Skip to contentSkip to conversation

xfs: _{attr,data}_map_shared should take ILOCK_EXCL until iread_extents is completely done

PATCH Superseded Darrick J. Wong 3 participants 5 messages Apr 11, 2023 – Apr 12, 2023 XFS filesystem

lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 xfs: _{attr,data}_map_shared should take ILOCK_EXCL until iread_extents is completely done

From: Darrick J. Wong <[email protected]>

While fuzzing the data fork extent count on a btree-format directory with xfs/375, I observed the following (excerpted) splat:

XFS: Assertion failed: xfs_isilocked(ip, XFS_ILOCK_EXCL), file: fs/xfs/libxfs/xfs_bmap.c, line: 1208

------------[ cut here ]------------
WARNING: CPU: 0 PID: 43192 at fs/xfs/xfs_message.c:104 assfail+0x46/0x4a [xfs]
Call Trace:
 <TASK>
 xfs_iread_extents+0x1af/0x210 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_dir_walk+0xb8/0x190 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_count_parent_dentries+0x41/0x80 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_validate+0x199/0x2e0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent+0xdf/0x130 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrub_metadata+0x2b8/0x730 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrubv_metadata+0x38b/0x4d0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_ioc_scrubv_metadata+0x111/0x160 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_file_ioctl+0x367/0xf50 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 __x64_sys_ioctl+0x82/0xa0
 do_syscall_64+0x2b/0x80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

The cause of this is a race condition in xfs_ilock_data_map_shared, which performs an unlocked access to the data fork to guess which lock mode it needs:

Thread 0 Thread 1

xfs_need_iread_extents
<observe no iext tree>
xfs_ilock(..., ILOCK_EXCL)
xfs_iread_extents
<observe no iext tree>
<check ILOCK_EXCL>
<load bmbt extents into iext>
<notice iext size doesn't
match nextents>
xfs_need_iread_extents <observe iext tree> xfs_ilock(..., ILOCK_SHARED)
<tear down iext tree>
xfs_iunlock(..., ILOCK_EXCL)
xfs_iread_extents <observe no iext tree> <check ILOCK_EXCL> *BOOM*

Fix this race by adding a flag to the xfs_ifork structure to indicate that we have not yet read in the extent records and changing the predicate to look at the flag state, not if_height. The memory barrier ensures that the flag will not be set until the very end of the function.

---
v2: use smp_store_release per reviewer comments

---
 fs/xfs/libxfs/xfs_bmap.c       |    6 ++++++
 fs/xfs/libxfs/xfs_inode_fork.c |   16 +++++++++++++++-
 fs/xfs/libxfs/xfs_inode_fork.h |    6 ++++--
 3 files changed, 25 insertions(+), 3 deletions(-)
fs/xfs/libxfs/xfs_bmap.c +6 -0
index 34de6e6898c4..f11ef331e5a4 100644
@@ -1171,6 +1171,12 @@ xfs_iread_extents(
 		goto out;
 	}
 	ASSERT(ir.loaded == xfs_iext_count(ifp));
+	/*
+	 * Use release semantics so that we can use acquire semantics in
+	 * xfs_need_iread_extents and be guaranteed to see a valid mapping tree
+	 * after that load.
+	 */
+	smp_store_release(&ifp->if_needextents, 0);
 	return 0;
 out:
 	xfs_iext_destroy(ifp);
fs/xfs/libxfs/xfs_inode_fork.c +15 -1
index 6b21760184d9..1bbe5ea3f00b 100644
@@ -226,10 +226,15 @@ xfs_iformat_data_fork(
 
 	/*
 	 * Initialize the extent count early, as the per-format routines may
-	 * depend on it.
+	 * depend on it.  Use release semantics to set needextents /after/ we
+	 * set the format. This ensures that we can use acquire semantics on
+	 * needextents in xfs_need_iread_extents() and be guaranteed to see a
+	 * valid format value after that load.
 	 */
 	ip->i_df.if_format = dip->di_format;
 	ip->i_df.if_nextents = xfs_dfork_data_extents(dip);
+	smp_store_release(&ip->i_df.if_needextents,
+			   ip->i_df.if_format == XFS_DINODE_FMT_BTREE ? 1 : 0);
 
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFIFO:
@@ -282,8 +287,17 @@ xfs_ifork_init_attr(
 	enum xfs_dinode_fmt	format,
 	xfs_extnum_t		nextents)
 {
+	/*
+	 * Initialize the extent count early, as the per-format routines may
+	 * depend on it.  Use release semantics to set needextents /after/ we
+	 * set the format. This ensures that we can use acquire semantics on
+	 * needextents in xfs_need_iread_extents() and be guaranteed to see a
+	 * valid format value after that load.
+	 */
 	ip->i_af.if_format = format;
 	ip->i_af.if_nextents = nextents;
+	smp_store_release(&ip->i_af.if_needextents,
+			   ip->i_af.if_format == XFS_DINODE_FMT_BTREE ? 1 : 0);
 }
 
 void
fs/xfs/libxfs/xfs_inode_fork.h +4 -2
index d3943d6ad0b9..96d307784c85 100644
@@ -24,6 +24,7 @@ struct xfs_ifork {
 	xfs_extnum_t		if_nextents;	/* # of extents in this fork */
 	short			if_broot_bytes;	/* bytes allocated for root */
 	int8_t			if_format;	/* format of this fork */
+	uint8_t			if_needextents;	/* extents have not been read */
 };
 
 /*
@@ -260,9 +261,10 @@ int xfs_iext_count_upgrade(struct xfs_trans *tp, struct xfs_inode *ip,
 		uint nr_to_add);
 
 /* returns true if the fork has extents but they are not read in yet. */
-static inline bool xfs_need_iread_extents(struct xfs_ifork *ifp)
+static inline bool xfs_need_iread_extents(const struct xfs_ifork *ifp)
 {
-	return ifp->if_format == XFS_DINODE_FMT_BTREE && ifp->if_height == 0;
+	/* see xfs_iformat_{data,attr}_fork() for needextents semantics */
+	return smp_load_acquire(&ifp->if_needextents) != 0;
 }
 
 #endif	/* __XFS_INODE_FORK_H__ */
#1 lore raw

PATCH v2 1/1 xfsprogs: _{attr,data}_map_shared should take ILOCK_EXCL until iread_extents is completely done

From: Darrick J. Wong <[email protected]>

While fuzzing the data fork extent count on a btree-format directory with xfs/375, I observed the following (excerpted) splat:

XFS: Assertion failed: xfs_isilocked(ip, XFS_ILOCK_EXCL), file: fs/xfs/libxfs/xfs_bmap.c, line: 1208

------------[ cut here ]------------
WARNING: CPU: 0 PID: 43192 at fs/xfs/xfs_message.c:104 assfail+0x46/0x4a [xfs]
Call Trace:
 <TASK>
 xfs_iread_extents+0x1af/0x210 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_dir_walk+0xb8/0x190 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_count_parent_dentries+0x41/0x80 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_validate+0x199/0x2e0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent+0xdf/0x130 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrub_metadata+0x2b8/0x730 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrubv_metadata+0x38b/0x4d0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_ioc_scrubv_metadata+0x111/0x160 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_file_ioctl+0x367/0xf50 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 __x64_sys_ioctl+0x82/0xa0
 do_syscall_64+0x2b/0x80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

The cause of this is a race condition in xfs_ilock_data_map_shared, which performs an unlocked access to the data fork to guess which lock mode it needs:

Thread 0 Thread 1

xfs_need_iread_extents
<observe no iext tree>
xfs_ilock(..., ILOCK_EXCL)
xfs_iread_extents
<observe no iext tree>
<check ILOCK_EXCL>
<load bmbt extents into iext>
<notice iext size doesn't
match nextents>
xfs_need_iread_extents <observe iext tree> xfs_ilock(..., ILOCK_SHARED)
<tear down iext tree>
xfs_iunlock(..., ILOCK_EXCL)
xfs_iread_extents <observe no iext tree> <check ILOCK_EXCL> *BOOM*

Fix this race by adding a flag to the xfs_ifork structure to indicate that we have not yet read in the extent records and changing the predicate to look at the flag state, not if_height. The memory barrier ensures that the flag will not be set until the very end of the function.

---
Here's the userspace version, which steals heavily from the kernel but otherwise uses liburcu underneath.

---
 include/atomic.h        |  100 +++++++++++++++++++++++++++++++++++++++++++++++
 libxfs/libxfs_priv.h    |    3 -
 libxfs/xfs_bmap.c       |    6 +++
 libxfs/xfs_inode_fork.c |   16 +++++++-
 libxfs/xfs_inode_fork.h |    6 ++-
 5 files changed, 125 insertions(+), 6 deletions(-)
include/atomic.h +100 -0
index 9c4aa5849ae..98889190bb0 100644
@@ -118,4 +118,104 @@ atomic64_set(atomic64_t *a, int64_t v)
 
 #endif /* HAVE_URCU_ATOMIC64 */
 
+#define __smp_mb()		cmm_smp_mb()
+
+/* from compiler_types.h */
+/*
+ * __unqual_scalar_typeof(x) - Declare an unqualified scalar type, leaving
+ *			       non-scalar types unchanged.
+ */
+/*
+ * Prefer C11 _Generic for better compile-times and simpler code. Note: 'char'
+ * is not type-compatible with 'signed char', and we define a separate case.
+ */
+#define __scalar_type_to_expr_cases(type)				\
+		unsigned type:	(unsigned type)0,			\
+		signed type:	(signed type)0
+
+#define __unqual_scalar_typeof(x) typeof(				\
+		_Generic((x),						\
+			 char:	(char)0,				\
+			 __scalar_type_to_expr_cases(char),		\
+			 __scalar_type_to_expr_cases(short),		\
+			 __scalar_type_to_expr_cases(int),		\
+			 __scalar_type_to_expr_cases(long),		\
+			 __scalar_type_to_expr_cases(long long),	\
+			 default: (x)))
+
+/* Is this type a native word size -- useful for atomic operations */
+#define __native_word(t) \
+	(sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || \
+	 sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
+
+#define compiletime_assert(foo, str)	BUILD_BUG_ON(!(foo))
+
+#define compiletime_assert_atomic_type(t)				\
+	compiletime_assert(__native_word(t),				\
+		"Need native word sized stores/loads for atomicity.")
+
+/* from barrier.h */
+#ifndef __smp_store_release
+#define __smp_store_release(p, v)					\
+do {									\
+	compiletime_assert_atomic_type(*p);				\
+	__smp_mb();							\
+	WRITE_ONCE(*p, v);						\
+} while (0)
+#endif
+
+#ifndef __smp_load_acquire
+#define __smp_load_acquire(p)						\
+({									\
+	__unqual_scalar_typeof(*p) ___p1 = READ_ONCE(*p);		\
+	compiletime_assert_atomic_type(*p);				\
+	__smp_mb();							\
+	(typeof(*p))___p1;						\
+})
+#endif
+
+#ifndef smp_store_release
+#define smp_store_release(p, v) __smp_store_release((p), (v))
+#endif
+
+#ifndef smp_load_acquire
+#define smp_load_acquire(p) __smp_load_acquire(p)
+#endif
+
+/* from rwonce.h */
+/*
+ * Yes, this permits 64-bit accesses on 32-bit architectures. These will
+ * actually be atomic in some cases (namely Armv7 + LPAE), but for others we
+ * rely on the access being split into 2x32-bit accesses for a 32-bit quantity
+ * (e.g. a virtual address) and a strong prevailing wind.
+ */
+#define compiletime_assert_rwonce_type(t)					\
+	compiletime_assert(__native_word(t) || sizeof(t) == sizeof(long long),	\
+		"Unsupported access size for {READ,WRITE}_ONCE().")
+
+/*
+ * Use __READ_ONCE() instead of READ_ONCE() if you do not require any
+ * atomicity. Note that this may result in tears!
+ */
+#ifndef __READ_ONCE
+#define __READ_ONCE(x)	(*(const volatile __unqual_scalar_typeof(x) *)&(x))
+#endif
+
+#define READ_ONCE(x)							\
+({									\
+	compiletime_assert_rwonce_type(x);				\
+	__READ_ONCE(x);							\
+})
+
+#define __WRITE_ONCE(x, val)						\
+do {									\
+	*(volatile typeof(x) *)&(x) = (val);				\
+} while (0)
+
+#define WRITE_ONCE(x, val)						\
+do {									\
+	compiletime_assert_rwonce_type(x);				\
+	__WRITE_ONCE(x, val);						\
+} while (0)
+
 #endif /* __ATOMIC_H__ */
libxfs/libxfs_priv.h +0 -3
index e5f37df28c0..5fdfeeea060 100644
@@ -222,9 +222,6 @@ static inline bool WARN_ON(bool expr) {
 #define percpu_counter_read_positive(x)	((*x) > 0 ? (*x) : 0)
 #define percpu_counter_sum(x)		(*x)
 
-#define READ_ONCE(x)			(x)
-#define WRITE_ONCE(x, val)		((x) = (val))
-
 /*
  * get_random_u32 is used for di_gen inode allocation, it must be zero for
  * libxfs or all sorts of badness can occur!
libxfs/xfs_bmap.c +6 -0
index 76591d07766..957fc96bfa0 100644
@@ -1164,6 +1164,12 @@ xfs_iread_extents(
 		goto out;
 	}
 	ASSERT(ir.loaded == xfs_iext_count(ifp));
+	/*
+	 * Use release semantics so that we can use acquire semantics in
+	 * xfs_need_iread_extents and be guaranteed to see a valid mapping tree
+	 * after that load.
+	 */
+	smp_store_release(&ifp->if_needextents, 0);
 	return 0;
 out:
 	xfs_iext_destroy(ifp);
libxfs/xfs_inode_fork.c +15 -1
index f16b8ccc059..af00479469b 100644
@@ -224,10 +224,15 @@ xfs_iformat_data_fork(
 
 	/*
 	 * Initialize the extent count early, as the per-format routines may
-	 * depend on it.
+	 * depend on it.  Use release semantics to set needextents /after/ we
+	 * set the format. This ensures that we can use acquire semantics on
+	 * needextents in xfs_need_iread_extents() and be guaranteed to see a
+	 * valid format value after that load.
 	 */
 	ip->i_df.if_format = dip->di_format;
 	ip->i_df.if_nextents = xfs_dfork_data_extents(dip);
+	smp_store_release(&ip->i_df.if_needextents,
+			   ip->i_df.if_format == XFS_DINODE_FMT_BTREE ? 1 : 0);
 
 	switch (inode->i_mode & S_IFMT) {
 	case S_IFIFO:
@@ -280,8 +285,17 @@ xfs_ifork_init_attr(
 	enum xfs_dinode_fmt	format,
 	xfs_extnum_t		nextents)
 {
+	/*
+	 * Initialize the extent count early, as the per-format routines may
+	 * depend on it.  Use release semantics to set needextents /after/ we
+	 * set the format. This ensures that we can use acquire semantics on
+	 * needextents in xfs_need_iread_extents() and be guaranteed to see a
+	 * valid format value after that load.
+	 */
 	ip->i_af.if_format = format;
 	ip->i_af.if_nextents = nextents;
+	smp_store_release(&ip->i_af.if_needextents,
+			   ip->i_af.if_format == XFS_DINODE_FMT_BTREE ? 1 : 0);
 }
 
 void
libxfs/xfs_inode_fork.h +4 -2
index d3943d6ad0b..96d307784c8 100644
@@ -24,6 +24,7 @@ struct xfs_ifork {
 	xfs_extnum_t		if_nextents;	/* # of extents in this fork */
 	short			if_broot_bytes;	/* bytes allocated for root */
 	int8_t			if_format;	/* format of this fork */
+	uint8_t			if_needextents;	/* extents have not been read */
 };
 
 /*
@@ -260,9 +261,10 @@ int xfs_iext_count_upgrade(struct xfs_trans *tp, struct xfs_inode *ip,
 		uint nr_to_add);
 
 /* returns true if the fork has extents but they are not read in yet. */
-static inline bool xfs_need_iread_extents(struct xfs_ifork *ifp)
+static inline bool xfs_need_iread_extents(const struct xfs_ifork *ifp)
 {
-	return ifp->if_format == XFS_DINODE_FMT_BTREE && ifp->if_height == 0;
+	/* see xfs_iformat_{data,attr}_fork() for needextents semantics */
+	return smp_load_acquire(&ifp->if_needextents) != 0;
 }
 
 #endif	/* __XFS_INODE_FORK_H__ */
Reviewed-by on v2 1/1
#2 lore raw

Re: [PATCH v2] xfsprogs: _{attr,data}_map_shared should take ILOCK_EXCL until iread_extents is completely done

Quoted 165 lines · Darrick J. Wong

From: Darrick J. Wong <[email protected]>

While fuzzing the data fork extent count on a btree-format directory with xfs/375, I observed the following (excerpted) splat:

XFS: Assertion failed: xfs_isilocked(ip, XFS_ILOCK_EXCL), file: fs/xfs/libxfs/xfs_bmap.c, line: 1208

------------[ cut here ]------------
WARNING: CPU: 0 PID: 43192 at fs/xfs/xfs_message.c:104 assfail+0x46/0x4a [xfs]
Call Trace:
 <TASK>
 xfs_iread_extents+0x1af/0x210 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_dir_walk+0xb8/0x190 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_count_parent_dentries+0x41/0x80 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_validate+0x199/0x2e0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent+0xdf/0x130 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrub_metadata+0x2b8/0x730 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrubv_metadata+0x38b/0x4d0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_ioc_scrubv_metadata+0x111/0x160 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_file_ioctl+0x367/0xf50 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 __x64_sys_ioctl+0x82/0xa0
 do_syscall_64+0x2b/0x80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

The cause of this is a race condition in xfs_ilock_data_map_shared, which performs an unlocked access to the data fork to guess which lock mode it needs:

Thread 0 Thread 1

xfs_need_iread_extents
<observe no iext tree>
xfs_ilock(..., ILOCK_EXCL)
xfs_iread_extents
<observe no iext tree>
<check ILOCK_EXCL>
<load bmbt extents into iext>
<notice iext size doesn't
match nextents>
xfs_need_iread_extents <observe iext tree> xfs_ilock(..., ILOCK_SHARED)
<tear down iext tree>
xfs_iunlock(..., ILOCK_EXCL)
xfs_iread_extents <observe no iext tree> <check ILOCK_EXCL> *BOOM*

Fix this race by adding a flag to the xfs_ifork structure to indicate that we have not yet read in the extent records and changing the predicate to look at the flag state, not if_height. The memory barrier ensures that the flag will not be set until the very end of the function.

---
Here's the userspace version, which steals heavily from the kernel but otherwise uses liburcu underneath.

---
 include/atomic.h        |  100 +++++++++++++++++++++++++++++++++++++++++++++++
 libxfs/libxfs_priv.h    |    3 -
 libxfs/xfs_bmap.c       |    6 +++
 libxfs/xfs_inode_fork.c |   16 +++++++-
 libxfs/xfs_inode_fork.h |    6 ++-
 5 files changed, 125 insertions(+), 6 deletions(-)
include/atomic.h +91 -0
index 9c4aa5849ae..98889190bb0 100644
@@ -118,4 +118,104 @@ atomic64_set(atomic64_t *a, int64_t v)
 
 #endif /* HAVE_URCU_ATOMIC64 */
 
+#define __smp_mb()		cmm_smp_mb()
+
+/* from compiler_types.h */
+/*
+ * __unqual_scalar_typeof(x) - Declare an unqualified scalar type, leaving
+ *			       non-scalar types unchanged.
+ */
+/*
+ * Prefer C11 _Generic for better compile-times and simpler code. Note: 'char'
+ * is not type-compatible with 'signed char', and we define a separate case.
+ */
+#define __scalar_type_to_expr_cases(type)				\
+		unsigned type:	(unsigned type)0,			\
+		signed type:	(signed type)0
+
+#define __unqual_scalar_typeof(x) typeof(				\
+		_Generic((x),						\
+			 char:	(char)0,				\
+			 __scalar_type_to_expr_cases(char),		\
+			 __scalar_type_to_expr_cases(short),		\
+			 __scalar_type_to_expr_cases(int),		\
+			 __scalar_type_to_expr_cases(long),		\
+			 __scalar_type_to_expr_cases(long long),	\
+			 default: (x)))
+
+/* Is this type a native word size -- useful for atomic operations */
+#define __native_word(t) \
+	(sizeof(t) == sizeof(char) || sizeof(t) == sizeof(short) || \
+	 sizeof(t) == sizeof(int) || sizeof(t) == sizeof(long))
+
+#define compiletime_assert(foo, str)	BUILD_BUG_ON(!(foo))
+
+#define compiletime_assert_atomic_type(t)				\
+	compiletime_assert(__native_word(t),				\
+		"Need native word sized stores/loads for atomicity.")
+
+/* from barrier.h */
+#ifndef __smp_store_release
+#define __smp_store_release(p, v)					\
+do {									\
+	compiletime_assert_atomic_type(*p);				\
+	__smp_mb();							\
+	WRITE_ONCE(*p, v);						\
+} while (0)
+#endif
+
+#ifndef __smp_load_acquire
+#define __smp_load_acquire(p)						\
+({									\
+	__unqual_scalar_typeof(*p) ___p1 = READ_ONCE(*p);		\
+	compiletime_assert_atomic_type(*p);				\
+	__smp_mb();							\
+	(typeof(*p))___p1;						\
+})
+#endif
+
+#ifndef smp_store_release
+#define smp_store_release(p, v) __smp_store_release((p), (v))
+#endif
+
+#ifndef smp_load_acquire
+#define smp_load_acquire(p) __smp_load_acquire(p)
+#endif
+
+/* from rwonce.h */
+/*
+ * Yes, this permits 64-bit accesses on 32-bit architectures. These will
+ * actually be atomic in some cases (namely Armv7 + LPAE), but for others we
+ * rely on the access being split into 2x32-bit accesses for a 32-bit quantity
+ * (e.g. a virtual address) and a strong prevailing wind.
+ */
+#define compiletime_assert_rwonce_type(t)					\
+	compiletime_assert(__native_word(t) || sizeof(t) == sizeof(long long),	\
+		"Unsupported access size for {READ,WRITE}_ONCE().")
+
+/*
+ * Use __READ_ONCE() instead of READ_ONCE() if you do not require any
+ * atomicity. Note that this may result in tears!
+ */
+#ifndef __READ_ONCE
+#define __READ_ONCE(x)	(*(const volatile __unqual_scalar_typeof(x) *)&(x))
+#endif
+
+#define READ_ONCE(x)							\
+({									\
+	compiletime_assert_rwonce_type(x);				\
+	__READ_ONCE(x);							\
+})
+
+#define __WRITE_ONCE(x, val)						\
+do {									\

include/atomic.h atomic64_set(atomic64_t *a, int64_t v) · line 221

Quoted 10 lines
include/atomic.h +9 -0
+	*(volatile typeof(x) *)&(x) = (val);				\
+} while (0)
+
+#define WRITE_ONCE(x, val)						\
+do {									\
+	compiletime_assert_rwonce_type(x);				\
+	__WRITE_ONCE(x, val);						\
+} while (0)
+
 #endif /* __ATOMIC_H__ */

I'd put READ/WRITE_ONCE above the barrier stuff as __smp_store_release/__smp_load_acquire use it, but other than that it looks OK.

libxfs/libxfs_priv.h static inline bool WARN_ON(bool expr) { · line 226

Quoted 10 lines
libxfs/libxfs_priv.h +0 -2
index e5f37df28c0..5fdfeeea060 100644
@@ -222,9 +222,6 @@ static inline bool WARN_ON(bool expr) {
 #define percpu_counter_read_positive(x)	((*x) > 0 ? (*x) : 0)
 #define percpu_counter_sum(x)		(*x)
 
-#define READ_ONCE(x)			(x)
-#define WRITE_ONCE(x, val)		((x) = (val))

Yay! One less nasty hack for the userspace wrappers.

Overall looks good to me.

Signature · 2 lines
Dave Chinner
[email protected]
Reviewed-by on v2 1/1
#3 lore raw
In reply to Darrick J. Wong
Quoted 54 lines · Darrick J. Wong

From: Darrick J. Wong <[email protected]>

While fuzzing the data fork extent count on a btree-format directory with xfs/375, I observed the following (excerpted) splat:

XFS: Assertion failed: xfs_isilocked(ip, XFS_ILOCK_EXCL), file: fs/xfs/libxfs/xfs_bmap.c, line: 1208

------------[ cut here ]------------
WARNING: CPU: 0 PID: 43192 at fs/xfs/xfs_message.c:104 assfail+0x46/0x4a [xfs]
Call Trace:
 <TASK>
 xfs_iread_extents+0x1af/0x210 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_dir_walk+0xb8/0x190 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_count_parent_dentries+0x41/0x80 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent_validate+0x199/0x2e0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xchk_parent+0xdf/0x130 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrub_metadata+0x2b8/0x730 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_scrubv_metadata+0x38b/0x4d0 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_ioc_scrubv_metadata+0x111/0x160 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 xfs_file_ioctl+0x367/0xf50 [xfs 09f66509ece4938760fac7de64732a0cbd3e39cd]
 __x64_sys_ioctl+0x82/0xa0
 do_syscall_64+0x2b/0x80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

The cause of this is a race condition in xfs_ilock_data_map_shared, which performs an unlocked access to the data fork to guess which lock mode it needs:

Thread 0 Thread 1

xfs_need_iread_extents
<observe no iext tree>
xfs_ilock(..., ILOCK_EXCL)
xfs_iread_extents
<observe no iext tree>
<check ILOCK_EXCL>
<load bmbt extents into iext>
<notice iext size doesn't
match nextents>
xfs_need_iread_extents <observe iext tree> xfs_ilock(..., ILOCK_SHARED)
<tear down iext tree>
xfs_iunlock(..., ILOCK_EXCL)
xfs_iread_extents <observe no iext tree> <check ILOCK_EXCL> *BOOM*

Fix this race by adding a flag to the xfs_ifork structure to indicate that we have not yet read in the extent records and changing the predicate to look at the flag state, not if_height. The memory barrier ensures that the flag will not be set until the very end of the function.

Quoted 3 lines

---
v2: use smp_store_release per reviewer comments

Looks fine to me.

Signature · 2 lines
Dave Chinner
[email protected]
#4 lore raw
In reply to Darrick J. Wong
Quoted 4 lines · Darrick J. Wong
fs/xfs/libxfs/xfs_inode_fork.c +0 -0
@@ -226,10 +226,15 @@ xfs_iformat_data_fork(
 
 	/*
 	 * Initialize the extent count early, as the per-format routines may

fs/xfs/libxfs/xfs_inode_fork.c xfs_iformat_data_fork( · line 237

Quoted 10 lines
fs/xfs/libxfs/xfs_inode_fork.c +6 -1
-	 * depend on it.
+	 * depend on it.  Use release semantics to set needextents /after/ we
+	 * set the format. This ensures that we can use acquire semantics on
+	 * needextents in xfs_need_iread_extents() and be guaranteed to see a
+	 * valid format value after that load.
 	 */
 	ip->i_df.if_format = dip->di_format;
 	ip->i_df.if_nextents = xfs_dfork_data_extents(dip);
+	smp_store_release(&ip->i_df.if_needextents,
+			   ip->i_df.if_format == XFS_DINODE_FMT_BTREE ? 1 : 0);

ip->i_df is memset to zero a little earlier (same for the attr fork), this only needs to be:

	if (ip->i_df.if_format == XFS_DINODE_FMT_BTREE)
		smp_store_release(&ip->i_df.if_needextents, 1);