diff mbox series

[RFC] locking/ww_mutex: Adjust to lockdep nest_lock requirements

Message ID 20230911090729.5287-1-thomas.hellstrom@linux.intel.com (mailing list archive)
State New, archived
Headers show
Series [RFC] locking/ww_mutex: Adjust to lockdep nest_lock requirements | expand

Commit Message

Thomas Hellstrom Sept. 11, 2023, 9:07 a.m. UTC
When using mutex_acquire_nest() with a nest_lock, lockdep refcounts the
number of acquired lockdep_maps of mutexes of the same class, and also
keeps a pointer to the first acquired lockdep_map of a class. That pointer
is then used for various comparison-, printing- and checking purposes,
but there is no mechanism to actively ensure that lockdep_map stays in
memory. Instead, a warning is printed if the lockdep_map is freed and
there are still held locks of the same lock class, even if the lockdep_map
itself has been released.

In the context of WW/WD transactions that means that if a user unlocks
and frees a ww_mutex from within an ongoing ww transaction, and that
mutex happens to be the first ww_mutex grabbed in the transaction,
such a warning is printed and there might be a risk of a UAF.

Note that this is only problem when lockdep is enabled and affects only
dereferences of struct lockdep_map.

Adjust to this by adding a fake lockdep_map to the acquired context and
make sure it is the first acquired lockdep map of the associated
ww_mutex class. Then hold it for the duration of the WW/WD transaction.

Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Will Deacon <will@kernel.org>
Cc: Waiman Long <longman@redhat.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Daniel Vetter <daniel.vetter@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: intel-xe@lists.freedesktop.org
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 include/linux/ww_mutex.h | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

Comments

kernel test robot Sept. 12, 2023, 5:32 a.m. UTC | #1
Hello,

kernel test robot noticed "WARNING:possible_recursive_locking_detected" on:

commit: bb043828b2d487832c946751ffcc4ebd80d2a624 ("[RFC PATCH] locking/ww_mutex: Adjust to lockdep nest_lock requirements")
url: https://github.com/intel-lab-lkp/linux/commits/Thomas-Hellstr-m/locking-ww_mutex-Adjust-to-lockdep-nest_lock-requirements/20230911-170838
base: git://anongit.freedesktop.org/drm/drm-misc drm-misc-next
patch link: https://lore.kernel.org/all/20230911090729.5287-1-thomas.hellstrom@linux.intel.com/
patch subject: [RFC PATCH] locking/ww_mutex: Adjust to lockdep nest_lock requirements

in testcase: boot

compiler: clang-16
test machine: qemu-system-x86_64 -enable-kvm -cpu SandyBridge -smp 2 -m 16G

(please refer to attached dmesg/kmsg for entire log/backtrace)



If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <oliver.sang@intel.com>
| Closes: https://lore.kernel.org/oe-lkp/202309121305.81d349e5-oliver.sang@intel.com


[  252.054712][    T1] WARNING: possible recursive locking detected
[  252.055056][    T1] 6.5.0-rc2-00590-gbb043828b2d4 #5 Not tainted
[  252.055056][    T1] --------------------------------------------
[  252.055056][    T1] swapper/1 is trying to acquire lock:
[ 252.055056][ T1] ffffc9000001f990 (ww_class_mutex){+.+.}-{0:0}, at: test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:77 kbuild/src/rand/kernel/locking/test-ww_mutex.c:632) 
[  252.055056][    T1]
[  252.055056][    T1] but task is already holding lock:
[ 252.055056][ T1] ffffc9000001f800 (ww_class_mutex){+.+.}-{0:0}, at: do_one_initcall (kbuild/src/rand/init/main.c:1232) 
[  252.055056][    T1]
[  252.055056][    T1] other info that might help us debug this:
[  252.055056][    T1]  Possible unsafe locking scenario:
[  252.055056][    T1]
[  252.055056][    T1]        CPU0
[  252.055056][    T1]        ----
[  252.055056][    T1]   lock(ww_class_mutex);
[  252.055056][    T1]
[  252.055056][    T1]  *** DEADLOCK ***
[  252.055056][    T1]
[  252.055056][    T1]  May be due to missing lock nesting notation
[  252.055056][    T1]
[  252.055056][    T1] 2 locks held by swapper/1:
[ 252.055056][ T1] #0: ffffc9000001f7d8 (ww_class_acquire){+.+.}-{0:0}, at: do_one_initcall (kbuild/src/rand/init/main.c:1232) 
[ 252.055056][ T1] #1: ffffc9000001f800 (ww_class_mutex){+.+.}-{0:0}, at: do_one_initcall (kbuild/src/rand/init/main.c:1232) 
[  252.055056][    T1]
[  252.055056][    T1] stack backtrace:
[  252.055056][    T1] CPU: 0 PID: 1 Comm: swapper Not tainted 6.5.0-rc2-00590-gbb043828b2d4 #5 9492b9dd86bdf82b2f3deb22196638d07aad10f2
[  252.055056][    T1] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.16.2-debian-1.16.2-1 04/01/2014
[  252.055056][    T1] Call Trace:
[  252.055056][    T1]  <TASK>
[ 252.055056][ T1] dump_stack_lvl (kbuild/src/rand/lib/dump_stack.c:?) 
[ 252.055056][ T1] __lock_acquire (kbuild/src/rand/kernel/locking/lockdep.c:?) 
[ 252.055056][ T1] ? finish_lock_switch (kbuild/src/rand/arch/x86/include/asm/irqflags.h:42 kbuild/src/rand/arch/x86/include/asm/irqflags.h:77 kbuild/src/rand/kernel/sched/sched.h:1378 kbuild/src/rand/kernel/sched/core.c:5133) 
[ 252.055056][ T1] ? __schedule (kbuild/src/rand/kernel/sched/core.c:6718) 
[ 252.055056][ T1] lock_acquire (kbuild/src/rand/kernel/locking/lockdep.c:5761) 
[ 252.055056][ T1] ? test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:77 kbuild/src/rand/kernel/locking/test-ww_mutex.c:632) 
[ 252.055056][ T1] __mutex_lock_common (kbuild/src/rand/kernel/locking/mutex.c:603) 
[ 252.055056][ T1] ? test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:77 kbuild/src/rand/kernel/locking/test-ww_mutex.c:632) 
[ 252.055056][ T1] ? mark_lock (kbuild/src/rand/arch/x86/include/asm/bitops.h:228 kbuild/src/rand/arch/x86/include/asm/bitops.h:240 kbuild/src/rand/include/asm-generic/bitops/instrumented-non-atomic.h:142 kbuild/src/rand/kernel/locking/lockdep.c:228 kbuild/src/rand/kernel/locking/lockdep.c:4663) 
[ 252.055056][ T1] ? test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:77 kbuild/src/rand/kernel/locking/test-ww_mutex.c:632) 
[ 252.055056][ T1] ww_mutex_lock (kbuild/src/rand/kernel/locking/mutex.c:754 kbuild/src/rand/kernel/locking/mutex.c:871) 
[ 252.055056][ T1] test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:77 kbuild/src/rand/kernel/locking/test-ww_mutex.c:632) 
[ 252.055056][ T1] ? __asan_memcpy (kbuild/src/rand/mm/kasan/shadow.c:105) 
[ 252.055056][ T1] ? do_one_initcall (kbuild/src/rand/init/main.c:1232) 
[ 252.055056][ T1] ? __pfx_test_mutex_work (kbuild/src/rand/kernel/locking/test-ww_mutex.c:41) 
[ 252.055056][ T1] ? __pfx_test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:622) 
[ 252.055056][ T1] do_one_initcall (kbuild/src/rand/init/main.c:1232) 
[ 252.055056][ T1] ? __pfx_test_ww_mutex_init (kbuild/src/rand/kernel/locking/test-ww_mutex.c:622) 
[ 252.055056][ T1] do_initcall_level (kbuild/src/rand/init/main.c:1293) 
[ 252.055056][ T1] do_initcalls (kbuild/src/rand/init/main.c:1307) 
[ 252.055056][ T1] kernel_init_freeable (kbuild/src/rand/init/main.c:1550) 
[ 252.055056][ T1] ? __pfx_kernel_init (kbuild/src/rand/init/main.c:1429) 
[ 252.055056][ T1] kernel_init (kbuild/src/rand/init/main.c:1439) 
[ 252.055056][ T1] ? __pfx_kernel_init (kbuild/src/rand/init/main.c:1429) 
[ 252.055056][ T1] ret_from_fork (kbuild/src/rand/arch/x86/kernel/process.c:151) 
[ 252.055056][ T1] ? __pfx_kernel_init (kbuild/src/rand/init/main.c:1429) 
[ 252.055056][ T1] ret_from_fork_asm (kbuild/src/rand/arch/x86/entry/entry_64.S:298) 
[  252.055056][    T1] RIP: 0000:0x0
[ 252.055056][ T1] Code: Unable to access opcode bytes at 0xffffffffffffffd6.

Code starting with the faulting instruction
===========================================
[  252.055056][    T1] RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX: 0000000000000000
[  252.055056][    T1] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
[  252.055056][    T1] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
[  252.055056][    T1] RBP: 0000000000000000 R08: 0000000000000000 R09: 0000000000000000
[  252.055056][    T1] R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000000000
[  252.055056][    T1] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[  252.055056][    T1]  </TASK>
[  258.685819][    T1] All ww mutex selftests passed


The kernel config and materials to reproduce are available at:
https://download.01.org/0day-ci/archive/20230912/202309121305.81d349e5-oliver.sang@intel.com
diff mbox series

Patch

diff --git a/include/linux/ww_mutex.h b/include/linux/ww_mutex.h
index bb763085479a..a401a2f31a77 100644
--- a/include/linux/ww_mutex.h
+++ b/include/linux/ww_mutex.h
@@ -65,6 +65,16 @@  struct ww_acquire_ctx {
 #endif
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 	struct lockdep_map dep_map;
+	/**
+	 * @first_lock_dep_map: fake lockdep_map for first locked ww_mutex.
+	 *
+	 * lockdep requires the lockdep_map for the first locked ww_mutex
+	 * in a ww transaction to remain in memory until all ww_mutexes of
+	 * the transaction have been unlocked. Ensure this by keeping a
+	 * fake locked ww_mutex lockdep map between ww_acquire_init() and
+	 * ww_acquire_fini().
+	 */
+	struct lockdep_map first_lock_dep_map;
 #endif
 #ifdef CONFIG_DEBUG_WW_MUTEX_SLOWPATH
 	unsigned int deadlock_inject_interval;
@@ -146,7 +156,10 @@  static inline void ww_acquire_init(struct ww_acquire_ctx *ctx,
 	debug_check_no_locks_freed((void *)ctx, sizeof(*ctx));
 	lockdep_init_map(&ctx->dep_map, ww_class->acquire_name,
 			 &ww_class->acquire_key, 0);
+	lockdep_init_map(&ctx->first_lock_dep_map, ww_class->mutex_name,
+			 &ww_class->mutex_key, 0);
 	mutex_acquire(&ctx->dep_map, 0, 0, _RET_IP_);
+	mutex_acquire_nest(&ctx->first_lock_dep_map, 0, 0, &ctx->dep_map, _RET_IP_);
 #endif
 #ifdef CONFIG_DEBUG_WW_MUTEX_SLOWPATH
 	ctx->deadlock_inject_interval = 1;
@@ -185,6 +198,7 @@  static inline void ww_acquire_done(struct ww_acquire_ctx *ctx)
 static inline void ww_acquire_fini(struct ww_acquire_ctx *ctx)
 {
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
+	mutex_release(&ctx->first_lock_dep_map, _THIS_IP_);
 	mutex_release(&ctx->dep_map, _THIS_IP_);
 #endif
 #ifdef DEBUG_WW_MUTEXES