#asahi-re · 2023-03-24 07:09
eiln_and++ there's prob a noisy performance counter in the "ane%d" range, pls shut it off#asahi-re · 2023-03-24 07:09
eiln_oh and trace_device("/arm-io/error-handler", False)#asahi-re · 2023-03-24 07:09
eiln_also curious how ane1/ane3 is handled (if at all) bc strings suggest they're probed?#asahi-re · 2023-03-24 07:09
eiln_substituting with whatever's available#asahi-re · 2023-03-24 07:09
eiln_trace_range(irange(pPMUAddress, 0x38), mode=TraceMode.SYNC, name="ane%d-pd")#asahi-re · 2023-03-24 07:09
eiln_trace_device("/arm-io/dart-ane%d", mode=TraceMode.SYNC)#asahi-re · 2023-03-24 07:09
eiln_trace_device("/arm-io/ane%d", mode=TraceMode.SYNC)#asahi-re · 2023-03-24 07:09
eiln_p.pmgr_adt_clocks_enable("/arm-io/dart-ane%d")#asahi-re · 2023-03-24 07:08
eiln_p.pmgr_adt_clocks_enable("/arm-io/ane%d")#asahi-re · 2023-03-24 07:08
eiln_i need#asahi-re · 2023-03-24 07:08
eiln_ane3: 0x228e684000#asahi-re · 2023-03-24 07:08
eiln_ane2: 0x228e08c000#asahi-re · 2023-03-24 07:08
eiln_ane1: 0x28e684000#asahi-re · 2023-03-24 07:08
eiln_ane0: 0x28e08c000#asahi-re · 2023-03-24 07:08
eiln_the above is equivalent for all h13 except the 'pPMUAddress' line, which is#asahi-re · 2023-03-24 07:08
eiln_false == GetPowerStatus(PowerDomain_Id_H13_ANE_SET4)#asahi-re · 2023-03-24 07:08
eiln_false == GetPowerStatus(PowerDomain_Id_H13_ANE_SET3)#asahi-re · 2023-03-24 07:08
eiln_false == GetPowerStatus(PowerDomain_Id_H13_ANE_SET2)#asahi-re · 2023-03-24 07:08
eiln_false == GetPowerStatus(PowerDomain_Id_H13_ANE_SET1)#asahi-re · 2023-03-24 07:08
eiln_false == GetPowerStatus(PowerDomain_Id_H13_ANE_BASE)#asahi-re · 2023-03-24 07:07
eiln_pPMUAddress == (size_t*)(0x28e08c000)#asahi-re · 2023-03-24 07:07
eiln_it goes like:#asahi-re · 2023-03-24 07:07
eiln_but the base address is hardcoded as string literals(!) in the fw#asahi-re · 2023-03-24 07:07
eiln_so the power domains aren't described in the dt#asahi-re · 2023-03-24 07:07
eiln_for reference i rebased with the t8103 nodes https://github.com/eiln/linux/tree/ane-dt#asahi-re · 2023-02-05 07:03
eiln_thanks, testing it rn#asahi-re · 2023-02-05 07:02
eiln_also not in dt :D#asahi-re · 2023-02-05 07:01
eiln_ISP has 12 of those that work very similarly #asahi-re · 2023-02-05 07:01
eiln_ive dumped these by triggering a fault: rPMGR_ANE0_BASE_PS, rPMGR_ANE0_SET1_PS, rPMGR_ANE0_SET2_PS, rPMGR_ANE0_SET3_PS, rPMGR_ANE0_SET4_PS #asahi-re · 2023-02-05 06:59
eiln_yeah, they're handled as a whole. turned on in reverse and off in nornal order, but doesn't really matter#asahi-re · 2023-02-05 06:56
eiln_and 0th is named POWER_BASE#asahi-re · 2023-02-05 06:56
eiln_**i forget#asahi-re · 2023-02-05 06:56
eiln_marcan: in re to be safe, they're logged as ANE_POWER_SET_N (or PMGR_SET if forget) where N is 1-6