Donate to e Foundation | Murena handsets with /e/OS | Own a part of Murena! Learn more

Commit bfab36e8 authored by Anton Altaparmakov's avatar Anton Altaparmakov Committed by Linus Torvalds
Browse files

NTFS: Fix a mount time deadlock. 



Big thanks go to Mathias Kolehmainen for reporting the bug, providing
debug output and testing the patches I sent him to get it working.

The fix was to stop calling ntfs_attr_set() at mount time as that causes
balance_dirty_pages_ratelimited() to be called which on systems with
little memory actually tries to go and balance the dirty pages which tries
to take the s_umount semaphore but because we are still in fill_super()
across which the VFS holds s_umount for writing this results in a
deadlock.

We now do the dirty work by hand by submitting individual buffers.  This
has the annoying "feature" that mounting can take a few seconds if the
journal is large as we have clear it all.  One day someone should improve
on this by deferring the journal clearing to a helper kernel thread so it
can be done in the background but I don't have time for this at the moment
and the current solution works fine so I am leaving it like this for now.

Signed-off-by: default avatarAnton Altaparmakov <aia21@cantab.net>
Signed-off-by: default avatarLinus Torvalds <torvalds@linux-foundation.org>
parent f26e51f6

Documentation/filesystems/ntfs.txt

+3 −1
Original line number Original line Diff line number Diff line
@@ -407,7 +407,7 @@ raiddev /dev/md0 
	device		/dev/hda5

	device		/dev/hda5

	raid-disk	0

	raid-disk	0

	device		/dev/hdb1

	device		/dev/hdb1

	raid-disl	1

	raid-disk	1





For linear raid, just change the raid-level above to "raid-level linear", for

For linear raid, just change the raid-level above to "raid-level linear", for

mirrors, change it to "raid-level 1", and for stripe sets with parity, change

mirrors, change it to "raid-level 1", and for stripe sets with parity, change
@@ -457,6 +457,8 @@ ChangeLog 




Note, a technical ChangeLog aimed at kernel hackers is in fs/ntfs/ChangeLog.

Note, a technical ChangeLog aimed at kernel hackers is in fs/ntfs/ChangeLog.





2.1.29:

	- Fix a deadlock when mounting read-write.

2.1.28:

2.1.28:

	- Fix a deadlock.

	- Fix a deadlock.

2.1.27:

2.1.27:

fs/ntfs/ChangeLog

+12 −0
Original line number Original line Diff line number Diff line
@@ -17,6 +17,18 @@ ToDo/Notes: 
	  happen is unclear however so it is worth waiting until someone hits

	  happen is unclear however so it is worth waiting until someone hits

	  the problem.

	  the problem.





2.1.29 - Fix a deadlock at mount time.



	- During mount the VFS holds s_umount lock on the superblock.  So when

	  we try to empty the journal $LogFile contents by calling

	  ntfs_attr_set() when the machine does not have much memory and the

	  journal is large ntfs_attr_set() results in the VM trying to balance

	  dirty pages which in turn tries to that the s_umount lock and thus we

	  get a deadlock.  The solution is to not use ntfs_attr_set() and

	  instead do the zeroing by hand at the block level rather than page

	  cache level.

	- Fix sparse warnings.



2.1.28 - Fix a deadlock.

2.1.28 - Fix a deadlock.





	- Fix deadlock in fs/ntfs/inode.c::ntfs_put_inode().  Thanks to Sergey

	- Fix deadlock in fs/ntfs/inode.c::ntfs_put_inode().  Thanks to Sergey

fs/ntfs/Makefile

+1 −1
Original line number Original line Diff line number Diff line
@@ -6,7 +6,7 @@ ntfs-objs := aops.o attrib.o collate.o compress.o debug.o dir.o file.o \ 
	     index.o inode.o mft.o mst.o namei.o runlist.o super.o sysctl.o \

	     index.o inode.o mft.o mst.o namei.o runlist.o super.o sysctl.o \

	     unistr.o upcase.o

	     unistr.o upcase.o





EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.28\"

EXTRA_CFLAGS = -DNTFS_VERSION=\"2.1.29\"





ifeq ($(CONFIG_NTFS_DEBUG),y)

ifeq ($(CONFIG_NTFS_DEBUG),y)

EXTRA_CFLAGS += -DDEBUG

EXTRA_CFLAGS += -DDEBUG

fs/ntfs/aops.c

+11 −11
Original line number Original line Diff line number Diff line
@@ -2,7 +2,7 @@ 
 * aops.c - NTFS kernel address space operations and page cache handling.

 * aops.c - NTFS kernel address space operations and page cache handling.

 *	    Part of the Linux-NTFS project.

 *	    Part of the Linux-NTFS project.

 *

 *

 * Copyright (c) 2001-2006 Anton Altaparmakov

 * Copyright (c) 2001-2007 Anton Altaparmakov

 * Copyright (c) 2002 Richard Russon

 * Copyright (c) 2002 Richard Russon

 *

 *

 * This program/include file is free software; you can redistribute it and/or

 * This program/include file is free software; you can redistribute it and/or
@@ -396,7 +396,7 @@ static int ntfs_readpage(struct file *file, struct page *page) 
	loff_t i_size;

	loff_t i_size;

	struct inode *vi;

	struct inode *vi;

	ntfs_inode *ni, *base_ni;

	ntfs_inode *ni, *base_ni;

	u8 *kaddr;

	u8 *addr;

	ntfs_attr_search_ctx *ctx;

	ntfs_attr_search_ctx *ctx;

	MFT_RECORD *mrec;

	MFT_RECORD *mrec;

	unsigned long flags;

	unsigned long flags;
@@ -491,15 +491,15 @@ static int ntfs_readpage(struct file *file, struct page *page) 
		/* Race with shrinking truncate. */

		/* Race with shrinking truncate. */

		attr_len = i_size;

		attr_len = i_size;

	}

	}

	kaddr = kmap_atomic(page, KM_USER0);

	addr = kmap_atomic(page, KM_USER0);

	/* Copy the data to the page. */

	/* Copy the data to the page. */

	memcpy(kaddr, (u8*)ctx->attr +

	memcpy(addr, (u8*)ctx->attr +

			le16_to_cpu(ctx->attr->data.resident.value_offset),

			le16_to_cpu(ctx->attr->data.resident.value_offset),

			attr_len);

			attr_len);

	/* Zero the remainder of the page. */

	/* Zero the remainder of the page. */

	memset(kaddr + attr_len, 0, PAGE_CACHE_SIZE - attr_len);

	memset(addr + attr_len, 0, PAGE_CACHE_SIZE - attr_len);

	flush_dcache_page(page);

	flush_dcache_page(page);

	kunmap_atomic(kaddr, KM_USER0);

	kunmap_atomic(addr, KM_USER0);

put_unm_err_out:

put_unm_err_out:

	ntfs_attr_put_search_ctx(ctx);

	ntfs_attr_put_search_ctx(ctx);

unm_err_out:

unm_err_out:
@@ -1344,7 +1344,7 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc) 
	loff_t i_size;

	loff_t i_size;

	struct inode *vi = page->mapping->host;

	struct inode *vi = page->mapping->host;

	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);

	ntfs_inode *base_ni = NULL, *ni = NTFS_I(vi);

	char *kaddr;

	char *addr;

	ntfs_attr_search_ctx *ctx = NULL;

	ntfs_attr_search_ctx *ctx = NULL;

	MFT_RECORD *m = NULL;

	MFT_RECORD *m = NULL;

	u32 attr_len;

	u32 attr_len;
@@ -1484,14 +1484,14 @@ static int ntfs_writepage(struct page *page, struct writeback_control *wbc) 
		/* Shrinking cannot fail. */

		/* Shrinking cannot fail. */

		BUG_ON(err);

		BUG_ON(err);

	}

	}

	kaddr = kmap_atomic(page, KM_USER0);

	addr = kmap_atomic(page, KM_USER0);

	/* Copy the data from the page to the mft record. */

	/* Copy the data from the page to the mft record. */

	memcpy((u8*)ctx->attr +

	memcpy((u8*)ctx->attr +

			le16_to_cpu(ctx->attr->data.resident.value_offset),

			le16_to_cpu(ctx->attr->data.resident.value_offset),

			kaddr, attr_len);

			addr, attr_len);

	/* Zero out of bounds area in the page cache page. */

	/* Zero out of bounds area in the page cache page. */

	memset(kaddr + attr_len, 0, PAGE_CACHE_SIZE - attr_len);

	memset(addr + attr_len, 0, PAGE_CACHE_SIZE - attr_len);

	kunmap_atomic(kaddr, KM_USER0);

	kunmap_atomic(addr, KM_USER0);

	flush_dcache_page(page);

	flush_dcache_page(page);

	flush_dcache_mft_record_page(ctx->ntfs_ino);

	flush_dcache_mft_record_page(ctx->ntfs_ino);

	/* We are done with the page. */

	/* We are done with the page. */

fs/ntfs/attrib.c

+6 −2
Original line number Original line Diff line number Diff line 
/**

/**

 * attrib.c - NTFS attribute operations.  Part of the Linux-NTFS project.

 * attrib.c - NTFS attribute operations.  Part of the Linux-NTFS project.

 *

 *

 * Copyright (c) 2001-2006 Anton Altaparmakov

 * Copyright (c) 2001-2007 Anton Altaparmakov

 * Copyright (c) 2002 Richard Russon

 * Copyright (c) 2002 Richard Russon

 *

 *

 * This program/include file is free software; you can redistribute it and/or

 * This program/include file is free software; you can redistribute it and/or
@@ -2500,7 +2500,7 @@ int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt, const u8 val) 
	struct page *page;

	struct page *page;

	u8 *kaddr;

	u8 *kaddr;

	pgoff_t idx, end;

	pgoff_t idx, end;

	unsigned int start_ofs, end_ofs, size;

	unsigned start_ofs, end_ofs, size;





	ntfs_debug("Entering for ofs 0x%llx, cnt 0x%llx, val 0x%hx.",

	ntfs_debug("Entering for ofs 0x%llx, cnt 0x%llx, val 0x%hx.",

			(long long)ofs, (long long)cnt, val);

			(long long)ofs, (long long)cnt, val);
@@ -2548,6 +2548,8 @@ int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt, const u8 val) 
		kunmap_atomic(kaddr, KM_USER0);

		kunmap_atomic(kaddr, KM_USER0);

		set_page_dirty(page);

		set_page_dirty(page);

		page_cache_release(page);

		page_cache_release(page);

		balance_dirty_pages_ratelimited(mapping);

		cond_resched();

		if (idx == end)

		if (idx == end)

			goto done;

			goto done;

		idx++;

		idx++;
@@ -2604,6 +2606,8 @@ int ntfs_attr_set(ntfs_inode *ni, const s64 ofs, const s64 cnt, const u8 val) 
		kunmap_atomic(kaddr, KM_USER0);

		kunmap_atomic(kaddr, KM_USER0);

		set_page_dirty(page);

		set_page_dirty(page);

		page_cache_release(page);

		page_cache_release(page);

		balance_dirty_pages_ratelimited(mapping);

		cond_resched();

	}

	}

done:

done:

	ntfs_debug("Done.");

	ntfs_debug("Done.");