guest-diag: cdt2 mapping scan is opt-in (CDT_SCAN=1); it can wedge the client

2026-09-30T12:40:17+00:00

guest-diag: cdt2 payload scan limited to anon mappings (device-mapping reads can block)

2026-09-30T12:25:49+00:00

guest-diag: cdt2 payload scan only touches anon/device mappings (avoid SIGBUS on truncated file maps)

2026-09-30T12:08:32+00:00

guest-diag: cdt2 scans its own mappings for the payload after HtoD (DtoH mechanism probe)

2026-09-30T11:25:45+00:00

sema-svc: decode GRCE inline-data DMA (small HtoD path)

2026-09-30T10:13:26+00:00

The 256-byte cuMemcpyHtoD payload never appears in any CE push: CUDA submits it on the compute channel as a compute-class DMA whose source is the push's own inline data (LINE_LENGTH_IN=256, OFFSET_OUT=, LAUNCH_DMA=0x41, LOAD_INLINE_DATA 64 dwords, then the channel's SET_REPORT_SEMAPHORE release). The CE-shaped decoder skipped the data run (opcode 3) and no GR work ran at all. - Decode NVC5C0 GRCE methods on subchannel 1 (LINE_LENGTH_IN, LINE_COUNT, OFFSET_OUT/UPPER, LAUNCH_DMA) and the opcode-3 LOAD_INLINE_DATA run. - svc_grce_inline_write resolves the destination VA through trusted sources (pins, map, published page directories, targeted rm_mem resolve), pins the page and writes into vidmem or guest RAM; budgeted sema-grce logs name the target or report UNRESOLVED. - cdt2 test binary gains CDT_CP_PAUSE / CDT_HP_PAUSE for live inspection of the copy path; pushed by push_guest_tools.sh.

open-iov.git/tools/test/cuda_drv_test2.c, branch gr-forwarding

guest-diag: cdt2 mapping scan is opt-in (CDT_SCAN=1); it can wedge the client

guest-diag: cdt2 payload scan limited to anon mappings (device-mapping reads can block)

guest-diag: cdt2 payload scan only touches anon/device mappings (avoid SIGBUS on truncated file maps)

guest-diag: cdt2 scans its own mappings for the payload after HtoD (DtoH mechanism probe)

sema-svc: decode GRCE inline-data DMA (small HtoD path)