Skip to contentSkip to conversation

usb: usbip: fix a refcount leak in stub_probe()

PATCH Superseded Hangyu Hua 2 participants 7 messages Apr 7, 2022 – Apr 12, 2022

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Hangyu Hua

author
#0 lore raw

PATCH v2 1/1 usb: usbip: fix a refcount leak in stub_probe()

usb_get_dev() is called in stub_device_alloc(). When stub_probe() fails after that, usb_put_dev() needs to be called to release the reference.

Fix this by moving usb_put_dev() to sdev_free error path handling.

Fixes 3ff67445750a ("usbip: fix error handling in stub_probe()")

---

v2: add more description of this patch.

 drivers/usb/usbip/stub_dev.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/usb/usbip/stub_dev.c +1 -1
index d8d3892e5a69..3c6d452e3bf4 100644
@@ -393,7 +393,6 @@ static int stub_probe(struct usb_device *udev)
 
 err_port:
 	dev_set_drvdata(&udev->dev, NULL);
-	usb_put_dev(udev);
 
 	/* we already have busid_priv, just lock busid_lock */
 	spin_lock(&busid_priv->busid_lock);
@@ -408,6 +407,7 @@ static int stub_probe(struct usb_device *udev)
 	put_busid_priv(busid_priv);
 
 sdev_free:
+	usb_put_dev(udev);
 	stub_device_free(sdev);
 
 	return rc;
Signature · 1 line
2.25.1
maintainer
#1 lore raw
Quoted 11 lines · Hangyu Hua

usb_get_dev() is called in stub_device_alloc(). When stub_probe() fails after that, usb_put_dev() needs to be called to release the reference.

Fix this by moving usb_put_dev() to sdev_free error path handling.

Fixes 3ff67445750a ("usbip: fix error handling in stub_probe()")

---

v2: add more description of this patch.

Still missing details on how this problem was found. Please add them to the commit log.

thanks,
-- Shuah

author
#2 lore raw

Hi Shuah,

I find this by code review. Do i really need to add this to commit log? This look like a little weird.

Thanks

Quoted 18 lines · Shuah Khan
Quoted 11 lines · Hangyu Hua

usb_get_dev() is called in stub_device_alloc(). When stub_probe() fails after that, usb_put_dev() needs to be called to release the reference.

Fix this by moving usb_put_dev() to sdev_free error path handling.

Fixes 3ff67445750a ("usbip: fix error handling in stub_probe()")

---

v2: add more description of this patch.

Still missing details on how this problem was found. Please add them to the commit log.

thanks,
-- Shuah

maintainer
#3 lore raw
Quoted 4 lines · Hangyu Hua

Hi Shuah,

I find this by code review. Do i really need to add this to commit log? This look like a little weird.

Great. Good find.

It is important to understand how the problem is found. Please add it the change log. We usually expect dmesg or such info. that revealed refcount leak, since this one is found during code review, we would like to see that information the commit log.

Also please remember to avoid top posting.

Quoted 13 lines · Shuah Khan
Quoted 11 lines · Hangyu Hua
Quoted 10 lines

usb_get_dev() is called in stub_device_alloc(). When stub_probe() fails after that, usb_put_dev() needs to be called to release the reference.

Fix this by moving usb_put_dev() to sdev_free error path handling.

Fixes 3ff67445750a ("usbip: fix error handling in stub_probe()")

---

Quoted 5 lines
Quoted 5 lines
Quoted 1 line

v2: add more description of this patch.

Still missing details on how this problem was found. Please add them to the commit log.

thanks,
-- Shuah

author
#4 lore raw
Quoted 6 lines · Shuah Khan
Quoted 4 lines · Hangyu Hua

Hi Shuah,

I find this by code review. Do i really need to add this to commit log? This look like a little weird.

Quoted 8 lines

Great. Good find.

It is important to understand how the problem is found. Please add it the change log. We usually expect dmesg or such info. that revealed refcount leak, since this one is found during code review, we would like to see that information the commit log.

Also please remember to avoid top posting.

I get what you meant now. But i don't know how to get a clear dmesg or any other log. The kernel will not crash because of this. I just used gdb to find that udev->dev->kobj->kref gets bigger and bigger whenever I call stub_probe with busid_priv->status = STUB_BUSID_REMOV.

Thanks for telling me the rules.

Quoted 23 lines
Quoted 14 lines · Shuah Khan
Quoted 11 lines · Hangyu Hua
Quoted 10 lines

usb_get_dev() is called in stub_device_alloc(). When stub_probe() fails after that, usb_put_dev() needs to be called to release the reference.

Fix this by moving usb_put_dev() to sdev_free error path handling.

Fixes 3ff67445750a ("usbip: fix error handling in stub_probe()")

---

Quoted 5 lines
Quoted 5 lines
Quoted 1 line

v2: add more description of this patch.

Still missing details on how this problem was found. Please add them to the commit log.

thanks,
-- Shuah

maintainer
#5 lore raw
Quoted 14 lines · Hangyu Hua
Quoted 13 lines · Shuah Khan
Quoted 4 lines · Hangyu Hua

Hi Shuah,

I find this by code review. Do i really need to add this to commit log? This look like a little weird.

Great. Good find.

It is important to understand how the problem is found. Please add it the change log. We usually expect dmesg or such info. that revealed refcount leak, since this one is found during code review, we would like to see that information the commit log.

Quoted 6 lines
Quoted 1 line

Also please remember to avoid top posting.

I get what you meant now. But i don't know how to get a clear dmesg or any other log. The kernel will not crash because of this. I just used gdb to find that udev->dev->kobj->kref gets bigger and bigger whenever I call stub_probe with busid_priv->status = STUB_BUSID_REMOV.

Thanks for telling me the rules.

There is no need to gather dmesg etc. Just add a note that you found the problem during code review. Having complete information about why a change is made will be helpful for future changes to this code and somebody new trying understand the changes made to this file/routine and why.

thanks,
-- Shuah

author
#6 lore raw
Quoted 22 lines · Shuah Khan
Quoted 14 lines · Hangyu Hua
Quoted 13 lines · Shuah Khan
Quoted 4 lines · Hangyu Hua

Hi Shuah,

I find this by code review. Do i really need to add this to commit log? This look like a little weird.

Great. Good find.

It is important to understand how the problem is found. Please add it the change log. We usually expect dmesg or such info. that revealed refcount leak, since this one is found during code review, we would like to see that information the commit log.

Quoted 6 lines
Quoted 1 line

Also please remember to avoid top posting.

I get what you meant now. But i don't know how to get a clear dmesg or any other log. The kernel will not crash because of this. I just used gdb to find that udev->dev->kobj->kref gets bigger and bigger whenever I call stub_probe with busid_priv->status = STUB_BUSID_REMOV.

Thanks for telling me the rules.

Quoted 8 lines

There is no need to gather dmesg etc. Just add a note that you found the problem during code review. Having complete information about why a change is made will be helpful for future changes to this code and somebody new trying understand the changes made to this file/routine and why.

thanks,
-- Shuah

I get it. I will submit a v3.

Thanks.