open-iov.git/tools/guest-diag/wall2/README-session77.md, branch masterOpen-IOV vGPU-IO: host-side GSP emulation for unmodified NVIDIA guest drivers
https://git.vgpu.io/open-iov.git/atom?h=master2026-10-25T18:28:04+00:00wall2 session 111: FB FULLY disproved (whole-256MB fill_vidmem no flip -> memdesc not FB-backed); refined assist negative; SYS-fill of active neighborhoods wedged the guest (cycle next; never bulk-overwrite active-set-adjacent pages)2026-10-25T18:28:04+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:28:04+00:00urn:sha1:d529c2584c41c47a6c50f0a922e1ad805ecadd1fwall2 session 110: the sec2 loop runs CONTINUOUSLY inside one long-lived ioctl that holds the RM write lock (closes the lock model: tid 1791 read-wait, new clients D-state); the 299-byte reader is a red herring (address never mapped, 5/s); PEBS unavailable; sema-address instruments exhausted short of blind assist2026-10-25T18:22:56+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:22:56+00:00urn:sha1:6274b8015f6cc729056b9cd1254e4b28f3d850ccwall2 session 109: zero-ioctl window vs 24% kernel profile = BURSTY retry rounds between 200ms heartbeats; the 299-byte reader named as a health-poll candidate; next = name the node and capture the burst with host-RPC correlation2026-10-25T18:21:04+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:21:04+00:00urn:sha1:32a0a80e75389b3a9db44369d321e0dff35ee58bwall2 session 108: content-diff found 9 active driver pages (counter/staging semantics catalogued); the stuck loop's read is constant -> the sema is a static cell; next = bounded zero-cell sweep among the constant pages adjacent to the active set2026-10-25T18:19:26+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:19:26+00:00urn:sha1:673e114ca3ade72c7f50f0b0d4bdac6e3ae36c04wall2 session 107: FB assist sweeps negative (320+1024 cells; perf still 23.13%) - memdesc read not FB-sourced; the byte-identical stream twins (0x1026ad000/0x12afba000) are the memdesc-backing candidates for the next assist sweep2026-10-25T18:18:04+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:18:04+00:00urn:sha1:15e087e7d1f8bb02132b2e8d7b8a56dc142ff667wall2 session 106: channel buffer is driver-internal (not in NV_CHANNEL_ALLOC_PARAMS; kprobe was the right path, notrace-blocked); finish-tail sweep + auth-tag hash scan negative; next = transfer-shadow properties or bounded FB assist2026-10-25T18:16:49+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:16:49+00:00urn:sha1:7f074eb89a2d312efa6008c137ca529364da1682wall2 session 105: kprobe blocked (channelReadChannelMemdesc is notrace; tracefs mounted, UVM kprobes pre-existing); BAR1-path analysis (pokes landed in the shadow yet no flip -> memdesc page != FB 0x660b10); channel-field offsets enumerated; next = CE channel alloc-params dump for the channel-buffer field2026-10-25T18:15:44+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:15:44+00:00urn:sha1:c50f2f4e8dd377f01e5e7bf97dda173451ac6a58wall2 session 104: sema VA 0x4b19b10 verified NOT the loop page (converged FB 0x660b10 pokes landed, no flip; perf still 24.52%); kprobe on channelReadChannelMemdesc names the real authTagBufSemaOffset (offset=%si) + the channel pointer -> memdesc2026-10-25T18:14:32+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:14:32+00:00urn:sha1:8e9f0fa727a7fd5720ba5d1a351bb548f04f713bwall2 session 103: scrub init stream content-located (pa=0x114e98000 + twins); release 0x4b17e30 delivered (cur=1); auth-tag sema 0x4b19b10 page ambiguous (FB 0x660b10 garbage did not flip; linear guesses zero); next = validate-by-value walk pick2026-10-25T18:12:30+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:12:30+00:00urn:sha1:721c35c8eca61c0301f9763ab87e007c9b520d13wall2 session 102: fd sampler found no transient (bursty reader); no nvidia debugfs; memdesc paths ranked (content-locate the channel buffer by scrub method shapes / scrubber sema-push decodes / narrower burst sampling)2026-10-25T18:11:01+00:00Arthur Hanson Rasmussonarthur@vgpu.io2026-10-25T18:11:01+00:00urn:sha1:8e0da192d2ee708ff44b7e69da8c6ca8bf89e3e1