mm: remove enum page_entry_size

Remove the unnecessary encoding of page order into an enum and pass the
page order directly.  That lets us get rid of pe_order().

The switch constructs have to be changed to if/else constructs to prevent
GCC from warning on builds with 3-level page tables where PMD_ORDER and
PUD_ORDER have the same value.

If you are looking at this commit because your driver stopped compiling,
look at the previous commit as well and audit your driver to be sure it
doesn't depend on mmap_lock being held in its ->huge_fault method.

[willy@infradead.org: use "order %u" to match the (non dev_t) style]
  Link: https://lkml.kernel.org/r/ZOUYekbtTv+n8hYf@casper.infradead.org
Link: https://lkml.kernel.org/r/20230818202335.2739663-4-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

drivers/dax/device.c

@@ -228,32 +228,26 @@ static vm_fault_t __dev_dax_pud_fault(struct dev_dax *dev_dax,
 }
 #endif /* !CONFIG_HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD */
 
-static vm_fault_t dev_dax_huge_fault(struct vm_fault *vmf,
-		enum page_entry_size pe_size)
+static vm_fault_t dev_dax_huge_fault(struct vm_fault *vmf, unsigned int order)
 {
 	struct file *filp = vmf->vma->vm_file;
 	vm_fault_t rc = VM_FAULT_SIGBUS;
 	int id;
 	struct dev_dax *dev_dax = filp->private_data;
 
-	dev_dbg(&dev_dax->dev, "%s: %s (%#lx - %#lx) size = %d\n", current->comm,
+	dev_dbg(&dev_dax->dev, "%s: %s (%#lx - %#lx) order:%d\n", current->comm,
 			(vmf->flags & FAULT_FLAG_WRITE) ? "write" : "read",
-			vmf->vma->vm_start, vmf->vma->vm_end, pe_size);
+			vmf->vma->vm_start, vmf->vma->vm_end, order);
 
 	id = dax_read_lock();
-	switch (pe_size) {
-	case PE_SIZE_PTE:
+	if (order == 0)
 		rc = __dev_dax_pte_fault(dev_dax, vmf);
-		break;
-	case PE_SIZE_PMD:
+	else if (order == PMD_ORDER)
 		rc = __dev_dax_pmd_fault(dev_dax, vmf);
-		break;
-	case PE_SIZE_PUD:
+	else if (order == PUD_ORDER)
 		rc = __dev_dax_pud_fault(dev_dax, vmf);
-		break;
-	default:
+	else
 		rc = VM_FAULT_SIGBUS;
-	}
 
 	dax_read_unlock(id);
 
@@ -262,7 +256,7 @@ static vm_fault_t dev_dax_huge_fault(struct vm_fault *vmf,
 
 static vm_fault_t dev_dax_fault(struct vm_fault *vmf)
 {
-	return dev_dax_huge_fault(vmf, PE_SIZE_PTE);
+	return dev_dax_huge_fault(vmf, 0);
 }
 
 static int dev_dax_may_split(struct vm_area_struct *vma, unsigned long addr)

fs/dax.c

@@ -30,17 +30,6 @@
 #define CREATE_TRACE_POINTS
 #include <trace/events/fs_dax.h>
 
-static inline unsigned int pe_order(enum page_entry_size pe_size)
-{
-	if (pe_size == PE_SIZE_PTE)
-		return PAGE_SHIFT - PAGE_SHIFT;
-	if (pe_size == PE_SIZE_PMD)
-		return PMD_SHIFT - PAGE_SHIFT;
-	if (pe_size == PE_SIZE_PUD)
-		return PUD_SHIFT - PAGE_SHIFT;
-	return ~0;
-}
-
 /* We choose 4096 entries - same as per-zone page wait tables */
 #define DAX_WAIT_TABLE_BITS 12
 #define DAX_WAIT_TABLE_ENTRIES (1 << DAX_WAIT_TABLE_BITS)
@@ -1905,7 +1894,7 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp,
 /**
  * dax_iomap_fault - handle a page fault on a DAX file
  * @vmf: The description of the fault
- * @pe_size: Size of the page to fault in
+ * @order: Order of the page to fault in
  * @pfnp: PFN to insert for synchronous faults if fsync is required
  * @iomap_errp: Storage for detailed error code in case of error
  * @ops: Iomap ops passed from the file system
@@ -1915,17 +1904,15 @@ static vm_fault_t dax_iomap_pmd_fault(struct vm_fault *vmf, pfn_t *pfnp,
  * has done all the necessary locking for page fault to proceed
  * successfully.
  */
-vm_fault_t dax_iomap_fault(struct vm_fault *vmf, enum page_entry_size pe_size,
+vm_fault_t dax_iomap_fault(struct vm_fault *vmf, unsigned int order,
 		    pfn_t *pfnp, int *iomap_errp, const struct iomap_ops *ops)
 {
-	switch (pe_size) {
-	case PE_SIZE_PTE:
+	if (order == 0)
 		return dax_iomap_pte_fault(vmf, pfnp, iomap_errp, ops);
-	case PE_SIZE_PMD:
+	else if (order == PMD_ORDER)
 		return dax_iomap_pmd_fault(vmf, pfnp, ops);
-	default:
+	else
 		return VM_FAULT_FALLBACK;
-	}
 }
 EXPORT_SYMBOL_GPL(dax_iomap_fault);
 
@@ -1976,19 +1963,18 @@ dax_insert_pfn_mkwrite(struct vm_fault *vmf, pfn_t pfn, unsigned int order)
 /**
  * dax_finish_sync_fault - finish synchronous page fault
  * @vmf: The description of the fault
- * @pe_size: Size of entry to be inserted
+ * @order: Order of entry to be inserted
  * @pfn: PFN to insert
  *
  * This function ensures that the file range touched by the page fault is
  * stored persistently on the media and handles inserting of appropriate page
  * table entry.
  */
-vm_fault_t dax_finish_sync_fault(struct vm_fault *vmf,
-		enum page_entry_size pe_size, pfn_t pfn)
+vm_fault_t dax_finish_sync_fault(struct vm_fault *vmf, unsigned int order,
+		pfn_t pfn)
 {
 	int err;
 	loff_t start = ((loff_t)vmf->pgoff) << PAGE_SHIFT;
-	unsigned int order = pe_order(pe_size);
 	size_t len = PAGE_SIZE << order;
 
 	err = vfs_fsync_range(vmf->vma->vm_file, start, start + len - 1, 1);

fs/erofs/data.c

@@ -413,14 +413,14 @@ const struct address_space_operations erofs_raw_access_aops = {
 
 #ifdef CONFIG_FS_DAX
 static vm_fault_t erofs_dax_huge_fault(struct vm_fault *vmf,
-		enum page_entry_size pe_size)
+		unsigned int order)
 {
-	return dax_iomap_fault(vmf, pe_size, NULL, NULL, &erofs_iomap_ops);
+	return dax_iomap_fault(vmf, order, NULL, NULL, &erofs_iomap_ops);
 }
 
 static vm_fault_t erofs_dax_fault(struct vm_fault *vmf)
 {
-	return erofs_dax_huge_fault(vmf, PE_SIZE_PTE);
+	return erofs_dax_huge_fault(vmf, 0);
 }
 
 static const struct vm_operations_struct erofs_dax_vm_ops = {

fs/ext2/file.c

@@ -103,7 +103,7 @@ static vm_fault_t ext2_dax_fault(struct vm_fault *vmf)
 	}
 	filemap_invalidate_lock_shared(inode->i_mapping);
 
-	ret = dax_iomap_fault(vmf, PE_SIZE_PTE, NULL, NULL, &ext2_iomap_ops);
+	ret = dax_iomap_fault(vmf, 0, NULL, NULL, &ext2_iomap_ops);
 
 	filemap_invalidate_unlock_shared(inode->i_mapping);
 	if (write)

fs/ext4/file.c

@@ -723,8 +723,7 @@ ext4_file_write_iter(struct kiocb *iocb, struct iov_iter *from)
 }
 
 #ifdef CONFIG_FS_DAX
-static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf,
-		enum page_entry_size pe_size)
+static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf, unsigned int order)
 {
 	int error = 0;
 	vm_fault_t result;
@@ -740,7 +739,7 @@ static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf,
 	 * read-only.
 	 *
 	 * We check for VM_SHARED rather than vmf->cow_page since the latter is
-	 * unset for pe_size != PE_SIZE_PTE (i.e. only in do_cow_fault); for
+	 * unset for order != 0 (i.e. only in do_cow_fault); for
 	 * other sizes, dax_iomap_fault will handle splitting / fallback so that
 	 * we eventually come back with a COW page.
 	 */
@@ -764,7 +763,7 @@ static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf,
 	} else {
 		filemap_invalidate_lock_shared(mapping);
 	}
-	result = dax_iomap_fault(vmf, pe_size, &pfn, &error, &ext4_iomap_ops);
+	result = dax_iomap_fault(vmf, order, &pfn, &error, &ext4_iomap_ops);
 	if (write) {
 		ext4_journal_stop(handle);
 
@@ -773,7 +772,7 @@ static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf,
 			goto retry;
 		/* Handling synchronous page fault? */
 		if (result & VM_FAULT_NEEDDSYNC)
-			result = dax_finish_sync_fault(vmf, pe_size, pfn);
+			result = dax_finish_sync_fault(vmf, order, pfn);
 		filemap_invalidate_unlock_shared(mapping);
 		sb_end_pagefault(sb);
 	} else {
@@ -785,7 +784,7 @@ static vm_fault_t ext4_dax_huge_fault(struct vm_fault *vmf,
 
 static vm_fault_t ext4_dax_fault(struct vm_fault *vmf)
 {
-	return ext4_dax_huge_fault(vmf, PE_SIZE_PTE);
+	return ext4_dax_huge_fault(vmf, 0);
 }
 
 static const struct vm_operations_struct ext4_dax_vm_ops = {

fs/fuse/dax.c

@@ -784,8 +784,8 @@ static int fuse_dax_writepages(struct address_space *mapping,
 	return dax_writeback_mapping_range(mapping, fc->dax->dev, wbc);
 }
 
-static vm_fault_t __fuse_dax_fault(struct vm_fault *vmf,
-				   enum page_entry_size pe_size, bool write)
+static vm_fault_t __fuse_dax_fault(struct vm_fault *vmf, unsigned int order,
+		bool write)
 {
 	vm_fault_t ret;
 	struct inode *inode = file_inode(vmf->vma->vm_file);
@@ -809,7 +809,7 @@ static vm_fault_t __fuse_dax_fault(struct vm_fault *vmf,
 	 * to populate page cache or access memory we are trying to free.
 	 */
 	filemap_invalidate_lock_shared(inode->i_mapping);
-	ret = dax_iomap_fault(vmf, pe_size, &pfn, &error, &fuse_iomap_ops);
+	ret = dax_iomap_fault(vmf, order, &pfn, &error, &fuse_iomap_ops);
 	if ((ret & VM_FAULT_ERROR) && error == -EAGAIN) {
 		error = 0;
 		retry = true;
@@ -818,7 +818,7 @@ static vm_fault_t __fuse_dax_fault(struct vm_fault *vmf,
 	}
 
 	if (ret & VM_FAULT_NEEDDSYNC)
-		ret = dax_finish_sync_fault(vmf, pe_size, pfn);
+		ret = dax_finish_sync_fault(vmf, order, pfn);
 	filemap_invalidate_unlock_shared(inode->i_mapping);
 
 	if (write)
@@ -829,24 +829,22 @@ static vm_fault_t __fuse_dax_fault(struct vm_fault *vmf,
 
 static vm_fault_t fuse_dax_fault(struct vm_fault *vmf)
 {
-	return __fuse_dax_fault(vmf, PE_SIZE_PTE,
-				vmf->flags & FAULT_FLAG_WRITE);
+	return __fuse_dax_fault(vmf, 0, vmf->flags & FAULT_FLAG_WRITE);
 }
 
-static vm_fault_t fuse_dax_huge_fault(struct vm_fault *vmf,
-			       enum page_entry_size pe_size)
+static vm_fault_t fuse_dax_huge_fault(struct vm_fault *vmf, unsigned int order)
 {
-	return __fuse_dax_fault(vmf, pe_size, vmf->flags & FAULT_FLAG_WRITE);
+	return __fuse_dax_fault(vmf, order, vmf->flags & FAULT_FLAG_WRITE);
 }
 
 static vm_fault_t fuse_dax_page_mkwrite(struct vm_fault *vmf)
 {
-	return __fuse_dax_fault(vmf, PE_SIZE_PTE, true);
+	return __fuse_dax_fault(vmf, 0, true);
 }
 
 static vm_fault_t fuse_dax_pfn_mkwrite(struct vm_fault *vmf)
 {
-	return __fuse_dax_fault(vmf, PE_SIZE_PTE, true);
+	return __fuse_dax_fault(vmf, 0, true);
 }
 
 static const struct vm_operations_struct fuse_dax_vm_ops = {

fs/xfs/xfs_file.c

@@ -1287,11 +1287,11 @@ xfs_file_llseek(
 static inline vm_fault_t
 xfs_dax_fault(
 	struct vm_fault		*vmf,
-	enum page_entry_size	pe_size,
+	unsigned int		order,
 	bool			write_fault,
 	pfn_t			*pfn)
 {
-	return dax_iomap_fault(vmf, pe_size, pfn, NULL,
+	return dax_iomap_fault(vmf, order, pfn, NULL,
 			(write_fault && !vmf->cow_page) ?
 				&xfs_dax_write_iomap_ops :
 				&xfs_read_iomap_ops);
@@ -1300,7 +1300,7 @@ xfs_dax_fault(
 static inline vm_fault_t
 xfs_dax_fault(
 	struct vm_fault		*vmf,
-	enum page_entry_size	pe_size,
+	unsigned int		order,
 	bool			write_fault,
 	pfn_t			*pfn)
 {
@@ -1322,14 +1322,14 @@ xfs_dax_fault(
 static vm_fault_t
 __xfs_filemap_fault(
 	struct vm_fault		*vmf,
-	enum page_entry_size	pe_size,
+	unsigned int		order,
 	bool			write_fault)
 {
 	struct inode		*inode = file_inode(vmf->vma->vm_file);
 	struct xfs_inode	*ip = XFS_I(inode);
 	vm_fault_t		ret;
 
-	trace_xfs_filemap_fault(ip, pe_size, write_fault);
+	trace_xfs_filemap_fault(ip, order, write_fault);
 
 	if (write_fault) {
 		sb_start_pagefault(inode->i_sb);
@@ -1340,9 +1340,9 @@ __xfs_filemap_fault(
 		pfn_t pfn;
 
 		xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
-		ret = xfs_dax_fault(vmf, pe_size, write_fault, &pfn);
+		ret = xfs_dax_fault(vmf, order, write_fault, &pfn);
 		if (ret & VM_FAULT_NEEDDSYNC)
-			ret = dax_finish_sync_fault(vmf, pe_size, pfn);
+			ret = dax_finish_sync_fault(vmf, order, pfn);
 		xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
 	} else {
 		if (write_fault) {
@@ -1373,29 +1373,29 @@ xfs_filemap_fault(
 	struct vm_fault		*vmf)
 {
 	/* DAX can shortcut the normal fault path on write faults! */
-	return __xfs_filemap_fault(vmf, PE_SIZE_PTE,
+	return __xfs_filemap_fault(vmf, 0,
 			IS_DAX(file_inode(vmf->vma->vm_file)) &&
 			xfs_is_write_fault(vmf));
 }
 
 static vm_fault_t
 xfs_filemap_huge_fault(
 	struct vm_fault		*vmf,
-	enum page_entry_size	pe_size)
+	unsigned int		order)
 {
 	if (!IS_DAX(file_inode(vmf->vma->vm_file)))
 		return VM_FAULT_FALLBACK;
 
 	/* DAX can shortcut the normal fault path on write faults! */
-	return __xfs_filemap_fault(vmf, pe_size,
+	return __xfs_filemap_fault(vmf, order,
 			xfs_is_write_fault(vmf));
 }
 
 static vm_fault_t
 xfs_filemap_page_mkwrite(
 	struct vm_fault		*vmf)
 {
-	return __xfs_filemap_fault(vmf, PE_SIZE_PTE, true);
+	return __xfs_filemap_fault(vmf, 0, true);
 }
 
 /*
@@ -1408,7 +1408,7 @@ xfs_filemap_pfn_mkwrite(
 	struct vm_fault		*vmf)
 {
 
-	return __xfs_filemap_fault(vmf, PE_SIZE_PTE, true);
+	return __xfs_filemap_fault(vmf, 0, true);
 }
 
 static const struct vm_operations_struct xfs_file_vm_ops = {

fs/xfs/xfs_trace.h

@@ -802,36 +802,28 @@ DEFINE_INODE_EVENT(xfs_inode_inactivating);
  * ring buffer.  Somehow this was only worth mentioning in the ftrace sample
  * code.
  */
-TRACE_DEFINE_ENUM(PE_SIZE_PTE);
-TRACE_DEFINE_ENUM(PE_SIZE_PMD);
-TRACE_DEFINE_ENUM(PE_SIZE_PUD);
-
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_SHARED);
 TRACE_DEFINE_ENUM(XFS_REFC_DOMAIN_COW);
 
 TRACE_EVENT(xfs_filemap_fault,
-	TP_PROTO(struct xfs_inode *ip, enum page_entry_size pe_size,
-		 bool write_fault),
-	TP_ARGS(ip, pe_size, write_fault),
+	TP_PROTO(struct xfs_inode *ip, unsigned int order, bool write_fault),
+	TP_ARGS(ip, order, write_fault),
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(xfs_ino_t, ino)
-		__field(enum page_entry_size, pe_size)
+		__field(unsigned int, order)
 		__field(bool, write_fault)
 	),
 	TP_fast_assign(
 		__entry->dev = VFS_I(ip)->i_sb->s_dev;
 		__entry->ino = ip->i_ino;
-		__entry->pe_size = pe_size;
+		__entry->order = order;
 		__entry->write_fault = write_fault;
 	),
-	TP_printk("dev %d:%d ino 0x%llx %s write_fault %d",
+	TP_printk("dev %d:%d ino 0x%llx order %u write_fault %d",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
-		  __print_symbolic(__entry->pe_size,
-			{ PE_SIZE_PTE,	"PTE" },
-			{ PE_SIZE_PMD,	"PMD" },
-			{ PE_SIZE_PUD,	"PUD" }),
+		  __entry->order,
 		  __entry->write_fault)
 )