Skip to contentSkip to conversation

cifs: Remove cifs_truncate_page() as it should be superfluous

PATCH In review David Howells 2 participants 2 messages Apr 1, 2025 – Apr 2, 2025

Subsystems (4)
lore
Share & tools
lore

David Howells

author maintainer
#0 lore raw

PATCH 1/1 cifs: Remove cifs_truncate_page() as it should be superfluous

The calls to cifs_truncate_page() should be superfluous as the places that call it also call truncate_setsize() or cifs_setsize() and therefore truncate_pagecache() which should also clear the tail part of the folio containing the EOF marker.

Further, smb3_simple_falloc() calls both cifs_setsize() and truncate_setsize() in addition to cifs_truncate_page().

Remove the superfluous calls.

This gets rid of another function referring to struct page.

[Should cifs_setsize() also set inode->i_blocks?]

---
 fs/smb/client/cifsfs.h  |    1 -
 fs/smb/client/inode.c   |   19 -------------------
 fs/smb/client/smb2ops.c |    2 --
 3 files changed, 22 deletions(-)
fs/smb/client/cifsfs.h +0 -1
index 8dea0cf3a8de..a769fa7ceece 100644
@@ -135,7 +135,6 @@ extern ssize_t cifs_file_copychunk_range(unsigned int xid,
 
 extern long cifs_ioctl(struct file *filep, unsigned int cmd, unsigned long arg);
 extern void cifs_setsize(struct inode *inode, loff_t offset);
-extern int cifs_truncate_page(struct address_space *mapping, loff_t from);
 
 struct smb3_fs_context;
 extern struct dentry *cifs_smb3_do_mount(struct file_system_type *fs_type,
fs/smb/client/inode.c +0 -19
index 3bb21aa58474..a00a9d91d0da 100644
@@ -2901,23 +2901,6 @@ int cifs_fiemap(struct inode *inode, struct fiemap_extent_info *fei, u64 start,
 	return -EOPNOTSUPP;
 }
 
-int cifs_truncate_page(struct address_space *mapping, loff_t from)
-{
-	pgoff_t index = from >> PAGE_SHIFT;
-	unsigned offset = from & (PAGE_SIZE - 1);
-	struct page *page;
-	int rc = 0;
-
-	page = grab_cache_page(mapping, index);
-	if (!page)
-		return -ENOMEM;
-
-	zero_user_segment(page, offset, PAGE_SIZE);
-	unlock_page(page);
-	put_page(page);
-	return rc;
-}
-
 void cifs_setsize(struct inode *inode, loff_t offset)
 {
 	struct cifsInodeInfo *cifs_i = CIFS_I(inode);
@@ -3012,8 +2995,6 @@ cifs_set_file_size(struct inode *inode, struct iattr *attrs,
 		 */
 		attrs->ia_ctime = attrs->ia_mtime = current_time(inode);
 		attrs->ia_valid |= ATTR_CTIME | ATTR_MTIME;
-
-		cifs_truncate_page(inode->i_mapping, inode->i_size);
 	}
 
 	return rc;
fs/smb/client/smb2ops.c +0 -2
index 4dd11eafb69d..13d060e0fc33 100644
@@ -3526,8 +3526,6 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
 		if (rc == 0) {
 			netfs_resize_file(&cifsi->netfs, new_eof, true);
 			cifs_setsize(inode, new_eof);
-			cifs_truncate_page(inode->i_mapping, inode->i_size);
-			truncate_setsize(inode, new_eof);
 		}
 		goto out;
 	}
Reviewed-by on v1 1/1
#1 lore raw
Quoted 5 lines · David Howells

The calls to cifs_truncate_page() should be superfluous as the places that call it also call truncate_setsize() or cifs_setsize() and therefore truncate_pagecache() which should also clear the tail part of the folio containing the EOF marker.

Quoted 8 lines

Further, smb3_simple_falloc() calls both cifs_setsize() and truncate_setsize() in addition to cifs_truncate_page().

Remove the superfluous calls.

This gets rid of another function referring to struct page.

[Should cifs_setsize() also set inode->i_blocks?]

I don't think that's necessary as the current inode will be marked for revalidation (e.g. cifsInodeInfo::time is set to 0), so next call to cifs_fattr_to_inode() will update inode->i_blocks. If it isn't being updated, then I'm missing something else.

Quoted 7 lines
Quoted 5 lines
---
 fs/smb/client/cifsfs.h  |    1 -
 fs/smb/client/inode.c   |   19 -------------------
 fs/smb/client/smb2ops.c |    2 --
 3 files changed, 22 deletions(-)

Looks good,