Since r349169, my Rock64 has consistently panic'd whilst attaching rockchip_dwmmc1. A kernel built at r349135 works OK. The relevant output looks like: rockchip_dwmmc0: <Synopsys DesignWare Mobile Storage Host Controller (RockChip)> mem 0xff500000-0xff503fff irq 40 on ofwbus0 rockchip_dwmmc0: Hardware version ID is 270a mmc0: <MMC/SD bus> on rockchip_dwmmc0 rockchip_dwmmc1: <Synopsys DesignWare Mobile Storage Host Controller (RockChip)> mem 0xff520000-0xff523fff irq 42 on ofwbus0 rockchip_dwmmc1: Hardware version ID is 270a panic: Duplicate alloc of 0xfffffd000089cf50 from zone 0xfffffd0000817540(16) slab 0xfffffd000089cf90(0) cpuid = 0 time = 1 KDB: stack backtrace: db_trace_self() at db_trace_self_wrapper+0x28 pc = 0xffff000000535d54 lr = 0xffff0000000df10c sp = 0xffff0000000104d0 fp = 0xffff0000000106e0 db_trace_self_wrapper() at vpanic+0x18c pc = 0xffff0000000df10c lr = 0xffff000000278218 sp = 0xffff0000000106f0 fp = 0xffff000000010790 vpanic() at panic+0x44 pc = 0xffff000000278218 lr = 0xffff000000277fc8 sp = 0xffff0000000107a0 fp = 0xffff000000010820 panic() at uma_dbg_alloc+0x144 pc = 0xffff000000277fc8 lr = 0xffff0000004fa4b0 sp = 0xffff000000010830 fp = 0xffff000000010850 uma_dbg_alloc() at uma_zalloc_arg+0x9b0 pc = 0xffff0000004fa4b0 lr = 0xffff0000004f9960 sp = 0xffff000000010860 fp = 0xffff0000000108e0 uma_zalloc_arg() at malloc+0x9c pc = 0xffff0000004f9960 lr = 0xffff000000252a8c sp = 0xffff0000000108f0 fp = 0xffff000000010920 malloc() at bounce_bus_dmamem_alloc+0x4c pc = 0xffff000000252a8c lr = 0xffff000000533b64 sp = 0xffff000000010930 fp = 0xffff000000010960 bounce_bus_dmamem_alloc() at dwmmc_attach+0x5fc pc = 0xffff000000533b64 lr = 0xffff000000556f14 sp = 0xffff000000010970 fp = 0xffff0000000109e0 dwmmc_attach() at device_attach+0x3f4 pc = 0xffff000000556f14 lr = 0xffff0000002abd8c sp = 0xffff0000000109f0 fp = 0xffff000000010a40 device_attach() at bus_generic_new_pass+0x12c pc = 0xffff0000002abd8c lr = 0xffff0000002adb40 sp = 0xffff000000010a50 fp = 0xffff000000010a80 ... I've looked through all the intervening commits and don't see any smoking gun. Does anyone have any suggestions? -- Peter Jeremy
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:21 UTC