ISP Camera Device Driver

Job Title :Senior Camera Driver / Embedded Linux Kernel EngineerJob Summary :We are seeking a highly skilled Camera Driver Engineer with expertise in Android Camera HAL3, V4L2, and MIPI CSI-2 integration to develop, optimize, and debug the end-to-end camera pipeline from Android Camera2 API through Camera HAL, V4L2 kernel drivers, and sensor hardware.The ideal candidate will have hands-on experience with Linux kernel driver development, multimedia subsystems, and performance optimization for camera capture, streaming, and ISP pipelines.Key Responsibilities :Camera Framework & HAL Development :- Develop, integrate, and optimize the Android Camera HAL3 layer.- Implement and manage the capture flow between Camera2 API, Camera Provider, and Camera HAL.- Ensure proper communication between Camera Service (framework) and hardware abstraction layer (HAL) using HIDL/AIDL.- Collaborate with SoC vendors and integrate V4L2-based drivers into the Android camera stack.Flow Summary (End-to-End) :- App Layer (Camera2 API) : Application creates a CaptureRequest ? sent to Camera Service.- Framework (Camera Service) : Routes requests to Camera Provider (HIDL/AIDL) ? loads HAL implementation.- Camera Provider : Manages available camera devices, connects to HAL shared library (camera.vendor.so).- Camera HAL : Converts framework requests into hardware operations configures streams, manages buffers, and communicates with kernel drivers via V4L2 IOCTLs.- Kernel Driver (V4L2) : Handles sensor, CSI-2 receiver, and ISP interaction; streams buffers using DMA/videobuf2.- Hardware Layer : MIPI CSI-2 sensor sends frames ? captured and processed by ISP ? delivered to user-space buffers.Kernel-Level Deferred Work (Tasklets vs Workqueues) :- Implement tasklets for fast, atomic bottom-half handling post-ISR (e.g., DMA completion notification).- Use workqueues for deferred tasks that require sleeping or longer operations (e.g., copying data to user buffers).- Example : ISR triggers a tasklet to signal frame-ready ? workqueue processes the frame metadata and prepares the next capture.Buffer Management in Kernel / Driver Context :- Design C/C++ data structures to manage streaming buffers in kernel space.- Manage a linked list of DMA buffers using list_head for efficient allocation, queueing, and freeing during streaming operations.- Implement vb2_queue operations (queue_setup, buf_prepare, start_streaming, stop_streaming) for V4L2 devices.V4L2 Camera Sensor Driver Development :- Write and register V4L2 subdevice drivers for camera sensors using I2C interface.- Implement essential V4L2 callbacks- Handle buffer flow using videobuf2 (vb2) framework :- Userspace : VIDIOC_REQBUFS, VIDIOC_QBUF, VIDIOC_STREAMON- Kernel : DMA streaming to frame buffers- Userspace : Dequeue frames with VIDIOC_DQBUF- Verify operation using tools like v4l2-ctl and media-ctl.MIPI CSI-2 Integration and Debug"- Integrate MIPI CSI-2 camera sensors via D-PHY lanes.- Configure timing parameters (e.g., HS_SETTLE, CLK_PREPARE, CLK_ZERO).- Debug and fix frame drop or latency issues :- Monitor dmesg logs for CSI errors or DMA underruns.- Validate frame rate using v4l2-ctl - stream-mmap.- Tune buffer queue depth, vb2 configuration, and ISP pipeline timing.- Analyze MIPI trace logs or PHY counters for signal integrity issues.- Optimize I2C register writes and CSI lane mapping in device tree.Technical Skills & Keywords :- Camera Driver Development (V4L2 / Camera HAL3 / Camera Provider)- Linux Kernel Driver Programming- Multimedia Subsystem (Video, ISP, Sensor)- Android Camera2 Framework- DMA / Videobuf2 / Buffer ManagementHardware Interfaces :- MIPI CSI-2, D-PHY- I2C / GPIO / Regmap- ISP / CSI Receiver / Sensor InterfaceDebugging Tools :- v4l2-ctl, media-ctl, dmesg, trace-cmd, perf, ftrace- Oscilloscope / Logic Analyzer for CSI signalsProgramming :- C, C++, Embedded C- Kernel Synchronization (Spinlocks, Tasklets, Workqueues)- Device Tree (DTS) Configuration and DebuggingPreferred Experience :- Camera bring-up on Android or embedded Linux platforms (Qualcomm, MediaTek, NVIDIA, NXP, TI, etc.)- Developing and debugging camera sensor subdevices- Hands-on experience with ISP pipelines and tuning- Performance optimization (latency, frame rate, bandwidth)- Familiarity with libcamera or GStreamer frameworks (ref:hirist.tech)