Skip to contentSkip to conversation

add support for fallocate mode 0

PATCH New Ronnie Sahlberg 4 participants 9 messages Jan 15, 2020 – Jan 17, 2020

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover add support for fallocate mode 0

Steve,
Please find a small patch that adds support for fallocate mode 0 to extend a non-sparse file.

#1 lore raw

PATCH 1/1 cifs: add support for fallocate mode 0 for non-sparse files

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -3
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
 		else if (i_size_read(inode) >= off + len)
 			/* not extending file and already not sparse */
 			rc = 0;
-		/* BB: in future add else clause to extend file */
-		else
-			rc = -EOPNOTSUPP;
+		/* extend file */
+		else {
+			eof = cpu_to_le64(off + len);
+			rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+					  cfile->fid.volatile_fid, cfile->pid,
+					  &eof);
+		}
 		if (rc)
 			trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
 				tcon->tid, tcon->ses->Suid, off, len, rc);
Signature · 1 line
2.13.6
#2 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

Does it affect (or enable) any xfstests?

Quoted 32 lines · Ronnie Sahlberg

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -4
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
                else if (i_size_read(inode) >= off + len)
                        /* not extending file and already not sparse */
                        rc = 0;
-               /* BB: in future add else clause to extend file */
-               else
-                       rc = -EOPNOTSUPP;
+               /* extend file */
+               else {
+                       eof = cpu_to_le64(off + len);
+                       rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+                                         cfile->fid.volatile_fid, cfile->pid,
+                                         &eof);
+               }
                if (rc)
                        trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
                                tcon->tid, tcon->ses->Suid, off, len, rc);
--

2.13.6

Signature · 3 lines
Thanks,

Steve
#3 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

Quoted 2 lines · Steve French

Does it affect (or enable) any xfstests?

It shouldn't affect any current tests.
It adds support for
xfs_io -c "falloc 0 512M" <file>

generic/071 now passes with this patch. Possibly other tests as well
that use "xfs_io -c falloc" as well

Quoted 40 lines · Ronnie Sahlberg
Quoted 30 lines

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -4
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
                else if (i_size_read(inode) >= off + len)
                        /* not extending file and already not sparse */
                        rc = 0;
-               /* BB: in future add else clause to extend file */
-               else
-                       rc = -EOPNOTSUPP;
+               /* extend file */
+               else {
+                       eof = cpu_to_le64(off + len);
+                       rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+                                         cfile->fid.volatile_fid, cfile->pid,
+                                         &eof);
+               }
                if (rc)
                        trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
                                tcon->tid, tcon->ses->Suid, off, len, rc);
--
Quoted 2 lines

2.13.6

--
Thanks,

Steve

#4 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

Tentatively merged into cifs-2.6.git for-next pending more testing

Quoted 53 lines · ronnie sahlberg
Quoted 2 lines · Steve French

Does it affect (or enable) any xfstests?

It shouldn't affect any current tests.
It adds support for
xfs_io -c "falloc 0 512M" <file>

generic/071 now passes with this patch. Possibly other tests as well
that use "xfs_io -c falloc" as well

Quoted 40 lines · Ronnie Sahlberg
Quoted 30 lines

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -4
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
                else if (i_size_read(inode) >= off + len)
                        /* not extending file and already not sparse */
                        rc = 0;
-               /* BB: in future add else clause to extend file */
-               else
-                       rc = -EOPNOTSUPP;
+               /* extend file */
+               else {
+                       eof = cpu_to_le64(off + len);
+                       rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+                                         cfile->fid.volatile_fid, cfile->pid,
+                                         &eof);
+               }
                if (rc)
                        trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
                                tcon->tid, tcon->ses->Suid, off, len, rc);
--
Quoted 2 lines

2.13.6

--
Thanks,

Steve

Signature · 3 lines
Thanks,

Steve
#5 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

temporarily removed to allow Ronnie to debug a test failure

Quoted 65 lines · Steve French

Tentatively merged into cifs-2.6.git for-next pending more testing

Quoted 53 lines · ronnie sahlberg
Quoted 2 lines · Steve French

Does it affect (or enable) any xfstests?

It shouldn't affect any current tests.
It adds support for
xfs_io -c "falloc 0 512M" <file>

generic/071 now passes with this patch. Possibly other tests as well
that use "xfs_io -c falloc" as well

Quoted 40 lines · Ronnie Sahlberg
Quoted 30 lines

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -4
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
                else if (i_size_read(inode) >= off + len)
                        /* not extending file and already not sparse */
                        rc = 0;
-               /* BB: in future add else clause to extend file */
-               else
-                       rc = -EOPNOTSUPP;
+               /* extend file */
+               else {
+                       eof = cpu_to_le64(off + len);
+                       rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+                                         cfile->fid.volatile_fid, cfile->pid,
+                                         &eof);
+               }
                if (rc)
                        trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
                                tcon->tid, tcon->ses->Suid, off, len, rc);
--
Quoted 2 lines

2.13.6

--
Thanks,

Steve

--
Thanks,

Steve

Signature · 3 lines
Thanks,

Steve
#6 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

The bug is basically that if we extend a file by fallocate mode==0 and immediately afterwards mmap() the file we will only mmap() as much as end-of-file was
prior to the truncate and then if we try to touch any address in this extended region userspace dies with bus error.

The patch added "extend a file with fallocate mode==0 for NON-Sparse files" and caused xfstest to fail.
The fix is to force us to revalidate the file attributes (the size is the important one) when we extend the file so
mmap() will work properly.
I have fixed this in the patch and will resend tomorrow after some more testing.

Looking for other SMB2_set_eof() callsites I see we already had the same bug for the case of extending a SPARSE
file using fallocate mode==0. I fixed that too and will audit all other plases where we use SMB2_set_eof()
to see if they are safe or not before resending.

Quoted 76 lines · Steve French

temporarily removed to allow Ronnie to debug a test failure

Quoted 57 lines · Steve French

Tentatively merged into cifs-2.6.git for-next pending more testing

Quoted 52 lines · ronnie sahlberg
Quoted 2 lines · Steve French

Does it affect (or enable) any xfstests?

It shouldn't affect any current tests.
It adds support for
xfs_io -c "falloc 0 512M" <file>

generic/071 now passes with this patch. Possibly other tests as well
that use "xfs_io -c falloc" as well

Quoted 39 lines · Ronnie Sahlberg
Quoted 30 lines

RHBZ 1336264

---
 fs/cifs/smb2ops.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
fs/cifs/smb2ops.c +7 -4
index 6250370c1170..91818f7c1b9c 100644
@@ -3106,9 +3106,13 @@ static long smb3_simple_falloc(struct file *file, struct cifs_tcon *tcon,
                else if (i_size_read(inode) >= off + len)
                        /* not extending file and already not sparse */
                        rc = 0;
-               /* BB: in future add else clause to extend file */
-               else
-                       rc = -EOPNOTSUPP;
+               /* extend file */
+               else {
+                       eof = cpu_to_le64(off + len);
+                       rc = SMB2_set_eof(xid, tcon, cfile->fid.persistent_fid,
+                                         cfile->fid.volatile_fid, cfile->pid,
+                                         &eof);
+               }
                if (rc)
                        trace_smb3_falloc_err(xid, cfile->fid.persistent_fid,
                                tcon->tid, tcon->ses->Suid, off, len, rc);
--
Quoted 2 lines

2.13.6

--
Thanks,

Quoted 8 lines
Quoted 1 line
Quoted 1 line

Steve

--
Thanks,

Steve

--
Thanks,

Steve

#7 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

чт, 16 янв. 2020 г. в 01:05, ronnie sahlberg <[email protected]>:

Quoted 14 lines

The bug is basically that if we extend a file by fallocate mode==0 and immediately afterwards mmap() the file we will only mmap() as much as end-of-file was
prior to the truncate and then if we try to touch any address in this extended region userspace dies with bus error.

The patch added "extend a file with fallocate mode==0 for NON-Sparse files" and caused xfstest to fail.
The fix is to force us to revalidate the file attributes (the size is the important one) when we extend the file so
mmap() will work properly.
I have fixed this in the patch and will resend tomorrow after some more testing.

Quoted 2 lines

Looking for other SMB2_set_eof() callsites I see we already had the same bug for the case of extending a SPARSE

I agree that regardless of file being sparse or not, we should somehow update a size in the VFS after calling SMB2_set_eof().

Quoted 3 lines

file using fallocate mode==0. I fixed that too and will audit all other plases where we use SMB2_set_eof()
to see if they are safe or not before resending.

One of those places where SMB2_set_eof() is called is cifs_set_file_size() which does call the following after getting a successful response from the server:

2250 >-------if (rc == 0) {
2251 >------->-------cifsInode->server_eof = attrs->ia_size; 2252 >------->-------cifs_setsize(inode, attrs->ia_size); 2253 >------->-------cifs_truncate_page(inode->i_mapping, inode->i_size); 2254 >-------}

This is called by cifs_setattr_[no]unix() which does the following afterwards:

2569 >-------if ((attrs->ia_valid & ATTR_SIZE) &&
2570 >------- attrs->ia_size != i_size_read(inode))
2571 >------->-------truncate_setsize(inode, attrs->ia_size);

truncate_setsize() does similar things as cifs_setsize() besides setting cinode->time to 0. This code path probably needs to be refactored. But putting this aside, for the current fallocate fix I think we should use the same existing mechanism to update a file size in the VFS without revalidating the inode.

--
Best regards,
Pavel Shilovsky

#8 lore raw

Re: [PATCH] cifs: add support for fallocate mode 0 for non-sparse files

Quoted 1 line · Pavel Shilovsky

Thanks.

Quoted 41 lines

чт, 16 янв. 2020 г. в 01:05, ronnie sahlberg <[email protected]>:

Quoted 14 lines

The bug is basically that if we extend a file by fallocate mode==0 and immediately afterwards mmap() the file we will only mmap() as much as end-of-file was
prior to the truncate and then if we try to touch any address in this extended region userspace dies with bus error.

The patch added "extend a file with fallocate mode==0 for NON-Sparse files" and caused xfstest to fail.
The fix is to force us to revalidate the file attributes (the size is the important one) when we extend the file so
mmap() will work properly.
I have fixed this in the patch and will resend tomorrow after some more testing.

Quoted 2 lines

Looking for other SMB2_set_eof() callsites I see we already had the same bug for the case of extending a SPARSE

I agree that regardless of file being sparse or not, we should somehow update a size in the VFS after calling SMB2_set_eof().

Quoted 3 lines

file using fallocate mode==0. I fixed that too and will audit all other plases where we use SMB2_set_eof()
to see if they are safe or not before resending.

One of those places where SMB2_set_eof() is called is cifs_set_file_size() which does call the following after getting a successful response from the server:

2250 >-------if (rc == 0) {
2251 >------->-------cifsInode->server_eof = attrs->ia_size; 2252 >------->-------cifs_setsize(inode, attrs->ia_size); 2253 >------->-------cifs_truncate_page(inode->i_mapping, inode->i_size); 2254 >-------}

This is called by cifs_setattr_[no]unix() which does the following afterwards:

2569 >-------if ((attrs->ia_valid & ATTR_SIZE) &&
2570 >------- attrs->ia_size != i_size_read(inode))
2571 >------->-------truncate_setsize(inode, attrs->ia_size);

Quoted 5 lines

truncate_setsize() does similar things as cifs_setsize() besides setting cinode->time to 0. This code path probably needs to be refactored. But putting this aside, for the current fallocate fix I think we should use the same existing mechanism to update a file size in the VFS without revalidating the inode.

I can do that change. Note however that since we are still setting cinode->time to 0
we are still going to trigger a revalidation at some stage.

Quoted 4 lines

--
Best regards,
Pavel Shilovsky