Re: [PATCH v5 1/2] dt-bindings: mtd: amlogic,meson-nand: support fields for boot ROM code

From: Arseniy Krasnov <avkrasnov@salutedevices.com>
To: Miquel Raynal <miquel.raynal@bootlin.com>
Cc: Richard Weinberger <richard@nod.at>,
	Vignesh Raghavendra <vigneshr@ti.com>,
	Rob Herring <robh+dt@kernel.org>,
	Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
	Conor Dooley <conor+dt@kernel.org>,
	Neil Armstrong <neil.armstrong@linaro.org>,
	Kevin Hilman <khilman@baylibre.com>,
	Jerome Brunet <jbrunet@baylibre.com>,
	Martin Blumenstingl <martin.blumenstingl@googlemail.com>,
	<linux-mtd@lists.infradead.org>, <devicetree@vger.kernel.org>,
	<linux-arm-kernel@lists.infradead.org>,
	<linux-amlogic@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>, <oxffffaa@gmail.com>,
	<kernel@sberdevices.ru>
Subject: Re: [PATCH v5 1/2] dt-bindings: mtd: amlogic,meson-nand: support fields for boot ROM code
Date: Tue, 7 May 2024 11:11:40 +0300	[thread overview]
Message-ID: <771a35e4-db15-8c4d-29e9-7a984cb34abc@salutedevices.com> (raw)
In-Reply-To: <20240507100553.31578d0d@xps-13>

On 07.05.2024 11:05, Miquel Raynal wrote:
> Hi Arseniy,
> 
> avkrasnov@salutedevices.com wrote on Tue, 7 May 2024 10:35:51 +0300:
> 
>> On 07.05.2024 10:27, Miquel Raynal wrote:
>>> Hi Arseniy,
>>>
>>> avkrasnov@salutedevices.com wrote on Tue, 7 May 2024 09:53:06 +0300:
>>>   
>>>> On 06.05.2024 16:48, Miquel Raynal wrote:  
>>>>> Hi Arseniy,
>>>>>
>>>>> avkrasnov@salutedevices.com wrote on Tue, 16 Apr 2024 11:51:00 +0300:
>>>>>     
>>>>>> Boot ROM code on Meson requires that some pages on NAND must be written
>>>>>> in special mode: "short" ECC mode where each block is 384 bytes and
>>>>>> scrambling mode is on.    
>>>>>
>>>>> Ok
>>>>>     
>>>>>> Such pages located with the specified interval within specified offset.    
>>>>>
>>>>> I'm sorry I don't get that sentence.    
>>>>
>>>> Sorry, I mean this (let me draw :) ) :
>>>>
>>>> [ page 0 ][ page 1 ][ page 2 ][ page 3 ][ page 4 ][ page 5 ][ page 6 ][ page 7 ][ page 8 ][ page 9 ]
>>>>
>>>> For example, we have 10 pages starting from the beginning of the chip - this is "within specified offset",
>>>> e.g. offset is 10. BootROM on axg needs that (for example) every third page must be written in "special"
>>>> mode: scrambling is on and ECC is 384 bytes. Such pages are 0, 2, 4, 6, 8. E.g. "specified interval" will
>>>> be 3.  
>>>
>>> Shall be 2, no?  
>>
>> yes, starting from 0 - then 2. e.g.
>> if (!(page_num % 2))
>>     boot ROM need this page
>>
>>>   
>>>>
>>>> So:
>>>>
>>>> amlogic,boot-pages: 10
>>>> amlogic,boot-page-step: 3  
>>>
>>> Ok I get it. Thanks for the explanation. I don't really understand the
>>> logic behind it though. Do you know why the bootROM would access only
>>> one page over 2 or 3? Is there a default value? Is this configurable?  
>>
>> No, boot rom source is closed, I don't have access to it. I get this logic
>> from old version of vendor's uboot - in practice they use non 2 or 3, they
>> use hardcoded 128 step value. And amlogic,boot-pages is 1024
> 
> Feels like they are trying to use only the first page of each block, no?
> 
> That's very weird but I understand better.

A little bit no, they use every 128 page in range [0, 1024] pages. E.g. there will
be 8 such pages:
0
128
256
512
640
768
896
1024

They write some metadata about SoC to such pages.

Thanks, Arseniy

> 
> Thanks,
> Miquèl

______________________________________________________
Linux MTD discussion mailing list
http://lists.infradead.org/mailman/listinfo/linux-mtd/