usb/hid: slab-out-of-bounds read in usbhid_parse

541 views
Skip to first unread message

Andrey Konovalov

unread,
Sep 18, 2017, 1:33:48 PM9/18/17
to Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Hi!

I've got the following crash while fuzzing the kernel with syzkaller.

On commit ebb2c2437d8008d46796902ff390653822af6cc4 (Sep 18).

It seems that there's no proper check on the hdesc->bNumDescriptors
value in usbhid_parse(). it iterates over hdesc->desc and accesses
hdesc->desc[n] fields, which might be out-of-bounds.

==================================================================
BUG: KASAN: slab-out-of-bounds in usbhid_parse+0x9b1/0xa20
Read of size 1 at addr ffff88006c5f8edf by task kworker/1:2/1261

CPU: 1 PID: 1261 Comm: kworker/1:2 Not tainted
4.14.0-rc1-42251-gebb2c2437d80 #169
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
Workqueue: usb_hub_wq hub_event
Call Trace:
__dump_stack lib/dump_stack.c:16
dump_stack+0x292/0x395 lib/dump_stack.c:52
print_address_description+0x78/0x280 mm/kasan/report.c:252
kasan_report_error mm/kasan/report.c:351
kasan_report+0x22f/0x340 mm/kasan/report.c:409
__asan_report_load1_noabort+0x19/0x20 mm/kasan/report.c:427
usbhid_parse+0x9b1/0xa20 drivers/hid/usbhid/hid-core.c:1004
hid_add_device+0x16b/0xb30 drivers/hid/hid-core.c:2944
usbhid_probe+0xc28/0x1100 drivers/hid/usbhid/hid-core.c:1369
usb_probe_interface+0x35d/0x8e0 drivers/usb/core/driver.c:361
really_probe drivers/base/dd.c:413
driver_probe_device+0x610/0xa00 drivers/base/dd.c:557
__device_attach_driver+0x230/0x290 drivers/base/dd.c:653
bus_for_each_drv+0x161/0x210 drivers/base/bus.c:463
__device_attach+0x26e/0x3d0 drivers/base/dd.c:710
device_initial_probe+0x1f/0x30 drivers/base/dd.c:757
bus_probe_device+0x1eb/0x290 drivers/base/bus.c:523
device_add+0xd0b/0x1660 drivers/base/core.c:1835
usb_set_configuration+0x104e/0x1870 drivers/usb/core/message.c:1932
generic_probe+0x73/0xe0 drivers/usb/core/generic.c:174
usb_probe_device+0xaf/0xe0 drivers/usb/core/driver.c:266
really_probe drivers/base/dd.c:413
driver_probe_device+0x610/0xa00 drivers/base/dd.c:557
__device_attach_driver+0x230/0x290 drivers/base/dd.c:653
bus_for_each_drv+0x161/0x210 drivers/base/bus.c:463
__device_attach+0x26e/0x3d0 drivers/base/dd.c:710
device_initial_probe+0x1f/0x30 drivers/base/dd.c:757
bus_probe_device+0x1eb/0x290 drivers/base/bus.c:523
device_add+0xd0b/0x1660 drivers/base/core.c:1835
usb_new_device+0x7b8/0x1020 drivers/usb/core/hub.c:2457
hub_port_connect drivers/usb/core/hub.c:4903
hub_port_connect_change drivers/usb/core/hub.c:5009
port_event drivers/usb/core/hub.c:5115
hub_event+0x194d/0x3740 drivers/usb/core/hub.c:5195
process_one_work+0xc7f/0x1db0 kernel/workqueue.c:2119
worker_thread+0x221/0x1850 kernel/workqueue.c:2253
kthread+0x3a1/0x470 kernel/kthread.c:231
ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:431

Allocated by task 1261:
save_stack_trace+0x1b/0x20 arch/x86/kernel/stacktrace.c:59
save_stack+0x43/0xd0 mm/kasan/kasan.c:447
set_track mm/kasan/kasan.c:459
kasan_kmalloc+0xad/0xe0 mm/kasan/kasan.c:551
__kmalloc+0x14e/0x310 mm/slub.c:3782
kmalloc ./include/linux/slab.h:498
usb_get_configuration+0x372/0x2a60 drivers/usb/core/config.c:848
usb_enumerate_device drivers/usb/core/hub.c:2290
usb_new_device+0xaae/0x1020 drivers/usb/core/hub.c:2426
hub_port_connect drivers/usb/core/hub.c:4903
hub_port_connect_change drivers/usb/core/hub.c:5009
port_event drivers/usb/core/hub.c:5115
hub_event+0x194d/0x3740 drivers/usb/core/hub.c:5195
process_one_work+0xc7f/0x1db0 kernel/workqueue.c:2119
worker_thread+0x221/0x1850 kernel/workqueue.c:2253
kthread+0x3a1/0x470 kernel/kthread.c:231
ret_from_fork+0x2a/0x40 arch/x86/entry/entry_64.S:431

Freed by task 2927:
save_stack_trace+0x1b/0x20 arch/x86/kernel/stacktrace.c:59
save_stack+0x43/0xd0 mm/kasan/kasan.c:447
set_track mm/kasan/kasan.c:459
kasan_slab_free+0x72/0xc0 mm/kasan/kasan.c:524
slab_free_hook mm/slub.c:1390
slab_free_freelist_hook mm/slub.c:1412
slab_free mm/slub.c:2988
kfree+0xf6/0x2f0 mm/slub.c:3919
free_rb_tree_fname+0x8a/0xe0 fs/ext4/dir.c:402
ext4_htree_free_dir_info fs/ext4/dir.c:424
ext4_release_dir+0x49/0x70 fs/ext4/dir.c:622
__fput+0x33e/0x800 fs/file_table.c:210
____fput+0x1a/0x20 fs/file_table.c:244
task_work_run+0x1af/0x280 kernel/task_work.c:112
tracehook_notify_resume ./include/linux/tracehook.h:191
exit_to_usermode_loop+0x1e1/0x220 arch/x86/entry/common.c:162
prepare_exit_to_usermode arch/x86/entry/common.c:197
syscall_return_slowpath+0x414/0x480 arch/x86/entry/common.c:266
entry_SYSCALL_64_fastpath+0xc0/0xc2 arch/x86/entry/entry_64.S:238

The buggy address belongs to the object at ffff88006c5f8ea0
which belongs to the cache kmalloc-64 of size 64
The buggy address is located 63 bytes inside of
64-byte region [ffff88006c5f8ea0, ffff88006c5f8ee0)
The buggy address belongs to the page:
page:ffffea0001b17e00 count:1 mapcount:0 mapping: (null) index:0x0
flags: 0x100000000000100(slab)
raw: 0100000000000100 0000000000000000 0000000000000000 00000001002a002a
raw: ffffea0001a83000 0000001500000015 ffff88006c403800 0000000000000000
page dumped because: kasan: bad access detected

Memory state around the buggy address:
ffff88006c5f8d80: fb fb fb fb fb fb fb fb fc fc fc fc 00 00 00 00
ffff88006c5f8e00: 00 fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
>ffff88006c5f8e80: fc fc fc fc 00 00 00 00 00 00 00 07 fc fc fc fc
^
ffff88006c5f8f00: fb fb fb fb fb fb fb fb fc fc fc fc fb fb fb fb
ffff88006c5f8f80: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
==================================================================

Kim Jaejoong

unread,
Sep 19, 2017, 7:47:36 AM9/19/17
to Andrey Konovalov, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Hi, Andrey Konovalov

Thanks for the report.

2017-09-19 2:33 GMT+09:00 Andrey Konovalov <andre...@google.com>:
> Hi!
>
> I've got the following crash while fuzzing the kernel with syzkaller.
>
> On commit ebb2c2437d8008d46796902ff390653822af6cc4 (Sep 18).
>
> It seems that there's no proper check on the hdesc->bNumDescriptors
> value in usbhid_parse(). it iterates over hdesc->desc and accesses
> hdesc->desc[n] fields, which might be out-of-bounds.

The bNumDescriptors in hid descriptor means 'numeric expression
specifying the number of class descriptors'.
The value bNumberDescriptors identifies the number of additional class
specific descriptors present.
(refer to the 6.1.2 HID Descriptor in hid documents :
http://www.usb.org/developers/hidpage/HID1_11.pdf)

So, it can be out-of-bounds in hdesc->desc[n] if there is an
additional class descriptor.

Does the patch below fix this?

Thanks, jaejoong
------

diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
index 089bad8..7b6a0b6 100644
--- a/drivers/hid/usbhid/hid-core.c
+++ b/drivers/hid/usbhid/hid-core.c
@@ -974,7 +974,7 @@ static int usbhid_parse(struct hid_device *hid)
u32 quirks = 0;
unsigned int rsize = 0;
char *rdesc;
- int ret, n;
+ int ret;

quirks = usbhid_lookup_quirk(le16_to_cpu(dev->descriptor.idVendor),
le16_to_cpu(dev->descriptor.idProduct));
@@ -1000,9 +1000,8 @@ static int usbhid_parse(struct hid_device *hid)
hid->version = le16_to_cpu(hdesc->bcdHID);
hid->country = hdesc->bCountryCode;

- for (n = 0; n < hdesc->bNumDescriptors; n++)
- if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
- rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
+ if (hdesc->desc[0].bDescriptorType == HID_DT_REPORT)
+ rsize = le16_to_cpu(hdesc->desc[0].wDescriptorLength);

if (!rsize || rsize > HID_MAX_DESCRIPTOR_SIZE) {
dbg_hid("weird size of report descriptor (%u)\n", rsize);
> --
> To unsubscribe from this list: send the line "unsubscribe linux-input" in
> the body of a message to majo...@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
0001-HID-usbhid-out-of-bound-in-hdesc-desc.patch

Andrey Konovalov

unread,
Sep 19, 2017, 8:38:33 AM9/19/17
to Kim Jaejoong, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
On Tue, Sep 19, 2017 at 1:47 PM, Kim Jaejoong <climb...@gmail.com> wrote:
> Hi, Andrey Konovalov
>
> Thanks for the report.
>
> 2017-09-19 2:33 GMT+09:00 Andrey Konovalov <andre...@google.com>:
>> Hi!
>>
>> I've got the following crash while fuzzing the kernel with syzkaller.
>>
>> On commit ebb2c2437d8008d46796902ff390653822af6cc4 (Sep 18).
>>
>> It seems that there's no proper check on the hdesc->bNumDescriptors
>> value in usbhid_parse(). it iterates over hdesc->desc and accesses
>> hdesc->desc[n] fields, which might be out-of-bounds.
>
> The bNumDescriptors in hid descriptor means 'numeric expression
> specifying the number of class descriptors'.
> The value bNumberDescriptors identifies the number of additional class
> specific descriptors present.
> (refer to the 6.1.2 HID Descriptor in hid documents :
> http://www.usb.org/developers/hidpage/HID1_11.pdf)
>
> So, it can be out-of-bounds in hdesc->desc[n] if there is an
> additional class descriptor.
>
> Does the patch below fix this?

Hi Kim,

I'm not sure. Is there a check on the bLength field of a
hid_descriptor struct? Can it be less than sizeof(struct
hid_descriptor)? If so, we still do an out-of-bounds access to
hdesc->desc[0] (or some other fields).

Thanks!
> --
> You received this message because you are subscribed to the Google Groups "syzkaller" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to syzkaller+...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

Kim Jaejoong

unread,
Sep 20, 2017, 12:57:24 AM9/20/17
to Andrey Konovalov, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Hi Andrey
You are right. I add hid descriptr size from HID device with bLength field.

Could you test and review below patch?

To. usb & input guys.

While dig this report, i was wondering about bNumDescriptors in HID descriptor.
HID document from usb.org said, 'this number must be at least one (1)
as a Report descriptor will always be present.'

There is no mention of the order of class descriptors. Suppose you
have a HID device with a report descriptor and a physical descriptor.

If you have the following hid descriptor in this case,
HID descriptor
bLength: 12
bDescriptor Type: HID
.. skip
bNumDescriptors: 2
bDescriptorType: physical
bDescriptorLength: any
bDescriptorType: Report
bDescriptorLength: any

If the order of the report descriptor is the second as above,
usbhid_parse () will fail because my patch is only check the first
bDescriptor Type.
But If the order of the report descriptor is always first, there is no
problem. How do you think this?

Thanks, jaejoong

-----------------

diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
index 089bad8..94c3805 100644
--- a/drivers/hid/usbhid/hid-core.c
+++ b/drivers/hid/usbhid/hid-core.c
@@ -974,7 +974,7 @@ static int usbhid_parse(struct hid_device *hid)
u32 quirks = 0;
unsigned int rsize = 0;
char *rdesc;
- int ret, n;
+ int ret;

quirks = usbhid_lookup_quirk(le16_to_cpu(dev->descriptor.idVendor),
le16_to_cpu(dev->descriptor.idProduct));
@@ -997,12 +997,16 @@ static int usbhid_parse(struct hid_device *hid)
return -ENODEV;
}

+ if (hdesc->bLength < sizeof(*hdesc)) {
+ dbg_hid("hid descriptor is too short\n");
+ return -EINVAL;
+ }
+
hid->version = le16_to_cpu(hdesc->bcdHID);
hid->country = hdesc->bCountryCode;

- for (n = 0; n < hdesc->bNumDescriptors; n++)
- if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
- rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
+ if (hdesc->desc[0].bDescriptorType == HID_DT_REPORT)
+ rsize = le16_to_cpu(hdesc->desc[0].wDescriptorLength);

if (!rsize || rsize > HID_MAX_DESCRIPTOR_SIZE) {
dbg_hid("weird size of report descriptor (%u)\n", rsize);
-----------

Andrey Konovalov

unread,
Sep 20, 2017, 6:49:30 AM9/20/17
to Kim Jaejoong, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
On Wed, Sep 20, 2017 at 6:57 AM, Kim Jaejoong <climb...@gmail.com> wrote:
> Hi Andrey
>
> 2017-09-19 21:38 GMT+09:00 Andrey Konovalov <andre...@google.com>:
>> Hi Kim,
>>
>> I'm not sure. Is there a check on the bLength field of a
>> hid_descriptor struct? Can it be less than sizeof(struct
>> hid_descriptor)? If so, we still do an out-of-bounds access to
>> hdesc->desc[0] (or some other fields).
>
> You are right. I add hid descriptr size from HID device with bLength field.
>
> Could you test and review below patch?

It seems that this patch fixes the issue. I'll keep fuzzing to make sure.

Tested-by: Andrey Konovalov <andre...@google.com>

Thanks!

Alan Stern

unread,
Sep 20, 2017, 11:50:52 AM9/20/17
to Kim Jaejoong, Andrey Konovalov, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
On Wed, 20 Sep 2017, Kim Jaejoong wrote:

> To. usb & input guys.
>
> While dig this report, i was wondering about bNumDescriptors in HID descriptor.
> HID document from usb.org said, 'this number must be at least one (1)
> as a Report descriptor will always be present.'
>
> There is no mention of the order of class descriptors. Suppose you
> have a HID device with a report descriptor and a physical descriptor.
>
> If you have the following hid descriptor in this case,
> HID descriptor
> bLength: 12
> bDescriptor Type: HID
> .. skip
> bNumDescriptors: 2
> bDescriptorType: physical
> bDescriptorLength: any
> bDescriptorType: Report
> bDescriptorLength: any
>
> If the order of the report descriptor is the second as above,
> usbhid_parse () will fail because my patch is only check the first
> bDescriptor Type.
> But If the order of the report descriptor is always first, there is no
> problem. How do you think this?

The descriptors can appear in any order. You should not assume that
the report descriptor will always come first.

Alan Stern

Jaejoong Kim

unread,
Sep 20, 2017, 9:42:13 PM9/20/17
to Alan Stern, Andrey Konovalov, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Hi Alan
Thanks for clarifying. I will resend patch with modification.

Jaejoong

>
> Alan Stern
>

Jaejoong Kim

unread,
Sep 26, 2017, 3:31:56 AM9/26/17
to Alan Stern, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany, Jaejoong Kim
The starting address of the hid descriptor is obtained via
usb_get_extra_descriptor(). If the hid descriptor has the wrong size, it
is possible to access the wrong address. So, before accessing the hid
descriptor, we need to check the entire size through the bLength field.

It also shows how many class descriptors it has through the bNumDescriptors
of the hid descriptor. Assuming that the connected hid descriptor has two
class descriptors(report and physical descriptors), the code below can
cause OOB because hdesc->desc is an array of size 1.

for (n = 0; n < hdesc->bNumDescriptors; n++)
if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);

Since we know the starting address of the hid descriptor and the value of
the bNumDescriptors is variable, we directly access the buffer containing
the hid descriptor without usbing hdesc->desc to obtain the size of the
report descriptor.
Reported-by: Alexander Potapenko <gli...@google.com>
Signed-off-by: Jaejoong Kim <climb...@gmail.com>
---
drivers/hid/usbhid/hid-core.c | 39 +++++++++++++++++++++++++++------------
include/linux/hid.h | 2 ++
2 files changed, 29 insertions(+), 12 deletions(-)

diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
index 089bad8..7bad173 100644
--- a/drivers/hid/usbhid/hid-core.c
+++ b/drivers/hid/usbhid/hid-core.c
@@ -970,12 +970,19 @@ static int usbhid_parse(struct hid_device *hid)
struct usb_interface *intf = to_usb_interface(hid->dev.parent);
struct usb_host_interface *interface = intf->cur_altsetting;
struct usb_device *dev = interface_to_usbdev (intf);
- struct hid_descriptor *hdesc;
+ unsigned char *buffer = intf->altsetting->extra;
+ int buflen = intf->altsetting->extralen;
+ int length;
u32 quirks = 0;
unsigned int rsize = 0;
char *rdesc;
int ret, n;

+ if (!buffer) {
+ dbg_hid("class descriptor not present\n");
+ return -ENODEV;
+ }
+
quirks = usbhid_lookup_quirk(le16_to_cpu(dev->descriptor.idVendor),
le16_to_cpu(dev->descriptor.idProduct));

@@ -990,19 +997,27 @@ static int usbhid_parse(struct hid_device *hid)
quirks |= HID_QUIRK_NOGET;
}

- if (usb_get_extra_descriptor(interface, HID_DT_HID, &hdesc) &&
- (!interface->desc.bNumEndpoints ||
- usb_get_extra_descriptor(&interface->endpoint[0], HID_DT_HID, &hdesc))) {
- dbg_hid("class descriptor not present\n");
- return -ENODEV;
- }
+ while (buflen > 2) {
+ length = buffer[0];
+ if (!length || length < HID_DESCRIPTOR_MIN_SIZE)
+ goto next_desc;

- hid->version = le16_to_cpu(hdesc->bcdHID);
- hid->country = hdesc->bCountryCode;
+ if (buffer[1] == HID_DT_HID) {
+ hid->version = get_unaligned_le16(&buffer[2]);
+ hid->country = buffer[4];

- for (n = 0; n < hdesc->bNumDescriptors; n++)
- if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
- rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
+ for (n = 0; n < buffer[5]; n++) {
+ /* we are just interested in report descriptor */
+ if (buffer[6+3*n] != HID_DT_REPORT)
+ continue;
+ rsize = get_unaligned_le16(&buffer[7+3*n]);
+ }
+ }
+
+next_desc:
+ buflen -= length;
+ buffer += length;
+ }

if (!rsize || rsize > HID_MAX_DESCRIPTOR_SIZE) {
dbg_hid("weird size of report descriptor (%u)\n", rsize);
diff --git a/include/linux/hid.h b/include/linux/hid.h
index ab05a86..2d53c0f 100644
--- a/include/linux/hid.h
+++ b/include/linux/hid.h
@@ -638,6 +638,8 @@ struct hid_descriptor {
struct hid_class_descriptor desc[1];
} __attribute__ ((packed));

+#define HID_DESCRIPTOR_MIN_SIZE 9
+
#define HID_DEVICE(b, g, ven, prod) \
.bus = (b), .group = (g), .vendor = (ven), .product = (prod)
#define HID_USB_DEVICE(ven, prod) \
--
2.7.4

Alan Stern

unread,
Sep 26, 2017, 10:18:33 AM9/26/17
to Jaejoong Kim, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
On Tue, 26 Sep 2017, Jaejoong Kim wrote:

> The starting address of the hid descriptor is obtained via
> usb_get_extra_descriptor(). If the hid descriptor has the wrong size, it
> is possible to access the wrong address. So, before accessing the hid
> descriptor, we need to check the entire size through the bLength field.
>
> It also shows how many class descriptors it has through the bNumDescriptors
> of the hid descriptor. Assuming that the connected hid descriptor has two
> class descriptors(report and physical descriptors), the code below can
> cause OOB because hdesc->desc is an array of size 1.
>
> for (n = 0; n < hdesc->bNumDescriptors; n++)
> if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
> rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
>
> Since we know the starting address of the hid descriptor and the value of
> the bNumDescriptors is variable, we directly access the buffer containing
> the hid descriptor without usbing hdesc->desc to obtain the size of the
> report descriptor.

I do not like this patch at all.
It's silly to duplicate the code that is already present in
usb_get_extra_descriptor(). There's no reason to avoid using that
function here.

> - hid->version = le16_to_cpu(hdesc->bcdHID);
> - hid->country = hdesc->bCountryCode;
> + if (buffer[1] == HID_DT_HID) {
> + hid->version = get_unaligned_le16(&buffer[2]);
> + hid->country = buffer[4];

It's also silly not to use the preformatted hid_descriptor structure and
instead put error-prone byte offsets directly into the code.

> - for (n = 0; n < hdesc->bNumDescriptors; n++)
> - if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
> - rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
> + for (n = 0; n < buffer[5]; n++) {
> + /* we are just interested in report descriptor */
> + if (buffer[6+3*n] != HID_DT_REPORT)
> + continue;
> + rsize = get_unaligned_le16(&buffer[7+3*n]);

Finally, this patch doesn't fix the actual problem! You don't check
here whether 8+3*n >= length.

This whole thing could be fixed with a much smaller change to the
original code. Just do something like this:

num_descriptors = min_t(int, hdesc->bNumDescriptors,
(hdesc->bLength - 6) / 3);
for (n = 0; n < num_descriptors; n++)

Alan Stern

Jaejoong Kim

unread,
Sep 26, 2017, 3:59:55 PM9/26/17
to Alan Stern, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Hi, Alan,

Thanks for the review.
To be honest, I was fooled. You are right. That is a duplicate code
with usb_get_extra_descriptor().

>
>> - hid->version = le16_to_cpu(hdesc->bcdHID);
>> - hid->country = hdesc->bCountryCode;
>> + if (buffer[1] == HID_DT_HID) {
>> + hid->version = get_unaligned_le16(&buffer[2]);
>> + hid->country = buffer[4];
>
> It's also silly not to use the preformatted hid_descriptor structure and
> instead put error-prone byte offsets directly into the code.

Right.

>
>> - for (n = 0; n < hdesc->bNumDescriptors; n++)
>> - if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
>> - rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);
>> + for (n = 0; n < buffer[5]; n++) {
>> + /* we are just interested in report descriptor */
>> + if (buffer[6+3*n] != HID_DT_REPORT)
>> + continue;
>> + rsize = get_unaligned_le16(&buffer[7+3*n]);
>
> Finally, this patch doesn't fix the actual problem! You don't check
> here whether 8+3*n >= length.
>
> This whole thing could be fixed with a much smaller change to the
> original code. Just do something like this:
>
> num_descriptors = min_t(int, hdesc->bNumDescriptors,
> (hdesc->bLength - 6) / 3);
> for (n = 0; n < num_descriptors; n++)

Right. Your change is better and clear.
With your change, I think we need check the size of hdesc before
accessing hdesc->desc[n] as Andrey Konovalov has already pointed out,

diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
index 089bad8..0b41d44 100644
--- a/drivers/hid/usbhid/hid-core.c
+++ b/drivers/hid/usbhid/hid-core.c
@@ -975,6 +975,7 @@ static int usbhid_parse(struct hid_device *hid)
unsigned int rsize = 0;
char *rdesc;
int ret, n;
+ int num_descriptors;

quirks = usbhid_lookup_quirk(le16_to_cpu(dev->descriptor.idVendor),
le16_to_cpu(dev->descriptor.idProduct));
@@ -997,10 +998,17 @@ static int usbhid_parse(struct hid_device *hid)
return -ENODEV;
}

+ if (hdesc->bLength < sizeof(struct hid_descriptor)) {
+ dbg_hid("hid descriptor is too short\n");
+ return -EINVAL;
+ }
+
hid->version = le16_to_cpu(hdesc->bcdHID);
hid->country = hdesc->bCountryCode;

- for (n = 0; n < hdesc->bNumDescriptors; n++)
+ num_descriptors = min_t(int, hdesc->bNumDescriptors,
+ (hdesc->bLength - 6) / 3);
+ for (n = 0; n < num_descriptors; n++)
if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);

Jaejoong

Alan Stern

unread,
Sep 26, 2017, 4:23:00 PM9/26/17
to Jaejoong Kim, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Dmitry Vyukov, Kostya Serebryany
Yes, this is a lot better.

Alan Stern

Michel Hermier

unread,
Sep 27, 2017, 2:27:02 AM9/27/17
to Alan Stern, Kostya Serebryany, syzkaller, Jiri Kosina, USB list, Jaejoong Kim, Andrey Konovalov, Benjamin Tissoires, Dmitry Vyukov, linux...@vger.kernel.org, LKML
Is it possible to explicit the magic number 6 and 3 in the code. Currently, it looks like it comes from no where.

I'm also wondering if this change will not affect some devices in the wild, by rejecting hid descriptors with num descriptors == 0 ?


Alan Stern

Alan Stern

unread,
Sep 27, 2017, 10:29:21 AM9/27/17
to Michel Hermier, Kostya Serebryany, syzkaller, Jiri Kosina, USB list, Jaejoong Kim, Andrey Konovalov, Benjamin Tissoires, Dmitry Vyukov, linux...@vger.kernel.org, LKML
Yes, it is possible. The 6 is equal to

offsetof(struct hid_descriptor, desc)

and the 3 is equal to

sizeof(struct hid_class_descriptor)

(at least, I think it is -- the structure is marked as packed so its
size should be 3).

In this case I found the numbers to be more readable, but other people
may have different opinions.

> I'm also wondering if this change will not affect some devices in the wild,
> by rejecting hid descriptors with num descriptors == 0 ?

It's possible, but I doubt it. If such devices do exist, they should
never have worked in the first place. Certainly they would generate
warnings or errors during enumeration because of their invalid
descriptors.

Alan Stern

Jaejoong Kim

unread,
Sep 28, 2017, 4:44:24 AM9/28/17
to Alan Stern, Michel Hermier, Kostya Serebryany, syzkaller, Jiri Kosina, USB list, Andrey Konovalov, Benjamin Tissoires, Dmitry Vyukov, linux...@vger.kernel.org, LKML
2017-09-27 23:29 GMT+09:00 Alan Stern <st...@rowland.harvard.edu>:
> On Wed, 27 Sep 2017, Michel Hermier wrote:
>
>> Le 27 sept. 2017 07:42, "Alan Stern" <st...@rowland.harvard.edu> a écrit :
>
>> > - for (n = 0; n < hdesc->bNumDescriptors; n++)
>> > + num_descriptors = min_t(int, hdesc->bNumDescriptors,
>> > + (hdesc->bLength - 6) / 3);
>> > + for (n = 0; n < num_descriptors; n++)
>> > if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
>> > rsize = le16_to_cpu(hdesc->desc[n].
>> wDescriptorLength);
>>
>> Yes, this is a lot better.

OK.

>>
>>
>> Is it possible to explicit the magic number 6 and 3 in the code. Currently,
>> it looks like it comes from no where.

I gree with you.

>
> Yes, it is possible. The 6 is equal to
>
> offsetof(struct hid_descriptor, desc)
>
> and the 3 is equal to
>
> sizeof(struct hid_class_descriptor)
>
> (at least, I think it is -- the structure is marked as packed so its
> size should be 3).
>
> In this case I found the numbers to be more readable, but other people
> may have different opinions.

I will post V2 shortly.

>
>> I'm also wondering if this change will not affect some devices in the wild,
>> by rejecting hid descriptors with num descriptors == 0 ?
>
> It's possible, but I doubt it. If such devices do exist, they should
> never have worked in the first place. Certainly they would generate
> warnings or errors during enumeration because of their invalid
> descriptors.
>
> Alan Stern
>

Jaejoong

Jaejoong Kim

unread,
Sep 28, 2017, 6:16:44 AM9/28/17
to Alan Stern, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Jaejoong Kim
The hid descriptor identifies the length and type of subordinate
descriptors for a device. If the received hid descriptor is smaller than
the size of the struct hid_descriptor, it is possible to cause
out-of-bounds.

In addition, if bNumDescriptors of the hid descriptor have an incorrect
value, this can also cause out-of-bounds while approaching hdesc->desc[n].

So check the size of hid descriptor and bNumDescriptors.
Reported-by: Andrey Konovalov <andre...@google.com>
Signed-off-by: Jaejoong Kim <climb...@gmail.com>
---

Changes in v2:
- write a new commit message because orginal version is wrong approach
- add check hid descriptor size
- get proper value for bNumDescriptors as suggested by Alan Stern
- fix the Reported-by

drivers/hid/usbhid/hid-core.c | 12 +++++++++++-
1 file changed, 11 insertions(+), 1 deletion(-)

diff --git a/drivers/hid/usbhid/hid-core.c b/drivers/hid/usbhid/hid-core.c
index 089bad8..045b5da 100644
--- a/drivers/hid/usbhid/hid-core.c
+++ b/drivers/hid/usbhid/hid-core.c
@@ -975,6 +975,8 @@ static int usbhid_parse(struct hid_device *hid)
unsigned int rsize = 0;
char *rdesc;
int ret, n;
+ int num_descriptors;
+ size_t offset = offsetof(struct hid_descriptor, desc);

quirks = usbhid_lookup_quirk(le16_to_cpu(dev->descriptor.idVendor),
le16_to_cpu(dev->descriptor.idProduct));
@@ -997,10 +999,18 @@ static int usbhid_parse(struct hid_device *hid)
return -ENODEV;
}

+ if (hdesc->bLength < sizeof(struct hid_descriptor)) {
+ dbg_hid("hid descriptor is too short\n");
+ return -EINVAL;
+ }
+
hid->version = le16_to_cpu(hdesc->bcdHID);
hid->country = hdesc->bCountryCode;

- for (n = 0; n < hdesc->bNumDescriptors; n++)
+ num_descriptors = min_t(int, hdesc->bNumDescriptors,
+ (hdesc->bLength - offset) / sizeof(struct hid_class_descriptor));
+
+ for (n = 0; n < num_descriptors; n++)
if (hdesc->desc[n].bDescriptorType == HID_DT_REPORT)
rsize = le16_to_cpu(hdesc->desc[n].wDescriptorLength);

--
2.7.4

Jaejoong Kim

unread,
Oct 10, 2017, 3:25:47 AM10/10/17
to Alan Stern, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller, Jaejoong Kim
Hi,

To. Jiri, Alan,

Could you please review this patch?

To. Andey,

Could you please test with this patch for KASAN OOB error?

Thanks, jaejoong

Andrey Konovalov

unread,
Oct 10, 2017, 7:23:05 AM10/10/17
to Jaejoong Kim, Alan Stern, Jiri Kosina, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller
On Tue, Oct 10, 2017 at 9:25 AM, Jaejoong Kim <climb...@gmail.com> wrote:
> Hi,
>
> To. Jiri, Alan,
>
> Could you please review this patch?
>
> To. Andey,
>
> Could you please test with this patch for KASAN OOB error?

Hi!

Yes, your patch fixes the issue.

Thanks!

Tested-by: Andrey Konovalov <andre...@google.com>

Alan Stern

unread,
Oct 10, 2017, 10:22:24 AM10/10/17
to Jaejoong Kim, Jiri Kosina, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller
Acked-by: Alan Stern <st...@rowland.harvard.edu>

Jiri Kosina

unread,
Oct 11, 2017, 9:06:02 AM10/11/17
to Jaejoong Kim, Alan Stern, Andrey Konovalov, Benjamin Tissoires, USB list, linux...@vger.kernel.org, LKML, syzkaller
On Thu, 28 Sep 2017, Jaejoong Kim wrote:

> The hid descriptor identifies the length and type of subordinate
> descriptors for a device. If the received hid descriptor is smaller than
> the size of the struct hid_descriptor, it is possible to cause
> out-of-bounds.
>
> In addition, if bNumDescriptors of the hid descriptor have an incorrect
> value, this can also cause out-of-bounds while approaching hdesc->desc[n].
>
> So check the size of hid descriptor and bNumDescriptors.

Applied to for-4.14/upstream-fixes. Thanks,

--
Jiri Kosina
SUSE Labs

Reply all
Reply to author
Forward
0 new messages