[PATCH 0/5] Cleanup patches of vmalloc

[PATCH 0/5] Cleanup patches of vmalloc

[Baoquan He]

Some cleanup patches found when reading vmalloc code.

Built kernel and run on x86_64 machine.

Baoquan He (5):
  mm/vmalloc: remove the unnecessary type check
  mm/vmalloc: remove the redundant boundary check
  mm/vmalloc: fix typo in local variable name
  mm/vmalloc: Add code comment for find_vmap_area_exceed_addr()
  mm: ioremap: rename ioremap_page_range() to ioremap_range()

 Documentation/ia64/aliasing.rst  |  2 +-
 arch/alpha/kernel/proto.h        |  2 +-
 arch/arc/mm/ioremap.c            |  2 +-
 arch/arm/mm/ioremap.c            |  6 +++---
 arch/arm64/mm/ioremap.c          |  2 +-
 arch/hexagon/mm/ioremap.c        |  2 +-
 arch/ia64/mm/ioremap.c           |  2 +-
 arch/mips/loongson64/init.c      |  2 +-
 arch/mips/mm/ioremap.c           |  2 +-
 arch/openrisc/mm/ioremap.c       |  2 +-
 arch/parisc/mm/ioremap.c         |  2 +-
 arch/powerpc/kernel/isa-bridge.c |  2 +-
 arch/powerpc/kernel/pci_64.c     |  2 +-
 arch/powerpc/mm/ioremap.c        |  2 +-
 arch/s390/pci/pci.c              |  2 +-
 arch/sh/kernel/cpu/sh4/sq.c      |  2 +-
 arch/sh/mm/ioremap.c             |  2 +-
 arch/x86/mm/ioremap.c            |  2 +-
 arch/xtensa/mm/ioremap.c         |  2 +-
 drivers/pci/pci.c                |  2 +-
 include/linux/io.h               |  4 ++--
 mm/ioremap.c                     |  2 +-
 mm/vmalloc.c                     | 27 ++++++++++++---------------
 23 files changed, 37 insertions(+), 40 deletions(-)

-- 
2.34.1

[PATCH 1/5] mm/vmalloc: remove the unnecessary type check

[Baoquan He]

In function adjust_va_to_fit_type(), it checks all values of passed
in fit type, including NOTHING_FIT in the else branch. In fact, it's
unnecessary to check NOTHING_FIT since the check has been done before
adjust_va_to_fit_type() is called in all call sites. So clean it up.

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/vmalloc.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 07db42455dd4..b7a138ab7b79 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -1418,8 +1418,6 @@ adjust_va_to_fit_type(struct vmap_area *va,
 		 * Shrink this VA to remaining size.
 		 */
 		va->va_start = nva_start_addr + size;
-	} else {
-		return -1;
 	}
 
 	if (type != FL_FIT_TYPE) {
-- 
2.34.1

Re: [PATCH 1/5] mm/vmalloc: remove the unnecessary type check

[Uladzislau Rezki]

> In function adjust_va_to_fit_type(), it checks all values of passed
> in fit type, including NOTHING_FIT in the else branch. In fact, it's
> unnecessary to check NOTHING_FIT since the check has been done before
> adjust_va_to_fit_type() is called in all call sites. So clean it up.
> 
> Signed-off-by: Baoquan He <bhe@redhat.com>
> ---
>  mm/vmalloc.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 07db42455dd4..b7a138ab7b79 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1418,8 +1418,6 @@ adjust_va_to_fit_type(struct vmap_area *va,
>  		 * Shrink this VA to remaining size.
>  		 */
>  		va->va_start = nva_start_addr + size;
> -	} else {
> -		return -1;
>  	}
>  
>  	if (type != FL_FIT_TYPE) {
> -- 
> 2.34.1
> 
Why not just invoke the classify_va_fit_type() inside the adjust_va_to_fit_type()?
In this case we do not need to rely on upper-stack checks and the split logic 
becomes fully implemented in one solid function.

--
Uladzislau Rezki

Re: [PATCH 1/5] mm/vmalloc: remove the unnecessary type check

[Baoquan He]

On 06/06/22 at 10:36pm, Uladzislau Rezki wrote:
> > In function adjust_va_to_fit_type(), it checks all values of passed
> > in fit type, including NOTHING_FIT in the else branch. In fact, it's
> > unnecessary to check NOTHING_FIT since the check has been done before
> > adjust_va_to_fit_type() is called in all call sites. So clean it up.
> > 
> > Signed-off-by: Baoquan He <bhe@redhat.com>
> > ---
> >  mm/vmalloc.c | 2 --
> >  1 file changed, 2 deletions(-)
> > 
> > diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> > index 07db42455dd4..b7a138ab7b79 100644
> > --- a/mm/vmalloc.c
> > +++ b/mm/vmalloc.c
> > @@ -1418,8 +1418,6 @@ adjust_va_to_fit_type(struct vmap_area *va,
> >  		 * Shrink this VA to remaining size.
> >  		 */
> >  		va->va_start = nva_start_addr + size;
> > -	} else {
> > -		return -1;
> >  	}
> >  
> >  	if (type != FL_FIT_TYPE) {
> > -- 
> > 2.34.1
> > 
> Why not just invoke the classify_va_fit_type() inside the adjust_va_to_fit_type()?
> In this case we do not need to rely on upper-stack checks and the split logic 
> becomes fully implemented in one solid function.

Ah, that's much better, let me post a new one. Thanks a lot.

[PATCH 2/5] mm/vmalloc: remove the redundant boundary check

[Baoquan He]

In function find_va_links(), when traversing the vmap_area tree, the
comparing to check if the passed in 'va' is above or below 'tmp_va'
is redundant, assuming both 'va' and 'tmp_va' has ->va_start <= ->va_end.

Here, to simplify the checking as code change.

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/vmalloc.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index b7a138ab7b79..c7e1634ff2b9 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -874,11 +874,9 @@ find_va_links(struct vmap_area *va,
 		 * Trigger the BUG() if there are sides(left/right)
 		 * or full overlaps.
 		 */
-		if (va->va_start < tmp_va->va_end &&
-				va->va_end <= tmp_va->va_start)
+		if (va->va_end <= tmp_va->va_start)
 			link = &(*link)->rb_left;
-		else if (va->va_end > tmp_va->va_start &&
-				va->va_start >= tmp_va->va_end)
+		else if (va->va_start >= tmp_va->va_end)
 			link = &(*link)->rb_right;
 		else {
 			WARN(1, "vmalloc bug: 0x%lx-0x%lx overlaps with 0x%lx-0x%lx\n",
-- 
2.34.1

Re: [PATCH 2/5] mm/vmalloc: remove the redundant boundary check

[Uladzislau Rezki]

On Mon, Jun 06, 2022 at 04:39:06PM +0800, Baoquan He wrote:
> In function find_va_links(), when traversing the vmap_area tree, the
> comparing to check if the passed in 'va' is above or below 'tmp_va'
> is redundant, assuming both 'va' and 'tmp_va' has ->va_start <= ->va_end.
> 
> Here, to simplify the checking as code change.
> 
> Signed-off-by: Baoquan He <bhe@redhat.com>
> ---
>  mm/vmalloc.c | 6 ++----
>  1 file changed, 2 insertions(+), 4 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index b7a138ab7b79..c7e1634ff2b9 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -874,11 +874,9 @@ find_va_links(struct vmap_area *va,
>  		 * Trigger the BUG() if there are sides(left/right)
>  		 * or full overlaps.
>  		 */
> -		if (va->va_start < tmp_va->va_end &&
> -				va->va_end <= tmp_va->va_start)
> +		if (va->va_end <= tmp_va->va_start)
>  			link = &(*link)->rb_left;
> -		else if (va->va_end > tmp_va->va_start &&
> -				va->va_start >= tmp_va->va_end)
> +		else if (va->va_start >= tmp_va->va_end)
>  			link = &(*link)->rb_right;
>  		else {
>  			WARN(1, "vmalloc bug: 0x%lx-0x%lx overlaps with 0x%lx-0x%lx\n",
> -- 
> 2.34.1
> 
Make sense to me: Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>

--
Uladzislau Rezki

[PATCH 3/5] mm/vmalloc: fix typo in local variable name

[Baoquan He]

In __purge_vmap_area_lazy(), rename local_pure_list to local_purge_list.

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/vmalloc.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index c7e1634ff2b9..11dfc897de40 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -1673,32 +1673,32 @@ static void purge_fragmented_blocks_allcpus(void);
 static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end)
 {
 	unsigned long resched_threshold;
-	struct list_head local_pure_list;
+	struct list_head local_purge_list;
 	struct vmap_area *va, *n_va;
 
 	lockdep_assert_held(&vmap_purge_lock);
 
 	spin_lock(&purge_vmap_area_lock);
 	purge_vmap_area_root = RB_ROOT;
-	list_replace_init(&purge_vmap_area_list, &local_pure_list);
+	list_replace_init(&purge_vmap_area_list, &local_purge_list);
 	spin_unlock(&purge_vmap_area_lock);
 
-	if (unlikely(list_empty(&local_pure_list)))
+	if (unlikely(list_empty(&local_purge_list)))
 		return false;
 
 	start = min(start,
-		list_first_entry(&local_pure_list,
+		list_first_entry(&local_purge_list,
 			struct vmap_area, list)->va_start);
 
 	end = max(end,
-		list_last_entry(&local_pure_list,
+		list_last_entry(&local_purge_list,
 			struct vmap_area, list)->va_end);
 
 	flush_tlb_kernel_range(start, end);
 	resched_threshold = lazy_max_pages() << 1;
 
 	spin_lock(&free_vmap_area_lock);
-	list_for_each_entry_safe(va, n_va, &local_pure_list, list) {
+	list_for_each_entry_safe(va, n_va, &local_purge_list, list) {
 		unsigned long nr = (va->va_end - va->va_start) >> PAGE_SHIFT;
 		unsigned long orig_start = va->va_start;
 		unsigned long orig_end = va->va_end;
-- 
2.34.1

Re: [PATCH 3/5] mm/vmalloc: fix typo in local variable name

[Uladzislau Rezki]

On Mon, Jun 06, 2022 at 04:39:07PM +0800, Baoquan He wrote:
> In __purge_vmap_area_lazy(), rename local_pure_list to local_purge_list.
> 
> Signed-off-by: Baoquan He <bhe@redhat.com>
> ---
>  mm/vmalloc.c | 12 ++++++------
>  1 file changed, 6 insertions(+), 6 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index c7e1634ff2b9..11dfc897de40 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -1673,32 +1673,32 @@ static void purge_fragmented_blocks_allcpus(void);
>  static bool __purge_vmap_area_lazy(unsigned long start, unsigned long end)
>  {
>  	unsigned long resched_threshold;
> -	struct list_head local_pure_list;
> +	struct list_head local_purge_list;
>  	struct vmap_area *va, *n_va;
>  
>  	lockdep_assert_held(&vmap_purge_lock);
>  
>  	spin_lock(&purge_vmap_area_lock);
>  	purge_vmap_area_root = RB_ROOT;
> -	list_replace_init(&purge_vmap_area_list, &local_pure_list);
> +	list_replace_init(&purge_vmap_area_list, &local_purge_list);
>  	spin_unlock(&purge_vmap_area_lock);
>  
> -	if (unlikely(list_empty(&local_pure_list)))
> +	if (unlikely(list_empty(&local_purge_list)))
>  		return false;
>  
>  	start = min(start,
> -		list_first_entry(&local_pure_list,
> +		list_first_entry(&local_purge_list,
>  			struct vmap_area, list)->va_start);
>  
>  	end = max(end,
> -		list_last_entry(&local_pure_list,
> +		list_last_entry(&local_purge_list,
>  			struct vmap_area, list)->va_end);
>  
>  	flush_tlb_kernel_range(start, end);
>  	resched_threshold = lazy_max_pages() << 1;
>  
>  	spin_lock(&free_vmap_area_lock);
> -	list_for_each_entry_safe(va, n_va, &local_pure_list, list) {
> +	list_for_each_entry_safe(va, n_va, &local_purge_list, list) {
>  		unsigned long nr = (va->va_end - va->va_start) >> PAGE_SHIFT;
>  		unsigned long orig_start = va->va_start;
>  		unsigned long orig_end = va->va_end;
> -- 
> 2.34.1
> 
Makes sense to me: Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>

--
Uladzislau Rezki

[PATCH 4/5] mm/vmalloc: Add code comment for find_vmap_area_exceed_addr()

[Baoquan He]

Its behaviour is like find_vma() which finds an area above the specified
address, add comment to make it easier to understand.

And also fix two places of grammer mistake/typo.

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 mm/vmalloc.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 11dfc897de40..860ed9986775 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -790,6 +790,7 @@ unsigned long vmalloc_nr_pages(void)
 	return atomic_long_read(&nr_vmalloc_pages);
 }
 
+/* Look up the first VA which satisfies  addr < va_end,  NULL if none. */
 static struct vmap_area *find_vmap_area_exceed_addr(unsigned long addr)
 {
 	struct vmap_area *va = NULL;
@@ -929,7 +930,7 @@ link_va(struct vmap_area *va, struct rb_root *root,
 		 * Some explanation here. Just perform simple insertion
 		 * to the tree. We do not set va->subtree_max_size to
 		 * its current size before calling rb_insert_augmented().
-		 * It is because of we populate the tree from the bottom
+		 * It is because we populate the tree from the bottom
 		 * to parent levels when the node _is_ in the tree.
 		 *
 		 * Therefore we set subtree_max_size to zero after insertion,
@@ -1659,7 +1660,7 @@ static atomic_long_t vmap_lazy_nr = ATOMIC_LONG_INIT(0);
 
 /*
  * Serialize vmap purging.  There is no actual critical section protected
- * by this look, but we want to avoid concurrent calls for performance
+ * by this lock, but we want to avoid concurrent calls for performance
  * reasons and to make the pcpu_get_vm_areas more deterministic.
  */
 static DEFINE_MUTEX(vmap_purge_lock);
-- 
2.34.1

Re: [PATCH 4/5] mm/vmalloc: Add code comment for find_vmap_area_exceed_addr()

[Uladzislau Rezki]

On Mon, Jun 06, 2022 at 04:39:08PM +0800, Baoquan He wrote:
> Its behaviour is like find_vma() which finds an area above the specified
> address, add comment to make it easier to understand.
> 
> And also fix two places of grammer mistake/typo.
> 
> Signed-off-by: Baoquan He <bhe@redhat.com>
> ---
>  mm/vmalloc.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/vmalloc.c b/mm/vmalloc.c
> index 11dfc897de40..860ed9986775 100644
> --- a/mm/vmalloc.c
> +++ b/mm/vmalloc.c
> @@ -790,6 +790,7 @@ unsigned long vmalloc_nr_pages(void)
>  	return atomic_long_read(&nr_vmalloc_pages);
>  }
>  
> +/* Look up the first VA which satisfies  addr < va_end,  NULL if none. */
>  static struct vmap_area *find_vmap_area_exceed_addr(unsigned long addr)
>  {
>  	struct vmap_area *va = NULL;
> @@ -929,7 +930,7 @@ link_va(struct vmap_area *va, struct rb_root *root,
>  		 * Some explanation here. Just perform simple insertion
>  		 * to the tree. We do not set va->subtree_max_size to
>  		 * its current size before calling rb_insert_augmented().
> -		 * It is because of we populate the tree from the bottom
> +		 * It is because we populate the tree from the bottom
>  		 * to parent levels when the node _is_ in the tree.
>  		 *
>  		 * Therefore we set subtree_max_size to zero after insertion,
> @@ -1659,7 +1660,7 @@ static atomic_long_t vmap_lazy_nr = ATOMIC_LONG_INIT(0);
>  
>  /*
>   * Serialize vmap purging.  There is no actual critical section protected
> - * by this look, but we want to avoid concurrent calls for performance
> + * by this lock, but we want to avoid concurrent calls for performance
>   * reasons and to make the pcpu_get_vm_areas more deterministic.
>   */
>  static DEFINE_MUTEX(vmap_purge_lock);
> -- 
> 2.34.1
> 
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>

--
Uladzislau Rezki

[PATCH 5/5] mm: ioremap: rename ioremap_page_range() to ioremap_range()

[Baoquan He]

Because the current ioremap_page_range() only maps IO address to kernel
virtual address, no struct page pointer passed in or page handling related.
So rename it here.

The renaming is done with below command:
sed -i "s/ioremap_page_range/ioremap_range/g" `git grep -l ioremap_page_range`

Signed-off-by: Baoquan He <bhe@redhat.com>
---
 Documentation/ia64/aliasing.rst  | 2 +-
 arch/alpha/kernel/proto.h        | 2 +-
 arch/arc/mm/ioremap.c            | 2 +-
 arch/arm/mm/ioremap.c            | 6 +++---
 arch/arm64/mm/ioremap.c          | 2 +-
 arch/hexagon/mm/ioremap.c        | 2 +-
 arch/ia64/mm/ioremap.c           | 2 +-
 arch/mips/loongson64/init.c      | 2 +-
 arch/mips/mm/ioremap.c           | 2 +-
 arch/openrisc/mm/ioremap.c       | 2 +-
 arch/parisc/mm/ioremap.c         | 2 +-
 arch/powerpc/kernel/isa-bridge.c | 2 +-
 arch/powerpc/kernel/pci_64.c     | 2 +-
 arch/powerpc/mm/ioremap.c        | 2 +-
 arch/s390/pci/pci.c              | 2 +-
 arch/sh/kernel/cpu/sh4/sq.c      | 2 +-
 arch/sh/mm/ioremap.c             | 2 +-
 arch/x86/mm/ioremap.c            | 2 +-
 arch/xtensa/mm/ioremap.c         | 2 +-
 drivers/pci/pci.c                | 2 +-
 include/linux/io.h               | 4 ++--
 mm/ioremap.c                     | 2 +-
 mm/vmalloc.c                     | 2 +-
 23 files changed, 26 insertions(+), 26 deletions(-)

diff --git a/Documentation/ia64/aliasing.rst b/Documentation/ia64/aliasing.rst
index a08b36aba015..1736be3e9820 100644
--- a/Documentation/ia64/aliasing.rst
+++ b/Documentation/ia64/aliasing.rst
@@ -165,7 +165,7 @@ ioremap()
 
 	If the granule contains non-WB memory, but we can cover the
 	region safely with kernel page table mappings, we can use
-	ioremap_page_range() as most other architectures do.
+	ioremap_range() as most other architectures do.
 
 	Failing all of the above, we have to fall back to a UC mapping.
 
diff --git a/arch/alpha/kernel/proto.h b/arch/alpha/kernel/proto.h
index 5816a31c1b38..7a1aad5f7777 100644
--- a/arch/alpha/kernel/proto.h
+++ b/arch/alpha/kernel/proto.h
@@ -187,7 +187,7 @@ __alpha_remap_area_pages(unsigned long address, unsigned long phys_addr,
 
 	prot = __pgprot(_PAGE_VALID | _PAGE_ASM | _PAGE_KRE
 			| _PAGE_KWE | flags);
-	return ioremap_page_range(address, address + size, phys_addr, prot);
+	return ioremap_range(address, address + size, phys_addr, prot);
 }
 
 /* irq.c */
diff --git a/arch/arc/mm/ioremap.c b/arch/arc/mm/ioremap.c
index 0ee75aca6e10..1b9d6ba7e46a 100644
--- a/arch/arc/mm/ioremap.c
+++ b/arch/arc/mm/ioremap.c
@@ -85,7 +85,7 @@ void __iomem *ioremap_prot(phys_addr_t paddr, unsigned long size,
 		return NULL;
 	area->phys_addr = paddr;
 	vaddr = (unsigned long)area->addr;
-	if (ioremap_page_range(vaddr, vaddr + size, paddr, prot)) {
+	if (ioremap_range(vaddr, vaddr + size, paddr, prot)) {
 		vunmap((void __force *)vaddr);
 		return NULL;
 	}
diff --git a/arch/arm/mm/ioremap.c b/arch/arm/mm/ioremap.c
index 576c0e6c92fc..6c942409f188 100644
--- a/arch/arm/mm/ioremap.c
+++ b/arch/arm/mm/ioremap.c
@@ -110,7 +110,7 @@ void __init add_static_vm_early(struct static_vm *svm)
 int ioremap_page(unsigned long virt, unsigned long phys,
 		 const struct mem_type *mtype)
 {
-	return ioremap_page_range(virt, virt + PAGE_SIZE, phys,
+	return ioremap_range(virt, virt + PAGE_SIZE, phys,
 				  __pgprot(mtype->prot_pte));
 }
 EXPORT_SYMBOL(ioremap_page);
@@ -312,7 +312,7 @@ static void __iomem * __arm_ioremap_pfn_caller(unsigned long pfn,
 		err = remap_area_sections(addr, pfn, size, type);
 	} else
 #endif
-		err = ioremap_page_range(addr, addr + size, paddr,
+		err = ioremap_range(addr, addr + size, paddr,
 					 __pgprot(type->prot_pte));
 
 	if (err) {
@@ -473,7 +473,7 @@ int pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr)
 	if (res->end > IO_SPACE_LIMIT)
 		return -EINVAL;
 
-	return ioremap_page_range(vaddr, vaddr + resource_size(res), phys_addr,
+	return ioremap_range(vaddr, vaddr + resource_size(res), phys_addr,
 				  __pgprot(get_mem_type(pci_ioremap_mem_type)->prot_pte));
 }
 EXPORT_SYMBOL(pci_remap_iospace);
diff --git a/arch/arm64/mm/ioremap.c b/arch/arm64/mm/ioremap.c
index b21f91cd830d..573621e7b173 100644
--- a/arch/arm64/mm/ioremap.c
+++ b/arch/arm64/mm/ioremap.c
@@ -52,7 +52,7 @@ static void __iomem *__ioremap_caller(phys_addr_t phys_addr, size_t size,
 	addr = (unsigned long)area->addr;
 	area->phys_addr = phys_addr;
 
-	err = ioremap_page_range(addr, addr + size, phys_addr, prot);
+	err = ioremap_range(addr, addr + size, phys_addr, prot);
 	if (err) {
 		vunmap((void *)addr);
 		return NULL;
diff --git a/arch/hexagon/mm/ioremap.c b/arch/hexagon/mm/ioremap.c
index 255c5b1ee1a7..90a6cd75d1e1 100644
--- a/arch/hexagon/mm/ioremap.c
+++ b/arch/hexagon/mm/ioremap.c
@@ -30,7 +30,7 @@ void __iomem *ioremap(unsigned long phys_addr, unsigned long size)
 	area = get_vm_area(size, VM_IOREMAP);
 	addr = (unsigned long)area->addr;
 
-	if (ioremap_page_range(addr, addr+size, phys_addr, prot)) {
+	if (ioremap_range(addr, addr+size, phys_addr, prot)) {
 		vunmap((void *)addr);
 		return NULL;
 	}
diff --git a/arch/ia64/mm/ioremap.c b/arch/ia64/mm/ioremap.c
index 55fd3eb753ff..ef2e52591247 100644
--- a/arch/ia64/mm/ioremap.c
+++ b/arch/ia64/mm/ioremap.c
@@ -86,7 +86,7 @@ ioremap (unsigned long phys_addr, unsigned long size)
 
 		area->phys_addr = phys_addr;
 		addr = (void __iomem *) area->addr;
-		if (ioremap_page_range((unsigned long) addr,
+		if (ioremap_range((unsigned long) addr,
 				(unsigned long) addr + size, phys_addr, prot)) {
 			vunmap((void __force *) addr);
 			return NULL;
diff --git a/arch/mips/loongson64/init.c b/arch/mips/loongson64/init.c
index ee8de1735b7c..53f148522663 100644
--- a/arch/mips/loongson64/init.c
+++ b/arch/mips/loongson64/init.c
@@ -162,7 +162,7 @@ static int __init add_legacy_isa_io(struct fwnode_handle *fwnode, resource_size_
 
 	vaddr = PCI_IOBASE + range->io_start;
 
-	ioremap_page_range(vaddr, vaddr + size, hw_start, pgprot_device(PAGE_KERNEL));
+	ioremap_range(vaddr, vaddr + size, hw_start, pgprot_device(PAGE_KERNEL));
 
 	return 0;
 }
diff --git a/arch/mips/mm/ioremap.c b/arch/mips/mm/ioremap.c
index b6dad2fd5575..b1cfeee648a5 100644
--- a/arch/mips/mm/ioremap.c
+++ b/arch/mips/mm/ioremap.c
@@ -101,7 +101,7 @@ void __iomem *ioremap_prot(phys_addr_t phys_addr, unsigned long size,
 	vaddr = (unsigned long)area->addr;
 
 	flags |= _PAGE_GLOBAL | _PAGE_PRESENT | __READABLE | __WRITEABLE;
-	if (ioremap_page_range(vaddr, vaddr + size, phys_addr,
+	if (ioremap_range(vaddr, vaddr + size, phys_addr,
 			__pgprot(flags))) {
 		free_vm_area(area);
 		return NULL;
diff --git a/arch/openrisc/mm/ioremap.c b/arch/openrisc/mm/ioremap.c
index daae13a76743..f5dc775f5715 100644
--- a/arch/openrisc/mm/ioremap.c
+++ b/arch/openrisc/mm/ioremap.c
@@ -64,7 +64,7 @@ void __iomem *__ref ioremap(phys_addr_t addr, unsigned long size)
 		fixmaps_used += (size >> PAGE_SHIFT);
 	}
 
-	if (ioremap_page_range(v, v + size, p,
+	if (ioremap_range(v, v + size, p,
 			__pgprot(pgprot_val(PAGE_KERNEL) | _PAGE_CI))) {
 		if (likely(mem_init_done))
 			vfree(area->addr);
diff --git a/arch/parisc/mm/ioremap.c b/arch/parisc/mm/ioremap.c
index 345ff0b66499..74c940a70b82 100644
--- a/arch/parisc/mm/ioremap.c
+++ b/arch/parisc/mm/ioremap.c
@@ -80,7 +80,7 @@ void __iomem *ioremap(unsigned long phys_addr, unsigned long size)
 		return NULL;
 
 	addr = (void __iomem *) area->addr;
-	if (ioremap_page_range((unsigned long)addr, (unsigned long)addr + size,
+	if (ioremap_range((unsigned long)addr, (unsigned long)addr + size,
 			       phys_addr, pgprot)) {
 		vunmap(addr);
 		return NULL;
diff --git a/arch/powerpc/kernel/isa-bridge.c b/arch/powerpc/kernel/isa-bridge.c
index dc746611ebc0..0eb13278a96e 100644
--- a/arch/powerpc/kernel/isa-bridge.c
+++ b/arch/powerpc/kernel/isa-bridge.c
@@ -46,7 +46,7 @@ static void remap_isa_base(phys_addr_t pa, unsigned long size)
 	WARN_ON_ONCE(size & ~PAGE_MASK);
 
 	if (slab_is_available()) {
-		if (ioremap_page_range(ISA_IO_BASE, ISA_IO_BASE + size, pa,
+		if (ioremap_range(ISA_IO_BASE, ISA_IO_BASE + size, pa,
 				pgprot_noncached(PAGE_KERNEL)))
 			vunmap_range(ISA_IO_BASE, ISA_IO_BASE + size);
 	} else {
diff --git a/arch/powerpc/kernel/pci_64.c b/arch/powerpc/kernel/pci_64.c
index 19b03ddf5631..7adee5cc58a4 100644
--- a/arch/powerpc/kernel/pci_64.c
+++ b/arch/powerpc/kernel/pci_64.c
@@ -138,7 +138,7 @@ void __iomem *ioremap_phb(phys_addr_t paddr, unsigned long size)
 		return NULL;
 
 	addr = (unsigned long)area->addr;
-	if (ioremap_page_range(addr, addr + size, paddr,
+	if (ioremap_range(addr, addr + size, paddr,
 			pgprot_noncached(PAGE_KERNEL))) {
 		vunmap_range(addr, addr + size);
 		return NULL;
diff --git a/arch/powerpc/mm/ioremap.c b/arch/powerpc/mm/ioremap.c
index 4f12504fb405..0cac4e5c8a8f 100644
--- a/arch/powerpc/mm/ioremap.c
+++ b/arch/powerpc/mm/ioremap.c
@@ -89,7 +89,7 @@ void __iomem *do_ioremap(phys_addr_t pa, phys_addr_t offset, unsigned long size,
 	area->phys_addr = pa;
 	va = (unsigned long)area->addr;
 
-	ret = ioremap_page_range(va, va + size, pa, prot);
+	ret = ioremap_range(va, va + size, pa, prot);
 	if (!ret)
 		return (void __iomem *)area->addr + offset;
 
diff --git a/arch/s390/pci/pci.c b/arch/s390/pci/pci.c
index bc980fd313d5..f0568be4ac4b 100644
--- a/arch/s390/pci/pci.c
+++ b/arch/s390/pci/pci.c
@@ -252,7 +252,7 @@ static void __iomem *__ioremap(phys_addr_t addr, size_t size, pgprot_t prot)
 		return NULL;
 
 	vaddr = (unsigned long) area->addr;
-	if (ioremap_page_range(vaddr, vaddr + size, addr, prot)) {
+	if (ioremap_range(vaddr, vaddr + size, addr, prot)) {
 		free_vm_area(area);
 		return NULL;
 	}
diff --git a/arch/sh/kernel/cpu/sh4/sq.c b/arch/sh/kernel/cpu/sh4/sq.c
index a76b94e41e91..a85f0c05e9d6 100644
--- a/arch/sh/kernel/cpu/sh4/sq.c
+++ b/arch/sh/kernel/cpu/sh4/sq.c
@@ -110,7 +110,7 @@ static int __sq_remap(struct sq_mapping *map, pgprot_t prot)
 
 	vma->phys_addr = map->addr;
 
-	if (ioremap_page_range((unsigned long)vma->addr,
+	if (ioremap_range((unsigned long)vma->addr,
 			       (unsigned long)vma->addr + map->size,
 			       vma->phys_addr, prot)) {
 		vunmap(vma->addr);
diff --git a/arch/sh/mm/ioremap.c b/arch/sh/mm/ioremap.c
index 21342581144d..60a582e5b4f3 100644
--- a/arch/sh/mm/ioremap.c
+++ b/arch/sh/mm/ioremap.c
@@ -132,7 +132,7 @@ __ioremap_caller(phys_addr_t phys_addr, unsigned long size,
 	area->phys_addr = phys_addr;
 	orig_addr = addr = (unsigned long)area->addr;
 
-	if (ioremap_page_range(addr, addr + size, phys_addr, pgprot)) {
+	if (ioremap_range(addr, addr + size, phys_addr, pgprot)) {
 		vunmap((void *)orig_addr);
 		return NULL;
 	}
diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
index 1ad0228f8ceb..0c245f4eaa60 100644
--- a/arch/x86/mm/ioremap.c
+++ b/arch/x86/mm/ioremap.c
@@ -286,7 +286,7 @@ __ioremap_caller(resource_size_t phys_addr, unsigned long size,
 	if (memtype_kernel_map_sync(phys_addr, size, pcm))
 		goto err_free_area;
 
-	if (ioremap_page_range(vaddr, vaddr + size, phys_addr, prot))
+	if (ioremap_range(vaddr, vaddr + size, phys_addr, prot))
 		goto err_free_area;
 
 	ret_addr = (void __iomem *) (vaddr + offset);
diff --git a/arch/xtensa/mm/ioremap.c b/arch/xtensa/mm/ioremap.c
index a400188c16b9..8e914a16c04f 100644
--- a/arch/xtensa/mm/ioremap.c
+++ b/arch/xtensa/mm/ioremap.c
@@ -33,7 +33,7 @@ static void __iomem *xtensa_ioremap(unsigned long paddr, unsigned long size,
 	vaddr = (unsigned long)area->addr;
 	area->phys_addr = paddr;
 
-	err = ioremap_page_range(vaddr, vaddr + size, paddr, prot);
+	err = ioremap_range(vaddr, vaddr + size, paddr, prot);
 
 	if (err) {
 		vunmap((void *)vaddr);
diff --git a/drivers/pci/pci.c b/drivers/pci/pci.c
index cfaf40a540a8..d77219be295d 100644
--- a/drivers/pci/pci.c
+++ b/drivers/pci/pci.c
@@ -4234,7 +4234,7 @@ int pci_remap_iospace(const struct resource *res, phys_addr_t phys_addr)
 	if (res->end > IO_SPACE_LIMIT)
 		return -EINVAL;
 
-	return ioremap_page_range(vaddr, vaddr + resource_size(res), phys_addr,
+	return ioremap_range(vaddr, vaddr + resource_size(res), phys_addr,
 				  pgprot_device(PAGE_KERNEL));
 #else
 	/*
diff --git a/include/linux/io.h b/include/linux/io.h
index 5fc800390fe4..80974fa37d53 100644
--- a/include/linux/io.h
+++ b/include/linux/io.h
@@ -21,10 +21,10 @@ void __ioread32_copy(void *to, const void __iomem *from, size_t count);
 void __iowrite64_copy(void __iomem *to, const void *from, size_t count);
 
 #ifdef CONFIG_MMU
-int ioremap_page_range(unsigned long addr, unsigned long end,
+int ioremap_range(unsigned long addr, unsigned long end,
 		       phys_addr_t phys_addr, pgprot_t prot);
 #else
-static inline int ioremap_page_range(unsigned long addr, unsigned long end,
+static inline int ioremap_range(unsigned long addr, unsigned long end,
 				     phys_addr_t phys_addr, pgprot_t prot)
 {
 	return 0;
diff --git a/mm/ioremap.c b/mm/ioremap.c
index 5fe598ecd9b7..d08b30db332b 100644
--- a/mm/ioremap.c
+++ b/mm/ioremap.c
@@ -33,7 +33,7 @@ void __iomem *ioremap_prot(phys_addr_t addr, size_t size, unsigned long prot)
 		return NULL;
 	vaddr = (unsigned long)area->addr;
 
-	if (ioremap_page_range(vaddr, vaddr + size, addr, __pgprot(prot))) {
+	if (ioremap_range(vaddr, vaddr + size, addr, __pgprot(prot))) {
 		free_vm_area(area);
 		return NULL;
 	}
diff --git a/mm/vmalloc.c b/mm/vmalloc.c
index 860ed9986775..32a18ae16bf5 100644
--- a/mm/vmalloc.c
+++ b/mm/vmalloc.c
@@ -312,7 +312,7 @@ static int vmap_range_noflush(unsigned long addr, unsigned long end,
 	return err;
 }
 
-int ioremap_page_range(unsigned long addr, unsigned long end,
+int ioremap_range(unsigned long addr, unsigned long end,
 		phys_addr_t phys_addr, pgprot_t prot)
 {
 	int err;
-- 
2.34.1

Re: [PATCH 5/5] mm: ioremap: rename ioremap_page_range() to ioremap_range()

[Christoph Hellwig]

On Mon, Jun 06, 2022 at 04:39:09PM +0800, Baoquan He wrote:
> Because the current ioremap_page_range() only maps IO address to kernel
> virtual address, no struct page pointer passed in or page handling related.
> So rename it here.
> 
> The renaming is done with below command:
> sed -i "s/ioremap_page_range/ioremap_range/g" `git grep -l ioremap_page_range`

This creates a lot of churn without much of a benefit.  If you want
to get rid of the name please convert most architectures to the
generioc ioremap code first so that all these callers go away..

Re: [PATCH 5/5] mm: ioremap: rename ioremap_page_range() to ioremap_range()

[Baoquan He]

On 06/06/22 at 10:27pm, Christoph Hellwig wrote:
> On Mon, Jun 06, 2022 at 04:39:09PM +0800, Baoquan He wrote:
> > Because the current ioremap_page_range() only maps IO address to kernel
> > virtual address, no struct page pointer passed in or page handling related.
> > So rename it here.
> > 
> > The renaming is done with below command:
> > sed -i "s/ioremap_page_range/ioremap_range/g" `git grep -l ioremap_page_range`
> 
> This creates a lot of churn without much of a benefit.  If you want
> to get rid of the name please convert most architectures to the
> generioc ioremap code first so that all these callers go away..

Thanks for checking.

Yeah, I didn't manually adjust those indentation after replacing via
command. While the name of ioremap_page_range() is misleading, relative
to its implementation.

Converting to use generic ioremap code on most of architectures sounds
like a good idea, just like what the arm/arm64 converting patchset is
doing. From a quick look, not all places can take the converting, and
there's one in pci_remap_iospace() under drivers. So what I need to do
is to:

1)take back this pach;
2)convert as many architectures to the generic ioremap code as possible;
3)rename the left places of ioremap_page_range() to ioremap_range().

Please check if this is what you are suggesting.

Thanks
Baoquan