *mapping) +static inline int mapping_mapped (const struct address_space *const mapping) {return !RB_EMPTY_ROOT (&mapping->i_mmap.rb_root); }@@ -599,7 +600,7 @@ static inline int mapping_mapped(struct address_space *mapping)* If i_mmap_writable is negative, no new writable mappings are allowed. You * can only deny writable mappings, if none exists right now. */ -static inline int mapping_writably_mapped (struct address_space *mapping) +static inline int mapping_writably_mapped (const struct address_space *const mapping) {return atomic_read (&mapping->i_mmap_writable) >0 ; } --2.47.2
PATCH v5 5/12 mm: constify process_shares_mm() for improved const-correctness
In reply to Max KellermannThis function only reads from the pointer arguments.
Local (loop) variables are also annotated with const to clarify that these will not be written to.
--- include/linux/mm.h | 2 +- mm/oom_kill.c | 7 ++++--- 2 files changed, 5 insertions(+), 4 deletions(-)
include/linux/mm.h +1 -1
@@ -3840,7 +3840,7 @@ static inline int in_gate_area(struct mm_struct *mm, unsigned long addr) }#endif /* __HAVE_ARCH_GATE_AREA */ -extern bool process_shares_mm (struct task_struct *p,struct mm_struct *mm); +bool process_shares_mm (const struct task_struct *p,const struct mm_struct *mm);void drop_slab (void );
mm/oom_kill.c +4 -3
@@ -490,12 +490,13 @@ static bool oom_killer_disabled __read_mostly;* task's threads: if one of those is using this mm then this task was also * using it. */ -bool process_shares_mm (struct task_struct *p,struct mm_struct *mm) +bool process_shares_mm (const struct task_struct *const p, +const struct mm_struct *const mm) { -struct task_struct *t; +const struct task_struct *t;for_each_thread (p, t) { -struct mm_struct *t_mm =READ_ONCE (t->mm); +const struct mm_struct *t_mm =READ_ONCE (t->mm);if (t_mm)return t_mm == mm; }
Signature · 1 line
2.47.2
Re: [PATCH v5 05/12] mm: constify process_shares_mm() for improved const-correctness
Quoted 6 lines · Max Kellermann
This function only reads from the pointer arguments.
Local (loop) variables are also annotated with
constto clarify that these will not be written to.Signed-off-by Max Kellermann <[email protected]>
This is nice thanks,
Quoted 42 lines
--- include/linux/mm.h | 2 +- mm/oom_kill.c | 7 ++++--- 2 files changed, 5 insertions(+), 4 deletions(-)include/linux/mm.h +1
-1@@ -3840,7 +3840,7 @@ static inline int in_gate_area(struct mm_struct *mm, unsigned long addr) }#endif /* __HAVE_ARCH_GATE_AREA */ -extern bool process_shares_mm (struct task_struct *p,struct mm_struct *mm); +bool process_shares_mm (const struct task_struct *p,const struct mm_struct *mm);void drop_slab (void );mm/oom_kill.c +4
-4@@ -490,12 +490,13 @@ static bool oom_killer_disabled __read_mostly;* task's threads: if one of those is using this mm then this task was also * using it. */ -bool process_shares_mm (struct task_struct *p,struct mm_struct *mm) +bool process_shares_mm (const struct task_struct *const p, +const struct mm_struct *const mm) { -struct task_struct *t; +const struct task_struct *t;for_each_thread (p, t) { -struct mm_struct *t_mm =READ_ONCE (t->mm); +const struct mm_struct *t_mm =READ_ONCE (t->mm);if (t_mm)return t_mm == mm; } --2.47.2
PATCH v5 6/12 mm, s390: constify mapping related test functions for improved const-correctness
In reply to Max KellermannWe select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
(Even though seemingly unrelated, this also constifies the pointer parameter of mmap_is_legacy() in arch/s390/mm/mmap.c because a copy of the function exists in mm/util.c.)
--- arch/s390/mm/mmap.c | 2 +- include/linux/mm.h | 6 +++--- include/linux/pagemap.h | 2 +- mm/util.c | 11 ++++++----- 4 files changed, 11 insertions(+), 10 deletions(-)
arch/s390/mm/mmap.c +1 -1
@@ -27,7 +27,7 @@ static unsigned long stack_maxrandom_size(void)return STACK_RND_MASK <<PAGE_SHIFT ; } -static inline int mmap_is_legacy (struct rlimit *rlim_stack) +static inline int mmap_is_legacy (const struct rlimit *const rlim_stack) {if (current->personality &ADDR_COMPAT_LAYOUT )return 1 ;
include/linux/mm.h +3 -3
@@ -986,7 +986,7 @@ static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return falsestatic inline bool vma_is_anon_shmem (const struct vm_area_struct *vma) {return false; }#endif -int vma_is_stack_for_current (struct vm_area_struct *vma); +int vma_is_stack_for_current (const struct vm_area_struct *vma);/* flush_tlb_range() takes a vma, not a mm, and can care about flags */ #define TLB_FLUSH_VMA (mm,flags) { .vm_mm = (mm), .vm_flags = (flags) }
@@ -2585,7 +2585,7 @@ void folio_add_pin(struct folio *folio);int account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc);int __account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc, -struct task_struct *task,bool bypass_rlim); +const struct task_struct *task,bool bypass_rlim);struct kvec ;struct page *get_dump_page (unsigned long addr,int *locked);
@@ -3348,7 +3348,7 @@ void anon_vma_interval_tree_verify(struct anon_vma_chain *node); avc; avc = anon_vma_interval_tree_iter_next(avc, start, last))/* mmap.c */ -extern int __vm_enough_memory (struct mm_struct *mm,long pages,int cap_sys_admin); +extern int __vm_enough_memory (const struct mm_struct *mm,long pages,int cap_sys_admin);extern int insert_vm_struct (struct mm_struct *,struct vm_area_struct *);extern void exit_mmap (struct mm_struct *); bool mmap_read_lock_maybe_expand(struct mm_struct *mm, struct vm_area_struct *vma,
include/linux/pagemap.h +1 -1
@@ -551,7 +551,7 @@ static inline void filemap_nr_thps_dec(struct address_space *mapping)#endif } -struct address_space *folio_mapping (struct folio *); +struct address_space *folio_mapping (const struct folio *folio);/** * folio_flush_mapping - Find the file mapping this folio belongs to.
mm/util.c +6 -5
@@ -315,7 +315,7 @@ void *memdup_user_nul(const void __user *src, size_t len)EXPORT_SYMBOL (memdup_user_nul);/* Check if the vma is being used as a stack by this task */ -int vma_is_stack_for_current (struct vm_area_struct *vma) +int vma_is_stack_for_current (const struct vm_area_struct *const vma) {struct task_struct *__maybe_unused t = current;
@@ -410,7 +410,7 @@ unsigned long arch_mmap_rnd(void)return rnd <<PAGE_SHIFT ; } -static int mmap_is_legacy (struct rlimit *rlim_stack) +static int mmap_is_legacy (const struct rlimit *const rlim_stack) {if (current->personality &ADDR_COMPAT_LAYOUT )return 1 ;
@@ -504,7 +504,7 @@ EXPORT_SYMBOL_IF_KUNIT(arch_pick_mmap_layout);* * -ENOMEM if RLIMIT_MEMLOCK would be exceeded. */ int __account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc, -struct task_struct *task,bool bypass_rlim) +const struct task_struct *const task,const bool bypass_rlim) {unsigned long locked_vm, limit;int ret =0 ;
@@ -688,7 +688,7 @@ struct anon_vma *folio_anon_vma(const struct folio *folio)* You can call this for folios which aren't in the swap cache or page * cache and it will return NULL. */ -struct address_space *folio_mapping (struct folio *folio) +struct address_space *folio_mapping (const struct folio *const folio) {struct address_space *mapping;
@@ -926,7 +926,8 @@ EXPORT_SYMBOL_GPL(vm_memory_committed);* Note this is a helper function intended to be used by LSMs which * wish to use this logic. */ -int __vm_enough_memory (struct mm_struct *mm,long pages,int cap_sys_admin) +int __vm_enough_memory (const struct mm_struct *const mm, +const long pages,const int cap_sys_admin) {long allowed;unsigned long bytes_failed;
Signature · 1 line
2.47.2
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
Quoted 7 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Quoted 6 lines
(Even though seemingly unrelated, this also constifies the pointer parameter of
mmap_is_legacy()inarch/s390/mm/mmap.cbecause a copy of the function exists inmm/util.c.)Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
Also here, some getters hiding.
Quoted 25 lines
--- arch/s390/mm/mmap.c | 2 +- include/linux/mm.h | 6 +++--- include/linux/pagemap.h | 2 +- mm/util.c | 11 ++++++----- 4 files changed, 11 insertions(+), 10 deletions(-)arch/s390/mm/mmap.c +1
-1@@ -27,7 +27,7 @@ static unsigned long stack_maxrandom_size(void)return STACK_RND_MASK <<PAGE_SHIFT ; } -static inline int mmap_is_legacy (struct rlimit *rlim_stack) +static inline int mmap_is_legacy (const struct rlimit *const rlim_stack) {if (current->personality &ADDR_COMPAT_LAYOUT )return 1 ;include/linux/mm.h +0
-0
include/linux/mm.h static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return false · line 989
Quoted 5 lines
include/linux/mm.h +1
-1@@ -986,7 +986,7 @@ static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return falsestatic inline bool vma_is_anon_shmem (const struct vm_area_struct *vma) {return false; }#endif -int vma_is_stack_for_current (struct vm_area_struct *vma); +int vma_is_stack_for_current (const struct vm_area_struct *vma);
Should this also be *const ?
include/linux/mm.h void folio_add_pin(struct folio *folio); · line 2589
Quoted 10 lines
include/linux/mm.h +1
-1/* flush_tlb_range() takes a vma, not a mm, and can care about flags */ #define TLB_FLUSH_VMA (mm,flags) { .vm_mm = (mm), .vm_flags = (flags) }@@ -2585,7 +2585,7 @@ void folio_add_pin(struct folio *folio);int account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc);int __account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc, -struct task_struct *task,bool bypass_rlim); +const struct task_struct *task,bool bypass_rlim);
Dito.
Quoted 16 lines
include/linux/mm.h +1
-1struct kvec ;struct page *get_dump_page (unsigned long addr,int *locked);@@ -3348,7 +3348,7 @@ void anon_vma_interval_tree_verify(struct anon_vma_chain *node); avc; avc = anon_vma_interval_tree_iter_next(avc, start, last))/* mmap.c */ -extern int __vm_enough_memory (struct mm_struct *mm,long pages,int cap_sys_admin); +extern int __vm_enough_memory (const struct mm_struct *mm,long pages,int cap_sys_admin);extern int insert_vm_struct (struct mm_struct *,struct vm_area_struct *);extern void exit_mmap (struct mm_struct *); bool mmap_read_lock_maybe_expand(struct mm_struct *mm, struct vm_area_struct *vma,include/linux/pagemap.h +0
-0
include/linux/pagemap.h static inline void filemap_nr_thps_dec(struct address_space *mapping) · line 554
Quoted 5 lines
include/linux/pagemap.h +1
-1@@ -551,7 +551,7 @@ static inline void filemap_nr_thps_dec(struct address_space *mapping)#endif } -struct address_space *folio_mapping (struct folio *); +struct address_space *folio_mapping (const struct folio *folio);
And this one?
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
Not going to add too much noise as same points as David here :)
I'm still definitely leaning towards us not doing the double-const unless thre's a good reason to or perhaps for larger functions.
But I'm open to being convinced otherwise :)
Quoted 86 lines · David Hildenbrand
Quoted 7 lines · Max Kellermann
We select certain test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Quoted 6 lines
(Even though seemingly unrelated, this also constifies the pointer parameter of
mmap_is_legacy()inarch/s390/mm/mmap.cbecause a copy of the function exists inmm/util.c.)Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>Also here, some getters hiding.
Quoted 24 lines
--- arch/s390/mm/mmap.c | 2 +- include/linux/mm.h | 6 +++--- include/linux/pagemap.h | 2 +- mm/util.c | 11 ++++++----- 4 files changed, 11 insertions(+), 10 deletions(-)arch/s390/mm/mmap.c +1
-1@@ -27,7 +27,7 @@ static unsigned long stack_maxrandom_size(void)return STACK_RND_MASK <<PAGE_SHIFT ; } -static inline int mmap_is_legacy (struct rlimit *rlim_stack) +static inline int mmap_is_legacy (const struct rlimit *const rlim_stack) {if (current->personality &ADDR_COMPAT_LAYOUT )return 1 ;include/linux/mm.h +0
-0include/linux/mm.h static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return false · line 988
Quoted 4 lines
include/linux/mm.h +1
-1@@ -986,7 +986,7 @@ static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return falsestatic inline bool vma_is_anon_shmem (const struct vm_area_struct *vma) {return false; }#endif -int vma_is_stack_for_current (struct vm_area_struct *vma); +int vma_is_stack_for_current (const struct vm_area_struct *vma);Should this also be *const ?
include/linux/mm.h void folio_add_pin(struct folio *folio); · line 2587
Quoted 7 lines
include/linux/mm.h +1
-1/* flush_tlb_range() takes a vma, not a mm, and can care about flags */ #define TLB_FLUSH_VMA (mm,flags) { .vm_mm = (mm), .vm_flags = (flags) }@@ -2585,7 +2585,7 @@ void folio_add_pin(struct folio *folio);int account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc);int __account_locked_vm (struct mm_struct *mm,unsigned long pages,bool inc, -struct task_struct *task,bool bypass_rlim); +const struct task_struct *task,bool bypass_rlim);Dito.
Quoted 15 lines
include/linux/mm.h +1
-1struct kvec ;struct page *get_dump_page (unsigned long addr,int *locked);@@ -3348,7 +3348,7 @@ void anon_vma_interval_tree_verify(struct anon_vma_chain *node); avc; avc = anon_vma_interval_tree_iter_next(avc, start, last))/* mmap.c */ -extern int __vm_enough_memory (struct mm_struct *mm,long pages,int cap_sys_admin); +extern int __vm_enough_memory (const struct mm_struct *mm,long pages,int cap_sys_admin);extern int insert_vm_struct (struct mm_struct *,struct vm_area_struct *);extern void exit_mmap (struct mm_struct *); bool mmap_read_lock_maybe_expand(struct mm_struct *mm, struct vm_area_struct *vma,include/linux/pagemap.h +0
-0include/linux/pagemap.h static inline void filemap_nr_thps_dec(struct address_space *mapping) · line 553
Quoted 4 lines
include/linux/pagemap.h +1
-1@@ -551,7 +551,7 @@ static inline void filemap_nr_thps_dec(struct address_space *mapping)#endif } -struct address_space *folio_mapping (struct folio *); +struct address_space *folio_mapping (const struct folio *folio);And this one?
--
CheersDavid / dhildenb
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
In reply to David HildenbrandQuoted 4 lines · David Hildenbrand
include/linux/mm.h static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return false · line 989
Quoted 2 lines
include/linux/mm.h +1
-1-int vma_is_stack_for_current (struct vm_area_struct *vma); +int vma_is_stack_for_current (const struct vm_area_struct *vma);Should this also be *const ?
No. These are function protoypes. A "const" on a parameter value (pointer address, not pointed-to memory) makes no sense on a prototype.
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
Quoted 9 lines · Max Kellermann
Quoted 4 lines · David Hildenbrand
include/linux/mm.h static inline bool vma_is_shmem(const struct vm_area_struct *vma) { return false · line 989
Quoted 2 lines
include/linux/mm.h +1
-1-int vma_is_stack_for_current (struct vm_area_struct *vma); +int vma_is_stack_for_current (const struct vm_area_struct *vma);Should this also be *const ?
No. These are function protoypes. A "const" on a parameter value (pointer address, not pointed-to memory) makes no sense on a prototype.
But couldn't you argue the same about variable names? In most (not all :) ) we keep declaration + definition in sync. So thus my confusion.
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
Quoted 8 lines · David Hildenbrand
Quoted 5 lines
Quoted 1 line
Should this also be *const ?
No. These are function protoypes. A "const" on a parameter value (pointer address, not pointed-to memory) makes no sense on a prototype.
But couldn't you argue the same about variable names? In most (not all :) ) we keep declaration + definition in sync. So thus my confusion.
Variable names in the prototypes have no effect either, but they serve as useful documentation.
Whereas the "const" on a parameter value documents nothing - it's an implementation detail whether the function would like to modify parameter values. That implementation detail has no effect for the caller.
Of course, we could have "const" in the prototype as well. This boils down to personal taste. It's not my taste (has no use, has no effect, documents nothing, only adds noise for no gain), so I didn't add it. If you prefer to have that, I'll leave my taste and home and add it, but only after you guys make up your minds about whether you want to have const parameters at all.
Re: [PATCH v5 06/12] mm, s390: constify mapping related test functions for improved const-correctness
Quoted 18 lines · Max Kellermann
Quoted 8 lines · David Hildenbrand
Quoted 5 lines
Quoted 1 line
Should this also be *const ?
No. These are function protoypes. A "const" on a parameter value (pointer address, not pointed-to memory) makes no sense on a prototype.
But couldn't you argue the same about variable names? In most (not all :) ) we keep declaration + definition in sync. So thus my confusion.
Variable names in the prototypes have no effect either, but they serve as useful documentation.
Whereas the "const" on a parameter value documents nothing - it's an implementation detail whether the function would like to modify parameter values. That implementation detail has no effect for the caller.
Quoted 6 lines
Of course, we could have "const" in the prototype as well. This boils down to personal taste. It's not my taste (has no use, has no effect, documents nothing, only adds noise for no gain), so I didn't add it. If you prefer to have that, I'll leave my taste and home and add it, but only after you guys make up your minds about whether you want to have const parameters at all.
Valid points. The problem is that it could very soon become inconsistent.
For example, when I write a new function I usually just copy what I have from the definition into the declaration.
For example, checkpatch complains about missing variable names and I think it complains when "extern" is used for functions.
If we were to decide to go that route (not keep them in perfect sync), I guess it would be reasonable to extend checkpatch to warn if "*const" is used in a declaration. (perl magic, no idea how hard that would be)
I'm sure there are false positives in the following:
$ git grep "\*const" *.h | grep -v inline | wc -l 403
Signature · 3 lines
Cheers David / dhildenb
PATCH v5 7/12 parisc: constify mmap_upper_limit() parameter for improved const-correctness
In reply to Max KellermannThis piece is necessary to make the rlim_stack parameter to mmap_base() const.
--- arch/parisc/include/asm/processor.h | 2 +- arch/parisc/kernel/sys_parisc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
arch/parisc/include/asm/processor.h +1 -1
@@ -48,7 +48,7 @@#ifndef __ASSEMBLER__struct rlimit ; -unsigned long mmap_upper_limit (struct rlimit *rlim_stack); +unsigned long mmap_upper_limit (const struct rlimit *rlim_stack);unsigned long calc_max_stack_size (unsigned long stack_max);/*
arch/parisc/kernel/sys_parisc.c +1 -1
@@ -77,7 +77,7 @@ unsigned long calc_max_stack_size(unsigned long stack_max)* indicating that "current" should be used instead of a passed-in * value from the exec bprm as done with arch_pick_mmap_layout(). */ -unsigned long mmap_upper_limit (struct rlimit *rlim_stack) +unsigned long mmap_upper_limit (const struct rlimit *const rlim_stack) {unsigned long stack_base;
Signature · 1 line
2.47.2
Re: [PATCH v5 07/12] parisc: constify mmap_upper_limit() parameter for improved const-correctness
Quoted 15 lines · Max Kellermann
This piece is necessary to make the
rlim_stackparameter tommap_base()const.Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>--- arch/parisc/include/asm/processor.h | 2 +- arch/parisc/kernel/sys_parisc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)arch/parisc/include/asm/processor.h +0
-0
arch/parisc/include/asm/processor.h line 52
Quoted 6 lines
arch/parisc/include/asm/processor.h +1
-1@@ -48,7 +48,7 @@#ifndef __ASSEMBLER__struct rlimit ; -unsigned long mmap_upper_limit (struct rlimit *rlim_stack); +unsigned long mmap_upper_limit (const struct rlimit *rlim_stack);unsigned long calc_max_stack_size (unsigned long stack_max);
*const like in the other case?
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 07/12] parisc: constify mmap_upper_limit() parameter for improved const-correctness
Quoted 11 lines · David Hildenbrand
Quoted 10 lines · Max Kellermann
This piece is necessary to make the
rlim_stackparameter tommap_base()const.Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>--- arch/parisc/include/asm/processor.h | 2 +- arch/parisc/kernel/sys_parisc.c | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-)
Quoted 12 lines
arch/parisc/include/asm/processor.h line 51
Quoted 10 lines
arch/parisc/include/asm/processor.h +1
-1@@ -48,7 +48,7 @@#ifndef __ASSEMBLER__struct rlimit ; -unsigned long mmap_upper_limit (struct rlimit *rlim_stack); +unsigned long mmap_upper_limit (const struct rlimit *rlim_stack);unsigned long calc_max_stack_size (unsigned long stack_max);*const like in the other case?
Ditto :>)
Quoted 6 lines
--
CheersDavid / dhildenb
PATCH v5 8/12 mm: constify arch_pick_mmap_layout() for improved const-correctness
In reply to Max KellermannThis function only reads from the rlimit pointer (but writes to the mm_struct pointer which is kept without const).
All callees are already const-ified or (internal functions) are being constified by this patch.
--- arch/s390/mm/mmap.c | 5 +++-- arch/sparc/kernel/sys_sparc_64.c | 3 ++- arch/x86/mm/mmap.c | 7 ++++--- include/linux/sched/mm.h | 4 ++-- mm/util.c | 9 ++++++--- 5 files changed, 17 insertions(+), 11 deletions(-)
arch/s390/mm/mmap.c +3 -2
@@ -47,7 +47,7 @@ static unsigned long mmap_base_legacy(unsigned long rnd) }static inline unsigned long mmap_base (unsigned long rnd, -struct rlimit *rlim_stack) +const struct rlimit *const rlim_stack) {unsigned long gap = rlim_stack->rlim_cur;unsigned long pad =stack_maxrandom_size () + stack_guard_gap;
@@ -169,7 +169,8 @@ unsigned long arch_get_unmapped_area_topdown(struct file *filp, unsigned long ad* This function, called very early during the creation of a new * process VM image, sets up which VM layout function to use: */ -void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) +void arch_pick_mmap_layout (struct mm_struct *const mm, +const struct rlimit *const rlim_stack) {unsigned long random_factor =0UL ;
arch/sparc/kernel/sys_sparc_64.c +2 -1
@@ -294,7 +294,8 @@ static unsigned long mmap_rnd(void)return rnd <<PAGE_SHIFT ; } -void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) +void arch_pick_mmap_layout (struct mm_struct *const mm, +const struct rlimit *const rlim_stack) {unsigned long random_factor =mmap_rnd ();unsigned long gap;
arch/x86/mm/mmap.c +4 -3
@@ -80,7 +80,7 @@ unsigned long arch_mmap_rnd(void) }static unsigned long mmap_base (unsigned long rnd,unsigned long task_size, -struct rlimit *rlim_stack) +const struct rlimit *const rlim_stack) {unsigned long gap = rlim_stack->rlim_cur;unsigned long pad =stack_maxrandom_size (task_size) + stack_guard_gap;
@@ -110,7 +110,7 @@ static unsigned long mmap_legacy_base(unsigned long rnd,*/ static void arch_pick_mmap_base (unsigned long *base,unsigned long *legacy_base,unsigned long random_factor,unsigned long task_size, -struct rlimit *rlim_stack) +const struct rlimit *const rlim_stack) { *legacy_base =mmap_legacy_base (random_factor, task_size);if (mmap_is_legacy ())
@@ -119,7 +119,8 @@ static void arch_pick_mmap_base(unsigned long *base, unsigned long *legacy_base, *base =mmap_base (random_factor, task_size, rlim_stack); } -void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) +void arch_pick_mmap_layout (struct mm_struct *const mm, +const struct rlimit *const rlim_stack) {if (mmap_is_legacy ())mm_flags_clear (MMF_TOPDOWN , mm);
include/linux/sched/mm.h +2 -2
@@ -178,7 +178,7 @@ static inline void mm_update_next_owner(struct mm_struct *mm)#endif extern void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack); +const struct rlimit *rlim_stack); unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,
@@ -211,7 +211,7 @@ generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, unsigned long flags, vm_flags_t vm_flags);#else static inline void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack) {} +const struct rlimit *rlim_stack) {}#endif static inline bool in_vfork (struct task_struct *tsk)
mm/util.c +6 -3
@@ -431,7 +431,8 @@ static int mmap_is_legacy(const struct rlimit *const rlim_stack)#define MIN_GAP (SZ_128M)#define MAX_GAP (STACK_TOP / 6 * 5) -static unsigned long mmap_base (unsigned long rnd,struct rlimit *rlim_stack) +static unsigned long mmap_base (const unsigned long rnd, +const struct rlimit *const rlim_stack) {#ifdef CONFIG_STACK_GROWSUP /*
@@ -462,7 +463,8 @@ static unsigned long mmap_base(unsigned long rnd, struct rlimit *rlim_stack)#endif } -void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) +void arch_pick_mmap_layout (struct mm_struct *const mm, +const struct rlimit *const rlim_stack) {unsigned long random_factor =0UL ;
@@ -478,7 +480,8 @@ void arch_pick_mmap_layout(struct mm_struct *mm, struct rlimit *rlim_stack) } }#elif defined(CONFIG_MMU) && !defined(HAVE_ARCH_PICK_MMAP_LAYOUT) -void arch_pick_mmap_layout (struct mm_struct *mm,struct rlimit *rlim_stack) +void arch_pick_mmap_layout (struct mm_struct *const mm, +const struct rlimit *const rlim_stack) { mm->mmap_base =TASK_UNMAPPED_BASE ;mm_flags_clear (MMF_TOPDOWN , mm);
Signature · 1 line
2.47.2
Re: [PATCH v5 08/12] mm: constify arch_pick_mmap_layout() for improved const-correctness
Quoted 9 lines · Max Kellermann
This function only reads from the rlimit pointer (but writes to the
mm_structpointer which is kept withoutconst).All callees are already const-ified or (internal functions) are being constified by this patch.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
[...]
Quoted 13 lines
index
2201da0afecc..0232d983b715100644include/linux/sched/mm.h +1
-1@@ -178,7 +178,7 @@ static inline void mm_update_next_owner(struct mm_struct *mm)#endif extern void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack); +const struct rlimit *rlim_stack); unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,
include/linux/sched/mm.h generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, · line 215
Quoted 6 lines
include/linux/sched/mm.h +1
-1@@ -211,7 +211,7 @@ generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, unsigned long flags, vm_flags_t vm_flags);#else static inline void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack) {} +const struct rlimit *rlim_stack) {}#endif
Should both these cases also use *const?
(for the latter we probably don't care either, but maybe just to be consistent)
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 08/12] mm: constify arch_pick_mmap_layout() for improved const-correctness
Quoted 33 lines · David Hildenbrand
Quoted 9 lines · Max Kellermann
This function only reads from the rlimit pointer (but writes to the
mm_structpointer which is kept withoutconst).All callees are already const-ified or (internal functions) are being constified by this patch.
Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
[...]
Quoted 11 lines
index
2201da0afecc..0232d983b715100644include/linux/sched/mm.h +1
-1@@ -178,7 +178,7 @@ static inline void mm_update_next_owner(struct mm_struct *mm)#endif extern void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack); +const struct rlimit *rlim_stack); unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,include/linux/sched/mm.h generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, · line 215
Quoted 6 lines
include/linux/sched/mm.h +1
-1@@ -211,7 +211,7 @@ generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, unsigned long flags, vm_flags_t vm_flags);#else static inline void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack) {} +const struct rlimit *rlim_stack) {}#endif Should both these cases also use *const?
Quoted 2 lines
(for the latter we probably don't care either, but maybe just to be consistent)
It'd be pretty hideous aesthetically. Really feel like we shouldn't be doing it...
Imagine
static inline void foo(const struct mm_struct *const mm, const struct vm_area_struct *const vma,
const struct rlimit *const rlim_stack);
I mean that's _just not readable_.
Quoted 6 lines
--
CheersDavid / dhildenb
Re: [PATCH v5 08/12] mm: constify arch_pick_mmap_layout() for improved const-correctness
In reply to David HildenbrandQuoted 22 lines · David Hildenbrand
Quoted 13 lines
index
2201da0afecc..0232d983b715100644include/linux/sched/mm.h +1
-1@@ -178,7 +178,7 @@ static inline void mm_update_next_owner(struct mm_struct *mm)#endif extern void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack); +const struct rlimit *rlim_stack); unsigned long arch_get_unmapped_area(struct file *filp, unsigned long addr,include/linux/sched/mm.h generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, · line 215
Quoted 6 lines
include/linux/sched/mm.h +1
-1@@ -211,7 +211,7 @@ generic_get_unmapped_area_topdown(struct file *filp, unsigned long addr, unsigned long flags, vm_flags_t vm_flags);#else static inline void arch_pick_mmap_layout (struct mm_struct *mm, -struct rlimit *rlim_stack) {} +const struct rlimit *rlim_stack) {}#endif Should both these cases also use *const?
Quoted 2 lines
(for the latter we probably don't care either, but maybe just to be consistent)
Actually, it would *only* make sense on the latter, because the former is a prototype...
PATCH v5 9/12 mm: constify ptdesc_pmd_pts_count() and folio_get_private()
In reply to Max KellermannThese functions from mm_types.h are trivial getters that should never write to the given pointers.
--- include/linux/mm_types.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
include/linux/mm_types.h +2 -2
@@ -632,7 +632,7 @@ static inline void ptdesc_pmd_pts_dec(struct ptdesc *ptdesc)atomic_dec (&ptdesc->pt_share_count); } -static inline int ptdesc_pmd_pts_count (struct ptdesc *ptdesc) +static inline int ptdesc_pmd_pts_count (const struct ptdesc *const ptdesc) {return atomic_read (&ptdesc->pt_share_count); }
@@ -660,7 +660,7 @@ static inline void set_page_private(struct page *page, unsigned long private) page->private = private; } -static inline void *folio_get_private (struct folio *folio) +static inline void *folio_get_private (const struct folio *const folio) {return folio->private; }
Signature · 1 line
2.47.2
Re: [PATCH v5 09/12] mm: constify ptdesc_pmd_pts_count() and folio_get_private()
Quoted 6 lines · Max Kellermann
These functions from
mm_types.hare trivial getters that should never write to the given pointers.Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 09/12] mm: constify ptdesc_pmd_pts_count() and folio_get_private()
In reply to Max KellermannQuoted 5 lines · Max Kellermann
These functions from
mm_types.hare trivial getters that should never write to the given pointers.Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>
(Again, on basis of us figuring out whether we want the double-const)
LGTM, so:
Quoted 29 lines
--- include/linux/mm_types.h | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)include/linux/mm_types.h +2
-3@@ -632,7 +632,7 @@ static inline void ptdesc_pmd_pts_dec(struct ptdesc *ptdesc)atomic_dec (&ptdesc->pt_share_count); } -static inline int ptdesc_pmd_pts_count (struct ptdesc *ptdesc) +static inline int ptdesc_pmd_pts_count (const struct ptdesc *const ptdesc) {return atomic_read (&ptdesc->pt_share_count); }@@ -660,7 +660,7 @@ static inline void set_page_private(struct page *page, unsigned long private) page->private = private; } -static inline void *folio_get_private (struct folio *folio) +static inline void *folio_get_private (const struct folio *const folio) {return folio->private; } --2.47.2
PATCH v5 10/12 mm: constify various inline test functions for improved const-correctness
In reply to Max KellermannWe select certain test functions from mm_inline.h which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
One exception is the function folio_migrate_refs() which does write to the "new" folio pointer; there, only the "old" folio pointer is being constified; only its "flags" field is read, but nothing written.
--- include/linux/mm_inline.h | 26 ++++++++++++++------------ 1 file changed, 14 insertions(+), 12 deletions(-)
include/linux/mm_inline.h +14 -12
@@ -25,7 +25,7 @@* 0 if @folio is a normal anonymous folio, a tmpfs folio or otherwise * ram or swap backed folio. */ -static inline int folio_is_file_lru (struct folio *folio) +static inline int folio_is_file_lru (const struct folio *const folio) {return !folio_test_swapbacked (folio); }
@@ -84,7 +84,7 @@ static __always_inline void __folio_clear_lru_flags(struct folio *folio)* Return: The LRU list a folio should be on, as an index * into the array of LRU lists. */ -static __always_inline enum lru_list folio_lru_list(struct folio *folio) +static __always_inline enum lru_list folio_lru_list(const struct folio *const folio) {enum lru_list lru;
@@ -141,7 +141,7 @@ static inline int lru_tier_from_refs(int refs, bool workingset)return workingset ?MAX_NR_TIERS -1 :order_base_2 (refs); } -static inline int folio_lru_refs (struct folio *folio) +static inline int folio_lru_refs (const struct folio *const folio) {unsigned long flags =READ_ONCE (folio->flags.f);
@@ -154,14 +154,14 @@ static inline int folio_lru_refs(struct folio *folio)return ((flags &LRU_REFS_MASK ) >>LRU_REFS_PGOFF ) +1 ; } -static inline int folio_lru_gen (struct folio *folio) +static inline int folio_lru_gen (const struct folio *folio) {unsigned long flags =READ_ONCE (folio->flags.f);return ((flags &LRU_GEN_MASK ) >>LRU_GEN_PGOFF ) -1 ; } -static inline bool lru_gen_is_active (struct lruvec *lruvec,int gen) +static inline bool lru_gen_is_active (const struct lruvec *const lruvec,const int gen) {unsigned long max_seq = lruvec->lrugen.max_seq;
@@ -217,12 +217,13 @@ static inline void lru_gen_update_size(struct lruvec *lruvec, struct folio *foliVM_WARN_ON_ONCE (lru_gen_is_active (lruvec, old_gen) && !lru_gen_is_active (lruvec, new_gen)); } -static inline unsigned long lru_gen_folio_seq (struct lruvec *lruvec,struct folio *folio, +static inline unsigned long lru_gen_folio_seq (const struct lruvec *const lruvec, +const struct folio *const folio,bool reclaiming) {int gen;int type =folio_is_file_lru (folio); -struct lru_gen_folio *lrugen = &lruvec->lrugen; +const struct lru_gen_folio *lrugen = &lruvec->lrugen;/* * +-----------------------------------+-----------------------------------+
@@ -302,7 +303,8 @@ static inline bool lru_gen_del_folio(struct lruvec *lruvec, struct folio *folio,return true; } -static inline void folio_migrate_refs (struct folio *new,struct folio *old) +static inline void folio_migrate_refs (struct folio *const new, +const struct folio *const old) {unsigned long refs =READ_ONCE (old->flags.f) &LRU_REFS_MASK ;
@@ -330,7 +332,7 @@ static inline bool lru_gen_del_folio(struct lruvec *lruvec, struct folio *folio,return false; } -static inline void folio_migrate_refs (struct folio *new,struct folio *old) +static inline void folio_migrate_refs (struct folio *new,const struct folio *old) { }
@@ -508,7 +510,7 @@ static inline void dec_tlb_flush_pending(struct mm_struct *mm)atomic_dec (&mm->tlb_flush_pending); } -static inline bool mm_tlb_flush_pending (struct mm_struct *mm) +static inline bool mm_tlb_flush_pending (const struct mm_struct *const mm) {/* * Must be called after having acquired the PTL; orders against that
@@ -521,7 +523,7 @@ static inline bool mm_tlb_flush_pending(struct mm_struct *mm)return atomic_read (&mm->tlb_flush_pending); } -static inline bool mm_tlb_flush_nested (struct mm_struct *mm) +static inline bool mm_tlb_flush_nested (const struct mm_struct *const mm) {/* * Similar to mm_tlb_flush_pending(), we must have acquired the PTL
@@ -605,7 +607,7 @@ pte_install_uffd_wp_if_needed(struct vm_area_struct *vma, unsigned long addr,return false; } -static inline bool vma_has_recency (struct vm_area_struct *vma) +static inline bool vma_has_recency (const struct vm_area_struct *const vma) {if (vma->vm_flags & (VM_SEQ_READ |VM_RAND_READ ))return false;
Signature · 1 line
2.47.2
Re: [PATCH v5 10/12] mm: constify various inline test functions for improved const-correctness
Quoted 8 lines · Max Kellermann
We select certain test functions from
mm_inline.hwhich either invoke each other, functions that are already const-ified, or no further functions.It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Quoted 7 lines
One exception is the function
folio_migrate_refs()which does write to the "new" folio pointer; there, only the "old" folio pointer is being constified; only its "flags" field is read, but nothing written.Signed-off-by Max Kellermann <[email protected]>Reviewed-by Vishal Moola (Oracle) <[email protected]>---
Also some getters hiding (and functions that actually implement logic -- folio_migrate_refs())
[...]
include/linux/mm_inline.h static inline int folio_lru_refs(struct folio *folio) · line 157
Quoted 3 lines
include/linux/mm_inline.h +1
-1-static inline int folio_lru_gen (struct folio *folio) +static inline int folio_lru_gen (const struct folio *folio)
*const ?
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 10/12] mm: constify various inline test functions for improved const-correctness
Quoted 2 lines · David Hildenbrand
Also some getters hiding (and functions that actually implement logic --
folio_migrate_refs())
Indeed; I mentioned that one in the message body, and I thought this was good enough, but I dropped the word "test" from the subject line to avoid confusion.
Quoted 4 lines
include/linux/mm_inline.h static inline int folio_lru_refs(struct folio *folio) · line 157
Quoted 2 lines
include/linux/mm_inline.h +1
-1-static inline int folio_lru_gen (struct folio *folio) +static inline int folio_lru_gen (const struct folio *folio)*const ?
Right. Added for next revision.
PATCH v5 11/12 mm: constify assert/test functions in mm.h
In reply to Max KellermannFor improved const-correctness.
We select certain assert and test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
--- include/linux/mm.h | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-)
include/linux/mm.h +20 -20
@@ -703,7 +703,7 @@ static inline void release_fault_lock(struct vm_fault *vmf)mmap_read_unlock (vmf->vma->vm_mm); } -static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (struct vm_fault *const vmf) {if (vmf->flags &FAULT_FLAG_VMA_LOCK )vma_assert_locked (vmf->vma);
@@ -716,7 +716,7 @@ static inline void release_fault_lock(struct vm_fault *vmf)mmap_read_unlock (vmf->vma->vm_mm); } -static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (const struct vm_fault *vmf) {mmap_assert_locked (vmf->vma->vm_mm); }
@@ -859,7 +859,7 @@ static inline bool vma_is_initial_stack(const struct vm_area_struct *vma) vma->vm_end >= vma->vm_mm->start_stack; } -static inline bool vma_is_temporary_stack (struct vm_area_struct *vma) +static inline bool vma_is_temporary_stack (const struct vm_area_struct *const vma) {int maybe_stack = vma->vm_flags & (VM_GROWSDOWN |VM_GROWSUP );
@@ -873,7 +873,7 @@ static inline bool vma_is_temporary_stack(struct vm_area_struct *vma)return false; } -static inline bool vma_is_foreign (struct vm_area_struct *vma) +static inline bool vma_is_foreign (const struct vm_area_struct *const vma) {if (!current->mm)return true;
@@ -884,7 +884,7 @@ static inline bool vma_is_foreign(struct vm_area_struct *vma)return false; } -static inline bool vma_is_accessible (struct vm_area_struct *vma) +static inline bool vma_is_accessible (const struct vm_area_struct *const vma) {return vma->vm_flags &VM_ACCESS_FLAGS ; }
@@ -895,7 +895,7 @@ static inline bool is_shared_maywrite(vm_flags_t vm_flags) (VM_SHARED |VM_MAYWRITE ); } -static inline bool vma_is_shared_maywrite (struct vm_area_struct *vma) +static inline bool vma_is_shared_maywrite (const struct vm_area_struct *const vma) {return is_shared_maywrite (vma->vm_flags); }
@@ -1839,7 +1839,7 @@ static inline struct folio *pfn_folio(unsigned long pfn) }#ifdef CONFIG_MMU -static inline pte_t mk_pte (struct page *page,pgprot_t pgprot) +static inline pte_t mk_pte (const struct page *page,pgprot_t pgprot) {return pfn_pte (page_to_pfn (page), pgprot); }
@@ -1854,7 +1854,7 @@ static inline pte_t mk_pte(struct page *page, pgprot_t pgprot)* * Return: A page table entry suitable for mapping this folio. */ -static inline pte_t folio_mk_pte (struct folio *folio,pgprot_t pgprot) +static inline pte_t folio_mk_pte (const struct folio *folio,pgprot_t pgprot) {return pfn_pte (folio_pfn (folio), pgprot); }
@@ -1870,7 +1870,7 @@ static inline pte_t folio_mk_pte(struct folio *folio, pgprot_t pgprot)* * Return: A page table entry suitable for mapping this folio. */ -static inline pmd_t folio_mk_pmd (struct folio *folio,pgprot_t pgprot) +static inline pmd_t folio_mk_pmd (const struct folio *folio,pgprot_t pgprot) {return pmd_mkhuge (pfn_pmd (folio_pfn (folio), pgprot)); }
@@ -1886,7 +1886,7 @@ static inline pmd_t folio_mk_pmd(struct folio *folio, pgprot_t pgprot)* * Return: A page table entry suitable for mapping this folio. */ -static inline pud_t folio_mk_pud (struct folio *folio,pgprot_t pgprot) +static inline pud_t folio_mk_pud (const struct folio *folio,pgprot_t pgprot) {return pud_mkhuge (pfn_pud (folio_pfn (folio), pgprot)); }
@@ -3488,7 +3488,7 @@ struct vm_area_struct *vma_lookup(struct mm_struct *mm, unsigned long addr)return mtree_load (&mm->mm_mt, addr); } -static inline unsigned long stack_guard_start_gap (struct vm_area_struct *vma) +static inline unsigned long stack_guard_start_gap (const struct vm_area_struct *const vma) {if (vma->vm_flags &VM_GROWSDOWN )return stack_guard_gap;
@@ -3500,7 +3500,7 @@ static inline unsigned long stack_guard_start_gap(struct vm_area_struct *vma)return 0 ; } -static inline unsigned long vm_start_gap (struct vm_area_struct *vma) +static inline unsigned long vm_start_gap (const struct vm_area_struct *const vma) {unsigned long gap =stack_guard_start_gap (vma);unsigned long vm_start = vma->vm_start;
@@ -3511,7 +3511,7 @@ static inline unsigned long vm_start_gap(struct vm_area_struct *vma)return vm_start; } -static inline unsigned long vm_end_gap (struct vm_area_struct *vma) +static inline unsigned long vm_end_gap (const struct vm_area_struct *const vma) {unsigned long vm_end = vma->vm_end;
@@ -3523,7 +3523,7 @@ static inline unsigned long vm_end_gap(struct vm_area_struct *vma)return vm_end; } -static inline unsigned long vma_pages (struct vm_area_struct *vma) +static inline unsigned long vma_pages (const struct vm_area_struct *const vma) {return (vma->vm_end - vma->vm_start) >>PAGE_SHIFT ; }
@@ -3540,7 +3540,7 @@ static inline struct vm_area_struct *find_exact_vma(struct mm_struct *mm,return vma; } -static inline bool range_in_vma (struct vm_area_struct *vma, +static inline bool range_in_vma (const struct vm_area_struct *const vma,unsigned long start,unsigned long end) {return (vma && vma->vm_start <= start && end <= vma->vm_end);
@@ -3656,7 +3656,7 @@ static inline int vm_fault_to_errno(vm_fault_t vm_fault, int foll_flags)* Indicates whether GUP can follow a PROT_NONE mapped page, or whether * a (NUMA hinting) fault is required. */ -static inline bool gup_can_follow_protnone (struct vm_area_struct *vma, +static inline bool gup_can_follow_protnone (const struct vm_area_struct *const vma,unsigned int flags) {/*
@@ -3786,7 +3786,7 @@ static inline bool debug_guardpage_enabled(void)return static_branch_unlikely (&_debug_guardpage_enabled); } -static inline bool page_is_guard (struct page *page) +static inline bool page_is_guard (const struct page *const page) {if (!debug_guardpage_enabled ())return false;
@@ -3817,7 +3817,7 @@ static inline void debug_pagealloc_map_pages(struct page *page, int numpages) {}static inline void debug_pagealloc_unmap_pages (struct page *page,int numpages) {}static inline unsigned int debug_guardpage_minorder (void ) {return 0 ; }static inline bool debug_guardpage_enabled (void ) {return false; } -static inline bool page_is_guard (struct page *page) {return false; } +static inline bool page_is_guard (const struct page *const page) {return false; }static inline bool set_page_guard (struct zone *zone,struct page *page,unsigned int order) {return false; } static inline void clear_page_guard(struct zone *zone, struct page *page,
@@ -3899,7 +3899,7 @@ void vmemmap_free(unsigned long start, unsigned long end,#endif #ifdef CONFIG_SPARSEMEM_VMEMMAP -static inline unsigned long vmem_altmap_offset (struct vmem_altmap *altmap) +static inline unsigned long vmem_altmap_offset (const struct vmem_altmap *altmap) {/* number of pfns from base where pfn_to_page() is valid */ if (altmap)
@@ -3913,7 +3913,7 @@ static inline void vmem_altmap_free(struct vmem_altmap *altmap, altmap->alloc -= nr_pfns; }#else -static inline unsigned long vmem_altmap_offset (struct vmem_altmap *altmap) +static inline unsigned long vmem_altmap_offset (const struct vmem_altmap *altmap) {return 0 ; }
Signature · 1 line
2.47.2
Re: [PATCH v5 11/12] mm: constify assert/test functions in mm.h
Quoted 29 lines · Max Kellermann
For improved const-correctness.
We select certain assert and test functions which either invoke each other, functions that are already const-ified, or no further functions.
It is therefore relatively trivial to const-ify them, which provides a basis for further const-ification further up the call stack.
Signed-off-by Max Kellermann <[email protected]>--- include/linux/mm.h | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-)include/linux/mm.h +1
-1@@ -703,7 +703,7 @@ static inline void release_fault_lock(struct vm_fault *vmf)mmap_read_unlock (vmf->vma->vm_mm); } -static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (struct vm_fault *const vmf) {if (vmf->flags &FAULT_FLAG_VMA_LOCK )vma_assert_locked (vmf->vma);
include/linux/mm.h static inline void release_fault_lock(struct vm_fault *vmf) · line 720
Quoted 6 lines
include/linux/mm.h +1
-1@@ -716,7 +716,7 @@ static inline void release_fault_lock(struct vm_fault *vmf)mmap_read_unlock (vmf->vma->vm_mm); } -static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (const struct vm_fault *vmf) {
This confused me a bit: in the upper variant it's "*const" and here it's "const *".
There are multiple such cases here, which might imply that it is not "relatively trivial to const-ify them". :)
Did you miss to convert some of these cases properly, or is there a reason for this inconsistency?
Signature · 3 lines
Cheers David / dhildenb
Re: [PATCH v5 11/12] mm: constify assert/test functions in mm.h
Quoted 6 lines · David Hildenbrand
include/linux/mm.h static inline void release_fault_lock(struct vm_fault *vmf) · line 720
Quoted 3 lines
include/linux/mm.h +1
-1-static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (const struct vm_fault *vmf) {This confused me a bit: in the upper variant it's "*const" and here it's "const *".
That was indeed a mistake. Both should be "const*const".
Quoted 2 lines
There are multiple such cases here, which might imply that it is not "relatively trivial to const-ify them". :)
I double-checked this patch and couldn't find any other such mistake. Or do you mean the function vs prototype thing on parameter values?
Re: [PATCH v5 11/12] mm: constify assert/test functions in mm.h
Quoted 8 lines · Max Kellermann
Quoted 6 lines · David Hildenbrand
include/linux/mm.h static inline void release_fault_lock(struct vm_fault *vmf) · line 720
Quoted 3 lines
include/linux/mm.h +1
-1-static inline void assert_fault_locked (struct vm_fault *vmf) +static inline void assert_fault_locked (const struct vm_fault *vmf) {This confused me a bit: in the upper variant it's "*const" and here it's "const *".
Quoted 7 lines
That was indeed a mistake. Both should be "const*const".
Quoted 2 lines
There are multiple such cases here, which might imply that it is not "relatively trivial to const-ify them". :)
I double-checked this patch and couldn't find any other such mistake. Or do you mean the function vs prototype thing on parameter values?
If there is a simple rule (declaration/definition), then it's trivial. Probably worth spelling out that rule somewhere (unless I missed it).
As raised in the other reply, not sure if we should just keep them in sync.
I'm, not the biggest fan of the *const in general here. I can see why Andrew suggested it, but only doing that for pointers is a bit weird. Anyhow, that discussion is likely happening elsewhere, and I don't think there is real harm when doing it, as long as we are consistent with what we're doing.
Signature · 3 lines
Cheers David / dhildenb
PATCH v5 12/12 mm: constify highmem related functions for improved const-correctness
In reply to Max KellermannLots of functions in mm/highmem.c do not write to the given pointers and do not call functions that take non-const pointers and can therefore be constified.
This includes functions like kunmap() which might be implemented in a way that writes to the pointer (e.g. to update reference counters or mapping fields), but currently are not.
kmap() on the other hand cannot be made const because it calls set_page_address() which is non-const in some architectures/configurations.
--- arch/arm/include/asm/highmem.h | 6 ++--- arch/xtensa/include/asm/highmem.h | 2 +- include/linux/highmem-internal.h | 44 +++++++++++++++++-------------- include/linux/highmem.h | 8 +++--- mm/highmem.c | 10 +++---- 5 files changed, 37 insertions(+), 33 deletions(-)
arch/arm/include/asm/highmem.h +3 -3
@@ -46,9 +46,9 @@ extern pte_t *pkmap_page_table;#endif #ifdef ARCH_NEEDS_KMAP_HIGH_GET -extern void *kmap_high_get (struct page *page); +extern void *kmap_high_get (const struct page *page); -static inline void *arch_kmap_local_high_get (struct page *page) +static inline void *arch_kmap_local_high_get (const struct page *page) {if (IS_ENABLED (CONFIG_DEBUG_HIGHMEM ) && !cache_is_vivt ())return NULL ;
@@ -57,7 +57,7 @@ static inline void *arch_kmap_local_high_get(struct page *page)#define arch_kmap_local_high_get arch_kmap_local_high_get#else /* ARCH_NEEDS_KMAP_HIGH_GET */ -static inline void *kmap_high_get (struct page *page) +static inline void *kmap_high_get (const struct page *const page) {return NULL ; }
arch/xtensa/include/asm/highmem.h +1 -1
@@ -29,7 +29,7 @@#if DCACHE_WAY_SIZE >PAGE_SIZE #define get_pkmap_color get_pkmap_color -static inline int get_pkmap_color (struct page *page) +static inline int get_pkmap_color (const struct page *const page) {return DCACHE_ALIAS (page_to_phys (page)); }
include/linux/highmem-internal.h +24 -20
140 lines not on this page · raw message
include/linux/highmem.h +4 -4
@@ -43,7 +43,7 @@ static inline void *kmap(struct page *page);* Counterpart to kmap(). A NOOP for CONFIG_HIGHMEM=n and for mappings of * pages in the low memory area. */ -static inline void kunmap (struct page *page); +static inline void kunmap (const struct page *page);/** * kmap_to_page - Get the page for a kmap'ed address
@@ -93,7 +93,7 @@ static inline void kmap_flush_unused(void);* disabling migration in order to keep the virtual address stable across * preemption. No caller of kmap_local_page() can rely on this side effect. */ -static inline void *kmap_local_page (struct page *page); +static inline void *kmap_local_page (const struct page *page);/** * kmap_local_folio - Map a page in this folio for temporary usage
@@ -129,7 +129,7 @@ static inline void *kmap_local_page(struct page *page);* Context: Can be invoked from any context. * Return: The virtual address of @offset. */ -static inline void *kmap_local_folio (struct folio *folio,size_t offset); +static inline void *kmap_local_folio (const struct folio *folio,size_t offset);/** * kmap_atomic - Atomically map a page for temporary usage - Deprecated!
@@ -176,7 +176,7 @@ static inline void *kmap_local_folio(struct folio *folio, size_t offset);* kunmap_atomic(vaddr2); * kunmap_atomic(vaddr1); */ -static inline void *kmap_atomic (struct page *page); +static inline void *kmap_atomic (const struct page *page);/* Highmem related interfaces for management code */ static inline unsigned long nr_free_highpages (void );
mm/highmem.c +5 -5
@@ -61,7 +61,7 @@ static inline int kmap_local_calc_idx(int idx)/* * Determine color of virtual address where the page should be mapped. */ -static inline unsigned int get_pkmap_color (struct page *page) +static inline unsigned int get_pkmap_color (const struct page *const page) {return 0 ; }
@@ -334,7 +334,7 @@ EXPORT_SYMBOL(kmap_high);* * This can be called from any context. */ -void *kmap_high_get (struct page *page) +void *kmap_high_get (const struct page *const page) {unsigned long vaddr, flags;
@@ -356,7 +356,7 @@ void *kmap_high_get(struct page *page)* If ARCH_NEEDS_KMAP_HIGH_GET is not defined then this may be called * only from user context. */ -void kunmap_high (struct page *page) +void kunmap_high (const struct page *const page) {unsigned long vaddr;unsigned long nr;
@@ -508,7 +508,7 @@ static inline void kmap_local_idx_pop(void)#endif #ifndef arch_kmap_local_high_get -static inline void *arch_kmap_local_high_get (struct page *page) +static inline void *arch_kmap_local_high_get (const struct page *const page) {return NULL ; }
@@ -572,7 +572,7 @@ void *__kmap_local_pfn_prot(unsigned long pfn, pgprot_t prot) }EXPORT_SYMBOL_GPL (__kmap_local_pfn_prot); -void *__kmap_local_page_prot (struct page *page,pgprot_t prot) +void *__kmap_local_page_prot (const struct page *const page,const pgprot_t prot) {void *kmap;
Signature · 1 line
2.47.2
Re: [PATCH v5 12/12] mm: constify highmem related functions for improved const-correctness
Quoted 8 lines · Max Kellermann
Lots of functions in
mm/highmem.cdo not write to the given pointers and do not call functions that take non-const pointers and can therefore be constified.This includes functions like
kunmap()which might be implemented in a way that writes to the pointer (e.g. to update reference counters or mapping fields), but currently are not.
Quoted 5 lines
kmap()on the other hand cannot be made const because it callsset_page_address()which is non-const in some architectures/configurations.Signed-off-by Max Kellermann <[email protected]>
Some inconsitencies on const vs. double-const here.
We need to figure this out :)
Since const <type> *val will pretty well work in all circumstances, I again think we ought to just do that instead of double-const.
Thanks
Quoted 292 lines
--- arch/arm/include/asm/highmem.h | 6 ++--- arch/xtensa/include/asm/highmem.h | 2 +- include/linux/highmem-internal.h | 44 +++++++++++++++++-------------- include/linux/highmem.h | 8 +++--- mm/highmem.c | 10 +++---- 5 files changed, 37 insertions(+), 33 deletions(-)arch/arm/include/asm/highmem.h +3
-3@@ -46,9 +46,9 @@ extern pte_t *pkmap_page_table;#endif #ifdef ARCH_NEEDS_KMAP_HIGH_GET -extern void *kmap_high_get (struct page *page); +extern void *kmap_high_get (const struct page *page); -static inline void *arch_kmap_local_high_get (struct page *page) +static inline void *arch_kmap_local_high_get (const struct page *page) {if (IS_ENABLED (CONFIG_DEBUG_HIGHMEM ) && !cache_is_vivt ())return NULL ;@@ -57,7 +57,7 @@ static inline void *arch_kmap_local_high_get(struct page *page)#define arch_kmap_local_high_get arch_kmap_local_high_get#else /* ARCH_NEEDS_KMAP_HIGH_GET */ -static inline void *kmap_high_get (struct page *page) +static inline void *kmap_high_get (const struct page *const page) {return NULL ; }arch/xtensa/include/asm/highmem.h +1
-1@@ -29,7 +29,7 @@#if DCACHE_WAY_SIZE >PAGE_SIZE #define get_pkmap_color get_pkmap_color -static inline int get_pkmap_color (struct page *page) +static inline int get_pkmap_color (const struct page *const page) {return DCACHE_ALIAS (page_to_phys (page)); }include/linux/highmem-internal.h +24
-20@@ -7,7 +7,7 @@*/ #ifdef CONFIG_KMAP_LOCAL void *__kmap_local_pfn_prot (unsigned long pfn,pgprot_t prot); -void *__kmap_local_page_prot (struct page *page,pgprot_t prot); +void *__kmap_local_page_prot (const struct page *page,pgprot_t prot);void kunmap_local_indexed (const void *vaddr);void kmap_local_fork (struct task_struct *tsk);void __kmap_local_sched_out (void );@@ -33,11 +33,11 @@ static inline void kmap_flush_tlb(unsigned long addr) { }#endif void *kmap_high (struct page *page); -void kunmap_high (struct page *page); +void kunmap_high (const struct page *page);void __kmap_flush_unused (void );struct page *__kmap_to_page (void *addr); -static inline void *kmap (struct page *page) +static inline void *kmap (struct page *const page) {void *addr;@@ -50,7 +50,7 @@ static inline void *kmap(struct page *page)return addr; } -static inline void kunmap (struct page *page) +static inline void kunmap (const struct page *const page) {might_sleep ();if (!PageHighMem (page))@@ -68,12 +68,12 @@ static inline void kmap_flush_unused(void)__kmap_flush_unused (); } -static inline void *kmap_local_page (struct page *page) +static inline void *kmap_local_page (const struct page *const page) {return __kmap_local_page_prot (page, kmap_prot); } -static inline void *kmap_local_page_try_from_panic (struct page *page) +static inline void *kmap_local_page_try_from_panic (const struct page *const page) {if (!PageHighMem (page))return page_address (page);@@ -81,13 +81,15 @@ static inline void *kmap_local_page_try_from_panic(struct page *page)return NULL ; } -static inline void *kmap_local_folio (struct folio *folio,size_t offset) +static inline void *kmap_local_folio (const struct folio *const folio, +const size_t offset) { -struct page *page =folio_page (folio, offset /PAGE_SIZE ); +const struct page *page =folio_page (folio, offset /PAGE_SIZE );return __kmap_local_page_prot (page, kmap_prot) + offset %PAGE_SIZE ; } -static inline void *kmap_local_page_prot (struct page *page,pgprot_t prot) +static inline void *kmap_local_page_prot (const struct page *const page, +const pgprot_t prot) {return __kmap_local_page_prot (page, prot); }@@ -102,7 +104,7 @@ static inline void __kunmap_local(const void *vaddr)kunmap_local_indexed (vaddr); } -static inline void *kmap_atomic_prot (struct page *page,pgprot_t prot) +static inline void *kmap_atomic_prot (const struct page *const page,const pgprot_t prot) {if (IS_ENABLED (CONFIG_PREEMPT_RT ))migrate_disable ();@@ -113,7 +115,7 @@ static inline void *kmap_atomic_prot(struct page *page, pgprot_t prot)return __kmap_local_page_prot (page, prot); } -static inline void *kmap_atomic (struct page *page) +static inline void *kmap_atomic (const struct page *const page) {return kmap_atomic_prot (page, kmap_prot); }@@ -167,38 +169,40 @@ static inline struct page *kmap_to_page(void *addr)return virt_to_page (addr); } -static inline void *kmap (struct page *page) +static inline void *kmap (struct page *const page) {might_sleep ();return page_address (page); } -static inline void kunmap_high (struct page *page) { } +static inline void kunmap_high (const struct page *const page) { }static inline void kmap_flush_unused (void ) { } -static inline void kunmap (struct page *page) +static inline void kunmap (const struct page *const page) {#ifdef ARCH_HAS_FLUSH_ON_KUNMAP kunmap_flush_on_unmap (page_address (page));#endif } -static inline void *kmap_local_page (struct page *page) +static inline void *kmap_local_page (const struct page *const page) {return page_address (page); } -static inline void *kmap_local_page_try_from_panic (struct page *page) +static inline void *kmap_local_page_try_from_panic (const struct page *const page) {return page_address (page); } -static inline void *kmap_local_folio (struct folio *folio,size_t offset) +static inline void *kmap_local_folio (const struct folio *const folio, +const size_t offset) {return folio_address (folio) + offset; } -static inline void *kmap_local_page_prot (struct page *page,pgprot_t prot) +static inline void *kmap_local_page_prot (const struct page *const page, +const pgprot_t prot) {return kmap_local_page (page); }@@ -215,7 +219,7 @@ static inline void __kunmap_local(const void *addr)#endif } -static inline void *kmap_atomic (struct page *page) +static inline void *kmap_atomic (const struct page *const page) {if (IS_ENABLED (CONFIG_PREEMPT_RT ))migrate_disable ();@@ -225,7 +229,7 @@ static inline void *kmap_atomic(struct page *page)return page_address (page); } -static inline void *kmap_atomic_prot (struct page *page,pgprot_t prot) +static inline void *kmap_atomic_prot (const struct page *const page,const pgprot_t prot) {return kmap_atomic (page); }include/linux/highmem.h +4
-4@@ -43,7 +43,7 @@ static inline void *kmap(struct page *page);* Counterpart to kmap(). A NOOP for CONFIG_HIGHMEM=n and for mappings of * pages in the low memory area. */ -static inline void kunmap (struct page *page); +static inline void kunmap (const struct page *page);/** * kmap_to_page - Get the page for a kmap'ed address @@ -93,7 +93,7 @@ static inline void kmap_flush_unused(void);* disabling migration in order to keep the virtual address stable across * preemption. No caller of kmap_local_page() can rely on this side effect. */ -static inline void *kmap_local_page (struct page *page); +static inline void *kmap_local_page (const struct page *page);/** * kmap_local_folio - Map a page in this folio for temporary usage @@ -129,7 +129,7 @@ static inline void *kmap_local_page(struct page *page);* Context: Can be invoked from any context. * Return: The virtual address of @offset. */ -static inline void *kmap_local_folio (struct folio *folio,size_t offset); +static inline void *kmap_local_folio (const struct folio *folio,size_t offset);/** * kmap_atomic - Atomically map a page for temporary usage - Deprecated! @@ -176,7 +176,7 @@ static inline void *kmap_local_folio(struct folio *folio, size_t offset);* kunmap_atomic(vaddr2); * kunmap_atomic(vaddr1); */ -static inline void *kmap_atomic (struct page *page); +static inline void *kmap_atomic (const struct page *page);/* Highmem related interfaces for management code */ static inline unsigned long nr_free_highpages (void );mm/highmem.c +5
-6@@ -61,7 +61,7 @@ static inline int kmap_local_calc_idx(int idx)/* * Determine color of virtual address where the page should be mapped. */ -static inline unsigned int get_pkmap_color (struct page *page) +static inline unsigned int get_pkmap_color (const struct page *const page) {return 0 ; }@@ -334,7 +334,7 @@ EXPORT_SYMBOL(kmap_high);* * This can be called from any context. */ -void *kmap_high_get (struct page *page) +void *kmap_high_get (const struct page *const page) {unsigned long vaddr, flags;@@ -356,7 +356,7 @@ void *kmap_high_get(struct page *page)* If ARCH_NEEDS_KMAP_HIGH_GET is not defined then this may be called * only from user context. */ -void kunmap_high (struct page *page) +void kunmap_high (const struct page *const page) {unsigned long vaddr;unsigned long nr;@@ -508,7 +508,7 @@ static inline void kmap_local_idx_pop(void)#endif #ifndef arch_kmap_local_high_get -static inline void *arch_kmap_local_high_get (struct page *page) +static inline void *arch_kmap_local_high_get (const struct page *const page) {return NULL ; }@@ -572,7 +572,7 @@ void *__kmap_local_pfn_prot(unsigned long pfn, pgprot_t prot) }EXPORT_SYMBOL_GPL (__kmap_local_pfn_prot); -void *__kmap_local_page_prot (struct page *page,pgprot_t prot) +void *__kmap_local_page_prot (const struct page *const page,const pgprot_t prot) {void *kmap; --2.47.2
Quoted 48 lines · Max Kellermann
For improved const-correctness.
This patch series systematically adds const qualifiers to pointer parameters throughout the memory management subsystem, establishing a foundation for improved const-correctness across the entire Linux kernel.
Const-correctness provides multiple benefits:
1. Type Safety: The compiler enforces that functions marked as taking const parameters cannot accidentally modify the data, catching potential bugs at compile time rather than runtime.
2. Compiler Optimizations: When the compiler knows data won't be modified, it can generate more efficient code through better register allocation, code motion, and aliasing analysis.
3. API Documentation: Const qualifiers serve as self-documenting code, making it immediately clear to developers which functions are read-only operations versus those that modify state.
4. Maintenance Safety: Future modifications to const-correct code are less likely to introduce subtle bugs, as the compiler will reject attempts to modify data that should remain unchanged.
The memory management subsystem is a fundamental building block of the kernel. Most higher-level kernel subsystems (filesystems, drivers, networking) depend on mm interfaces. By establishing const-correctness at this foundational level:
1. Enables Propagation: Higher-level subsystems can adopt const-correctness in their own interfaces. Without const-correct mm functions, filesystems cannot mark their own parameters const when they need to call mm functions.
2. Maximum Impact: Changes to core mm APIs benefit the entire kernel, as these functions are called from virtually every subsystem.
3. Prevents Impedance Mismatch: Without const-correctness in mm, other subsystems must either cast away const (dangerous) or avoid using const altogether (missing optimization opportunities).
Each patch focuses on a specific header or subsystem component to ease review and bisection.
This work was initially posted as a single large patch:
https://lore.kernel.org/lkml/[email protected]/
Quoted 6 lines
Following feedback from Lorenzo Stoakes and David Hildenbrand, it has been split into focused, reviewable chunks. The approach was validated with a smaller patch that received agreement:
https://lore.kernel.org/lkml/[email protected]/Signed-off-by Max Kellermann <[email protected]>
Seems fine,
Quoted 17 lines · Max Kellermann
For improved const-correctness.
This patch series systematically adds const qualifiers to pointer parameters throughout the memory management subsystem, establishing a foundation for improved const-correctness across the entire Linux kernel.
Const-correctness provides multiple benefits:
1. Type Safety: The compiler enforces that functions marked as taking const parameters cannot accidentally modify the data, catching potential bugs at compile time rather than runtime.
2. Compiler Optimizations: When the compiler knows data won't be modified, it can generate more efficient code through better register allocation, code motion, and aliasing analysis.
Quoted 7 lines
3. API Documentation: Const qualifiers serve as self-documenting code, making it immediately clear to developers which functions are read-only operations versus those that modify state.
4. Maintenance Safety: Future modifications to const-correct code are less likely to introduce subtle bugs, as the compiler will reject attempts to modify data that should remain unchanged.
I think all of the above is really a lot of noise that doesn't add a huge amount of value.
Please in your own words say why you are doing it, and also please mention why you feel it's justified to do:
const <type> *const param
As mentioned on review of 2/12.
I'm pretty well leaning towards - let's just not do the 2nd const at all, unless there's a really good reason to do so.
There are also legit cases where you might want to reassign a local variable.
Largely, granted, you shouldn't be reassinging params, but it's pretty constraining.
Quoted 14 lines
The memory management subsystem is a fundamental building block of the kernel. Most higher-level kernel subsystems (filesystems, drivers, networking) depend on mm interfaces. By establishing const-correctness at this foundational level:
1. Enables Propagation: Higher-level subsystems can adopt const-correctness in their own interfaces. Without const-correct mm functions, filesystems cannot mark their own parameters const when they need to call mm functions.
2. Maximum Impact: Changes to core mm APIs benefit the entire kernel, as these functions are called from virtually every subsystem.
Quoted 6 lines
3. Prevents Impedance Mismatch: Without const-correctness in mm, other subsystems must either cast away const (dangerous) or avoid using const altogether (missing optimization opportunities).
Each patch focuses on a specific header or subsystem component to ease review and bisection.
All this is unnecessary noise, can you summarise more succinctly.
More words than 'const-ify everything' doesn't mean 'several paragraphs of noise'.
Quoted 9 lines
This work was initially posted as a single large patch:
https://lore.kernel.org/lkml/[email protected]/Following feedback from Lorenzo Stoakes and David Hildenbrand, it has been split into focused, reviewable chunks. The approach was validated with a smaller patch that received agreement:
https://lore.kernel.org/lkml/[email protected]/
^--- All of this should be below the line and is associated with versions not the series as a whole really.
Quoted 20 lines
Signed-off-by Max Kellermann <[email protected]>---
v1 -> v2:
- made several parameter values const (i.e. the pointer address, not just the pointed-to memory), as suggested by Andrew Morton and Yuanchu Xie
- drop existing+obsolete "extern" keywords on lines modified by these patches (suggested by Vishal Moola)
- add missing parameter names on lines modified by these patches (suggested by Vishal Moola)
- more "const" pointers (e.g. thetask_structpassed toprocess_shares_mm())
- add missing "const" to s390, fixing s390 build failure
- moved themmap_is_legacy()change inarch/s390/mm/mmap.cfrom 08/12 to 06/12 (suggested by Vishal Moola)v2 -> v3:
- remove garbage from 06/12
- changed tags on subject line (suggested by Matthew Wilcox)
Quoted 7 lines
v3 -> v4:
- more verbose commit messages including a listing of function names (suggested by David Hildenbrand and Lorenzo Stoakes)v4 -> v5:
- back to shorter commit messages after an agreement between David Hildenbrand and Lorenzo Stoakes was found
I did ask you to do this in reverse order and add lore links :) I mean these aren't big things but it's REALLY helpful for reviewers.
Thanks.
Quoted 48 lines
Max Kellermann (12):
mm: constify shmem related test functions for improved
const-correctness
mm: constify pagemap related test functions for improved
const-correctness
mm: constify zone related test functions for improved
const-correctness
fs: constify mapping related test functions for improved
const-correctness
mm: constifyprocess_shares_mm()for improved const-correctness mm, s390: constify mapping related test functions for improved
const-correctness
parisc: constifymmap_upper_limit()parameter for improved
const-correctness
mm: constifyarch_pick_mmap_layout()for improved const-correctness
mm: constifyptdesc_pmd_pts_count()andfolio_get_private()
mm: constify various inline test functions for improved
const-correctness
mm: constify assert/test functions inmm.h
mm: constify highmem related functions for improved const-correctnessarch/arm/include/asm/highmem.h | 6 +-- arch/parisc/include/asm/processor.h | 2 +- arch/parisc/kernel/sys_parisc.c | 2 +- arch/s390/mm/mmap.c | 7 ++-- arch/sparc/kernel/sys_sparc_64.c | 3 +- arch/x86/mm/mmap.c | 7 ++-- arch/xtensa/include/asm/highmem.h | 2 +- include/linux/fs.h | 7 ++-- include/linux/highmem-internal.h | 44 +++++++++++---------- include/linux/highmem.h | 8 ++-- include/linux/mm.h | 56 +++++++++++++-------------- include/linux/mm_inline.h | 26 +++++++------ include/linux/mm_types.h | 4 +- include/linux/mmzone.h | 42 ++++++++++---------- include/linux/pagemap.h | 59 +++++++++++++++-------------- include/linux/sched/mm.h | 4 +- include/linux/shmem_fs.h | 4 +- mm/highmem.c | 10 ++--- mm/oom_kill.c | 7 ++-- mm/shmem.c | 6 +-- mm/util.c | 20 ++++++---- 21 files changed, 171 insertions(+), 155 deletions(-)--
2.47.2
Matthew Wilcox
maintainerQuoted 1 line · Max Kellermann
For improved const-correctness.
SLOW DOWN.
This series is unimportant churn. There's no way it should be up to v5 already. Wait a freaking week before you send another version.
Quoted 7 lines · Matthew Wilcox
Quoted 1 line · Max Kellermann
For improved const-correctness.
SLOW DOWN.
This series is unimportant churn. There's no way it should be up to v5 already. Wait a freaking week before you send another version.
... while we have v6 already on the list :D
In this case, it was reasonable to follow up rather sooner than later (too much drama and v6 now seems to make people happy, I wouldn't want that spanning over multiple weeks).
But yeah, too many revisions on a single day.
Signature · 3 lines
Cheers David / dhildenb