Re: [syzbot] [usb?] memory leak in class_create

From: Alan Stern
Date: Mon May 01 2023 - 15:16:55 EST


On Mon, May 01, 2023 at 09:53:45AM -0700, syzbot wrote:
> Hello,
>
> syzbot found the following issue on:
>
> HEAD commit: 22b8cc3e78f5 Merge tag 'x86_mm_for_6.4' of git://git.kerne..
> git tree: upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=16fc7958280000
> kernel config: https://syzkaller.appspot.com/x/.config?x=5046ebeca744dd40
> dashboard link: https://syzkaller.appspot.com/bug?extid=e7afd76ad060fa0d2605
> compiler: gcc (Debian 10.2.1-6) 10.2.1 20210110, GNU ld (GNU Binutils for Debian) 2.35.2
> syz repro: https://syzkaller.appspot.com/x/repro.syz?x=1599a2b4280000
> C reproducer: https://syzkaller.appspot.com/x/repro.c?x=14eb395fc80000
>
> Downloadable assets:
> disk image: https://storage.googleapis.com/syzbot-assets/3ad2088c196b/disk-22b8cc3e.raw.xz
> vmlinux: https://storage.googleapis.com/syzbot-assets/61919a5b89c6/vmlinux-22b8cc3e.xz
> kernel image: https://storage.googleapis.com/syzbot-assets/a7adb5503ac8/bzImage-22b8cc3e.xz
>
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+e7afd76ad060fa0d2605@xxxxxxxxxxxxxxxxxxxxxxxxx
>
> BUG: memory leak
> unreferenced object 0xffff88810af67080 (size 96):
> comm "kworker/0:2", pid 4402, jiffies 4294950769 (age 14.190s)
> hex dump (first 32 bytes):
> bf 03 9b 85 ff ff ff ff 00 00 00 00 00 00 00 00 ................
> 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
> backtrace:
> [<ffffffff81544e14>] kmalloc_trace+0x24/0x90 mm/slab_common.c:1057
> [<ffffffff82b65445>] kmalloc include/linux/slab.h:559 [inline]
> [<ffffffff82b65445>] kzalloc include/linux/slab.h:680 [inline]
> [<ffffffff82b65445>] class_create+0x25/0x90 drivers/base/class.c:261
> [<ffffffff831f7a80>] init_usb_class drivers/usb/core/file.c:91 [inline]
> [<ffffffff831f7a80>] usb_register_dev+0x290/0x3d0 drivers/usb/core/file.c:179
> [<ffffffff832cffc4>] usblp_probe+0x4e4/0x750 drivers/usb/class/usblp.c:1208
> [<ffffffff831f39a9>] usb_probe_interface+0x179/0x3c0 drivers/usb/core/driver.c:396
> [<ffffffff82b62d7d>] call_driver_probe drivers/base/dd.c:579 [inline]
> [<ffffffff82b62d7d>] really_probe+0x12d/0x430 drivers/base/dd.c:658
> [<ffffffff82b63141>] __driver_probe_device+0xc1/0x1a0 drivers/base/dd.c:800
> [<ffffffff82b6324a>] driver_probe_device+0x2a/0x120 drivers/base/dd.c:830
> [<ffffffff82b6343b>] __device_attach_driver+0xfb/0x150 drivers/base/dd.c:958
> [<ffffffff82b60191>] bus_for_each_drv+0xc1/0x110 drivers/base/bus.c:457
> [<ffffffff82b63962>] __device_attach+0x102/0x2a0 drivers/base/dd.c:1030
> [<ffffffff82b618fa>] bus_probe_device+0xca/0xd0 drivers/base/bus.c:532
> [<ffffffff82b5def3>] device_add+0x993/0xc60 drivers/base/core.c:3625
> [<ffffffff831f0a89>] usb_set_configuration+0x9a9/0xc90 drivers/usb/core/message.c:2211
> [<ffffffff832033a1>] usb_generic_driver_probe+0xa1/0x100 drivers/usb/core/generic.c:238
> [<ffffffff831f3080>] usb_probe_device+0x60/0x140 drivers/usb/core/driver.c:293

There is definitely a memory leak in usb_register_dev()'s error pathways
-- it doesn't call destroy_usb_class() -- but I don't think that is the
cause of this bug. Let's try some diagnostics.

Alan Stern

#syz test: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/ 22b8cc3e78f5

Index: usb-devel/drivers/usb/core/file.c
===================================================================
--- usb-devel.orig/drivers/usb/core/file.c
+++ usb-devel/drivers/usb/core/file.c
@@ -209,6 +209,8 @@ int usb_register_dev(struct usb_interfac
retval = PTR_ERR(intf->usb_dev);
}
up_write(&minor_rwsem);
+ dev_info(&intf->dev, "Post class create: refcount %d\n",
+ atomic_read(&usb_class->kref.refcount.refs));
return retval;
}
EXPORT_SYMBOL_GPL(usb_register_dev);
@@ -242,6 +244,8 @@ void usb_deregister_dev(struct usb_inter

intf->usb_dev = NULL;
intf->minor = -1;
+ dev_info(&intf->dev, "Pre class destroy: refcount %d\n",
+ atomic_read(&usb_class->kref.refcount.refs));
destroy_usb_class();
}
EXPORT_SYMBOL_GPL(usb_deregister_dev);