[i.MX93] temperature grade is wrong

[yoooh8668]

Hi,

I found the imx93 (rev.beta) does not match the temperature grade as it should be.

Here is the picture and its uboot log:

REV.beta

U-Boot 2023.04-00263-g7c7dd0dc2b0 (Apr 19 2024 - 15:50:35 +0800)

CPU:   i.MX93(52) rev1.0 1700 MHz (running at 1692 MHz)
CPU:   Consumer temperature grade (0C to 95C) at 51C

REV.GA

U-Boot 2023.04-00263-g7c7dd0dc2b0 (Apr 19 2024 - 15:50:35 +0800)

CPU:   i.MX93(52) rev1.1 1700 MHz (running at 1692 MHz)
CPU:   Consumer temperature grade (0C to 95C) at 35C

and here is the reference picture:

I'm afraid that the detect code in uboot is not correct.

//arch/arm/mach-imx/imx9/soc.c
u32 get_cpu_temp_grade(int *minc, int *maxc)
{
        int ret;
        u32 val;
        ret = fuse_read(2, 3, &val);
        if (ret)
                val = 0; /* If read fuse failed, return as blank fuse */

        val >>= 4;
        val &= 0x3;

        if (minc && maxc) {
                if (val == TEMP_AUTOMOTIVE) {
                        *minc = -40;
                        *maxc = 125;
                } else if (val == TEMP_INDUSTRIAL) {
                        *minc = -40;
                        *maxc = 105;
                } else if (val == TEMP_EXTCOMMERCIAL) {
                        if (is_imx93()){
                                /* imx93 only has extended industrial*/
                                *minc = -40;
                                *maxc = 125;
                        } else {
                                *minc = -20;
                                *maxc = 105;
                        }
                } else {
                        *minc = 0;
                        *maxc = 95;
                }
        }
        return val;
}

 THANKS~~

[opedis]

In the picture, the Q(Qualification leve) is "P".

Maybe that is problem. It is just a samples. Not a mass production.

[yoooh8668]

I think so, but wait for official reply.

[JorgeCas]

Hello,

That is strange since the part number indicates that correspond to an industrial part number.

Is this happening on other REV.beta devices?

Which are the software versions that you are using?

From what I have understood, the detect code in U-boot works as expected. It would be too strange if fuse values of device are wrong.

Best regards.

[yoooh8668]

Hi,
There are 4 piece of this sample(rev.beta) , all of them show the same result.

The branch is v2023.04_6.1.55_2.2.0 with our commits on it.
And there is no modifies on `arch/arm/mach-imx/imx9/soc.c` .

Maybe only the mass production can support this function(reg)?
Thanks~~

[JorgeCas · Accepted solution]

Hello,

The last release that is supported on A0 revision is Linux 6.1.22_2.0.0.

But I confirmed that behavior on our EVK. This could be caused by placeholder on the code, or this feature could not be implemented on the pre-production revision.

Best regards.