Re: [PATCH 2/2] tty: fix deadlock caused by calling printk() under tty_port->lock

From: Qi Zheng
Date: Wed May 11 2022 - 02:11:51 EST



On 2022/5/11 1:57 PM, Jiri Slaby wrote:
Hi,

LGTM:
Acked-by: Jiri Slaby <jirislaby@xxxxxxxxxx>

Thanks.


Minor comments below.

On 10. 05. 22, 13:38, Qi Zheng wrote:
The pty_write() invokes kmalloc() which may invoke a normal printk() to
print failure message. This can cause a deadlock in the scenario reported
by syz-bot below:

        CPU0              CPU1                    CPU2
        ----              ----                    ----
                          lock(console_owner);
                                                  lock(&port_lock_key);
   lock(&port->lock);
                          lock(&port_lock_key);
                                                  lock(&port->lock);
   lock(console_owner);

As commit dbdda842fe96 ("printk: Add console owner and waiter logic to
load balance console writes") said, such deadlock can be prevented by
using printk_deferred() in kmalloc() (which is invoked in the section
guarded by the port->lock). But there are too many printk() on the
kmalloc() path, and kmalloc() can be called from anywhere, so changing
printk() to printk_deferred() is too complicated and inelegant.

Therefore, this patch chooses to specify __GFP_NOWARN to kmalloc(), so
that printk() will not be called, and this deadlock problem can be avoided.

Syz-bot reported the following lockdep error:

======================================================
WARNING: possible circular locking dependency detected
5.4.143-00237-g08ccc19a-dirty #10 Not tainted
------------------------------------------------------
syz-executor.4/29420 is trying to acquire lock:
ffffffff8aedb2a0 (console_owner){....}-{0:0}, at: console_trylock_spinning kernel/printk/printk.c:1752 [inline]
ffffffff8aedb2a0 (console_owner){....}-{0:0}, at: vprintk_emit+0x2ca/0x470 kernel/printk/printk.c:2023

but task is already holding lock:
ffff8880119c9158 (&port->lock){-.-.}-{2:2}, at: pty_write+0xf4/0x1f0 drivers/tty/pty.c:120

which lock already depends on the new lock.

the existing dependency chain (in reverse order) is:

Maybe trim the stack traces a bit, the commit message is unnecessarily long...

Agree, will do.


-> #2 (&port->lock){-.-.}-{2:2}:
        __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
        _raw_spin_lock_irqsave+0x35/0x50 kernel/locking/spinlock.c:159
        tty_port_tty_get drivers/tty/tty_port.c:288 [inline]                  <-- lock(&port->lock);
        tty_port_default_wakeup+0x1d/0xb0 drivers/tty/tty_port.c:47
        serial8250_tx_chars+0x530/0xa80 drivers/tty/serial/8250/8250_port.c:1767
        serial8250_handle_irq.part.0+0x31f/0x3d0 drivers/tty/serial/8250/8250_port.c:1854
        serial8250_handle_irq drivers/tty/serial/8250/8250_port.c:1827 [inline]     <-- lock(&port_lock_key);
        serial8250_default_handle_irq+0xb2/0x220 drivers/tty/serial/8250/8250_port.c:1870
        serial8250_interrupt+0xfd/0x200 drivers/tty/serial/8250/8250_core.c:126
        __handle_irq_event_percpu+0x109/0xa50 kernel/irq/handle.c:156

Stop this trace here and trim the rest?

Will do.


        handle_irq_event_percpu+0x76/0x170 kernel/irq/handle.c:196
        handle_irq_event+0xa1/0x130 kernel/irq/handle.c:213
        handle_edge_irq+0x261/0xd00 kernel/irq/chip.c:833
        generic_handle_irq_desc include/linux/irqdesc.h:156 [inline]
        do_IRQ+0xf2/0x2e0 arch/x86/kernel/irq.c:250
        ret_from_intr+0x0/0x19
        native_safe_halt arch/x86/include/asm/irqflags.h:60 [inline]
        arch_safe_halt arch/x86/include/asm/irqflags.h:103 [inline]
        default_idle+0x2c/0x1a0 arch/x86/kernel/process.c:572
        cpuidle_idle_call kernel/sched/idle.c:184 [inline]
        do_idle+0x44c/0x590 kernel/sched/idle.c:294
        cpu_startup_entry+0x14/0x20 kernel/sched/idle.c:386
        start_secondary+0x2d1/0x3e0 arch/x86/kernel/smpboot.c:264
        secondary_startup_64+0xa4/0xb0 arch/x86/kernel/head_64.S:241

-> #1 (&port_lock_key){-.-.}-{2:2}:
        __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
        _raw_spin_lock_irqsave+0x35/0x50 kernel/locking/spinlock.c:159
        serial8250_console_write+0x184/0xa40 drivers/tty/serial/8250/8250_port.c:3198
                                        <-- lock(&port_lock_key);
        call_console_drivers kernel/printk/printk.c:1819 [inline]
        console_unlock+0x8cb/0xd00 kernel/printk/printk.c:2504
        vprintk_emit+0x1b5/0x470 kernel/printk/printk.c:2024            <-- lock(console_owner);
        vprintk_func+0x8d/0x250 kernel/printk/printk_safe.c:394
        printk+0xba/0xed kernel/printk/printk.c:2084
        register_console+0x8b3/0xc10 kernel/printk/printk.c:2829
        univ8250_console_init+0x3a/0x46 drivers/tty/serial/8250/8250_core.c:681
        console_init+0x49d/0x6d3 kernel/printk/printk.c:2915
        start_kernel+0x5e9/0x879 init/main.c:713
        secondary_startup_64+0xa4/0xb0 arch/x86/kernel/head_64.S:241

-> #0 (console_owner){....}-{0:0}:
        check_prev_add kernel/locking/lockdep.c:2600 [inline]
        check_prevs_add kernel/locking/lockdep.c:2705 [inline]
        validate_chain kernel/locking/lockdep.c:3095 [inline]
        __lock_acquire+0x27e6/0x4cc0 kernel/locking/lockdep.c:4200

Delete the above 4 lines?

Will do.


        lock_acquire+0x127/0x340 kernel/locking/lockdep.c:4734
        console_trylock_spinning kernel/printk/printk.c:1773 [inline]        <-- lock(console_owner);
        vprintk_emit+0x307/0x470 kernel/printk/printk.c:2023
        vprintk_func+0x8d/0x250 kernel/printk/printk_safe.c:394
        printk+0xba/0xed kernel/printk/printk.c:2084
        fail_dump lib/fault-inject.c:45 [inline]
        should_fail+0x67b/0x7c0 lib/fault-inject.c:144
        __should_failslab+0x152/0x1c0 mm/failslab.c:33
        should_failslab+0x5/0x10 mm/slab_common.c:1224
        slab_pre_alloc_hook mm/slab.h:468 [inline]
        slab_alloc_node mm/slub.c:2723 [inline]
        slab_alloc mm/slub.c:2807 [inline]
        __kmalloc+0x72/0x300 mm/slub.c:3871
        kmalloc include/linux/slab.h:582 [inline]
        tty_buffer_alloc+0x23f/0x2a0 drivers/tty/tty_buffer.c:175
        __tty_buffer_request_room+0x156/0x2a0 drivers/tty/tty_buffer.c:273
        tty_insert_flip_string_fixed_flag+0x93/0x250 drivers/tty/tty_buffer.c:318
        tty_insert_flip_string include/linux/tty_flip.h:37 [inline]
        pty_write+0x126/0x1f0 drivers/tty/pty.c:122                <-- lock(&port->lock);
        n_tty_write+0xa7a/0xfc0 drivers/tty/n_tty.c:2356
        do_tty_write drivers/tty/tty_io.c:961 [inline]
        tty_write+0x512/0x930 drivers/tty/tty_io.c:1045
        __vfs_write+0x76/0x100 fs/read_write.c:494

And stop here?

Will do.


        vfs_write+0x268/0x5c0 fs/read_write.c:558
        ksys_write+0x12d/0x250 fs/read_write.c:611
        do_syscall_64+0xd7/0x380 arch/x86/entry/common.c:291
        entry_SYSCALL_64_after_hwframe+0x49/0xbe

other info that might help us debug this:

Chain exists of:
   console_owner --> &port_lock_key --> &port->lock

Fixes: b6da31b2c07c ("tty: Fix data race in tty_insert_flip_string_fixed_flag")
Signed-off-by: Qi Zheng <zhengqi.arch@xxxxxxxxxxxxx>
---
  drivers/tty/tty_buffer.c | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/tty/tty_buffer.c b/drivers/tty/tty_buffer.c
index 646510476c30..bfa431a8e690 100644
--- a/drivers/tty/tty_buffer.c
+++ b/drivers/tty/tty_buffer.c
@@ -175,7 +175,8 @@ static struct tty_buffer *tty_buffer_alloc(struct tty_port *port, size_t size)
       */
      if (atomic_read(&port->buf.mem_used) > port->buf.mem_limit)
          return NULL;
-    p = kmalloc(sizeof(struct tty_buffer) + 2 * size, GFP_ATOMIC);
+    p = kmalloc(sizeof(struct tty_buffer) + 2 * size,
+            GFP_ATOMIC | __GFP_NOWARN);
      if (p == NULL)
          return NULL;



--
Thanks,
Qi