FlawAtlas
Search the atlas
UBUNTU-CVE-2025-38016 Moderate

UBUNTU-CVE-2025-38016

In the Linux kernel, the following vulnerability has been resolved: HID: bpf: abort dispatch if device destroyed The current HID bpf implementation assumes no output report/request will go through it after hid_bpf_destroy_device() has been called. This leads to a bug that unplugging certain types of HID devices causes a cleaned- up SRCU to be accessed. The bug was previously a hidden failure until a recent x86 percpu change [1] made it access not-present pages. The bug will be triggered if the conditions below are met: A) a device under the driver has some LEDs on B) hid_ll_driver->request() is uninplemented (e.g., logitech-djreceiver) If condition A is met, hidinput_led_worker() is always scheduled *after* hid_bpf_destroy_device(). hid_destroy_device ` hid_bpf_destroy_device ` cleanup_srcu_struct(&hdev->bpf.srcu) ` hid_remove_device ` ... ` led_classdev_unregister ` led_trigger_set(led_cdev, NULL) ` led_set_brightness(led_cdev, LED_OFF) ` ... ` input_inject_event ` input_event_dispose ` hidinput_input_event ` schedule_work(&hid->led_work) [hidinput_led_worker] This is fine when condition B is not met, where hidinput_led_worker() calls hid_ll_driver->request(). This is the case for most HID drivers, which implement it or use the generic one from usbhid. The driver itself or an underlying driver will then abort processing the request. Otherwise, hidinput_led_worker() tries hid_hw_output_report() and leads to the bug. hidinput_led_worker ` hid_hw_output_report ` dispatch_hid_bpf_output_report ` srcu_read_lock(&hdev->bpf.srcu) ` srcu_read_unlock(&hdev->bpf.srcu, idx) The bug has existed since the introduction [2] of dispatch_hid_bpf_output_report(). However, the same bug also exists in dispatch_hid_bpf_raw_requests(), and I've reproduced (no visible effect because of the lack of [1], but confirmed bpf.destroyed == 1) the bug against the commit (i.e., the Fixes:) introducing the function. This is because hidinput_led_worker() falls back to hid_hw_raw_request() when hid_ll_driver->output_report() is uninplemented (e.g., logitech- djreceiver). hidinput_led_worker ` hid_hw_output_report: -ENOSYS ` hid_hw_raw_request ` dispatch_hid_bpf_raw_requests ` srcu_read_lock(&hdev->bpf.srcu) ` srcu_read_unlock(&hdev->bpf.srcu, idx) Fix the issue by returning early in the two mentioned functions if hid_bpf has been marked as destroyed. Though dispatch_hid_bpf_device_event() handles input events, and there is no evidence that it may be called after the destruction, the same check, as a safety net, is also added to it to maintain the consistency among all dispatch functions. The impact of the bug on other architectures is unclear. Even if it acts as a hidden failure, this is still dangerous because it corrupts whatever is on the address calculated by SRCU. Thus, CC'ing the stable list. [1]: commit 9d7de2aa8b41 ("x86/percpu/64: Use relative percpu offsets") [2]: commit 9286675a2aed ("HID: bpf: add HID-BPF hooks for hid_hw_output_report")

Exploit probability Not scored
Published June 18, 2025
Required by Not available
Last source change August 17, 2026

02 / AFFECTED SOFTWARE

Affected packages

Ubuntu:22.04:LTS linux-starfive-6.5

11 explicit affected versions

Ubuntu:18.04:LTS linux-oem

50 explicit affected versions

Ubuntu:18.04:LTS linux-aws-5.0

6 explicit affected versions

Ubuntu:20.04:LTS linux-aws-5.11

12 explicit affected versions

Ubuntu:22.04:LTS linux-oem-6.5

19 explicit affected versions

Ubuntu:20.04:LTS linux-raspi2

6 explicit affected versions

Ubuntu:24.04:LTS linux-riscv

24 explicit affected versions

Ubuntu:22.04:LTS linux-riscv-6.5

10 explicit affected versions

Ubuntu:22.04:LTS linux-azure-5.19

7 explicit affected versions

Ubuntu:22.04:LTS linux-gcp-6.5

13 explicit affected versions

Ubuntu:20.04:LTS linux-oem-5.14

35 explicit affected versions

Ubuntu:20.04:LTS linux-gke-5.15

18 explicit affected versions

Ubuntu:20.04:LTS linux-aws-5.13

13 explicit affected versions

Ubuntu:20.04:LTS linux-aws-5.8

4 explicit affected versions

Ubuntu:16.04:LTS linux-hwe-edge

23 explicit affected versions

Ubuntu:24.04:LTS linux-raspi-realtime

1 explicit affected versions

Ubuntu:20.04:LTS linux-hwe-5.8

21 explicit affected versions

Ubuntu:20.04:LTS linux-hwe-5.13

16 explicit affected versions

Ubuntu:22.04:LTS linux-lowlatency-hwe-6.2

10 explicit affected versions

Ubuntu:24.04:LTS linux-hwe-6.11

8 explicit affected versions

Ubuntu:18.04:LTS linux-oracle-5.0

7 explicit affected versions

Ubuntu:20.04:LTS linux-riscv

12 explicit affected versions

Ubuntu:24.04:LTS linux-oracle-6.14

3 explicit affected versions

Ubuntu:24.04:LTS linux-azure-6.11

7 explicit affected versions

Ubuntu:20.04:LTS linux-gcp-5.11

12 explicit affected versions

Ubuntu:22.04:LTS linux-azure-6.5

14 explicit affected versions

Ubuntu:25.10 linux-azure-fde

10 explicit affected versions

Ubuntu:22.04:LTS linux-intel-iot-realtime

1 explicit affected versions

Ubuntu:20.04:LTS linux-intel-5.13

7 explicit affected versions

Ubuntu:20.04:LTS linux-oem-5.13

12 explicit affected versions

Ubuntu:20.04:LTS linux-oracle-5.11

12 explicit affected versions

Ubuntu:24.04:LTS linux-oem-6.14

5 explicit affected versions

Ubuntu:18.04:LTS linux-hwe-edge

10 explicit affected versions

Ubuntu:18.04:LTS linux-gke-5.4

37 explicit affected versions

Ubuntu:22.04:LTS linux-oracle-6.5

13 explicit affected versions

Ubuntu:22.04:LTS linux-azure-fde-6.2

10 explicit affected versions

Ubuntu:24.04:LTS linux-gcp-6.14

3 explicit affected versions

Ubuntu:20.04:LTS linux-oem-5.10

26 explicit affected versions

Ubuntu:20.04:LTS linux-gcp-5.13

13 explicit affected versions

Ubuntu:20.04:LTS linux-oem-5.6

29 explicit affected versions

Ubuntu:18.04:LTS linux-gcp-5.3

13 explicit affected versions

Ubuntu:24.04:LTS linux-riscv-6.14

4 explicit affected versions

Ubuntu:18.04:LTS linux-gke-4.15

33 explicit affected versions

Ubuntu:22.04:LTS linux-riscv

23 explicit affected versions

Ubuntu:18.04:LTS linux-aws-5.3

10 explicit affected versions

Ubuntu:Nvidia-BlueField:24.04:LTS linux-bluefield

11 explicit affected versions

Ubuntu:20.04:LTS linux-hwe-5.11

12 explicit affected versions

Ubuntu:20.04:LTS linux-azure-5.8

7 explicit affected versions

Ubuntu:22.04:LTS linux-gcp-6.2

12 explicit affected versions

Ubuntu:22.04:LTS linux-lowlatency-hwe-5.19

10 explicit affected versions

Ubuntu:22.04:LTS linux-oem-5.17

26 explicit affected versions

Ubuntu:24.04:LTS linux-aws-6.14

3 explicit affected versions

Ubuntu:22.04:LTS linux-azure-6.2

13 explicit affected versions

Ubuntu:22.04:LTS linux-oem-6.0

16 explicit affected versions

Ubuntu:22.04:LTS linux-azure-fde-5.19

3 explicit affected versions

Ubuntu:24.04:LTS linux-gcp-6.11

7 explicit affected versions

Ubuntu:20.04:LTS linux-gke

51 explicit affected versions

Ubuntu:24.04:LTS linux-oem-6.11

18 explicit affected versions

Ubuntu:20.04:LTS linux-azure-5.13

12 explicit affected versions

Ubuntu:22.04:LTS linux-oem-6.1

27 explicit affected versions

Ubuntu:22.04:LTS linux-gcp-5.19

8 explicit affected versions

Ubuntu:20.04:LTS linux-oracle-5.8

5 explicit affected versions

Ubuntu:22.04:LTS linux-riscv-5.19

8 explicit affected versions

Ubuntu:22.04:LTS linux-hwe-6.2

10 explicit affected versions

Ubuntu:24.04:LTS linux-lowlatency-hwe-6.11

7 explicit affected versions

Ubuntu:22.04:LTS linux-lowlatency-hwe-6.5

13 explicit affected versions

Ubuntu:20.04:LTS linux-azure-5.11

13 explicit affected versions

Ubuntu:22.04:LTS linux-starfive-5.19

7 explicit affected versions

Ubuntu:24.04:LTS linux-nvidia-6.11

6 explicit affected versions

Ubuntu:20.04:LTS linux-riscv-5.8

10 explicit affected versions

Ubuntu:20.04:LTS linux-oracle-5.13

12 explicit affected versions

Ubuntu:Pro:Realtime:24.04:LTS linux-realtime-6.14

1 explicit affected versions

Ubuntu:18.04:LTS linux-azure

44 explicit affected versions

Ubuntu:18.04:LTS linux-azure-5.3

16 explicit affected versions

Ubuntu:22.04:LTS linux-starfive-6.2

3 explicit affected versions

Ubuntu:20.04:LTS linux-riscv-5.11

13 explicit affected versions

Ubuntu:20.04:LTS linux-azure-fde

26 explicit affected versions

Ubuntu:20.04:LTS linux-gkeop-5.15

45 explicit affected versions

Ubuntu:18.04:LTS linux-gkeop-5.4

37 explicit affected versions

Ubuntu:22.04:LTS linux-aws-6.5

14 explicit affected versions

Ubuntu:18.04:LTS linux-gcp

38 explicit affected versions

Ubuntu:18.04:LTS linux-oracle-5.3

9 explicit affected versions

Ubuntu:22.04:LTS linux-hwe-5.19

11 explicit affected versions

Ubuntu:22.04:LTS linux-nvidia-6.2

7 explicit affected versions

Ubuntu:20.04:LTS linux-gcp-5.8

4 explicit affected versions

Ubuntu:24.04:LTS linux-realtime

1 explicit affected versions

Ubuntu:22.04:LTS linux-hwe-6.5

13 explicit affected versions

Ubuntu:18.04:LTS linux-hwe

43 explicit affected versions

Ubuntu:22.04:LTS linux-aws-5.19

10 explicit affected versions

Ubuntu:22.04:LTS linux-allwinner-5.19

8 explicit affected versions

Ubuntu:22.04:LTS linux-aws-6.2

14 explicit affected versions

Ubuntu:18.04:LTS linux-azure-edge

4 explicit affected versions

Ubuntu:20.04:LTS linux-gkeop

78 explicit affected versions

Ubuntu:22.04:LTS linux-nvidia-6.5

11 explicit affected versions

Ubuntu:22.04:LTS linux-realtime

1 explicit affected versions

03 / CONNECTIONS

Connected vulnerabilities

04 / EVIDENCE

Source records

Open Source Vulnerabilities UBUNTU-CVE-2025-38016

In the Linux kernel, the following vulnerability has been resolved: HID: bpf: abort dispatch if device destroyed The current HID bpf implementation assumes no output report/request will go through it after hid_bpf_destroy_device() has been called. This leads to a bug that unplugging certain types of HID devices causes a cleaned- up SRCU to be accessed. The bug was previously a hidden failure until a recent x86 percpu change [1] made it access not-present pages. The bug will be triggered if the conditions below are met: A) a device under the driver has some LEDs on B) hid_ll_driver->request() is uninplemented (e.g., logitech-djreceiver) If condition A is met, hidinput_led_worker() is always scheduled *after* hid_bpf_destroy_device(). hid_destroy_device ` hid_bpf_destroy_device ` cleanup_srcu_struct(&hdev->bpf.srcu) ` hid_remove_device ` ... ` led_classdev_unregister ` led_trigger_set(led_cdev, NULL) ` led_set_brightness(led_cdev, LED_OFF) ` ... ` input_inject_event ` input_event_dispose ` hidinput_input_event ` schedule_work(&hid->led_work) [hidinput_led_worker] This is fine when condition B is not met, where hidinput_led_worker() calls hid_ll_driver->request(). This is the case for most HID drivers, which implement it or use the generic one from usbhid. The driver itself or an underlying driver will then abort processing the request. Otherwise, hidinput_led_worker() tries hid_hw_output_report() and leads to the bug. hidinput_led_worker ` hid_hw_output_report ` dispatch_hid_bpf_output_report ` srcu_read_lock(&hdev->bpf.srcu) ` srcu_read_unlock(&hdev->bpf.srcu, idx) The bug has existed since the introduction [2] of dispatch_hid_bpf_output_report(). However, the same bug also exists in dispatch_hid_bpf_raw_requests(), and I've reproduced (no visible effect because of the lack of [1], but confirmed bpf.destroyed == 1) the bug against the commit (i.e., the Fixes:) introducing the function. This is because hidinput_led_worker() falls back to hid_hw_raw_request() when hid_ll_driver->output_report() is uninplemented (e.g., logitech- djreceiver). hidinput_led_worker ` hid_hw_output_report: -ENOSYS ` hid_hw_raw_request ` dispatch_hid_bpf_raw_requests ` srcu_read_lock(&hdev->bpf.srcu) ` srcu_read_unlock(&hdev->bpf.srcu, idx) Fix the issue by returning early in the two mentioned functions if hid_bpf has been marked as destroyed. Though dispatch_hid_bpf_device_event() handles input events, and there is no evidence that it may be called after the destruction, the same check, as a safety net, is also added to it to maintain the consistency among all dispatch functions. The impact of the bug on other architectures is unclear. Even if it acts as a hidden failure, this is still dangerous because it corrupts whatever is on the address calculated by SRCU. Thus, CC'ing the stable list. [1]: commit 9d7de2aa8b41 ("x86/percpu/64: Use relative percpu offsets") [2]: commit 9286675a2aed ("HID: bpf: add HID-BPF hooks for hid_hw_output_report")

View original source

05 / REFERENCES

Further evidence