On Fri, Apr 16, 2021 at 2:50 AM Jan Kiszka <jan.kis...@siemens.com> wrote:
> On 15.04.21 14:54, Greg Gallagher wrote: > > > > > > On Thu, Apr 15, 2021 at 8:49 AM Jan Kiszka <jan.kis...@siemens.com > > <mailto:jan.kis...@siemens.com>> wrote: > > > > Hi Greg, > > > > I debugged a bit between meetings: > > > > (gdb) > > 335 rc = dev->set_next_event((unsigned long) clc, dev); > > > > Here we crash in early boot because set_next_event is NULL: > > > > (gdb) p *dev > > $4 = {event_handler = 0xc03cc2e8 <tick_handle_periodic>, > > set_next_event = 0x0, set_next_ktime = 0x0, next_event = 10000000, > > max_delta_ns = 34359738352, min_delta_ns = 1000, mult = 268435456, > > shift = 32, > > state_use_accessors = CLOCK_EVT_STATE_ONESHOT, features = 2, > > retries = 0, set_state_periodic = 0x0, set_state_oneshot = 0x0, > > set_state_oneshot_stopped = 0xc0cd1178 <arch_timer_shutdown_virt>, > > set_state_shutdown = 0xc0cd1178 <arch_timer_shutdown_virt>, > > tick_resume = 0x0, broadcast = 0x0, suspend = 0x0, resume = 0x0, > > min_delta_ticks = 15, max_delta_ticks = 2147483647, name = > > 0xc1361bf4 "arch_sys_timer", rating = 450, > > irq = 18, bound_on = 0, cpumask = 0xc0f073f0 <cpu_bit_bitmap+4>, > > list = {next = 0xc17320cc <clockevent_devices>, prev = 0xc17320cc > > <clockevent_devices>}, owner = 0x0, ipipe_timer = 0xef7a6b40, > > ipipe_stolen = 0} > > (gdb) bt > > #0 clockevents_program_event (dev=0xef7ab340, expires=10000000, > > force=false) at ../kernel/time/clockevents.c:335 > > #1 0xc03cc610 in tick_setup_periodic (broadcast=<optimized out>, > > dev=<optimized out>) at ../kernel/time/tick-common.c:172 > > #2 tick_setup_periodic (dev=0xef7ab340, broadcast=<optimized out>) > > at ../kernel/time/tick-common.c:149 > > #3 0xc03cc8e4 in tick_check_new_device (newdev=0xef7ab340) at > > ../kernel/time/tick-common.c:370 > > #4 0xc03cb650 in clockevents_register_device (dev=0xef7ab340) at > > ../kernel/time/clockevents.c:468 > > #5 0xc03cbb54 in clockevents_config_and_register (dev=<optimized > > out>, freq=<optimized out>, min_delta=<optimized out>, > > max_delta=<optimized out>) at ../kernel/time/clockevents.c:515 > > #6 0xc0cd14f4 in __arch_timer_setup (clk=<optimized out>, > > type=<optimized out>) at ../drivers/clocksource/arm_arch_timer.c:873 > > #7 arch_timer_starting_cpu (cpu=<optimized out>) at > > ../drivers/clocksource/arm_arch_timer.c:963 > > #8 0xc03482b8 in cpuhp_invoke_callback (cpu=0, > > state=CPUHP_AP_ARM_ARCH_TIMER_STARTING, bringup=<optimized out>, > > node=<optimized out>, lastp=0x0) at ../kernel/cpu.c:173 > > #9 0xc0348c28 in cpuhp_invoke_ap_callback (node=<optimized out>, > > bringup=<optimized out>, state=<optimized out>, cpu=<optimized out>) > > at ../kernel/cpu.c:753 > > #10 cpuhp_issue_call (cpu=<optimized out>, state=<optimized out>, > > bringup=<optimized out>, node=<optimized out>) at > ../kernel/cpu.c:1660 > > #11 0xc0348f34 in __cpuhp_setup_state_cpuslocked (state=<optimized > > out>, name=<optimized out>, invoke=<optimized out>, > > startup=0xc0cd1388 <arch_timer_starting_cpu>, teardown=0xc0cd1244 > > <arch_timer_dying_cpu>, > > multi_instance=<optimized out>) at ../kernel/cpu.c:1807 > > #12 0xc0349128 in __cpuhp_setup_state > > (state=CPUHP_AP_ARM_ARCH_TIMER_STARTING, name=0xc1361b2c > > "clockevents/arm/arch_timer:starting", invoke=true, > > startup=0xc0cd1388 <arch_timer_starting_cpu>, teardown=0xc0cd1244 > > <arch_timer_dying_cpu>, > > multi_instance=false) at ../kernel/cpu.c:1836 > > #13 0xc15b1f58 in cpuhp_setup_state (teardown=<optimized out>, > > startup=<optimized out>, name=<optimized out>, state=<optimized > > out>) at ../include/linux/cpuhotplug.h:213 > > #14 arch_timer_register () at > > ../drivers/clocksource/arm_arch_timer.c:1209 > > #15 arch_timer_of_init (np=<optimized out>) at > > ../drivers/clocksource/arm_arch_timer.c:1386 > > #16 0xc15ad7cc in timer_probe () at > > ../drivers/clocksource/timer-probe.c:30 > > #17 0xc1500d80 in start_kernel () at ../init/main.c:699 > > #18 0x00000000 in ?? () > > Backtrace stopped: previous frame identical to this frame (corrupt > > stack?) > > > > > > Are we missing some config switch for this virt machine, or is there > > something else in I-pipe broken? > > > > Jan > > > > -- > > Siemens AG, T RDA IOT > > Corporate Competence Center Embedded Linux > > > > > > Let me take a quick, look I’m almost back at my desk. I’ll get back to > > you shortly. > > > > > > I compared arm_arch_timer.c to the upstream version and spotted > something which I would consider a mismerge of the I-pipe patch: > > diff --git a/drivers/clocksource/arm_arch_timer.c > b/drivers/clocksource/arm_arch_timer.c > index 6da6a12f6e2f..e6fc1c7e5fc8 100644 > --- a/drivers/clocksource/arm_arch_timer.c > +++ b/drivers/clocksource/arm_arch_timer.c > @@ -838,18 +838,7 @@ static void __arch_timer_setup(unsigned type, > BUG(); > } > > - arch_timer_check_ool_workaround(ate_match_local_cap_id, > NULL); > -#ifdef CONFIG_IPIPE > - clk->ipipe_timer = raw_cpu_ptr(&arch_itimer); > - if (arch_timer_uses_ppi == ARCH_TIMER_VIRT_PPI) { > - clk->ipipe_timer->irq = > arch_timer_ppi[ARCH_TIMER_VIRT_PPI]; > - clk->ipipe_timer->ack = arch_itimer_ack_virt; > - } else { > - clk->ipipe_timer->irq = > arch_timer_ppi[ARCH_TIMER_PHYS_SECURE_PPI]; > - clk->ipipe_timer->ack = arch_itimer_ack_phys; > - } > - clk->ipipe_timer->freq = arch_timer_rate; > -#endif > + clk->set_next_event = sne; > } else { > clk->features |= CLOCK_EVT_FEAT_DYNIRQ; > clk->name = "arch_mem_timer"; > > With that, I'm able to boot now. > > Please double-check that file. I'll throw that hunk into CI now. > > Jan > > -- > Siemens AG, T RDA IOT > Corporate Competence Center Embedded Linux That looks like the issue, I’ll apply this the 5.4.y Thanks Greg > >
