ext4: bmap & fiemap conversion to use iomap

IOMAP_F_MERGED needs to be set in case of non-extent based mapping. This is needed in later patches for conversion of ext4_fiemap to use iomap.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
 fs/ext4/inode.c | 4 ++++
 1 file changed, 4 insertions(+)

fs/ext4/inode.c +4 -0

index d035acab5b2a..6cf3b969dc86 100644
@@ -3335,6 +3335,10 @@ static void ext4_set_iomap(struct inode *inode, struct iomap *iomap,
 	iomap->offset = (u64) map->m_lblk << blkbits;
 	iomap->length = (u64) map->m_len << blkbits;
 
+	if ((map->m_flags & EXT4_MAP_MAPPED) &&
+	    !ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
+		iomap->flags |= IOMAP_F_MERGED;
+
 	/*
 	 * Flags passed to ext4_map_blocks() for direct I/O writes can result
 	 * in m_flags having both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN bits

Signature · 1 line

2.21.0

IOMAP_F_MERGED needs to be set in case of non-extent based mapping. This is needed in later patches for conversion of ext4_fiemap to use iomap.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

This patch avoids the memory alloc & free path when depth is 0, since anyway there is no extra caching done in that case. So on checking depth 0, simply return early.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
 fs/ext4/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

fs/ext4/inode.c +0 -0

index 6cf3b969dc86..81fccbae0aea 100644

fs/ext4/inode.c +0 -0

 }
 
 static int ext4_readpage(struct file *file, struct page *page)

Signature · 1 line

2.21.0

Quoted 18 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
  fs/ext4/inode.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

fs/ext4/inode.c +0 -0

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
  			return 0;
  	}
 

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3217

Quoted 2 lines

fs/ext4/inode.c +1 -1

-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
  fs/ext4/inode.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
  			return 0;
  	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Signed-off-by Ritesh Harjani <[email protected]>

---
 fs/ext4/inode.c | 12 +-----------
 1 file changed, 1 insertion(+), 11 deletions(-)

fs/ext4/inode.c +1 -11

index 6cf3b969dc86..fac8adbbb3f6 100644
@@ -3174,16 +3174,6 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
 	if (ext4_has_inline_data(inode))
 		return 0;
 
-	if (mapping_tagged(mapping, PAGECACHE_TAG_DIRTY) &&
-			test_opt(inode->i_sb, DELALLOC)) {
-		/*
-		 * With delalloc we want to sync the file
-		 * so that we can make sure we allocate
-		 * blocks for file
-		 */
-		filemap_write_and_wait(mapping);
-	}
-
 	if (EXT4_JOURNAL(inode) &&
 	    ext4_test_inode_state(inode, EXT4_STATE_JDATA)) {
 		/*

@@ -3214,7 +3204,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
 			return 0;
 	}
 
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);
 }
 
 static int ext4_readpage(struct file *file, struct page *page)

Signature · 1 line

2.21.0

Quoted 60 lines · Ritesh Harjani

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
  fs/ext4/inode.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
  			return 0;
  	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Quoted 1 line

Signed-off-by Ritesh Harjani <[email protected]>

Hmmm. I don't recall how jdata actually works, but I get the impression here that we're trying to flush dirty data out to the journal and then out to disk, and then drop the JDATA state from the inode. This mechanism exists (I guess?) so that dirty file pages get checkpointed out of jbd2 back into the filesystem so that bmap() returns meaningful results to lilo.

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Quoted 61 lines · Darrick J. Wong

Quoted 60 lines · Ritesh Harjani

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
  fs/ext4/inode.c | 2 +-
  1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
  			return 0;
  	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Quoted 8 lines

Quoted 1 line

Signed-off-by Ritesh Harjani <[email protected]>

Hmmm. I don't recall how jdata actually works, but I get the impression here that we're trying to flush dirty data out to the journal and then out to disk, and then drop the JDATA state from the inode. This mechanism exists (I guess?) so that dirty file pages get checkpointed out of jbd2 back into the filesystem so that bmap() returns meaningful results to lilo.

Exactly. E.g. when we are journalling data, we fill hole through mmap, we will have block allocated as unwritten and we need to write it out so that the data gets to the journal and then do journal flush to get the data to disk so that lilo can read it from the devices. So removing filemap_write_and_wait() when journalling data is wrong.

Quoted 6 lines

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Quoted 1 line

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Yeah, it should do that but that's only performance optimization so that we bother with journal flushing only when someone uses block mapping call on a file with journalled dirty data. So you can hardly notice the bug by testing...

Honza

Signature · 2 lines

Jan Kara <[email protected]>
SUSE Labs, CR

Quoted 8 lines

Quoted 7 lines

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Quoted 4 lines

Yeah, it should do that but that's only performance optimization so that we bother with journal flushing only when someone uses block mapping call on a file with journalled dirty data. So you can hardly notice the bug by testing...

If we ever decide to deprecate FIBMAP officially and push bootloaders to use FIEMAP, then we'll have to emulate all the flushing behaviors. But that's something for a separate patch.

Quoted 61 lines · Darrick J. Wong

Quoted 60 lines · Ritesh Harjani

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
   fs/ext4/inode.c | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
   			return 0;
   	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Quoted 8 lines

Quoted 1 line

Signed-off-by Ritesh Harjani <[email protected]>

Hmmm. I don't recall how jdata actually works, but I get the impression here that we're trying to flush dirty data out to the journal and then out to disk, and then drop the JDATA state from the inode. This mechanism exists (I guess?) so that dirty file pages get checkpointed out of jbd2 back into the filesystem so that bmap() returns meaningful results to lilo.

Exactly. E.g. when we are journalling data, we fill hole through mmap, we will have block allocated as unwritten and we need to write it out so that the data gets to the journal and then do journal flush to get the data to disk

so that lilo can read it from the devices. So removing filemap_write_and_wait() when journalling data is wrong.

Quoted 8 lines

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Yeah, it should do that but that's only performance optimization so that we bother with journal flushing only when someone uses block mapping call on a file with journalled dirty data. So you can hardly notice the bug by testing...

Honza

Quoted 71 lines · Jan Kara

Quoted 69 lines · Darrick J. Wong

Quoted 55 lines · Ritesh Harjani

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
   fs/ext4/inode.c | 2 +-
   1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
   			return 0;
   	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Quoted 6 lines

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Signed-off-by Ritesh Harjani <[email protected]>

Hmmm. I don't recall how jdata actually works, but I get the impression here that we're trying to flush dirty data out to the journal and then out to disk, and then drop the JDATA state from the inode. This mechanism exists (I guess?) so that dirty file pages get checkpointed out of jbd2 back into the filesystem so that bmap() returns meaningful results to lilo.

Quoted 3 lines

Exactly. E.g. when we are journalling data, we fill hole through mmap, we will have block allocated as unwritten and we need to write it out so that the data gets to the journal and then do journal flush to get the data to disk

So in data=journal case in ext4_page_mkwrite the data buffer will also be marked as, to be journalled. So does jbd2_journal_flush() itself don't take care of writing back any dirty page cache before it commit that transaction? and after then checkpoint it?

Sorry my knowledge about jbd2 is very naive.

Quoted 2 lines

so that lilo can read it from the devices. So removing filemap_write_and_wait() when journalling data is wrong.

Sure I understand this part. But was just curious on above query. Otherwise, IIUC, we will have to add filemap_write_and_wait() for JDATA case as well before calling for jbd2_journal_flush(). Will add this as a separate patch.

-ritesh

Quoted 16 lines

Quoted 8 lines

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Yeah, it should do that but that's only performance optimization so that we bother with journal flushing only when someone uses block mapping call on a file with journalled dirty data. So you can hardly notice the bug by testing...

Honza

Quoted 74 lines · Ritesh Harjani

Quoted 71 lines · Jan Kara

Quoted 69 lines · Darrick J. Wong

Quoted 55 lines · Ritesh Harjani

Quoted 11 lines · Darrick J. Wong

Quoted 10 lines · Ritesh Harjani

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
    fs/ext4/inode.c | 2 +-
    1 file changed, 1 insertion(+), 1 deletion(-)

Quoted 14 lines

fs/ext4/inode.c static sector_t ext4_bmap(struct address_space *mapping, sector_t block) · line 3216

Quoted 9 lines

fs/ext4/inode.c +1 -1

index 6cf3b969dc86..81fccbae0aea 100644
@@ -3214,7 +3214,7 @@ static sector_t ext4_bmap(struct address_space *mapping, sector_t block)
    			return 0;
    	}
-	return generic_block_bmap(mapping, block, ext4_get_block);
+	return iomap_bmap(mapping, block, &ext4_iomap_ops);

/me notes that iomap_bmap will filemap_write_and_wait for you, so one could optimize ext4_bmap to avoid the double-flush by moving the filemap_write_and_wait at the top of the function into the JDATA state clearing block.

IIUC, delalloc and data=journal mode are both mutually exclusive. So we could get rid of calling filemap_write_and_wait() all together from ext4_bmap().
And as you pointed filemap_write_and_wait() is called by default in iomap_bmap which should cover for delalloc case.

@Jan/Darrick,
Could you check if the attached patch looks good. If yes then will add your Reviewed-by and send a v6.

Thanks for the review!!

-ritesh

From 93f560d9a483b4f389056e543012d0941734a8f4 Mon Sep 17 00:00:00 2001
From: Ritesh Harjani <[email protected]>
Date: Tue, 20 Aug 2019 18:36:33 +0530
Subject: [PATCH 3/6] ext4: Move ext4 bmap to use iomap infrastructure.

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Quoted 6 lines

Also no need to call for filemap_write_and_wait() any more in ext4_bmap since data=journal mode anyway doesn't support delalloc and for all other cases iomap_bmap() anyway calls the same function, so no need for doing it twice.

Signed-off-by Ritesh Harjani <[email protected]>

Hmmm. I don't recall how jdata actually works, but I get the impression here that we're trying to flush dirty data out to the journal and then out to disk, and then drop the JDATA state from the inode. This mechanism exists (I guess?) so that dirty file pages get checkpointed out of jbd2 back into the filesystem so that bmap() returns meaningful results to lilo.

Quoted 8 lines

Quoted 3 lines

Exactly. E.g. when we are journalling data, we fill hole through mmap, we will have block allocated as unwritten and we need to write it out so that the data gets to the journal and then do journal flush to get the data to disk

So in data=journal case in ext4_page_mkwrite the data buffer will also be marked as, to be journalled. So does jbd2_journal_flush() itself don't take care of writing back any dirty page cache before it commit that transaction? and after then checkpoint it?

Er... this sentence is a little garbled, but I think the answer you're looking for is:

"Yes, writeback (i.e. filemap_write_and_wait) attaches the dirty blocks to a journal transaction; then jbd2_journal_flush forces the transaction data out to the on-disk journal; and it also checkpoints the journal so that the dirty blocks are then written back into the filesystem."

Quoted 10 lines

Sorry my knowledge about jbd2 is very naive.

Quoted 2 lines

so that lilo can read it from the devices. So removing filemap_write_and_wait() when journalling data is wrong.

Sure I understand this part. But was just curious on above query. Otherwise, IIUC, we will have to add filemap_write_and_wait() for JDATA case as well before calling for jbd2_journal_flush(). Will add this as a separate patch.

Well you could just move it...

bmap()

{
	/*
	 * In data=journal mode, we must checkpoint the journal to
	 * ensure that any dirty blocks in the journalare checkpointed
	 * to the location that we return to userspace.  Clear JDATA so
	 * that future writes will not be written through the journal.
	 */
	if (JDATA) {
		filemap_write_and_wait(...);
		clear JDATA
		jbd2_journal_flush(...);
	}

	return iomap_bmap(...);
}

(or did "Will add this as a separate patch" refer to fixing FIEMAP?)

--D

Quoted 20 lines

-ritesh

Quoted 16 lines

Quoted 8 lines

This makes me wonder if you still need the filemap_write_and_wait in the JDATA case because otherwise the journal flush won't have the effect of writing all the dirty pagecache back to the filesystem? OTOH I suppose the implicit write-and-wait call after we clear JDATA will not be writing to the journal.

Even more weirdly, the FIEMAP code doesn't drop JDATA at all...?

Yeah, it should do that but that's only performance optimization so that we bother with journal flushing only when someone uses block mapping call on a file with journalled dirty data. So you can hardly notice the bug by testing...

Honza

ext4_iomap_begin is already implemented which provides ext4_map_blocks, so just move the API from generic_block_bmap to iomap_bmap for iomap conversion.

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

For indirect block mapping if the i_block > max supported block in inode then ext4_ind_map_blocks() returns a -EIO error. But in case of fiemap this could be a valid query to ->iomap_begin call.
So check if the offset >= s_bitmap_maxbytes in ext4_iomap_begin_report(), then simply skip calling ext4_map_blocks().

Signed-off-by Ritesh Harjani <[email protected]>

Reviewed-by Jan Kara <[email protected]>

This patch moves ext4_fiemap to use iomap framework. For xattr a new 'ext4_iomap_xattr_ops' is added.

Signed-off-by Ritesh Harjani <[email protected]>

Reported-by kbuild test robot <[email protected]>

Reviewed-by Jan Kara <[email protected]>

---
 fs/ext4/extents.c | 290 ++++++++--------------------------------------
 fs/ext4/inline.c  |  41 -------
 2 files changed, 48 insertions(+), 283 deletions(-)

fs/ext4/extents.c +48 -242

index 0de548bb3c90..0e3dfea0c065 100644
@@ -28,6 +28,7 @@
 #include <linux/uaccess.h>
 #include <linux/fiemap.h>
 #include <linux/backing-dev.h>
+#include <linux/iomap.h>
 #include "ext4_jbd2.h"
 #include "ext4_extents.h"
 #include "xattr.h"

@@ -97,9 +98,6 @@ static int ext4_split_extent_at(handle_t *handle,
 			     int split_flag,
 			     int flags);
 
-static int ext4_find_delayed_extent(struct inode *inode,
-				    struct extent_status *newes);
-
 static int ext4_ext_trunc_restart_fn(struct inode *inode, int *dropped)
 {
 	/*

@@ -2176,155 +2174,6 @@ int ext4_ext_insert_extent(handle_t *handle, struct inode *inode,
 	return err;
 }
 
-static int ext4_fill_fiemap_extents(struct inode *inode,
-				    ext4_lblk_t block, ext4_lblk_t num,
-				    struct fiemap_extent_info *fieinfo)
-{
-	struct ext4_ext_path *path = NULL;
-	struct ext4_extent *ex;
-	struct extent_status es;
-	ext4_lblk_t next, next_del, start = 0, end = 0;
-	ext4_lblk_t last = block + num;
-	int exists, depth = 0, err = 0;
-	unsigned int flags = 0;
-	unsigned char blksize_bits = inode->i_sb->s_blocksize_bits;
-
-	while (block < last && block != EXT_MAX_BLOCKS) {
-		num = last - block;
-		/* find extent for this block */
-		down_read(&EXT4_I(inode)->i_data_sem);
-
-		path = ext4_find_extent(inode, block, &path, 0);
-		if (IS_ERR(path)) {
-			up_read(&EXT4_I(inode)->i_data_sem);
-			err = PTR_ERR(path);
-			path = NULL;
-			break;
-		}
-
-		depth = ext_depth(inode);
-		if (unlikely(path[depth].p_hdr == NULL)) {
-			up_read(&EXT4_I(inode)->i_data_sem);
-			EXT4_ERROR_INODE(inode, "path[%d].p_hdr == NULL", depth);
-			err = -EFSCORRUPTED;
-			break;
-		}
-		ex = path[depth].p_ext;
-		next = ext4_ext_next_allocated_block(path);
-
-		flags = 0;
-		exists = 0;
-		if (!ex) {
-			/* there is no extent yet, so try to allocate
-			 * all requested space */
-			start = block;
-			end = block + num;
-		} else if (le32_to_cpu(ex->ee_block) > block) {
-			/* need to allocate space before found extent */
-			start = block;
-			end = le32_to_cpu(ex->ee_block);
-			if (block + num < end)
-				end = block + num;
-		} else if (block >= le32_to_cpu(ex->ee_block)
-					+ ext4_ext_get_actual_len(ex)) {
-			/* need to allocate space after found extent */
-			start = block;
-			end = block + num;
-			if (end >= next)
-				end = next;
-		} else if (block >= le32_to_cpu(ex->ee_block)) {
-			/*
-			 * some part of requested space is covered
-			 * by found extent
-			 */
-			start = block;
-			end = le32_to_cpu(ex->ee_block)
-				+ ext4_ext_get_actual_len(ex);
-			if (block + num < end)
-				end = block + num;
-			exists = 1;
-		} else {
-			BUG();
-		}
-		BUG_ON(end <= start);
-
-		if (!exists) {
-			es.es_lblk = start;
-			es.es_len = end - start;
-			es.es_pblk = 0;
-		} else {
-			es.es_lblk = le32_to_cpu(ex->ee_block);
-			es.es_len = ext4_ext_get_actual_len(ex);
-			es.es_pblk = ext4_ext_pblock(ex);
-			if (ext4_ext_is_unwritten(ex))
-				flags |= FIEMAP_EXTENT_UNWRITTEN;
-		}
-
-		/*
-		 * Find delayed extent and update es accordingly. We call
-		 * it even in !exists case to find out whether es is the
-		 * last existing extent or not.
-		 */
-		next_del = ext4_find_delayed_extent(inode, &es);
-		if (!exists && next_del) {
-			exists = 1;
-			flags |= (FIEMAP_EXTENT_DELALLOC |
-				  FIEMAP_EXTENT_UNKNOWN);
-		}
-		up_read(&EXT4_I(inode)->i_data_sem);
-
-		if (unlikely(es.es_len == 0)) {
-			EXT4_ERROR_INODE(inode, "es.es_len == 0");
-			err = -EFSCORRUPTED;
-			break;
-		}
-
-		/*
-		 * This is possible iff next == next_del == EXT_MAX_BLOCKS.
-		 * we need to check next == EXT_MAX_BLOCKS because it is
-		 * possible that an extent is with unwritten and delayed
-		 * status due to when an extent is delayed allocated and
-		 * is allocated by fallocate status tree will track both of
-		 * them in a extent.
-		 *
-		 * So we could return a unwritten and delayed extent, and
-		 * its block is equal to 'next'.
-		 */
-		if (next == next_del && next == EXT_MAX_BLOCKS) {
-			flags |= FIEMAP_EXTENT_LAST;
-			if (unlikely(next_del != EXT_MAX_BLOCKS ||
-				     next != EXT_MAX_BLOCKS)) {
-				EXT4_ERROR_INODE(inode,
-						 "next extent == %u, next "
-						 "delalloc extent = %u",
-						 next, next_del);
-				err = -EFSCORRUPTED;
-				break;
-			}
-		}
-
-		if (exists) {
-			err = fiemap_fill_next_extent(fieinfo,
-				(__u64)es.es_lblk << blksize_bits,
-				(__u64)es.es_pblk << blksize_bits,
-				(__u64)es.es_len << blksize_bits,
-				flags);
-			if (err < 0)
-				break;
-			if (err == 1) {
-				err = 0;
-				break;
-			}
-		}
-
-		block = es.es_lblk + es.es_len;
-	}
-
-	ext4_ext_drop_refs(path);
-	kfree(path);
-	return err;
-}
-
 static int ext4_fill_es_cache_info(struct inode *inode,
 				   ext4_lblk_t block, ext4_lblk_t num,
 				   struct fiemap_extent_info *fieinfo)

@@ -5058,64 +4907,13 @@ int ext4_convert_unwritten_io_end_vec(handle_t *handle, ext4_io_end_t *io_end)
 	return ret < 0 ? ret : err;
 }
 
-/*
- * If newes is not existing extent (newes->ec_pblk equals zero) find
- * delayed extent at start of newes and update newes accordingly and
- * return start of the next delayed extent.
- *
- * If newes is existing extent (newes->ec_pblk is not equal zero)
- * return start of next delayed extent or EXT_MAX_BLOCKS if no delayed
- * extent found. Leave newes unmodified.
- */
-static int ext4_find_delayed_extent(struct inode *inode,
-				    struct extent_status *newes)
-{
-	struct extent_status es;
-	ext4_lblk_t block, next_del;
-
-	if (newes->es_pblk == 0) {
-		ext4_es_find_extent_range(inode, &ext4_es_is_delayed,
-					  newes->es_lblk,
-					  newes->es_lblk + newes->es_len - 1,
-					  &es);
-
-		/*
-		 * No extent in extent-tree contains block @newes->es_pblk,
-		 * then the block may stay in 1)a hole or 2)delayed-extent.
-		 */
-		if (es.es_len == 0)
-			/* A hole found. */
-			return 0;
-
-		if (es.es_lblk > newes->es_lblk) {
-			/* A hole found. */
-			newes->es_len = min(es.es_lblk - newes->es_lblk,
-					    newes->es_len);
-			return 0;
-		}
-
-		newes->es_len = es.es_lblk + es.es_len - newes->es_lblk;
-	}
-
-	block = newes->es_lblk + newes->es_len;
-	ext4_es_find_extent_range(inode, &ext4_es_is_delayed, block,
-				  EXT_MAX_BLOCKS, &es);
-	if (es.es_len == 0)
-		next_del = EXT_MAX_BLOCKS;
-	else
-		next_del = es.es_lblk;
-
-	return next_del;
-}
-
-static int ext4_xattr_fiemap(struct inode *inode,
-				struct fiemap_extent_info *fieinfo)
+static int ext4_iomap_xattr_fiemap(struct inode *inode, struct iomap *iomap)
 {
 	__u64 physical = 0;
-	__u64 length;
-	__u32 flags = FIEMAP_EXTENT_LAST;
+	__u64 length = 0;
 	int blockbits = inode->i_sb->s_blocksize_bits;
 	int error = 0;
+	u16 iomap_type;
 
 	/* in-inode? */
 	if (ext4_test_inode_state(inode, EXT4_STATE_XATTR)) {

@@ -5130,40 +4928,49 @@ static int ext4_xattr_fiemap(struct inode *inode,
 				EXT4_I(inode)->i_extra_isize;
 		physical += offset;
 		length = EXT4_SB(inode->i_sb)->s_inode_size - offset;
-		flags |= FIEMAP_EXTENT_DATA_INLINE;
 		brelse(iloc.bh);
-	} else { /* external block */
+		iomap_type = IOMAP_INLINE;
+	} else if (EXT4_I(inode)->i_file_acl) { /* external block */
 		physical = (__u64)EXT4_I(inode)->i_file_acl << blockbits;
 		length = inode->i_sb->s_blocksize;
+		iomap_type = IOMAP_MAPPED;
+	} else {
+		/* no in-inode or external block for xattr, so return -ENOENT */
+		error = -ENOENT;
+		goto out;
 	}
 
-	if (physical)
-		error = fiemap_fill_next_extent(fieinfo, 0, physical,
-						length, flags);
-	return (error < 0 ? error : 0);
+	iomap->addr = physical;
+	iomap->offset = 0;
+	iomap->length = length;
+	iomap->type = iomap_type;
+	iomap->flags = 0;
+out:
+	return error;
 }
 
-static int _ext4_fiemap(struct inode *inode,
-			struct fiemap_extent_info *fieinfo,
-			__u64 start, __u64 len,
-			int (*fill)(struct inode *, ext4_lblk_t,
-				    ext4_lblk_t,
-				    struct fiemap_extent_info *))
+static int ext4_iomap_xattr_begin(struct inode *inode, loff_t offset,
+				  loff_t length, unsigned flags,
+				  struct iomap *iomap, struct iomap *srcmap)
 {
-	ext4_lblk_t start_blk;
-	u32 ext4_fiemap_flags = FIEMAP_FLAG_SYNC|FIEMAP_FLAG_XATTR;
+	int error;
 
-	int error = 0;
-
-	if (ext4_has_inline_data(inode)) {
-		int has_inline = 1;
+	error = ext4_iomap_xattr_fiemap(inode, iomap);
+	if (error == 0 && (offset >= iomap->length))
+		error = -ENOENT;
+	return error;
+}
 
-		error = ext4_inline_data_fiemap(inode, fieinfo, &has_inline,
-						start, len);
+static const struct iomap_ops ext4_iomap_xattr_ops = {
+	.iomap_begin		= ext4_iomap_xattr_begin,
+};
 
-		if (has_inline)
-			return error;
-	}
+static int _ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
+			__u64 start, __u64 len, bool from_es_cache)
+{
+	ext4_lblk_t start_blk;
+	u32 ext4_fiemap_flags = FIEMAP_FLAG_SYNC | FIEMAP_FLAG_XATTR;
+	int error = 0;
 
 	if (fieinfo->fi_flags & FIEMAP_FLAG_CACHE) {
 		error = ext4_ext_precache(inode);

@@ -5172,19 +4979,19 @@ static int _ext4_fiemap(struct inode *inode,
 		fieinfo->fi_flags &= ~FIEMAP_FLAG_CACHE;
 	}
 
-	/* fallback to generic here if not in extents fmt */
-	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) &&
-	    fill == ext4_fill_fiemap_extents)
-		return generic_block_fiemap(inode, fieinfo, start, len,
-			ext4_get_block);
-
-	if (fill == ext4_fill_es_cache_info)
+	if (from_es_cache)
 		ext4_fiemap_flags &= FIEMAP_FLAG_XATTR;
+
 	if (fiemap_check_flags(fieinfo, ext4_fiemap_flags))
 		return -EBADR;
 
 	if (fieinfo->fi_flags & FIEMAP_FLAG_XATTR) {
-		error = ext4_xattr_fiemap(inode, fieinfo);
+		fieinfo->fi_flags &= ~FIEMAP_FLAG_XATTR;
+		error = iomap_fiemap(inode, fieinfo, start, len,
+				     &ext4_iomap_xattr_ops);
+	} else if (!from_es_cache) {
+		error = iomap_fiemap(inode, fieinfo, start, len,
+				     &ext4_iomap_report_ops);
 	} else {
 		ext4_lblk_t len_blks;
 		__u64 last_blk;

@@ -5199,7 +5006,8 @@ static int _ext4_fiemap(struct inode *inode,
 		 * Walk the extent tree gathering extent information
 		 * and pushing extents back to the user.
 		 */
-		error = fill(inode, start_blk, len_blks, fieinfo);
+		error = ext4_fill_es_cache_info(inode, start_blk, len_blks,
+						fieinfo);
 	}
 	return error;
 }

@@ -5207,8 +5015,7 @@ static int _ext4_fiemap(struct inode *inode,
 int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		__u64 start, __u64 len)
 {
-	return _ext4_fiemap(inode, fieinfo, start, len,
-			    ext4_fill_fiemap_extents);
+	return _ext4_fiemap(inode, fieinfo, start, len, false);
 }
 
 int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo,

@@ -5224,8 +5031,7 @@ int ext4_get_es_cache(struct inode *inode, struct fiemap_extent_info *fieinfo,
 			return 0;
 	}
 
-	return _ext4_fiemap(inode, fieinfo, start, len,
-			    ext4_fill_es_cache_info);
+	return _ext4_fiemap(inode, fieinfo, start, len, true);
 }
 
 

fs/ext4/inline.c +0 -41

index e61603f47035..b8b99634c832 100644
@@ -1857,47 +1857,6 @@ int ext4_inline_data_iomap(struct inode *inode, struct iomap *iomap)
 	return error;
 }
 
-int ext4_inline_data_fiemap(struct inode *inode,
-			    struct fiemap_extent_info *fieinfo,
-			    int *has_inline, __u64 start, __u64 len)
-{
-	__u64 physical = 0;
-	__u64 inline_len;
-	__u32 flags = FIEMAP_EXTENT_DATA_INLINE | FIEMAP_EXTENT_NOT_ALIGNED |
-		FIEMAP_EXTENT_LAST;
-	int error = 0;
-	struct ext4_iloc iloc;
-
-	down_read(&EXT4_I(inode)->xattr_sem);
-	if (!ext4_has_inline_data(inode)) {
-		*has_inline = 0;
-		goto out;
-	}
-	inline_len = min_t(size_t, ext4_get_inline_size(inode),
-			   i_size_read(inode));
-	if (start >= inline_len)
-		goto out;
-	if (start + len < inline_len)
-		inline_len = start + len;
-	inline_len -= start;
-
-	error = ext4_get_inode_loc(inode, &iloc);
-	if (error)
-		goto out;
-
-	physical = (__u64)iloc.bh->b_blocknr << inode->i_sb->s_blocksize_bits;
-	physical += (char *)ext4_raw_inode(&iloc) - iloc.bh->b_data;
-	physical += offsetof(struct ext4_inode, i_block);
-
-	brelse(iloc.bh);
-out:
-	up_read(&EXT4_I(inode)->xattr_sem);
-	if (physical)
-		error = fiemap_fill_next_extent(fieinfo, start, physical,
-						inline_len, flags);
-	return (error < 0 ? error : 0);
-}
-
 int ext4_inline_data_truncate(struct inode *inode, int *has_inline)
 {
 	handle_t *handle;

Signature · 1 line

2.21.0

This patch moves ext4_fiemap to use iomap framework. For xattr a new 'ext4_iomap_xattr_ops' is added.

Signed-off-by Ritesh Harjani <[email protected]>

Reported-by kbuild test robot <[email protected]>

Reviewed-by Jan Kara <[email protected]>

Currently FIEMAP_EXTENT_LAST is not working consistently across different filesystem's fiemap implementations. So add more information about how else this flag could set in other implementation.

Also in general, user should not completely rely on this flag as such since it could return false value for e.g.
when there is a delalloc extent which might get converted during writeback, immediately after the fiemap calls return.

Signed-off-by Ritesh Harjani <[email protected]>

---
 Documentation/filesystems/fiemap.txt | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

Documentation/filesystems/fiemap.txt +5 -5

index f6d9c99103a4..fedfa9b9dde5 100644
@@ -71,8 +71,7 @@ allocated is less than would be required to map the requested range,
 the maximum number of extents that can be mapped in the fm_extent[]
 array will be returned and fm_mapped_extents will be equal to
 fm_extent_count. In that case, the last extent in the array will not
-complete the requested range and will not have the FIEMAP_EXTENT_LAST
-flag set (see the next section on extent flags).
+complete the requested range.
 
 Each extent is described by a single fiemap_extent structure as
 returned in fm_extents.

@@ -96,7 +95,7 @@ block size of the file system.  With the exception of extents flagged as
 FIEMAP_EXTENT_MERGED, adjacent extents will not be merged.
 
 The fe_flags field contains flags which describe the extent returned.
-A special flag, FIEMAP_EXTENT_LAST is always set on the last extent in
+A special flag, FIEMAP_EXTENT_LAST *may be* set on the last extent in
 the file so that the process making fiemap calls can determine when no
 more extents are available, without having to call the ioctl again.
 

@@ -115,8 +114,9 @@ data. Note that the opposite is not true - it would be valid for
 FIEMAP_EXTENT_NOT_ALIGNED to appear alone.
 
 * FIEMAP_EXTENT_LAST
-This is the last extent in the file. A mapping attempt past this
-extent will return nothing.
+This is generally the last extent in the file. A mapping attempt past this
+extent may return nothing. In some implementations this flag is also set on
+the last dataset queried by the user (via fiemap->fm_length).
 
 * FIEMAP_EXTENT_UNKNOWN
 The location of this extent is currently unknown. This may indicate

Signature · 1 line

2.21.0

Currently FIEMAP_EXTENT_LAST is not working consistently across different filesystem's fiemap implementations. So add more information about how else this flag could set in other implementation.

Also in general, user should not completely rely on this flag as such since it could return false value for e.g.
when there is a delalloc extent which might get converted during writeback, immediately after the fiemap calls return.

Signed-off-by Ritesh Harjani <[email protected]>

---
 Documentation/filesystems/fiemap.txt | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

Documentation/filesystems/fiemap.txt +0 -0

index f6d9c99103a4..fedfa9b9dde5 100644

Quoted 14 lines · Ritesh Harjani

Currently FIEMAP_EXTENT_LAST is not working consistently across different filesystem's fiemap implementations. So add more information about how else this flag could set in other implementation.

Also in general, user should not completely rely on this flag as such since it could return false value for e.g.
when there is a delalloc extent which might get converted during writeback, immediately after the fiemap calls return.

Signed-off-by Ritesh Harjani <[email protected]>

---
  Documentation/filesystems/fiemap.txt | 10 +++++-----
  1 file changed, 5 insertions(+), 5 deletions(-)

Documentation/filesystems/fiemap.txt allocated is less than would be required to map the requested range, · line 74

Quoted 11 lines

Documentation/filesystems/fiemap.txt +1 -2

index f6d9c99103a4..fedfa9b9dde5 100644
@@ -71,8 +71,7 @@ allocated is less than would be required to map the requested range,
  the maximum number of extents that can be mapped in the fm_extent[]
  array will be returned and fm_mapped_extents will be equal to
  fm_extent_count. In that case, the last extent in the array will not
-complete the requested range and will not have the FIEMAP_EXTENT_LAST
-flag set (see the next section on extent flags).
+complete the requested range.

This sentence still seems like it should be true. If the filesystem knows there are more extents to come, it will definitely not set the LAST flag.

Documentation/filesystems/fiemap.txt block size of the file system. With the exception of extents flagged as · line 100

Quoted 8 lines

Documentation/filesystems/fiemap.txt +1 -1

@@ -96,7 +95,7 @@ block size of the file system.  With the exception of extents flagged as
  FIEMAP_EXTENT_MERGED, adjacent extents will not be merged.
 
  The fe_flags field contains flags which describe the extent returned.
-A special flag, FIEMAP_EXTENT_LAST is always set on the last extent in
+A special flag, FIEMAP_EXTENT_LAST *may be* set on the last extent in
  the file so that the process making fiemap calls can determine when no
  more extents are available, without having to call the ioctl again.

I'm not sure I'd highlight 'may be' here.

Documentation/filesystems/fiemap.txt data. Note that the opposite is not true - it would be valid for · line 119

Quoted 10 lines

Documentation/filesystems/fiemap.txt +3 -2

@@ -115,8 +114,9 @@ data. Note that the opposite is not true - it would be valid for
  FIEMAP_EXTENT_NOT_ALIGNED to appear alone.
 
  * FIEMAP_EXTENT_LAST
-This is the last extent in the file. A mapping attempt past this
-extent will return nothing.
+This is generally the last extent in the file. A mapping attempt past this
+extent may return nothing. In some implementations this flag is also set on
+the last dataset queried by the user (via fiemap->fm_length).

The word 'dataset' is used nowhere else in this document. How about

"Some filesystems set this flag to indicate this extent is the last one in the range queried by the user"

Thanks for the review.
Will make the suggested changes and send a v6.