linux-kernel - Re: [syzbot] KASAN: slab-out-of-bounds Read in do_wait_for

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20210720221445.7d022a9e@gmail.com>
Date:   Tue, 20 Jul 2021 22:14:45 +0300
From:   Pavel Skripkin <paskripkin@...il.com>
To:     syzbot <syzbot+cc699626e48a6ebaf295@...kaller.appspotmail.com>
Cc:     Larry.Finger@...inger.net, florian.c.schilhabel@...glemail.com,
        gregkh@...uxfoundation.org, hridayhegde1999@...il.com,
        linux-kernel@...r.kernel.org, linux-staging@...ts.linux.dev,
        rkovhaev@...il.com, straube.linux@...il.com,
        syzkaller-bugs@...glegroups.com
Subject: Re: [syzbot] KASAN: slab-out-of-bounds Read in do_wait_for_common

On Tue, 20 Jul 2021 08:21:06 -0700
syzbot <syzbot+cc699626e48a6ebaf295@...kaller.appspotmail.com> wrote:

> Hello,
> 
> syzbot has tested the proposed patch but the reproducer is still
> triggering an issue: possible deadlock in flush_workqueue
> 
> usb 6-1: r8712u: Loading firmware from "rtlwifi/rtl8712u.bin"
> usb 6-1: Direct firmware load for rtlwifi/rtl8712u.bin failed with
> error -2 usb 6-1: Falling back to sysfs fallback for:
> rtlwifi/rtl8712u.bin usb 6-1: r8712u: Firmware request failed
> ============================================
> WARNING: possible recursive locking detected
> 5.14.0-rc2-syzkaller #0 Not tainted
> --------------------------------------------
> kworker/0:3/3159 is trying to acquire lock:
> ffff888011064d38 ((wq_completion)events){+.+.}-{0:0}, at:
> flush_workqueue+0x15c/0x1750 kernel/workqueue.c:2787
> 
> but task is already holding lock:
> ffff888011064d38 ((wq_completion)events){+.+.}-{0:0}, at:
> process_one_work+0x7aa/0x10c0 kernel/workqueue.c:2249
> 
> other info that might help us debug this:
>  Possible unsafe locking scenario:
> 
>        CPU0
>        ----
>   lock((wq_completion)events);
>   lock((wq_completion)events);
> 
>  *** DEADLOCK ***
> 
>  May be due to missing lock nesting notation
> 
> 5 locks held by kworker/0:3/3159:
>  #0: ffff888011064d38 ((wq_completion)events){+.+.}-{0:0}, at:
> process_one_work+0x7aa/0x10c0 kernel/workqueue.c:2249 #1:
> ffffc900021d7d20 ((work_completion)(&fw_work->work)){+.+.}-{0:0}, at:
> process_one_work+0x7e8/0x10c0 kernel/workqueue.c:2251 #2:
> ffff8881467d4220 (&dev->mutex){....}-{3:3}, at: device_lock
> include/linux/device.h:754 [inline] #2: ffff8881467d4220
> (&dev->mutex){....}-{3:3}, at: rtl871x_load_fw_fail
> drivers/staging/rtl8712/hal_init.c:43 [inline] #2: ffff8881467d4220
> (&dev->mutex){....}-{3:3}, at: rtl871x_load_fw_cb+0x102/0x130
> drivers/staging/rtl8712/hal_init.c:56 #3: ffff8880363da220
> (&dev->mutex){....}-{3:3}, at: device_lock include/linux/device.h:754
> [inline] #3: ffff8880363da220 (&dev->mutex){....}-{3:3}, at:
> __device_driver_lock drivers/base/dd.c:1028 [inline] #3:
> ffff8880363da220 (&dev->mutex){....}-{3:3}, at:
> device_release_driver_internal+0xc1/0x7b0 drivers/base/dd.c:1229 #4:
> ffff8880308211a8 (&dev->mutex){....}-{3:3}, at: device_lock
> include/linux/device.h:754 [inline] #4: ffff8880308211a8
> (&dev->mutex){....}-{3:3}, at: __device_driver_lock
> drivers/base/dd.c:1028 [inline] #4: ffff8880308211a8
> (&dev->mutex){....}-{3:3}, at:
> device_release_driver_internal+0xc1/0x7b0 drivers/base/dd.c:1229
> 
> stack backtrace:
> CPU: 0 PID: 3159 Comm: kworker/0:3 Not tainted 5.14.0-rc2-syzkaller #0
> Hardware name: Google Google Compute Engine/Google Compute Engine,
> BIOS Google 01/01/2011 Workqueue: events request_firmware_work_func
> Call Trace:
>  __dump_stack lib/dump_stack.c:88 [inline]
>  dump_stack_lvl+0x1ae/0x29f lib/dump_stack.c:105
>  __lock_acquire+0x2615/0x6100 kernel/locking/lockdep.c:4853
>  lock_acquire+0x182/0x4a0 kernel/locking/lockdep.c:5625
>  flush_workqueue+0x178/0x1750 kernel/workqueue.c:2787
>  flush_scheduled_work include/linux/workqueue.h:597 [inline]
>  r871xu_dev_remove+0x159/0x420 drivers/staging/rtl8712/usb_intf.c:604
>  usb_unbind_interface+0x1f2/0x860 drivers/usb/core/driver.c:458
>  __device_release_driver drivers/base/dd.c:1201 [inline]
>  device_release_driver_internal+0x51e/0x7b0 drivers/base/dd.c:1232
>  bus_remove_device+0x2fd/0x410 drivers/base/bus.c:529
>  device_del+0x6e1/0xc10 drivers/base/core.c:3540
>  usb_disable_device+0x407/0x800 drivers/usb/core/message.c:1419
>  usb_set_configuration+0x42b/0x2100 drivers/usb/core/message.c:2027
>  usb_unbind_device+0x6b/0x170 drivers/usb/core/driver.c:309
>  __device_release_driver drivers/base/dd.c:1201 [inline]
>  device_release_driver_internal+0x51e/0x7b0 drivers/base/dd.c:1232
>  rtl871x_load_fw_fail drivers/staging/rtl8712/hal_init.c:45 [inline]
>  rtl871x_load_fw_cb+0x10a/0x130 drivers/staging/rtl8712/hal_init.c:56
>  request_firmware_work_func+0x175/0x250
> drivers/base/firmware_loader/main.c:1081
> process_one_work+0x833/0x10c0 kernel/workqueue.c:2276
> worker_thread+0xac1/0x1320 kernel/workqueue.c:2422
> kthread+0x453/0x480 kernel/kthread.c:319 ret_from_fork+0x1f/0x30
> arch/x86/entry/entry_64.S:295
> 


Hmm, did't notice this could happen. From my point of view,
flush_scheduled_work() can be replaced with flush_work() for all
scheduled works in this driver.



#syz test
git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master



 
With regards,
Pavel Skripkin

View attachment "0001-staging-rtl8712-rewrite-error-handling.patch" of type "text/x-patch" (6387 bytes)