Hi Heinrich, On Sat, 20 Dec 2025 at 18:59, Heinrich Schuchardt <[email protected]> wrote: > > qemu-riscv64_smode_defconfig with function tracing enables ends in an > endless loop in notify_dynamic. When iterating over state->spy_head, > list_for_each_entry_safe() does not discover the end of the list. > > The reason is that the address of state->spy_head has been relocated but > state->spy_head.next and state>spy_head.prev still point to the unrelocated > address. Therefore list_for_entry_safe() does not discover the end of the > list. > > We cannot start tracing before initr_reloc().
There is a step missing here...why is initr_reloc() sending an event? Also it isn't clear what this has to do with tracing. > > When placing INITCALL(initr_trace) directly after INITCALL(initr_reloc) the > timer driver panics with > > Could not initialize timer (err -11). > > Tracing uses the timer driver to add timestamps. The driver model needs > to be initialized to use the timer driver. > > INITCALL(initr_trace) needs to be placed after INITCALL(initr_dm) in > initcall_run_r() to let the timer driver work properly. > > Signed-off-by: Heinrich Schuchardt <[email protected]> > --- > common/board_r.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/common/board_r.c b/common/board_r.c > index 76f9fc090fb..2b8f5a423a6 100644 > --- a/common/board_r.c > +++ b/common/board_r.c > @@ -604,7 +604,6 @@ static void initcall_run_r(void) > * Please do not add logic to this function (variables, if (), etc.). > * For simplicity it should remain an ordered list of function calls. > */ > - INITCALL(initr_trace); > INITCALL(initr_reloc); > INITCALL(event_init); > /* TODO: could x86/PPC have this also perhaps? */ > @@ -635,6 +634,7 @@ static void initcall_run_r(void) > #if CONFIG_IS_ENABLED(DM) > INITCALL(initr_dm); > #endif > + INITCALL(initr_trace); No, this is far too late. > #if CONFIG_IS_ENABLED(ADDR_MAP) > INITCALL(init_addr_map); > #endif > -- > 2.51.0 > Regards, Simon
