media: em28xx: fix memory leak

PATCH In review Muhammad Usama Anjum 3 participants 5 messages Mar 24, 2021 – Apr 6, 2021

Subsystems (3)

Stable branch Media (V4L/DVB) EM28XX VIDEO4LINUX driver

lore

Share & tools

lore

Muhammad Usama Anjum

author

2021-03-24 18:09 UTC #0 lore raw

PATCH 1/1 media: em28xx: fix memory leak

If some error occurs, URB buffers should also be freed. If they aren't freed with the dvb here, the em28xx_dvb_fini call doesn't frees the URB buffers as dvb is set to NULL. The function in which error occurs should do all the cleanup for the allocations it had done.

Tested the patch with the reproducer provided by syzbot. This patch fixes the memleak.

Reported-by [email protected]

Signed-off-by Muhammad Usama Anjum <[email protected]>

---
Resending the same path as some email addresses were missing from the earlier email.

syzbot found the following issue on:

HEAD commit:    1a4431a5 Merge tag 'afs-fixes-20210315' of git://git.kerne..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=11013a7cd00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ff6b8b2e9d5a1227
dashboard link: https://syzkaller.appspot.com/bug?extid=889397c820fa56adf25d
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1559ae3ad00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176985c6d00000

 drivers/media/usb/em28xx/em28xx-dvb.c | 1 +
 1 file changed, 1 insertion(+)

drivers/media/usb/em28xx/em28xx-dvb.c +1 -0

index 526424279637..471bd74667e3 100644
@@ -2010,6 +2010,7 @@ static int em28xx_dvb_init(struct em28xx *dev)
 	return result;
 
 out_free:
+	em28xx_uninit_usb_xfer(dev, EM28XX_DIGITAL_MODE);
 	kfree(dvb);
 	dev->dvb = NULL;
 	goto ret;

Signature · 1 line

2.25.1

Muhammad Usama Anjum

author

2021-03-31 08:23 UTC #1 lore raw

Quoted 32 lines · Muhammad Usama Anjum

If some error occurs, URB buffers should also be freed. If they aren't freed with the dvb here, the em28xx_dvb_fini call doesn't frees the URB buffers as dvb is set to NULL. The function in which error occurs should do all the cleanup for the allocations it had done.
Tested the patch with the reproducer provided by syzbot. This patch fixes the memleak.
Reported-by [email protected]
Signed-off-by Muhammad Usama Anjum <[email protected]>
---
Resending the same path as some email addresses were missing from the earlier email.
syzbot found the following issue on:
HEAD commit:    1a4431a5 Merge tag 'afs-fixes-20210315' of git://git.kerne..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=11013a7cd00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ff6b8b2e9d5a1227
dashboard link: https://syzkaller.appspot.com/bug?extid=889397c820fa56adf25d
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1559ae3ad00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176985c6d00000
 drivers/media/usb/em28xx/em28xx-dvb.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/usb/em28xx/em28xx-dvb.c +0 -0
index 526424279637..471bd74667e3 100644

drivers/media/usb/em28xx/em28xx-dvb.c static int em28xx_dvb_init(struct em28xx *dev) · line 2016

Quoted 7 lines

drivers/media/usb/em28xx/em28xx-dvb.c +1 -0

@@ -2010,6 +2010,7 @@ static int em28xx_dvb_init(struct em28xx *dev)
 	return result;
 
 out_free:
+	em28xx_uninit_usb_xfer(dev, EM28XX_DIGITAL_MODE);
 	kfree(dvb);
 	dev->dvb = NULL;
 	goto ret;

I'd received the following notice and waiting for the review:

Quoted 10 lines · Patchwork

Hello,
The following patch (submitted by you) has been updated in Patchwork:
* linux-media: media: em28xx: fix memory leak
- http://patchwork.linuxtv.org/project/linux-media/patch/20210324180753.GA410359@LEGION/
- for: Linux Media kernel patches
was: New
now: Under Review

Quoted 4 lines

This email is a notification only - you do not need to respond.
Happy patchworking.

Thanks,
Usama

Dan Carpenter

2021-03-31 08:52 UTC #2 lore raw

Quoted 35 lines · Muhammad Usama Anjum

Quoted 34 lines · Muhammad Usama Anjum
If some error occurs, URB buffers should also be freed. If they aren't freed with the dvb here, the em28xx_dvb_fini call doesn't frees the URB buffers as dvb is set to NULL. The function in which error occurs should do all the cleanup for the allocations it had done.
Tested the patch with the reproducer provided by syzbot. This patch fixes the memleak.
Reported-by [email protected]
Signed-off-by Muhammad Usama Anjum <[email protected]>
---
Resending the same path as some email addresses were missing from the earlier email.
syzbot found the following issue on:
HEAD commit:    1a4431a5 Merge tag 'afs-fixes-20210315' of git://git.kerne..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=11013a7cd00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ff6b8b2e9d5a1227
dashboard link: https://syzkaller.appspot.com/bug?extid=889397c820fa56adf25d
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1559ae3ad00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176985c6d00000
 drivers/media/usb/em28xx/em28xx-dvb.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/usb/em28xx/em28xx-dvb.c +0 -0
index 526424279637..471bd74667e3 100644
@@ -2010,6 +2010,7 @@ static int em28xx_dvb_init(struct em28xx *dev)
 	return result;
 

Quoted 7 lines

drivers/media/usb/em28xx/em28xx-dvb.c static int em28xx_dvb_init(struct em28xx *dev) · line 2016
Quoted 5 lines
drivers/media/usb/em28xx/em28xx-dvb.c +1 -0
 out_free:
+	em28xx_uninit_usb_xfer(dev, EM28XX_DIGITAL_MODE);
 	kfree(dvb);
 	dev->dvb = NULL;
 	goto ret;
I'd received the following notice and waiting for the review:

Please wait a minimum of two weeks before asking for updates.

regards,
dan carpenter

Muhammad Usama Anjum

author

2021-04-06 09:44 UTC #3 lore raw

In reply to Muhammad Usama Anjum

Quoted 35 lines · Muhammad Usama Anjum

Quoted 34 lines · Muhammad Usama Anjum
If some error occurs, URB buffers should also be freed. If they aren't freed with the dvb here, the em28xx_dvb_fini call doesn't frees the URB buffers as dvb is set to NULL. The function in which error occurs should do all the cleanup for the allocations it had done.
Tested the patch with the reproducer provided by syzbot. This patch fixes the memleak.
Reported-by [email protected]
Signed-off-by Muhammad Usama Anjum <[email protected]>
---
Resending the same path as some email addresses were missing from the earlier email.
syzbot found the following issue on:
HEAD commit:    1a4431a5 Merge tag 'afs-fixes-20210315' of git://git.kerne..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=11013a7cd00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ff6b8b2e9d5a1227
dashboard link: https://syzkaller.appspot.com/bug?extid=889397c820fa56adf25d
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1559ae3ad00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176985c6d00000
 drivers/media/usb/em28xx/em28xx-dvb.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/usb/em28xx/em28xx-dvb.c +0 -0
index 526424279637..471bd74667e3 100644
@@ -2010,6 +2010,7 @@ static int em28xx_dvb_init(struct em28xx *dev)
 	return result;
 

Quoted 16 lines

drivers/media/usb/em28xx/em28xx-dvb.c static int em28xx_dvb_init(struct em28xx *dev) · line 2016
Quoted 5 lines
drivers/media/usb/em28xx/em28xx-dvb.c +1 -0
 out_free:
+	em28xx_uninit_usb_xfer(dev, EM28XX_DIGITAL_MODE);
 	kfree(dvb);
 	dev->dvb = NULL;
 	goto ret;
I'd received the following notice and waiting for the review:
Quoted 8 lines · Patchwork
Hello,
The following patch (submitted by you) has been updated in Patchwork:
* linux-media: media: em28xx: fix memory leak
- http://patchwork.linuxtv.org/project/linux-media/patch/20210324180753.GA410359@LEGION/
- for: Linux Media kernel patches

This patch has been accepted. This bug was introduced by 27ba0dac. Will it be backported and submitted for inclusion in stable release by maintainer automatically?

Quoted 6 lines

Quoted 1 line
Thanks,
Usama

Hans Verkuil

2021-04-06 10:07 UTC #4 lore raw

Quoted 48 lines · Muhammad Usama Anjum

Quoted 47 lines · Muhammad Usama Anjum
Quoted 32 lines · Muhammad Usama Anjum
If some error occurs, URB buffers should also be freed. If they aren't freed with the dvb here, the em28xx_dvb_fini call doesn't frees the URB buffers as dvb is set to NULL. The function in which error occurs should do all the cleanup for the allocations it had done.
Tested the patch with the reproducer provided by syzbot. This patch fixes the memleak.
Reported-by [email protected]
Signed-off-by Muhammad Usama Anjum <[email protected]>
---
Resending the same path as some email addresses were missing from the earlier email.
syzbot found the following issue on:
HEAD commit:    1a4431a5 Merge tag 'afs-fixes-20210315' of git://git.kerne..
git tree:       upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=11013a7cd00000
kernel config:  https://syzkaller.appspot.com/x/.config?x=ff6b8b2e9d5a1227
dashboard link: https://syzkaller.appspot.com/bug?extid=889397c820fa56adf25d
syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=1559ae3ad00000
C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=176985c6d00000
 drivers/media/usb/em28xx/em28xx-dvb.c | 1 +
 1 file changed, 1 insertion(+)
drivers/media/usb/em28xx/em28xx-dvb.c +0 -0
index 526424279637..471bd74667e3 100644
drivers/media/usb/em28xx/em28xx-dvb.c static int em28xx_dvb_init(struct em28xx *dev) · line 2016
Quoted 7 lines
drivers/media/usb/em28xx/em28xx-dvb.c +1 -0
@@ -2010,6 +2010,7 @@ static int em28xx_dvb_init(struct em28xx *dev)
 	return result;
 
 out_free:
+	em28xx_uninit_usb_xfer(dev, EM28XX_DIGITAL_MODE);
 	kfree(dvb);
 	dev->dvb = NULL;
 	goto ret;
I'd received the following notice and waiting for the review:
Quoted 4 lines · Patchwork
Hello,
The following patch (submitted by you) has been updated in Patchwork:

Quoted 7 lines

Quoted 4 lines
Quoted 4 lines
* linux-media: media: em28xx: fix memory leak
- http://patchwork.linuxtv.org/project/linux-media/patch/20210324180753.GA410359@LEGION/
- for: Linux Media kernel patches
This patch has been accepted. This bug was introduced by 27ba0dac. Will it be backported and submitted for inclusion in stable release by maintainer automatically?

That might not happen since there was no 'Fixes:' tag. Without that it will depend on the stable tree maintainers whether they'll pick it up or not.

Regards,

Hans