Where RM Runs: Host RM, GSP-RM, and Firmware Boot
A boundary-oriented view of how NVIDIA moves physical resource management onto GSP, bootstraps the firmware, and reconnects it to host RM over RPC.
hans@machine:~$ whoami
GPU systems, from the kernel up.
Field notes on drivers, firmware, memory, and Linux.
I'm Hans. I work on GPU systems at a fabless semiconductor company. This is where I trace the boundaries between hardware, firmware, the kernel, and the runtime.
Current focus
- GPU driver architecture
- Linux kernel integration
- firmware & bring-up
- memory management
Latest notes
View allWhere DKMS Ends in the NVIDIA Linux Driver
Following an NVIDIA driver build across RM, Kbuild, module ABI checks, GSP firmware, UVM, and DRM to see what DKMS does—and what it cannot do.
From GPU Fault to Replay: UVM, HMM, and Page Migration
Tracing a replayable GPU fault through UVM batching, residency policy, Linux HMM invalidation, migrate_vma, GPU page-table updates, and replay.
Selected writing
View allWhat We Talk About When We Talk About Refactoring
Refactoring becomes avoidance when structural change substitutes for naming the real constraint, making the hard decision, or investigating behavior.