media: rc: redrat3: KASAN use-after-free in rc_dev_uevent after probe failure
Hi,
I am reporting a KASAN slab-use-after-free triggered by a syzkaller USB reproducer on Linux 7.2. The issue is reproducible with HEAD commit 45c13f3f9e3bb15fd89ff2864c6f627a3b4b4229.
The reproducer connects a crafted USB device that is bound to the RedRat3 infrared receiver driver. During probe, the device reports invalid endpoint properties and then fails several hardware communication steps. The driver eventually fails probe with -EIO.
After the failed probe, systemd-journal reads the device uevent file. This reaches rc_dev_uevent(), then add_uevent_var(), vsnprintf(), and string(). KASAN reports a slab-use-after-free while string() reads a pointer that was allocated and freed in redrat3_dev_probe().
This looks like a lifetime issue in the RedRat3 probe failure path. The driver appears to free an object or string while it can still be referenced from the rc/input/device uevent path. A failed probe should make sure that no sysfs/uevent-visible field points to memory that has already been freed.
Reproducer:
syz reproducer:
syz_usb_connect(0x0, 0x39, &(0x7f0000000100)=ANY=[@ANYBLOB="12010002000000402a1101000001010203010902270001010080fa0904000003ffffff00070581024000000705820240000007050302400000"], 0x0)
console output: https://pastebin.com/raw/M2wZXb12
kernel config: https://pastebin.com/raw/vDCnLVLv
Kernel:
HEAD commit: 45c13f3f9e3bb15fd89ff2864c6f627a3b4b4229 git tree: torvalds/linux
kernel version: 7.2.0-14827-g45c13f3f9e3b #1 PREEMPT(full)
Relevant log:
usb 5-1: New USB device found, idVendor=112a, idProduct=0001, bcdDevice= 1.00 usb 5-1: config 1 interface 0 altsetting 0 bulk endpoint 0x81 has invalid maxpacket 64 usb 5-1: config 1 interface 0 altsetting 0 bulk endpoint 0x82 has invalid maxpacket 64 usb 5-1: config 1 interface 0 altsetting 0 bulk endpoint 0x3 has invalid maxpacket 64 redrat3 5-1:1.0: Problem fetching firmware ID
redrat3 5-1:1.0: Failed to read timeout from hardware
redrat3 5-1:1.0: redrat3_send_cmd: Error sending rr3 cmd res -71, data 0 redrat3 5-1:1.0: redrat3_enable_detector: detector status: -5, should be 1 redrat3 5-1:1.0: probe with driver redrat3 failed with error -5
==================================================================
BUG: KASAN: slab-use-after-free in string+0x231/0x2b0 Read of size 1 at addr ffff88804c7fb758 by task systemd-journal/5039 CPU: 1 UID: 0 PID: 5039 Comm: systemd-journal Tainted: G W Hardware name: QEMU Standard PC (Q35 + ICH9, 2009) Call Trace: <TASK> dump_stack_lvl+0x10e/0x190 print_report+0x153/0x7a0 kasan_report+0x147/0x180 string+0x231/0x2b0 vsnprintf+0x704/0xea0 add_uevent_var+0x1b2/0x430 rc_dev_uevent+0x15d/0x1a0 dev_uevent+0x51e/0x860 uevent_show+0x19f/0x320 dev_attr_show+0x58/0xc0 sysfs_kf_seq_show+0x331/0x480 seq_read_iter+0x475/0xe20 vfs_read+0x59d/0xa70 ksys_read+0x14a/0x250 do_syscall_64+0xfa/0x480 entry_SYSCALL_64_after_hwframe+0x76/0x7e </TASK> Allocated by task 11585: __kmalloc_cache_noprof+0x32e/0x630 redrat3_dev_probe+0x47a/0x2700 usb_probe_interface+0x71e/0xe00 really_probe+0x267/0xb20 __driver_probe_device+0x1f9/0x420 driver_probe_device+0x4f/0x240 __device_attach_driver+0x286/0x430 bus_for_each_drv+0x251/0x2e0 __device_attach+0x2b3/0x440 device_initial_probe+0xa1/0xd0 bus_probe_device+0x12a/0x220 device_add+0x7ec/0xb90 usb_new_device+0xbc6/0x1ae0 hub_event+0x2c04/0x4ef0 Freed by task 11585: kfree+0x19a/0x660 redrat3_dev_probe+0x1590/0x2700 usb_probe_interface+0x71e/0xe00 really_probe+0x267/0xb20 __driver_probe_device+0x1f9/0x420 driver_probe_device+0x4f/0x240 __device_attach_driver+0x286/0x430 bus_for_each_drv+0x251/0x2e0 __device_attach+0x2b3/0x440 device_initial_probe+0xa1/0xd0 bus_probe_device+0x12a/0x220 device_add+0x7ec/0xb90 usb_new_device+0xbc6/0x1ae0 hub_event+0x2c04/0x4ef0 The buggy address belongs to the object at ffff88804c7fb000
which belongs to the cache kmalloc-2k of size 2048 The buggy address is located 1880 bytes inside of freed 2048-byte region [ffff88804c7fb000, ffff88804c7fb800)
Quoted 18 lines · CJ Liu
Hi,
I am reporting a KASAN slab-use-after-free triggered by a syzkaller USB reproducer on Linux 7.2. The issue is reproducible with HEAD commit
45c13f3f9e3bb15fd89ff2864c6f627a3b4b4229.The reproducer connects a crafted USB device that is bound to the RedRat3 infrared receiver driver. During probe, the device reports invalid endpoint properties and then fails several hardware communication steps. The driver eventually fails probe with
-EIO.After the failed probe, systemd-journal reads the device uevent file. This reaches
rc_dev_uevent(), thenadd_uevent_var(),vsnprintf(), andstring(). KASAN reports a slab-use-after-free whilestring()reads a pointer that was allocated and freed inredrat3_dev_probe().
Quoted 4 lines
This looks like a lifetime issue in the RedRat3 probe failure path. The driver appears to free an object or string while it can still be referenced from the rc/input/device uevent path. A failed probe should make sure that no sysfs/uevent-visible field points to memory that has already been freed.
I am fairly sure this issue is fixed by this commit:
It was only merged a few hours ago.
Thank you for reporting.
Sean