[Letux-kernel] [PATCH 00/20] A bunch of JZ4730 fixups for letux-kernel

[H. Nikolaus Schaller]

Hi Paul,

> Am 18.12.2020 um 21:40 schrieb Paul Boddie <paul at boddie.org.uk>:
> 
> Nikolaus,
> 
> On Friday, 18 December 2020 10:19:14 CET H. Nikolaus Schaller wrote:
>> 
>> Mainly here are bigger diffs in gpio, almost none in dmac (since it is not
>> used) and some in msc:
>> 10021008
> 
> This is the clock rate setting which the new kernel has set to 0x6, meaning 
> 1/64 of the input clock, whereas the old kernel has it set to 0x0, meaning the 
> same as the input clock.

Ok.

> Here, we might need to know what the input clock is set to, and this doesn't 
> seem to be in the register dump.

I have made a new dump including more interesting secions like cpm, intc, ost,
rtc, wdt, i2c. I am not sure if all of them are initialized when mmc is
probed, but it is likely since looking for mmc is the last step before /sbin/init.

> The CPM registers start at 0x10000000 with 
> CFCR which contains the MSC bit (24) determining whether the input clock is 
> 24MHz or 16MHz. I can't imagine it being useful to divide this clock unless 
> SDIO is being used.
> 
> At 0x10000020, MSCR contains the MSC clock gating bit (13) for the peripheral 
> which must be clear for the clock to be propagated. Investigating this, my 
> clock driver might be missing the multiplexer for CFCR.MSC, so I have attached 
> a patch to remedy this.

With your patch registers 0x10000020 are now equal.

> 
>> 1002102c
>> 10021030
> 
> These are the command index and argument registers. I will aim to add more 
> details to the MMC page in the coming days.

This time there are more differences. Anyways shouldn't they be the same
right after completing probe? No communication should have taken place and
the controller should be idling here. Unless some other thread has already
started.

> 
>> BTW: thank you very much for your memory map in our wiki! I took base
>> address and range from there to read out the gpio, msc, dmac sections. If
>> anything else is missing it is just a matter of minutes to repeat.
>> 
>> So I think I have now a tool setup to compare working and work-in-progress
>> kernels.
> 
> It should be helpful. Thanks for uploading some data, mostly to get me to 
> think about the issues in a more structured way. Certainly, the fix I have 
> made with my patch will be necessary, even if it isn't sufficient, to get this 
> working.

If I remember correctly it has a positive effect but the card never left
busy state remains. Are there any gpio or other cgu or even intc settings
involved?

Attached are the new register dumps.

BR and thanks,
Nikolaus

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: regs-2.6.txt
URL: <http://lists.goldelico.com/pipermail/letux-kernel/attachments/20201219/4ddacf7b/attachment.txt>
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: regs-5.10.txt
URL: <http://lists.goldelico.com/pipermail/letux-kernel/attachments/20201219/4ddacf7b/attachment-0001.txt>