Enabling GDS support

I had system where GDS used to work , now running gdscheck after updating kernel it is showing NVME not supported

6.17.0-19-generic 
nvidia-fs loaded - 1.17.0.44
Driver Version: 590.48.01
CUDA Version: 13.1

NVIDIA H200 NVL  X2
e3:00.0 3D controller: NVIDIA Corporation Device 233b (rev a1)
e4:00.0 3D controller: NVIDIA Corporation Device 233b (rev a1)

nvidia driver loaded with NVreg_RegistryDwords="RMForceStaticBar1=1;ForceP2P=0;RmForceDisableIomapWC=1;"

cat /proc/driver/nvidia/params | grep -i static
RegistryDwords: "RMForceStaticBar1=1;ForceP2P=0;RmForceDisableIomapWC=1;"



p2p dma support in kernel
 cat /proc/kallsyms | grep -i p2pdma_pgmap_ops
0000000000000000 d p2pdma_pgmap_ops


-TNRT2:~/alok/gds-nvidia-fs/src$ sudo lspci -s e3:00.0 -vvvvv | grep "BAR 1"
                BAR 1: current size: 8GB, supported: 64MB 128MB 256MB 512MB 1GB 2GB 4GB 8GB 16GB 32GB 64GB 128GB 256GB 256TB 512TB 1PB 2PB 4PB 8PB 16PB 32PB 64PB 128PB 256PB 512PB 1EB 2EB 4EB 8EB


TNRT2:~/alok/gds-nvidia-fs/src$ /usr/local/cuda/gds/tools/gdscheck -p
 GDS release version: 1.11.1.6
 nvidia_fs version:  2.28 libcufile version: 2.12
 Platform: x86_64
 ============
 ENVIRONMENT:
 ============
 =====================
 DRIVER CONFIGURATION:
 =====================
 NVMe               : Unsupported
 NVMeOF             : Unsupported
 SCSI               : Unsupported
 ScaleFlux CSD      : Unsupported
 NVMesh             : Unsupported
 DDN EXAScaler      : Unsupported
 IBM Spectrum Scale : Unsupported
 NFS                : Unsupported
 BeeGFS             : Unsupported
 WekaFS             : Unsupported
 Userspace RDMA     : Unsupported
 --Mellanox PeerDirect : Disabled
 --rdma library        : Not Loaded (libcufile_rdma.so)
 --rdma devices        : Not configured
 --rdma_device_status  : Up: 0 Down: 0
 =====================
 CUFILE CONFIGURATION:
 =====================
 properties.use_compat_mode : false
 properties.force_compat_mode : false
 properties.gds_rdma_write_support : true
 properties.use_poll_mode : false
 properties.poll_mode_max_size_kb : 4
 properties.max_batch_io_size : 128
 properties.max_batch_io_timeout_msecs : 5
 properties.max_direct_io_size_kb : 16384
 properties.max_device_cache_size_kb : 131072
 properties.max_device_pinned_mem_size_kb : 33554432
 properties.posix_pool_slab_size_kb : 4 1024 16384
 properties.posix_pool_slab_count : 128 64 64
 properties.rdma_peer_affinity_policy : RoundRobin
 properties.rdma_dynamic_routing : 0
 fs.generic.posix_unaligned_writes : false
 fs.lustre.posix_gds_min_kb: 0
 fs.beegfs.posix_gds_min_kb: 0
 fs.weka.rdma_write_support: false
 fs.gpfs.gds_write_support: false
 profile.nvtx : false
 profile.cufile_stats : 0
 miscellaneous.api_check_aggressive : false
 execution.max_io_threads : 4
 execution.max_io_queue_depth : 128
 execution.parallel_io : true
 execution.min_io_threshold_size_kb : 8192
 execution.max_request_parallelism : 4
 properties.force_odirect_mode : true
 properties.prefer_iouring : false
 =========
 GPU INFO:
 =========
 GPU index 0 NVIDIA H200 NVL: Model Not Supported
 GPU index 1 NVIDIA H200 NVL: Model Not Supported
 ==============
 PLATFORM INFO:
 ==============
 Found ACS enabled for switch 0000:e2:00.0
 Found ACS enabled for switch 0000:e0:01.1
 Found ACS enabled for switch 0000:e2:01.0
 Found ACS enabled for switch 0000:e0:01.1
 IOMMU: disabled
 Nvidia Driver Info Status: Supported(Nvidia Open Driver Installed)
 Cuda Driver Version Installed:  13010
 Platform: AS -5126GS-TNRT2, Arch: x86_64(Linux 6.17.0-19-generic)
 Platform verification succeeded

Am i missing here anything , same config used to work .let me know if any other log is required

Running GDSIO on ordered mounted nvme throws before error

export CUFILE_ALLOW_COMPAT_MODE=false; export CUFILE_FORCE_COMPAT_MODE=false; export CUFILE_USE_PCIP2PDMA=true; export CUFILE_LOGGING_LEVEL=TRACE; /usr/local/cuda/gds/tools/gdsio -x 0 -d 0 -s 32K -i 4K -f /tmp/nvme/gds_file -I 1

file register error: GPUDirect Storage not supported on current file filename :/tmp/nvme/gds_file

root@S-5126GS-TNRT2:/home//alok/gds-nvidia-fs/src# cat cufile.log | grep -i error
15-04-2026 12:47:29:218 [pid=698457 tid=698457] ERROR cufio-fs:199 NVMe Driver not registered with nvidia-fs!!!
15-04-2026 12:47:29:219 [pid=698457 tid=698457] ERROR cufio-fs:199 NVMe Driver not registered with nvidia-fs!!!
15-04-2026 12:47:29:219 [pid=698457 tid=698457] ERROR cufio:297 cuFileHandleRegister error, file checks failed
15-04-2026 12:47:29:219 [pid=698457 tid=698457] ERROR cufio:339 cuFileHandleRegister error: GPUDirect Storage not supported on current file

I guess this is latest kernel so its support p2pdma hence we dont need to patch nvme kernel module and infact nvidia-fs is not even requesired

cufile.log also attached

cufile.log (317.5 KB)