Skip to contentSkip to conversation

hfsplus journal detection - try 4

PATCH In review Warren Turkal 4 participants 15 messages Nov 11, 2008 – Nov 20, 2008

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

Here's another shot at this patch. I found a small issue with sparse. If this is suitable for inclusion, could someone please pull it into their tree so that it can be merged upstream?

Thanks,
wt

#1 lore raw

PATCH 1/1 Identify journal info block in volume header.

---
 fs/hfsplus/hfsplus_raw.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
fs/hfsplus/hfsplus_raw.h +1 -1
index fe99fe8..14f1dd8 100644
@@ -94,7 +94,7 @@ struct hfsplus_vh {
 	__be16 version;
 	__be32 attributes;
 	__be32 last_mount_vers;
-	u32 reserved;
+	__be32 journal_info_block;
 
 	__be32 create_date;
 	__be32 modify_date;
Signature · 1 line
1.5.6.3
#2 lore raw

PATCH 1/1 Fix journal detection on HFS+.

The code was unconditionally assumming that the volume had a jounal if the jounal attribute was set in the volume header. However, the volume also has to have a non-zero journal info block to actually have a journal.

---
 fs/hfsplus/super.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +9 -2
index eb74531..128101b 100644
@@ -17,9 +17,16 @@
 
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
 #include "hfsplus_fs.h"
 
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
+{
+	return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&
+				vhdr->journal_info_block);
+}
+
 struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino)
 {
 	struct hfs_find_data fd;
@@ -260,7 +267,7 @@ static int hfsplus_remount(struct super_block *sb, int *flags, char *data)
 			printk(KERN_WARNING "hfs: filesystem is marked locked, leaving read-only.\n");
 			sb->s_flags |= MS_RDONLY;
 			*flags |= MS_RDONLY;
-		} else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) {
+		} else if (hfsplus_vol_has_journal(vhdr)) {
 			printk(KERN_WARNING "hfs: filesystem is marked journaled, leaving read-only.\n");
 			sb->s_flags |= MS_RDONLY;
 			*flags |= MS_RDONLY;
@@ -356,7 +363,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent)
 	} else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_SOFTLOCK)) {
 		printk(KERN_WARNING "hfs: Filesystem is marked locked, mounting read-only.\n");
 		sb->s_flags |= MS_RDONLY;
-	} else if ((vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) && !(sb->s_flags & MS_RDONLY)) {
+	} else if (hfsplus_vol_has_journal(vhdr) && !(sb->s_flags & MS_RDONLY)) {
 		printk(KERN_WARNING "hfs: write access to a journaled filesystem is not supported, "
 		       "use the force option at your own risk, mounting read-only.\n");
 		sb->s_flags |= MS_RDONLY;
Signature · 1 line
1.5.6.3
#3 lore raw

PATCH 1/1 Fix header include.

---
 fs/hfsplus/super.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +2 -2
index 128101b..4f00a84 100644
@@ -15,12 +15,12 @@
 #include <linux/vfs.h>
 #include <linux/nls.h>
 
+#include "hfsplus_fs.h"
+
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
-#include "hfsplus_fs.h"
-
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
 {
 	return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&
Signature · 1 line
1.5.6.3
#4 lore raw

Re: [PATCH] Fix header include.

Quoted 14 lines · On Tue, 11 Nov 2008 12:56:48 -0800 Warren Turkal <[email protected]> wrote:
---
 fs/hfsplus/super.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +1 -0
index 128101b..4f00a84 100644
@@ -15,12 +15,12 @@
 #include <linux/vfs.h>
 #include <linux/nls.h>
 
+#include "hfsplus_fs.h"

fs/hfsplus/super.c line 26

Quoted 10 lines
fs/hfsplus/super.c +1 -2
+
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
-#include "hfsplus_fs.h"
-
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
 {
 	return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&

again: what was wrong with the old code??

If it was causing some compilation problem then please quote the compiler output in the changelog.

#5 lore raw

Re: [PATCH] Fix header include.

Quoted 4 lines · Andrew Morton

again: what was wrong with the old code??

If it was causing some compilation problem then please quote the compiler output in the changelog.

I didn't write this patch, but I was the original author of this driver. The issue here is that Apple added a journal feature after the fact and the current Linux driver doesn't support that. The current code just checks for the equivalent of a feature bit, while this change makes it also confirm that there really is a journal file. Without this change, an HFS+ file system that has been setup to do transactional changes but doesn't currently have an actual journal file allocated will not be mountable with the Linux driver.

Just as a side note, although I am the original author I have never claimed to maintain this code. In fact, I wasn't the one that submitted it for inclusion. In any case, I don't plan on adding myself as maintainer.

Brad Boyer
[email protected]

#6 lore raw

Re: [PATCH] Fix header include.

In reply to Andrew Morton
Quoted 12 lines · Andrew Morton
Quoted 10 lines · On Tue, 11 Nov 2008 12:56:48 -0800 Warren Turkal <[email protected]> wrote:
fs/hfsplus/super.c +2 -0
@@ -15,12 +15,12 @@
 #include <linux/vfs.h>
 #include <linux/nls.h>
 
+#include "hfsplus_fs.h"
+
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
Quoted 7 lines

fs/hfsplus/super.c line 26

Quoted 5 lines
fs/hfsplus/super.c +0 -2
-#include "hfsplus_fs.h"
-
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
 {
 	return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&

again: what was wrong with the old code??

I just realized what you were asking after I sent my other reply. The include needs to move so 'struct hfsplus_vh' is declared before the function prototype. That happens in hfsplus_fs.h.

My apologies for the noise, although I stand by my statement that I have no intention of actively maintaining this code myself.

Quoted 2 lines

If it was causing some compilation problem then please quote the compiler output in the changelog.

The original submitter would need to include the actual compiler output.

Brad Boyer
[email protected]

#7 lore raw

Re: [PATCH] Fix header include.

I just submitted a patchset with more detailed commit comments.. It's the same change wrapped in a, hopefully, prettier candy coating.

wt

Quoted 37 lines · Brad Boyer
Quoted 12 lines · Andrew Morton
Quoted 10 lines · On Tue, 11 Nov 2008 12:56:48 -0800 Warren Turkal <[email protected]> wrote:
fs/hfsplus/super.c +2 -0
@@ -15,12 +15,12 @@
 #include <linux/vfs.h>
 #include <linux/nls.h>
 
+#include "hfsplus_fs.h"
+
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
Quoted 7 lines

fs/hfsplus/super.c line 26

Quoted 5 lines
fs/hfsplus/super.c +0 -2
-#include "hfsplus_fs.h"
-
 static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
 {
    return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&

again: what was wrong with the old code??

I just realized what you were asking after I sent my other reply. The include needs to move so 'struct hfsplus_vh' is declared before the function prototype. That happens in hfsplus_fs.h.

My apologies for the noise, although I stand by my statement that I have no intention of actively maintaining this code myself.

Quoted 2 lines

If it was causing some compilation problem then please quote the compiler output in the changelog.

The original submitter would need to include the actual compiler output.

Brad Boyer
[email protected]

#8 lore raw

Re: [PATCH] Fix journal detection on HFS+.

In reply to Warren Turkal
Quoted 3 lines · On Tue, 11 Nov 2008 12:56:47 -0800 Warren Turkal <[email protected]> wrote:

The code was unconditionally assumming that the volume had a jounal if the jounal attribute was set in the volume header. However, the volume also has to have a non-zero journal info block to actually have a journal.

OK, but so what?

Presumably there is some situation in which this is causing you a problem, but what is that situation, and what was the kernel's behaviour in that situation?

Quoted 36 lines
---
 fs/hfsplus/super.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +8 -1
index eb74531..128101b 100644
@@ -17,9 +17,16 @@
 
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
 #include "hfsplus_fs.h"
 
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
+{
+	return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&
+				vhdr->journal_info_block);
+}
+
 struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino)
 {
 	struct hfs_find_data fd;
@@ -260,7 +267,7 @@ static int hfsplus_remount(struct super_block *sb, int *flags, char *data)
 			printk(KERN_WARNING "hfs: filesystem is marked locked, leaving read-only.\n");
 			sb->s_flags |= MS_RDONLY;
 			*flags |= MS_RDONLY;
-		} else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) {
+		} else if (hfsplus_vol_has_journal(vhdr)) {
 			printk(KERN_WARNING "hfs: filesystem is marked journaled, leaving read-only.\n");
 			sb->s_flags |= MS_RDONLY;
 			*flags |= MS_RDONLY;

fs/hfsplus/super.c static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) · line 369

Quoted 8 lines
fs/hfsplus/super.c +1 -1
@@ -356,7 +363,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent)
 	} else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_SOFTLOCK)) {
 		printk(KERN_WARNING "hfs: Filesystem is marked locked, mounting read-only.\n");
 		sb->s_flags |= MS_RDONLY;
-	} else if ((vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) && !(sb->s_flags & MS_RDONLY)) {
+	} else if (hfsplus_vol_has_journal(vhdr) && !(sb->s_flags & MS_RDONLY)) {
 		printk(KERN_WARNING "hfs: write access to a journaled filesystem is not supported, "
 		       "use the force option at your own risk, mounting read-only.\n");
 		sb->s_flags |= MS_RDONLY;

The patch itself looks OK to me.

#9 lore raw

Re: [PATCH] Fix journal detection on HFS+.

The kernel was mounting an HFS+ FS ro if the volume header indicated that it had a journal and the journal info block was 0. The kernel should just treat the volume as not having a journal in that case.

wt

Quoted 60 lines · Andrew Morton
Quoted 3 lines · On Tue, 11 Nov 2008 12:56:47 -0800 Warren Turkal <[email protected]> wrote:

The code was unconditionally assumming that the volume had a jounal if the jounal attribute was set in the volume header. However, the volume also has to have a non-zero journal info block to actually have a journal.

OK, but so what?

Presumably there is some situation in which this is causing you a problem, but what is that situation, and what was the kernel's behaviour in that situation?

Quoted 36 lines
---
 fs/hfsplus/super.c |   11 +++++++++--
 1 files changed, 9 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +8 -1
index eb74531..128101b 100644
@@ -17,9 +17,16 @@
 
 static struct inode *hfsplus_alloc_inode(struct super_block *sb);
 static void hfsplus_destroy_inode(struct inode *inode);
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr);
 
 #include "hfsplus_fs.h"
 
+static bool hfsplus_vol_has_journal(struct hfsplus_vh *vhdr)
+{
+     return (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED) &&
+                             vhdr->journal_info_block);
+}
+
 struct inode *hfsplus_iget(struct super_block *sb, unsigned long ino)
 {
      struct hfs_find_data fd;
@@ -260,7 +267,7 @@ static int hfsplus_remount(struct super_block *sb, int *flags, char *data)
                      printk(KERN_WARNING "hfs: filesystem is marked locked, leaving read-only.\n");
                      sb->s_flags |= MS_RDONLY;
                      *flags |= MS_RDONLY;
-             } else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) {
+             } else if (hfsplus_vol_has_journal(vhdr)) {
                      printk(KERN_WARNING "hfs: filesystem is marked journaled, leaving read-only.\n");
                      sb->s_flags |= MS_RDONLY;
                      *flags |= MS_RDONLY;

fs/hfsplus/super.c static int hfsplus_fill_super(struct super_block *sb, void *data, int silent) · line 369

Quoted 8 lines
fs/hfsplus/super.c +1 -1
@@ -356,7 +363,7 @@ static int hfsplus_fill_super(struct super_block *sb, void *data, int silent)
      } else if (vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_SOFTLOCK)) {
              printk(KERN_WARNING "hfs: Filesystem is marked locked, mounting read-only.\n");
              sb->s_flags |= MS_RDONLY;
-     } else if ((vhdr->attributes & cpu_to_be32(HFSPLUS_VOL_JOURNALED)) && !(sb->s_flags & MS_RDONLY)) {
+     } else if (hfsplus_vol_has_journal(vhdr) && !(sb->s_flags & MS_RDONLY)) {
              printk(KERN_WARNING "hfs: write access to a journaled filesystem is not supported, "
                     "use the force option at your own risk, mounting read-only.\n");
              sb->s_flags |= MS_RDONLY;

The patch itself looks OK to me.

#10 lore raw

Re: [PATCH] Identify journal info block in volume header.

In reply to Warren Turkal
Quoted 10 lines · On Tue, 11 Nov 2008 12:56:46 -0800 Warren Turkal <[email protected]> wrote:
---
 fs/hfsplus/hfsplus_raw.h |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
fs/hfsplus/hfsplus_raw.h +0 -0
index fe99fe8..14f1dd8 100644

fs/hfsplus/hfsplus_raw.h struct hfsplus_vh { · line 100

Quoted 8 lines
fs/hfsplus/hfsplus_raw.h +1 -1
@@ -94,7 +94,7 @@ struct hfsplus_vh {
 	__be16 version;
 	__be32 attributes;
 	__be32 last_mount_vers;
-	u32 reserved;
+	__be32 journal_info_block;
 
 	__be32 create_date;
 	__be32 modify_date;

Please include a changelog which fully describes the reason for making this change.

#11 lore raw
In reply to Warren Turkal

Did anyone pick up this patch set?

wt

Quoted 7 lines · Warren Turkal

Here's another shot at this patch. I found a small issue with sparse. If this is suitable for inclusion, could someone please pull it into their tree so that it can be merged upstream?

Thanks,
wt

#12 lore raw
Quoted 2 lines · Warren Turkal

Did anyone pick up this patch set?

You could try sending it to someone who cares. Roman apparently doesn't, hfsplus is officially unmaintained. Guess that leaves Andrew Morton.

Jörn

Signature · 7 lines
There's nothing better for promoting creativity in a medium than
making an audience feel "Hmm ­ I could do better than that!"
-- Douglas Adams in a slashdot interview
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#13 lore raw
In reply to Warren Turkal

Andrew,

I am sending this to you since it's a small patchset and I'm not getting any traction with the alleged HFS+ maintainer. Is this small enough to make it into 2.6.28 or will it have to wait for the next cycle?

BTW, I am the guy who just came to your cube to bug you.

Thanks,
wt

Quoted 3 lines · Warren Turkal

Here's another shot at this patch. I found a small issue with sparse. If this is suitable for inclusion, could someone please pull it into their tree so that it can be merged upstream?

#14 lore raw
In reply to Warren Turkal
Quoted 3 lines · On Tue, 11 Nov 2008 12:56:45 -0800 Warren Turkal <[email protected]> wrote:

Here's another shot at this patch. I found a small issue with sparse. If this is suitable for inclusion, could someone please pull it into their tree so that it can be merged upstream?

I don't follow linux-fsdevel much. cc'ing linux-kernel gets my attention.

This patchset is missing an overall description of the problem which is being solved. I can kinda see what the changes do, but it would be much better if we were provided with a description of the kernel behaviour before and after the patches.