diff options
| author | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-09-21 21:08:06 -0400 |
|---|---|---|
| committer | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-09-21 21:08:06 -0400 |
| commit | 55ffe494a749cf0ebf10b3146b8c59a4ab81aa5f (patch) | |
| tree | 8e961e579950c64ccee564323257342c2862b42b /kernel/gsp_emulation.c | |
| parent | 92fa88178956b94439857b8ec97da7f29f42bf9f (diff) | |
| download | open-iov-55ffe494a749cf0ebf10b3146b8c59a4ab81aa5f.tar.gz open-iov-55ffe494a749cf0ebf10b3146b8c59a4ab81aa5f.zip | |
kernel: TLB invalidate auto-complete (MMU_INVALIDATE trigger) + fn-10 queue logging
- NV_VIRTUAL_FUNCTION_PRIV_MMU_INVALIDATE (BAR0+0x30B0, TRIGGER=bit31):
driver sets TRIGGER then spins in kgmmuCheckPendingInvalidates_TU102
polling for it to clear. Our shadow held TRIGGER forever -> nv_open_q
hard-wedged in RmInitAdapter (TLB invalidate during dmaAllocMapping).
Auto-clear the trigger bit in gsp_emu_poll_boot (instant completion).
- fn-10/47 handler logs hObjectOld + cmd_read/stat_write/drv_rp/stat_seq.
Confirmed: drv_rp tracks stat_write 1:1 (driver drains cleanly, no
lapping); frees complete; stat_seq increments per record.
- Health-check/crashcat loop (kgspHealthCheck -> crashcatEngineGetNext
CrashReport never returns NULL) is a SECONDARY symptom triggered by an
RPC failure during RmInitAdapter - under investigation.
Diffstat (limited to 'kernel/gsp_emulation.c')
| -rw-r--r-- | kernel/gsp_emulation.c | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/kernel/gsp_emulation.c b/kernel/gsp_emulation.c index bca8a52..7b3b261 100644 --- a/kernel/gsp_emulation.c +++ b/kernel/gsp_emulation.c @@ -178,6 +178,17 @@ static void gsp_emu_poll_boot(struct gsp_emu_context *ctx) u32 gsp_cpuctl = gsp_emu_shadow_read32(ctx, GSP_EMU_REG_GSP_CPUCTL); u32 sec2_cpuctl = gsp_emu_shadow_read32(ctx, GSP_EMU_REG_SEC2_CPUCTL); u32 engine = gsp_emu_shadow_read32(ctx, GSP_EMU_REG_GSP_ENGINE); + u32 mmu_inv = gsp_emu_shadow_read32(ctx, GSP_EMU_REG_MMU_INVALIDATE); + + /* TLB invalidate auto-complete: the driver sets TRIGGER then spins + * polling for it to clear (kgmmuCheckPendingInvalidates). Report + * instant completion or the driver wedges the open/init path. */ + if (mmu_inv & GSP_EMU_MMU_INVALIDATE_TRIGGER) { + gsp_emu_shadow_write32(ctx, GSP_EMU_REG_MMU_INVALIDATE, + mmu_inv & ~GSP_EMU_MMU_INVALIDATE_TRIGGER); + pr_info_ratelimited("GSP-BOOT: MMU_INVALIDATE trigger auto-cleared (val=0x%x)\n", + mmu_inv); + } /* GSP ENGINE reset asserted -> falcon rebooted, restart stage tracking */ if ((engine & 0x1) && ctx->boot_stage != GSP_BOOT_STAGE_IDLE) { |