From: Vicki Pfau <vi@endrift.com>
To: Dmitry Torokhov <dmitry.torokhov@gmail.com>, linux-input@vger.kernel.org
Cc: Vicki Pfau <vi@endrift.com>
Subject: [PATCH 2/2] Input: uinput - Release mutex while unregistering input device
Date: Wed, 6 Dec 2023 22:34:06 -0800 [thread overview]
Message-ID: <20231207063406.556770-3-vi@endrift.com> (raw)
In-Reply-To: <20231207063406.556770-1-vi@endrift.com>
Any pending requests may be holding a mutex from its own subsystem, e.g.
evdev, while waiting to be able to claim the uinput device mutex.
However, unregistering the device may try to claim that mutex, leading
to a deadlock. To prevent this from happening, we need to temporarily
give up the lock before calling input_unregister_device.
Fixes: e8b95728f724 ("Input: uinput - avoid FF flush when destroying device")
Signed-off-by: Vicki Pfau <vi@endrift.com>
---
drivers/input/misc/uinput.c | 28 +++++++++++++++++++++++++++-
1 file changed, 27 insertions(+), 1 deletion(-)
diff --git a/drivers/input/misc/uinput.c b/drivers/input/misc/uinput.c
index 0330e72798db..ac6e5baa2093 100644
--- a/drivers/input/misc/uinput.c
+++ b/drivers/input/misc/uinput.c
@@ -296,17 +296,34 @@ static void uinput_destroy_device(struct uinput_device *udev)
udev->state = UIST_NEW_DEVICE;
if (dev) {
+ udev->dev = NULL;
name = dev->name;
phys = dev->phys;
if (old_state == UIST_CREATED) {
uinput_flush_requests(udev);
+
+ /*
+ * Any pending requests may be holding a mutex from its
+ * own subsystem, e.g. evdev, while waiting to be able
+ * to claim the uinput device mutex. However,
+ * unregistering the device may try to claim that
+ * mutex, leading to a deadlock. To prevent this from
+ * happening, we need to temporarily give up the lock.
+ *
+ * Since this function is only called immediately
+ * before the caller exits the critical section without
+ * doing any further operations on the device, this
+ * is safe and we can immediately reclaim the mutex
+ * when done so the unlock is still balanced.
+ */
+ mutex_unlock(&udev->mutex);
input_unregister_device(dev);
+ mutex_lock(&udev->mutex);
} else {
input_free_device(dev);
}
kfree(name);
kfree(phys);
- udev->dev = NULL;
}
}
@@ -745,7 +762,16 @@ static int uinput_release(struct inode *inode, struct file *file)
{
struct uinput_device *udev = file->private_data;
+ int retval;
+
+ retval = mutex_lock_interruptible(&udev->mutex);
+ if (retval)
+ return retval;
+
uinput_destroy_device(udev);
+
+ mutex_unlock(&udev->mutex);
+
kfree(udev);
return 0;
--
2.43.0
next prev parent reply other threads:[~2023-12-07 6:44 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-12-07 6:34 [PATCH 0/2] Input: uinput - Multiple concurrency fixes in ff request handling Vicki Pfau
2023-12-07 6:34 ` [PATCH 1/2] Input: uinput - Allow uinput_request_submit wait interrupting Vicki Pfau
2023-12-08 19:32 ` Dmitry Torokhov
2023-12-09 3:24 ` Vicki Pfau
2023-12-15 3:04 ` Vicki Pfau
2024-01-12 1:37 ` Vi Pfau
2024-01-22 4:03 ` Dmitry Torokhov
2025-12-10 3:37 ` Vicki Pfau
2023-12-07 6:34 ` Vicki Pfau [this message]
2023-12-08 19:58 ` [PATCH 2/2] Input: uinput - Release mutex while unregistering input device Dmitry Torokhov
2023-12-09 3:24 ` Vicki Pfau
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20231207063406.556770-3-vi@endrift.com \
--to=vi@endrift.com \
--cc=dmitry.torokhov@gmail.com \
--cc=linux-input@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).