[PATCH] common/board_r: make sure to call initr_dm() before initr_trace()

[Heinrich Schuchardt <xypron.glpk@gmx.de>]

On 11/12/20 12:18 PM, Pragnesh Patel wrote:
> Tracing need timer ticks and initr_dm() will make gd->timer and
> gd->dm_root is equal to NULL, so make sure that initr_dm() to
> call before tracing got enabled.
>
> Signed-off-by: Pragnesh Patel <pragnesh.patel@sifive.com>
> ---
>  common/board_r.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/common/board_r.c b/common/board_r.c
> index 29dd7d26d9..7140a39947 100644
> --- a/common/board_r.c
> +++ b/common/board_r.c
> @@ -693,6 +693,9 @@ static int run_main_loop(void)
>   * TODO: perhaps reset the watchdog in the initcall function after each call?
>   */
>  static init_fnc_t init_sequence_r[] = {
> +#ifdef CONFIG_DM
> +	initr_dm,
> +#endif
>  	initr_trace,
>  	initr_reloc,
>  	/* TODO: could x86/PPC have this also perhaps? */
> @@ -718,9 +721,6 @@ static init_fnc_t init_sequence_r[] = {
>  	initr_noncached,
>  #endif
>  	initr_of_live,
> -#ifdef CONFIG_DM
> -	initr_dm,
> -#endif

You are moving initr_of_live before initr_of_live. I doubt this will
work for boards that have CONFIG_OF_LIVE=y.

Can't we move initr_trace down in the code to after both initr_of_live
and initr_dm?

@Simon:
Please, advise.

Best regards

Heinrich

>  #if defined(CONFIG_ARM) || defined(CONFIG_NDS32) || defined(CONFIG_RISCV) || \
>  	defined(CONFIG_SANDBOX)
>  	board_init,	/* Setup chipselects */
>