renode(TEST-PIX-019): wire the real NXP MU (IPCF doorbell) on the M7↔M4 link#105
Merged
Conversation
…->M4 link, replacing the stub "do the doorbell": the M7<->M4 link now carries the real NXP MU model (Miscellaneous.IMXRT700_MessagingUnit — same MU IP family as RT1176) instead of a plain-memory stub: per-core instances (aInstance=cpu_m7 @0x40C4C000, bInstance= cpu_m4 @0x40C4D000) + the cross-core doorbell IRQ lines (AInstanceIRQ->nvic_m7, BInstanceIRQ->nvic_m4). This models the NXP IPCF/RPMsg mechanism faithfully: MU = the doorbell, the SHMEM ring @0x20400000 = the RPMsg vring/payload (the relay-bus carrier relay#177 maps onto exactly this on the on-die link). 3-core harness STILL PASSES with the MU wired (M7 banner + M4 SHMEM heartbeat, no regression). HONEST: an interrupt-DRIVEN doorbell (firmware drives the MU's two-sided TR/RX-IE protocol) needs the RT1170/RT700 MU register map (TR offset, RX-IE bit, IRQ number) — not extractable from this Renode AOT build (no model source, no robot exercising it), so NOT blind-guessed. Asked gale for the MU register access + the gust ipc-rx consumer (gale#65). Until then: doorbell hardware wired + SHMEM ring is the live link; reactive handoff stays orchestration-modeled. rivet validate (v0.18.0): PASS. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
"do the doorbell" — the M7↔M4 link now carries the real NXP MU model (
Miscellaneous.IMXRT700_MessagingUnit, same MU IP family as RT1176) instead of a plain-memory stub:aInstance=cpu_m7 @0x40C4C000,bInstance=cpu_m4 @0x40C4D000)AInstanceIRQ→nvic_m7,BInstanceIRQ→nvic_m4)This models the NXP IPCF/RPMsg mechanism faithfully (your IPCF point): MU = the doorbell, the SHMEM ring @0x20400000 = the RPMsg vring/payload — the relay-bus carrier (relay#177) maps onto exactly this on the on-die link.
3-core harness still PASSES with the MU wired (M7 banner + M4 SHMEM heartbeat, no regression).
Honest constraint: an interrupt-driven doorbell (firmware drives the MU's two-sided TR/RX-IE protocol) needs the RT1170/RT700 MU register map (TR offset, RX-interrupt-enable bit, IRQ number) — not extractable from this Renode AOT build (no model source, no robot exercising it), so I did not blind-guess and claim a working IRQ. Asked gale for the MU register access + the gust
ipc-rxconsumer (gale#65). Until then: doorbell hardware wired + SHMEM ring is the live link; the reactive handoff stays orchestration-modeled.rivet validate(v0.18.0): PASS.🤖 Generated with Claude Code