* [PATCH] uio: Fix use-after-free in uio_open
@ 2023-12-21 4:26 Guanghui Feng
2023-12-21 8:08 ` Greg KH
2023-12-21 9:57 ` [PATCH v2] " Guanghui Feng
0 siblings, 2 replies; 3+ messages in thread
From: Guanghui Feng @ 2023-12-21 4:26 UTC (permalink / raw)
To: gregkh, linux-kernel; +Cc: baolin.wang
core-1 core-2
-------------------------------------------------------------
uio_unregister_device uio_open
idev = idr_find()
device_unregister(&idev->dev)
put_device(&idev->dev)
uio_device_release
get_device(&idev->dev)
kfree(idev)
uio_free_minor(minor)
uio_release
put_device(&idev->dev)
kfree(idev)
-------------------------------------------------------------
In the core-1 uio_unregister_device(), the device_unregister will kfree idev
when the idev->dev kobject ref is one. But after core-1 device_unregister,
put_device and before doing kfree, the core-2 may get_device. Then:
1. After core-1 kfree idev, the core-2 will do use-after-free for idev.
2. When core-2 do uio_release and put_device, the idev will be double freed.
To address this issue, we can get idev atomic & inc idev reference with minor_lock.
Signed-off-by: Guanghui Feng <guanghuifeng@linux.alibaba.com>
---
drivers/uio/uio.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/uio/uio.c b/drivers/uio/uio.c
index 62082d6..2d572f6 100644
--- a/drivers/uio/uio.c
+++ b/drivers/uio/uio.c
@@ -466,13 +466,13 @@ static int uio_open(struct inode *inode, struct file *filep)
mutex_lock(&minor_lock);
idev = idr_find(&uio_idr, iminor(inode));
- mutex_unlock(&minor_lock);
if (!idev) {
ret = -ENODEV;
+ mutex_unlock(&minor_lock);
goto out;
}
-
get_device(&idev->dev);
+ mutex_unlock(&minor_lock);
if (!try_module_get(idev->owner)) {
ret = -ENODEV;
@@ -1064,9 +1064,8 @@ void uio_unregister_device(struct uio_info *info)
wake_up_interruptible(&idev->wait);
kill_fasync(&idev->async_queue, SIGIO, POLL_HUP);
- device_unregister(&idev->dev);
-
uio_free_minor(minor);
+ device_unregister(&idev->dev);
return;
}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] uio: Fix use-after-free in uio_open
2023-12-21 4:26 [PATCH] uio: Fix use-after-free in uio_open Guanghui Feng
@ 2023-12-21 8:08 ` Greg KH
2023-12-21 9:57 ` [PATCH v2] " Guanghui Feng
1 sibling, 0 replies; 3+ messages in thread
From: Greg KH @ 2023-12-21 8:08 UTC (permalink / raw)
To: Guanghui Feng; +Cc: linux-kernel, baolin.wang
On Thu, Dec 21, 2023 at 12:26:48PM +0800, Guanghui Feng wrote:
> core-1 core-2
> -------------------------------------------------------------
> uio_unregister_device uio_open
> idev = idr_find()
> device_unregister(&idev->dev)
> put_device(&idev->dev)
> uio_device_release
> get_device(&idev->dev)
> kfree(idev)
> uio_free_minor(minor)
> uio_release
> put_device(&idev->dev)
> kfree(idev)
> -------------------------------------------------------------
>
> In the core-1 uio_unregister_device(), the device_unregister will kfree idev
> when the idev->dev kobject ref is one. But after core-1 device_unregister,
> put_device and before doing kfree, the core-2 may get_device. Then:
> 1. After core-1 kfree idev, the core-2 will do use-after-free for idev.
> 2. When core-2 do uio_release and put_device, the idev will be double freed.
>
> To address this issue, we can get idev atomic & inc idev reference with minor_lock.
Nit, can you wrap your lines at 72 columns?
Anyway, nice fix, how did you find this?
> Signed-off-by: Guanghui Feng <guanghuifeng@linux.alibaba.com>
What commit id does this fix?
thanks,
greg k-h
^ permalink raw reply [flat|nested] 3+ messages in thread
* [PATCH v2] uio: Fix use-after-free in uio_open
2023-12-21 4:26 [PATCH] uio: Fix use-after-free in uio_open Guanghui Feng
2023-12-21 8:08 ` Greg KH
@ 2023-12-21 9:57 ` Guanghui Feng
1 sibling, 0 replies; 3+ messages in thread
From: Guanghui Feng @ 2023-12-21 9:57 UTC (permalink / raw)
To: gregkh; +Cc: linux-kernel, baolin.wang, alikernel-developer
core-1 core-2
-------------------------------------------------------
uio_unregister_device uio_open
idev = idr_find()
device_unregister(&idev->dev)
put_device(&idev->dev)
uio_device_release
get_device(&idev->dev)
kfree(idev)
uio_free_minor(minor)
uio_release
put_device(&idev->dev)
kfree(idev)
-------------------------------------------------------
In the core-1 uio_unregister_device(), the device_unregister will kfree
idev when the idev->dev kobject ref is 1. But after core-1
device_unregister, put_device and before doing kfree, the core-2 may
get_device. Then:
1. After core-1 kfree idev, the core-2 will do use-after-free for idev.
2. When core-2 do uio_release and put_device, the idev will be double
freed.
To address this issue, we can get idev atomic & inc idev reference with
minor_lock.
Fixes: 57c5f4df0a5a ("uio: fix crash after the device is unregistered")
Signed-off-by: Guanghui Feng <guanghuifeng@linux.alibaba.com>
Reviewed-by: Baolin Wang <baolin.wang@linux.alibaba.com>
---
drivers/uio/uio.c | 7 +++----
1 file changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/uio/uio.c b/drivers/uio/uio.c
index 62082d6..2d572f6 100644
--- a/drivers/uio/uio.c
+++ b/drivers/uio/uio.c
@@ -466,13 +466,13 @@ static int uio_open(struct inode *inode, struct file *filep)
mutex_lock(&minor_lock);
idev = idr_find(&uio_idr, iminor(inode));
- mutex_unlock(&minor_lock);
if (!idev) {
ret = -ENODEV;
+ mutex_unlock(&minor_lock);
goto out;
}
-
get_device(&idev->dev);
+ mutex_unlock(&minor_lock);
if (!try_module_get(idev->owner)) {
ret = -ENODEV;
@@ -1064,9 +1064,8 @@ void uio_unregister_device(struct uio_info *info)
wake_up_interruptible(&idev->wait);
kill_fasync(&idev->async_queue, SIGIO, POLL_HUP);
- device_unregister(&idev->dev);
-
uio_free_minor(minor);
+ device_unregister(&idev->dev);
return;
}
--
1.8.3.1
^ permalink raw reply related [flat|nested] 3+ messages in thread
end of thread, other threads:[~2023-12-21 9:58 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-12-21 4:26 [PATCH] uio: Fix use-after-free in uio_open Guanghui Feng
2023-12-21 8:08 ` Greg KH
2023-12-21 9:57 ` [PATCH v2] " Guanghui Feng
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.