On 2022-03-11 02:18, Marek Vasut wrote:
On 3/11/22 01:53, Angus Ainslie wrote:
Based off the imx8mm u-boot driver and the linux kernel driver
What does this patch do again ? (it is impossible to tell from the
one-line commit message above unless you have it back-to-back with
Subject).
Also, if these tables come from Linux, please include the upstream
linux kernel commit ID from which these tables come from, so next time
the tables need to be synchronized from Linux to U-Boot, we could only
pick the new commits from Linux since the base commit.
Ok I'll update the description.
[...]
+static const char *imx8mq_a53_core_sels[] = {"arm_a53_div",
"arm_pll_out", };
+static const char *imx8mq_a53_sels[] = {"clock-osc-25m",
"arm_pll_out", "sys_pll2_500m",
+ "sys_pll2_1000m", "sys_pll1_800m",
"sys_pll1_400m",
+ "audio_pll1_out", "sys_pll3_out", };
+
+static const char *imx8mq_ahb_sels[] = {"clock-osc-25m",
"sys_pll1_133m", "sys_pll1_800m",
+ "sys_pll1_400m", "sys_pll2_125m",
"sys_pll3_out",
+ "audio_pll1_out", "video_pll1_out", };
+
+static const char *imx8mq_enet_axi_sels[] = {"clock-osc-25m",
"sys_pll1_266m", "sys_pll1_800m",
+ "sys_pll2_250m", "sys_pll2_200m",
"audio_pll1_out",
+ "video_pll1_out", "sys_pll3_out",
};
+
+#ifndef CONFIG_SPL_BUILD
You might want to invert the logic -- ifdef CONFIG_SPL_BUILD -- and
then NOT include a lot of these clock in SPL to reduce the size of the
SPL.
I don't use it in the SPL so just mirrored what was in the imx8mm
driver.
I think I prefer to just drop them and then if there is an SPL user they
can figure out where to disable SPL clocks based on the subsystems
enabled rather than the SPL_BUILD flag.
[...]
The rest looks good to me, thanks.
Thanks can I add a Reviewed-by ?
