On 01/03/21, Stefano Babic wrote:
Hi Oliver,
I have two open points with the patch, that I want finally merge:
On 19.02.21 13:34, Oliver Graute wrote:
case 0: + /* onboard eMMC */ +if (!imx8_power_domain_lookup_name("conn_sdhc0", &pd)) + power_domain_on(&pd); + + imx8_iomux_setup_multiple_pads(emmc0, ARRAY_SIZE(emmc0)); + init_clk_usdhc(0);
I hit this becuase the compiler is raising a warning due to implicit decalaration, but something more is coming out. Linker is ok, but there is no init_clk_usdhc() for this SOC, and the default "weak" function should be called, but this does nothing. So is everything ok ? Or what am I missing ?
I'm not sure about this. Perhaps it can use the implementation from imx8m/clock_imx8mm.c or there is a implementation needed explicit for imx8qm? @Peng can you comment this?
Until further notice I would just add these prototypes to fix the warnings:
+void init_clk_usdhc(u32 index); +int fsl_esdhc_initialize(struct bd_info *bis, struct fsl_esdhc_cfg *cfg);
- env_set("sec_boot", "no");
+#ifdef CONFIG_AHAB_BOOT
- env_set("sec_boot", "yes");
+#endif
+#ifdef CONFIG_ENV_IS_IN_MMC
- board_late_mmc_env_init();
You moved this function in common, but prototype is missing, and then a warning is raised.
I will just add the prototype to fix it:
+void board_late_mmc_env_init(void);
thx for your review efforts.
Best regards,
Oliver