Re: get, prepare, enable a clock not in DT?

From: Nicolas Ferre
Date: Wed Sep 04 2024 - 13:57:03 EST


On 04/09/2024 at 16:43, Alexander Dahl wrote:
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe

Hello Claudiu, Nicolas,

Am Wed, Sep 04, 2024 at 03:26:45PM +0200 schrieb Nicolas Ferre:
On 04/09/2024 at 09:33, claudiu beznea wrote:
EXTERNAL EMAIL: Do not click links or open attachments unless you know the content is safe

Hi, Alexander,

On 02.09.2024 11:24, Alexander Dahl wrote:
Hello Claudiu,

Am Sat, Aug 31, 2024 at 06:49:59PM +0300 schrieb claudiu beznea:
Hi, Alexander,

On 28.08.2024 09:55, Alexander Dahl wrote:
Hello Claudiu,

Am Fri, Aug 23, 2024 at 05:29:44PM +0300 schrieb claudiu beznea:


On 20.08.2024 15:17, Alexander Dahl wrote:
By chance: I don't have a sama7g5 based board at hand for testing.
The datasheet says the same as for sam9x60.
Does the nvmem_microchip_otpc driver actually work without timeout on
sama7g5?

Yes! This should be because system bus is clocked from MCK0 (as mentioned
in peripheral identifiers table) which is enabled by bootloader.

Not sure I can follow. Citing the SAMA7G5 datasheet section 30.4
(OTPC Product Dependencies):

"The OTPC is clocked through the Power Management Controller
(PMC). The user must power on the main RC oscillator and enable
the peripheral clock of the OTPC prior to reading or writing the
OTP memory."

I don't see this in [1]. Only:

"The OTPC is clocked through the Power Management Controller (PMC), so the
programmer must first to configure the PMC."

From this I got that it is about the MCK0 listed in table Table 8-11.
Peripheral Identifiers.

[1]
https://ww1.microchip.com/downloads/aemDocuments/documents/MPU32/ProductDocuments/DataSheets/SAMA7G5-Series-Data-Sheet-DS60001765A.pdf

Well, this seems to be an older version revision A from 03/2022.
I have DS60001765B (revision B) from 12/2023 and got this here (note
the missing 'A' in the filename):

https://ww1.microchip.com/downloads/aemDocuments/documents/MPU32/ProductDocuments/DataSheets/SAMA7G5-Series-Data-Sheet-DS60001765.pdf

This version clearly express your findings. The unknown now is the
"peripheral clock" that need to be enabled along with the
main_rc_oscillator. For that you may want to play around with PMC
Peripheral Control Register, PMC peripheral clock status register and see
if OTPC fails to work when disabling the peripheral clock with the OTPC ID
as there is no information about peripheral clock for OTPC in the
peripheral identifers table.

The SAM9X60 seems to be different here, than the SAMA7G5.

(I only have SAM9X60 here, I can not test on SAMA7G5.)

To make it work I did not enable or disable a certain PMC peripheral
clocks on SAM9X60. The peripheral ID for OTPC is 46 on SAM9X60, and
in the Peripheral Identifiers table there's no checkmark under "PMC
Clock Control". The ID is also not part of the sam9x60 clock drivers.
This tells me SAM9X60 does not need a PMC peripheral clock enabled for
the OTPC to work and my experiments confirm this. Enabling the main
RC oscillator was sufficient.

The feedback that I have from our internal teams is that for all the products, the main RC oscillator, the OTPC peripheral clock and the MCKx clocks associated to OTP must be enabled.

I cannot tell why you don't experience this on your board. But probably reading from OTP and writing/updating/locking... operations could be fairly different in terms of clock requirements and (internal to OTPC + its associated memory) synchronisation between clock domains: so my advice is that, even though you can "use/read" OTP on sam9x60 without the other clocks, it's probably a bit fragile.

As a summary, in addition to RC oscillator, the peripheral identifier for OTP should be:
SAM9X60 : 46
SAM9X70 : 46
SAMA7G54 : 67
SAMA7D65 : 63
and the associated MCK :
SAM9X60 : MCK (needs a documentation update)
SAM9X70 : MCK
SAMA7G54 : MCK0
SAMA7D65 : MCK0

Basically the datasheet for sam9x60 hasn't been updated too recently and the same sentence as the one you noticed hereunder, on latest versions of sama7g5/sam9x7 datasheet, applies as well: "The user must power on the main RC oscillator and enable the peripheral clock of the OTPC prior to reading or writing the OTP memory".
Note that internal documentation update request is entered in our system thanks to your feedback.

Hope this helps.

FYI, I asked internally. I'll keep you posted.

Thanks Nicolas, I appreciate this.

Thanks to you and Claudiu for digging into this.

Best regards,
Nicolas

Linked here:

https://www.microchip.com/en-us/product/sama7g54

The revision history is not very specific, it only says "Updated Power
Management". Errata sheet has nothing interesting on that topic.

We both cited what we saw in the datasheets. Revision A has the
section you cited, revision B has the section I cited.

Table from section 8.5 (Peripheral Clocks …) has no check mark at "PMC
clock control" but indeed lists MCK0 as main system bus clock.

This is what I was taking about.

If it
works on SAMA7G5 without explicitly enabling main RC oscillator, then
either that clock is on accidentally, or the datasheet is wrong in the
OTPC section.

Might be.

I don't have a SAMA7G5 at hand. Someone who has could test if OTPC
works with/without MCK0, and with/without main RC osc, all possible
combinations would be most helpful: with none of those, with only one,
only the other, both.

Hope we get this clock stuff sorted out?!

Greets
Alex


Thank you,
Claudiu Beznea


Personally I find the "clocked through PMC" part in the OTPC
section suspicious, because in the peripheral identifiers table OTPC
has no "PMC Clock Control" mark.

Not sure what's the difference between SAM9X60 and SAMA7G5 internally,
though. From a user's POV it's possible one of them requires the
main RC osc, and the other does not, but currently you can't tell from
the datasheets.

Here is a snapshot of reading the NVMEM on a SAMA7G5 with bootconfig and
thermal calibration packets:
https://www.linux4sam.org/bin/view/Linux4SAM/ThermalFaq

Greets
Alex