Segment 41

The assistant tested the per-IOMMU-group identity domain approach to restore GPU P2P DMA with IOMMU in full translation mode, but discovered that the nvidia Blackwell FSP boot sequence fails with error code 0x177 when IOMMU is in identity mode, as the FSP requires specific DMA mappings set up by the kernel's DMA API in translation mode. This definitively blocks P2P restoration via identity domains. The assistant immediately reverted by removing the modprobe hook and rebooting, restoring the working DMA-FQ configuration. The MTP/NEXTN speculation enabled in the previous session survived the reboot and continues to provide 12-45% per-request throughput improvement at low concurrency. The SGLang service for Qwen3.5-122B-A10B BF16 on 4 GPUs with TP=4 remains stable with NCCL_P2P_DISABLE=1.

The Blackwell FSP Barrier: A Definitive End to the IOMMU Identity Domain Path 4082 words

Chunks

• The Blackwell FSP Barrier: How IOMMU Identity Domains Met Their Match in NVIDIA's Firmware Security Processor
  2712 words
• The Blackwell FSP Barrier: A Synthesis of the IOMMU Identity Domain Experiment and the Pivot to MTP Speculation
  2542 words