thunderbolt: Fix use-after-free in remove_unplugged_switch()

PATCH In review Mika Westerberg 2 participants 3 messages Nov 19, 2020

Subsystems (2)

USB Thunderbolt driver

lore

Share & tools

lore

Mika Westerberg

author

2020-11-19 08:34 UTC #0 lore raw

PATCH v2 1/1 thunderbolt: Fix use-after-free in remove_unplugged_switch()

Paulian reported a crash that happens when a dock is unplugged during hibernation:

[78436.228217] thunderbolt 0-1: device disconnected
[78436.228365] BUG: kernel NULL pointer dereference, address: 00000000000001e0
...
[78436.228397] RIP: 0010:icm_free_unplugged_children+0x109/0x1a0
...
[78436.228432] Call Trace:
[78436.228439]  icm_rescan_work+0x24/0x30
[78436.228444]  process_one_work+0x1a3/0x3a0
[78436.228449]  worker_thread+0x30/0x370
[78436.228454]  ? process_one_work+0x3a0/0x3a0
[78436.228457]  kthread+0x13d/0x160
[78436.228461]  ? kthread_park+0x90/0x90
[78436.228465]  ret_from_fork+0x1f/0x30

This happens because remove_unplugged_switch() calls tb_switch_remove() that releases the memory pointed by sw so the following lines reference to a memory that might be released already.

Fix this by saving pointer to the parent device before calling tb_switch_remove().

Reported-by Paulian Bogdan Marinca <[email protected]>

Fixes 4f7c2e0d8765 ("thunderbolt: Make sure device runtime resume completes before taking domain lock")

Cc [email protected]

Signed-off-by Mika Westerberg <[email protected]>

---
Previous version (v1) can be found:

https://lore.kernel.org/linux-usb/[email protected]/

v2: Use get_device()/put_device() for the parent device.

 drivers/thunderbolt/icm.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

drivers/thunderbolt/icm.c +7 -3

index b51fc3f62b1f..cd985f334b9b 100644
@@ -1976,7 +1976,9 @@ static int complete_rpm(struct device *dev, void *data)
 
 static void remove_unplugged_switch(struct tb_switch *sw)
 {
-	pm_runtime_get_sync(sw->dev.parent);
+	struct device *parent = get_device(sw->dev.parent);
+
+	pm_runtime_get_sync(parent);
 
 	/*
 	 * Signal this and switches below for rpm_complete because

@@ -1987,8 +1989,10 @@ static void remove_unplugged_switch(struct tb_switch *sw)
 	bus_for_each_dev(&tb_bus_type, &sw->dev, NULL, complete_rpm);
 	tb_switch_remove(sw);
 
-	pm_runtime_mark_last_busy(sw->dev.parent);
-	pm_runtime_put_autosuspend(sw->dev.parent);
+	pm_runtime_mark_last_busy(parent);
+	pm_runtime_put_autosuspend(parent);
+
+	put_device(parent);
 }
 
 static void icm_free_unplugged_children(struct tb_switch *sw)

Signature · 1 line

2.29.2

Greg Kroah-Hartman

maintainer Reviewed-by on v2 1/1

2020-11-19 08:40 UTC #1 lore raw

Quoted 21 lines · Mika Westerberg

Paulian reported a crash that happens when a dock is unplugged during hibernation:
[78436.228217] thunderbolt 0-1: device disconnected
[78436.228365] BUG: kernel NULL pointer dereference, address: 00000000000001e0
...
[78436.228397] RIP: 0010:icm_free_unplugged_children+0x109/0x1a0
...
[78436.228432] Call Trace:
[78436.228439]  icm_rescan_work+0x24/0x30
[78436.228444]  process_one_work+0x1a3/0x3a0
[78436.228449]  worker_thread+0x30/0x370
[78436.228454]  ? process_one_work+0x3a0/0x3a0
[78436.228457]  kthread+0x13d/0x160
[78436.228461]  ? kthread_park+0x90/0x90
[78436.228465]  ret_from_fork+0x1f/0x30
This happens because remove_unplugged_switch() calls tb_switch_remove() that releases the memory pointed by sw so the following lines reference to a memory that might be released already.

Quoted 8 lines

Fix this by saving pointer to the parent device before calling tb_switch_remove().
Reported-by Paulian Bogdan Marinca <[email protected]>
Fixes 4f7c2e0d8765 ("thunderbolt: Make sure device runtime resume completes before taking domain lock")
Cc [email protected]
Signed-off-by Mika Westerberg <[email protected]>
---

Reviewed-by Greg Kroah-Hartman <[email protected]>

Mika Westerberg

author Applied

2020-11-19 14:48 UTC #2 lore raw

Quoted 25 lines · Greg Kroah-Hartman

Quoted 24 lines · Mika Westerberg
Paulian reported a crash that happens when a dock is unplugged during hibernation:
[78436.228217] thunderbolt 0-1: device disconnected
[78436.228365] BUG: kernel NULL pointer dereference, address: 00000000000001e0
...
[78436.228397] RIP: 0010:icm_free_unplugged_children+0x109/0x1a0
...
[78436.228432] Call Trace:
[78436.228439]  icm_rescan_work+0x24/0x30
[78436.228444]  process_one_work+0x1a3/0x3a0
[78436.228449]  worker_thread+0x30/0x370
[78436.228454]  ? process_one_work+0x3a0/0x3a0
[78436.228457]  kthread+0x13d/0x160
[78436.228461]  ? kthread_park+0x90/0x90
[78436.228465]  ret_from_fork+0x1f/0x30
This happens because remove_unplugged_switch() calls tb_switch_remove() that releases the memory pointed by sw so the following lines reference to a memory that might be released already.
Fix this by saving pointer to the parent device before calling tb_switch_remove().

Quoted 7 lines

Quoted 5 lines
Reported-by Paulian Bogdan Marinca <[email protected]>
Fixes 4f7c2e0d8765 ("thunderbolt: Make sure device runtime resume completes before taking domain lock")
Cc [email protected]
Signed-off-by Mika Westerberg <[email protected]>
---
Reviewed-by Greg Kroah-Hartman <[email protected]>

Thanks!

Applied to thunderbolt.git/fixes.