diff options
| author | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-10-16 14:56:36 -0400 |
|---|---|---|
| committer | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-10-16 14:56:36 -0400 |
| commit | 7d123630eac15b1dc50b67d85cdc859331ee9ffa (patch) | |
| tree | a62c94601022c8661f15a83f8ba24dfaa03b0c42 /scripts/test_cuda_workloads.sh | |
| parent | 8c0eacb6eaadaf9436d1d8bda7c7d18bcc3b2e65 (diff) | |
| download | open-iov-7d123630eac15b1dc50b67d85cdc859331ee9ffa.tar.gz open-iov-7d123630eac15b1dc50b67d85cdc859331ee9ffa.zip | |
guest-diag: fix CE-window demand-pin offset mask (& 0xFFFFF folded every 1 MB rank to offset 0)
Demand pins for 0x120700000/0x120800000/0x120a00000 all resolved to page 0 of
the first nv_vm_map_pages area (the same wrong PA), so the host read garbage
for every rank but the first and repeat clients stalled at DEVICE. Use the
full 16 MB window offset (& 0xFFFFFF); the per-area page bound rejects areas
too small. Live: r161 ran victory_mail green as clients 1 AND 2 in one boot,
then matmul and smiley - the repeat-client stall class is gone for these.
Diffstat (limited to 'scripts/test_cuda_workloads.sh')
0 files changed, 0 insertions, 0 deletions