On 10/30/25 4:58 AM, Alice Guo (OSS) wrote:

Hello Alice,

diff --git a/drivers/firmware/scmi/smt.c
b/drivers/firmware/scmi/smt.c index
237871559f0..cd1c0801f72 100644
--- a/drivers/firmware/scmi/smt.c
+++ b/drivers/firmware/scmi/smt.c
@@ -61,20 +61,6 @@ int scmi_dt_get_smt_buffer(struct udevice *dev,
struct scmi_smt *smt)
        if (device_is_compatible(dev, "arm,scmi") &&
ofnode_has_property(dev_ofnode(dev), "mboxes"))
                scmi_smt_enable_intr(smt, true);

-#ifdef CONFIG_ARM
-       if (dcache_status()) {
-               u32 align_size;
-
-               if (IS_ENABLED(CONFIG_ARM64))
-                       align_size = PAGE_SIZE;
-               else
-                       align_size = MMU_SECTION_SIZE;
-
-
        mmu_set_region_dcache_behaviour(ALIGN_DOWN((uintptr_t)smt->buf,
align_size),
-                                               ALIGN(smt->size, align_size), 
DCACHE_OFF);
-       }
-#endif
-
        return 0;
   }

--
2.51.0

Tested-by: Alice Guo <[email protected]>

Tested on two boards: IMX94 and IMX95. No regressions were observed
during boot. Looks good to me.
Thank you !

Hi Marek,

Applied this patch on the U-Boot NXP tree and found that it affects the 
i.MX8ULP. On i.MX8ULP, the SCMI buffer is in SRAM and mapped as normal memory.

Thank you for testing, this helps greatly !

Would it be possible to fix up the SRAM mapping in:

arch/arm/mach-imx/imx8ulp/soc.c:static struct mm_region imx8ulp_arm64_mem_map[] = {

?

--
Best regards,
Marek Vasut

Reply via email to