CVE-2023-52773

What it actually does

Null pointer dereference in amdgpu_dm_i2c_xfer() when ddc_service_construct() finds no hw_supported pin due to missing link or pin after driver unload/reload, causing kernel crash.

Why it matters to the business

Local denial of service on AMD GPU Linux systems leading to crashes and reboots, with limited operational impact given local access requirement and stable EPSS trend.

How an attack unfolds

Local attacker unloads and reloads the amdgpu module to reach an uninitialized link state then calls amdgpu_dm_i2c_xfer to trigger the NULL dereference.

How to fix it

Update Linux kernel to include commit fb5c134ca589fe670430acc9e7ebf2691ca2476d which adds the required NULL checks before I2C transfer.

How to confirm the fix worked

Confirm commit fb5c134ca589fe670430acc9e7ebf2691ca2476d is present in kernel source or verify updated distribution packages; reproduce driver reload followed by I2C operations without crash.

Typical attack path

1. Obtain local shell on affected Linux host
2. Unload and reload amdgpu kernel module
3. Perform I2C transfer via amdgpu_dm_i2c_xfer
4. Trigger NULL pointer dereference and kernel panic

How to detect it

Search kernel logs and dmesg for NULL pointer dereference messages referencing amdgpu_dm_i2c_xfer or ddc_service_construct during module reload events.

• amdgpu_dm_i2c_xfer
• NULL pointer dereference
• ddc_service_construct

What your security team should do

• Apply kernel update containing fix commit fb5c134ca589fe670430acc9e7ebf2691ca2476d
• Monitor kernel logs for amdgpu-related NULL dereference panics
• Limit unprivileged local access on systems with AMD GPUs