Hi,
On Sat, Jun 24, 2017 at 09:38:48AM +0000, chenhong (N) wrote:
> Description of problem:
>
> Encounterd BUG case:
> serio: i8042 KBD port at 0x60,0x64 irq 1
> BUG: unable to handle kernel NULL pointer dereference at 0000000000000050
> IP: [<ffffffff8150feaf>] _spin_lock_irqsave+0x1f/0x40
> PGD 0
> Oops: 0002 [#1] SMP
> last sysfs file:
> CPU 0
> Modules linked in:
>
> Pid: 1, comm: swapper Not tainted 2.6.32-358.el6.x86_64 #1 QEMU Standard PC
> (i440FX + PIIX, 1996)
> RIP: 0010:[<ffffffff8150feaf>] [<ffffffff8150feaf>]
> _spin_lock_irqsave+0x1f/0x40
> RSP: 0018:ffff880028203cc0 EFLAGS: 00010082
> RAX: 0000000000010000 RBX: 0000000000000000 RCX: 0000000000000000
> RDX: 0000000000000282 RSI: 0000000000000098 RDI: 0000000000000050
> RBP: ffff880028203cc0 R08: ffff88013e79c000 R09: ffff880028203ee0
> R10: 0000000000000298 R11: 0000000000000282 R12: 0000000000000050
> R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000098
> FS: 0000000000000000(0000) GS:ffff880028200000(0000) knlGS:0000000000000000
> CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b
> CR2: 0000000000000050 CR3: 0000000001a85000 CR4: 00000000001407f0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> Process swapper (pid: 1, threadinfo ffff88013e79c000, task ffff88013e79b500)
> Stack:
> ffff880028203d00 ffffffff813de186 ffffffffffffff02 0000000000000000
> <d> 0000000000000000 0000000000000000 0000000000000000 0000000000000098
> <d> ffff880028203d70 ffffffff813e0162 ffff880028203d20 ffffffff8103b8ac
> Call Trace:
> <IRQ>
> [<ffffffff813de186>] serio_interrupt+0x36/0xa0
> [<ffffffff813e0162>] i8042_interrupt+0x132/0x3a0
> [<ffffffff8103b8ac>] ? kvm_clock_read+0x1c/0x20
> [<ffffffff8103b8b9>] ? kvm_clock_get_cycles+0x9/0x10
> [<ffffffff810e1640>] handle_IRQ_event+0x60/0x170
> [<ffffffff8103b154>] ? kvm_guest_apic_eoi_write+0x44/0x50
> [<ffffffff810e3d8e>] handle_edge_irq+0xde/0x180
> [<ffffffff8100de89>] handle_irq+0x49/0xa0
> [<ffffffff81516c8c>] do_IRQ+0x6c/0xf0
> [<ffffffff8100b9d3>] ret_from_intr+0x0/0x11
> [<ffffffff81076f63>] ? __do_softirq+0x73/0x1e0
> [<ffffffff8109b75b>] ? hrtimer_interrupt+0x14b/0x260
> [<ffffffff8100c1cc>] ? call_softirq+0x1c/0x30
> [<ffffffff8100de05>] ? do_softirq+0x65/0xa0
> [<ffffffff81076d95>] ? irq_exit+0x85/0x90
> [<ffffffff81516d80>] ? smp_apic_timer_interrupt+0x70/0x9b
> [<ffffffff8100bb93>] ? apic_timer_interrupt+0x13/0x20
> Version-Release number of selected component (if applicable):
>
> RELEASE: 2.6.32-358.el6.x86_64 (Red Hat Enterprise Linux 6.4)
> VERSION: #1 SMP Thu Feb 21 02:37:52 EST 2013
>
> The problem was also reproduced on Red Hat Enterprise Linux
> 7.3-x86_64(3.10.0-514.el7)
>
> Cause of the problem:
> static irqreturn_t i8042_interrupt(int irq, void *dev_id)
> {
> ......
> serio = port->exists ? port->serio : NULL;
> ......
> if (likely(port->exists && !filtered))
> serio_interrupt(serio, data, dfl);
> ......
> }
>
> static int i8042_start(struct serio *serio)
> {
> struct i8042_port *port = serio->port_data;
> port->exists = true;
> mb();
> return 0;
> }
>
>
> i8042_probe
> |
> i8042_setup_kbd --> request_irq(i8042_interrupt)
> |
> i8042_register_ports --> serio_queue_event(SERIO_REGISTER_PORT) -->
> serio_handle_event --> serio_add_port --> i8042_start
>
>
> i8042_start which set port->exists be true may be called during the
> i8042_interrupt function according to the analysis of initialization order.
> If port->exists is set to be true after the assignment of serio and before
> calling of serio_interrupt, a NULL pointer will be passed to serio_interrupt.
> The latest upstream kernel still has this problem.
>
Thank you for your report and the correct analysis of the problem.
>
> Solution for this problem
> Add a check for serio to prevent NULL pointer.
>
> --- a/drivers/input/serio/i8042.c 2017-06-19 16:44:50.890078100 +0800
> +++ b/drivers/input/serio/i8042.c 2017-06-19 16:45:20.230022700 +0800
> @@ -524,7 +524,7 @@ i8042_interruptspin_unlock_irqrestore(&i8042_lock, flags);
> - if (likely(port->exists && !filtered))
> + if (likely(port->exists && !filtered && serio))
I do not think we need to check port->exists here, just checking serio
should be enough.
Also, please check Documentation/SubmittingPatches when submitting
patches in the future.
I believe we want the version of the patch below.
Thanks!
--
Dmitry
Input: i8042 - fix crash at boot time
From: Dmitry Torokhov <dmitry.torok...@gmail.com>
The driver checks port->exists twice in i8042_interrupt(), first when
trying to assign temporary "serio" variable, and second time when deciding
whether it should call serio_interrupt(). The value of port->exists may
change between the 2 checks, and we may end up calling serio_interrupt()
with a NULL pointer:
BUG: unable to handle kernel NULL pointer dereference at 0000000000000050
IP: [<ffffffff8150feaf>] _spin_lock_irqsave+0x1f/0x40
PGD 0
Oops: 0002 [#1] SMP
last sysfs file:
CPU 0
Modules linked in:
Pid: 1, comm: swapper Not tainted 2.6.32-358.el6.x86_64 #1 QEMU Standard PC
(i440FX + PIIX, 1996)
RIP: 0010:[<ffffffff8150feaf>] [<ffffffff8150feaf>]
_spin_lock_irqsave+0x1f/0x40
RSP: 0018:ffff880028203cc0 EFLAGS: 00010082
RAX: 0000000000010000 RBX: 0000000000000000 RCX: 0000000000000000
RDX: 0000000000000282 RSI: 0000000000000098 RDI: 0000000000000050
RBP: ffff880028203cc0 R08: ffff88013e79c000 R09: ffff880028203ee0
R10: 0000000000000298 R11: 0000000000000282 R12: 0000000000000050
R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000098
FS: 0000000000000000(0000) GS:ffff880028200000(0000) knlGS:0000000000000000
CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b
CR2: 0000000000000050 CR3: 0000000001a85000 CR4: 00000000001407f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
Process swapper (pid: 1, threadinfo ffff88013e79c000, task ffff88013e79b500)
Stack:
ffff880028203d00 ffffffff813de186 ffffffffffffff02 0000000000000000
<d> 0000000000000000 0000000000000000 0000000000000000 0000000000000098
<d> ffff880028203d70 ffffffff813e0162 ffff880028203d20 ffffffff8103b8ac
Call Trace:
<IRQ>
[<ffffffff813de186>] serio_interrupt+0x36/0xa0
[<ffffffff813e0162>] i8042_interrupt+0x132/0x3a0
[<ffffffff8103b8ac>] ? kvm_clock_read+0x1c/0x20
[<ffffffff8103b8b9>] ? kvm_clock_get_cycles+0x9/0x10
[<ffffffff810e1640>] handle_IRQ_event+0x60/0x170
[<ffffffff8103b154>] ? kvm_guest_apic_eoi_write+0x44/0x50
[<ffffffff810e3d8e>] handle_edge_irq+0xde/0x180
[<ffffffff8100de89>] handle_irq+0x49/0xa0
[<ffffffff81516c8c>] do_IRQ+0x6c/0xf0
[<ffffffff8100b9d3>] ret_from_intr+0x0/0x11
[<ffffffff81076f63>] ? __do_softirq+0x73/0x1e0
[<ffffffff8109b75b>] ? hrtimer_interrupt+0x14b/0x260
[<ffffffff8100c1cc>] ? call_softirq+0x1c/0x30
[<ffffffff8100de05>] ? do_softirq+0x65/0xa0
[<ffffffff81076d95>] ? irq_exit+0x85/0x90
[<ffffffff81516d80>] ? smp_apic_timer_interrupt+0x70/0x9b
[<ffffffff8100bb93>] ? apic_timer_interrupt+0x13/0x20
To avoid the issue let's change the second check to test whether serio is
NULL or not.
Also, let's take i8042_lock in i8042_start() and i8042_stop() instead of
trying to be overly smart and using memory barriers and synchronize_irq().
Reported-by: Chen Hong <chenho...@huawei.com>
Signed-off-by: Dmitry Torokhov <dmitry.torok...@gmail.com>
---
drivers/input/serio/i8042.c | 16 ++++++----------
1 file changed, 6 insertions(+), 10 deletions(-)
diff --git a/drivers/input/serio/i8042.c b/drivers/input/serio/i8042.c
index c52da651269b..2f19bdbf64b5 100644
--- a/drivers/input/serio/i8042.c
+++ b/drivers/input/serio/i8042.c
@@ -436,8 +436,10 @@ static int i8042_start(struct serio *serio)
{
struct i8042_port *port = serio->port_data;
+ spin_lock_irq(&i8042_lock);
port->exists = true;
- mb();
+ spin_unlock_irq(&i8042_lock);
+
return 0;
}
@@ -450,16 +452,10 @@ static void i8042_stop(struct serio *serio)
{
struct i8042_port *port = serio->port_data;
+ spin_lock_irq(&i8042_lock);
port->exists = false;
-
- /*
- * We synchronize with both AUX and KBD IRQs because there is
- * a (very unlikely) chance that AUX IRQ is raised for KBD port
- * and vice versa.
- */
- synchronize_irq(I8042_AUX_IRQ);
- synchronize_irq(I8042_KBD_IRQ);
port->serio = NULL;
+ spin_unlock_irq(&i8042_lock);
}
/*
@@ -576,7 +572,7 @@ static irqreturn_t i8042_interrupt(int irq, void *dev_id)
spin_unlock_irqrestore(&i8042_lock, flags);
- if (likely(port->exists && !filtered))
+ if (likely(serio && !filtered))
serio_interrupt(serio, data, dfl);
out:
