Hi Simon,
Subject: Re: [PATCH v1] arm: global_data: add scmi pointer
Hi Peng,
On Mon, 7 Oct 2024 at 20:59, Peng Fan peng.fan@nxp.com wrote:
Subject: Re: [PATCH v1] arm: global_data: add scmi pointer
On Sun, Oct 06, 2024 at 05:52:28PM +0800,
alice.guo@oss.nxp.com
wrote:
From: Peng Fan peng.fan@nxp.com
Add scmi device and channel pointer for i.MX9.
Signed-off-by: Peng Fan peng.fan@nxp.com Signed-off-by: Alice Guo alice.guo@nxp.com Reviewed-by: Ye Li ye.li@nxp.com
arch/arm/include/asm/global_data.h | 3 +++ 1 file changed, 3 insertions(+)
diff --git a/arch/arm/include/asm/global_data.h b/arch/arm/include/asm/global_data.h index 45401d5e3c..d04359f167 100644 --- a/arch/arm/include/asm/global_data.h +++ b/arch/arm/include/asm/global_data.h @@ -108,6 +108,9 @@ struct arch_global_data { #ifdef
CONFIG_SMBIOS
ulong smbios_start; /* Start address of SMBIOStable */
#endif +#if defined(CONFIG_SCMI_FIRMWARE) &&
defined(CONFIG_IMX9)
- struct udevice *scmi_dev;
+#endif };
#include <asm-generic/global_data.h>
But why is this only needed for i.MX9 families and not other users?
i.MX95 has some code that communicates with SCMI platform, but
no scmi
driver for those code, saying rom pass info. So we need a global scmi_dev.
Would it be OK to find the device using uclass_first_device(UCLASS_SMCI) ?
i.e. can we do without a pointer in global_data ?
It should be doable to drop the pointer. We will give a look to use the other approach.
Thanks, Peng.
If we do need it, I think it is better to have the condition just be SCMI_FIRMWARE (and live with the extra 8 bytes in global_data). We really should not have SoC-specific options in global_data, or even drivers for that matter.
Regards, Simon e