From: Sean Anderson <seanga2@gmail.com>
To: Damien Le Moal <Damien.LeMoal@wdc.com>, Anup Patel <anup@brainfault.org>
Cc: linux-riscv <linux-riscv@lists.infradead.org>,
Anup Patel <Anup.Patel@wdc.com>,
Palmer Dabbelt <palmer@dabbelt.com>,
Paul Walmsley <paul.walmsley@sifive.com>
Subject: Re: [PATCH 08/10] riscv: Add Kendryte K210 device tree
Date: Fri, 6 Mar 2020 19:18:38 -0500 [thread overview]
Message-ID: <c8197767-c76a-efc2-1fe2-250840bee605@gmail.com> (raw)
In-Reply-To: <BYAPR04MB58160D1A2B74D22332E498C1E7E70@BYAPR04MB5816.namprd04.prod.outlook.com>
On 3/2/20 12:08 AM, Damien Le Moal wrote:
> On 2020/03/02 14:05, Anup Patel wrote:
>> On Mon, Mar 2, 2020 at 10:21 AM Damien Le Moal <Damien.LeMoal@wdc.com> wrote:
>>>
>>> On 2020/03/02 13:22, Anup Patel wrote:
>>>> On Mon, Mar 2, 2020 at 9:46 AM Damien Le Moal <Damien.LeMoal@wdc.com> wrote:
>>>>>
>>>>> On 2020/03/02 13:07, Anup Patel wrote:
>>>>> [...]
>>>>>>> + sram0: memory@80000000 {
>>>>>>> + device_type = "memory";
>>>>>>> + reg = <0x80000000 0x400000>;
>>>>>>> + };
>>>>>>> +
>>>>>>> + sram1: memory@80400000 {
>>>>>>> + device_type = "memory";
>>>>>>> + reg = <0x80400000 0x200000>;
>>>>>>> + };
>>>>>>> +
>>>>>>> + kpu_sram: memory@80600000 {
>>>>>>> + device_type = "memory";
>>>>>>> + reg = <0x80600000 0x200000>;
>>>>>>> + };
>>>>>>
>>>>>> No need to have separate DT node for each RAM bank. This can be
>>>>>> express as single DT node as follows:
>>>>>>
>>>>>> sram: memory@80000000 {
>>>>>> device_type = "memory";
>>>>>> reg = <0x80000000 0x400000>,
>>>>>> <0x80400000 0x200000>,
>>>>>> <0x80600000 0x200000>;
>>>>>> };
>>>>>
>>>>> This is to match the U-boot device tree that Sean wrote. So I would rather keep
>>>>> it like this. And strictly speaking, if one wants to add a driver for the KPU,
>>>>> having the kpu memory segment for it separate makes it easy to reference it from
>>>>> a kpu device entry. But granted, the two sram segments can be declared with a
>>>>> single memory entry.
There is no clear documentation on how to do this, so I have been mostly
just trying things until they work. In U-Boot, separate memory device
nodes are treated as different "banks".
>>>>
>>>> But, that's not the preferred way of describing memory banks on the
>>>> same machine.
>>>> Usually, we create multiple memory DT nodes for NUMA systems.
>>>>
>>>> You can also refer various ARM/ARM64 DTS files.
>>>
>>> Oops... Sent an answer to this to the wrong email... Here it is again:
>>>
>>> Yes, I understand. But in the case of the K210, that last 2MB segment is really
>>> special as by default it is not usable as regular SRAM. I think it may be better
>>> to reflect that in the device tree. The K210 soc_early_init() call back can
>>> probe for that special entry too to see if it has to be turned on for use as
>>> regular memory or not (i.e. if a kpu driver will use it).
>>>
>>> Since booting Linux with 6MB of memory will be even more challenging than with
>>> 8, I agree that we may never see the case of a kpu driver being used. But I
>>> personally like making that special case clear in the device tree. No strong
>>> objection to your simplification though. So if you really object, I will go with it.
>>>
>>
>> I understand that it is helping you to distinguish last 2MB segment but this is
>> also possible using with single memory DT node as follows:
>>
>> sram: memory@80000000 {
>> device_type = "memory";
>> reg = <0x80000000 0x400000>,
>> <0x80400000 0x200000>,
>> <0x80600000 0x200000>;
>> reg-names = "sram0", "sram1", "kpu_sram";
>> };
>
> Nice trick. I did not know about it. Will use that then !
>>
>> The K210 soc_early_init() can do the following:
>> 1. Find memory DT node having device_type = "memory"
>> 2. Find bank number for "kpu_sram" based on "reg-names DT property
>> 3. Get based address of KPU SRAM from "reg" property based on bank
>> number found in step2 above.
>>
>> The reg-names is a standard DT property used to distinguish multiple
>> memory regions of device. Same can be used to distinguish multiple
>> banks of memory DT node.
>>
>> I am not adamant on having single memory DT node but just wanted
>> to let you know that this is not a preferred way for non-NUMA system.
Anup, do you have any suggestions on how to describe clocks for each
bank? I think the kpu sram may need some clock manipulation to work
properly. Perhaps something like
sram: memory@80000000 {
device_type = "memory";
reg = <0x80000000 0x400000>,
<0x80400000 0x200000>,
<0x80600000 0x200000>;
reg-names = "sram0", "sram1", "kpu_sram";
clocks = <&sysclk K210_CLK_SRAM0>,
<&sysclk K210_CLK_SRAM1>,
<&sysclk K210_CLK_PLL1>;
clock-names = "sram0", "sram1", "kpu";
};
--Sean
next prev parent reply other threads:[~2020-03-07 0:18 UTC|newest]
Thread overview: 89+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-02-12 10:34 [PATCH 00/10] Kendryte k210 SoC boards support Damien Le Moal
2020-02-12 10:34 ` [PATCH 01/10] riscv: Fix gitignore Damien Le Moal
2020-02-20 0:15 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 02/10] riscv: Force flat memory model with no-mmu Damien Le Moal
2020-02-14 20:18 ` Sean Anderson
2020-02-15 2:15 ` Damien Le Moal
2020-02-15 2:26 ` Sean Anderson
2020-02-15 2:40 ` Damien Le Moal
2020-03-02 3:48 ` Anup Patel
2020-03-04 18:38 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 03/10] riscv: Unaligned load/store handling for M_MODE Damien Le Moal
2020-03-02 3:57 ` Anup Patel
2020-03-04 19:28 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 04/10] riscv: Add BUILTIN_DTB support Damien Le Moal
2020-03-02 3:58 ` Anup Patel
2020-03-04 19:28 ` Palmer Dabbelt
2020-03-05 4:58 ` Anup Patel
2020-03-05 5:14 ` Damien Le Moal
2020-03-05 5:37 ` Anup Patel
2020-03-05 6:13 ` Damien Le Moal
2020-03-08 6:10 ` Anup Patel
2020-03-05 8:18 ` Atish Patra
2020-03-07 0:02 ` Sean Anderson
2020-03-07 1:51 ` Atish Patra
2020-03-07 2:08 ` Sean Anderson
2020-03-06 23:56 ` Sean Anderson
2020-02-12 10:34 ` [PATCH 05/10] riscv: Add SOC early init support Damien Le Moal
2020-03-04 19:28 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 06/10] riscv: Add Kendryte K210 SoC support Damien Le Moal
2020-02-14 20:31 ` Sean Anderson
2020-03-04 19:38 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 07/10] riscv: Select required drivers for Kendryte SOC Damien Le Moal
2020-03-02 3:59 ` Anup Patel
2020-03-04 19:44 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 08/10] riscv: Add Kendryte K210 device tree Damien Le Moal
2020-02-14 20:51 ` Sean Anderson
2020-02-15 2:34 ` Damien Le Moal
2020-02-15 2:48 ` Sean Anderson
2020-02-15 3:00 ` Damien Le Moal
2020-02-18 14:12 ` Carlos Eduardo de Paula
2020-02-18 14:18 ` Sean Anderson
2020-02-18 14:30 ` Carlos Eduardo de Paula
2020-02-18 17:48 ` Sean Anderson
2020-02-18 19:26 ` Carlos Eduardo de Paula
2020-02-19 9:06 ` Wladimir J. van der Laan
2020-02-19 22:28 ` Sean Anderson
2020-02-20 10:48 ` Wladimir J. van der Laan
2020-02-22 19:07 ` Wladimir J. van der Laan
2020-04-01 17:55 ` Drew Fustini
2020-04-02 2:24 ` Damien Le Moal
2020-02-19 8:50 ` Wladimir J. van der Laan
2020-02-27 19:43 ` Sean Anderson
2020-03-02 4:06 ` Anup Patel
2020-03-02 4:15 ` Damien Le Moal
2020-03-02 4:22 ` Anup Patel
2020-03-02 4:51 ` Damien Le Moal
2020-03-02 5:05 ` Anup Patel
2020-03-02 5:08 ` Damien Le Moal
2020-03-07 0:18 ` Sean Anderson [this message]
2020-03-07 4:11 ` Anup Patel
2020-03-04 19:44 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 09/10] riscv: Kendryte K210 default config Damien Le Moal
2020-03-02 4:07 ` Anup Patel
2020-03-04 19:44 ` Palmer Dabbelt
2020-02-12 10:34 ` [PATCH 10/10] riscv: create a loader.bin for the kendryte kflash.py tool Damien Le Moal
2020-03-02 4:08 ` Anup Patel
2020-03-04 19:44 ` Palmer Dabbelt
2020-02-14 15:05 ` [PATCH 00/10] Kendryte k210 SoC boards support Carlos Eduardo de Paula
2020-02-15 2:02 ` Damien Le Moal
2020-02-17 13:28 ` Carlos Eduardo de Paula
2020-02-26 21:31 ` Carlos Eduardo de Paula
2020-02-27 2:18 ` Damien Le Moal
2020-02-28 20:32 ` Sean Anderson
2020-03-02 3:01 ` Damien Le Moal
2020-03-02 3:53 ` Sean Anderson
2020-03-02 4:11 ` Damien Le Moal
2020-03-02 4:18 ` Sean Anderson
2020-03-02 4:54 ` Damien Le Moal
2020-03-02 4:56 ` Sean Anderson
2020-03-02 5:03 ` Damien Le Moal
2020-03-02 4:17 ` Anup Patel
2020-03-02 4:21 ` Sean Anderson
2020-03-02 4:48 ` Damien Le Moal
2020-03-02 4:51 ` Damien Le Moal
2020-03-02 5:02 ` Sean Anderson
2020-03-02 5:11 ` Damien Le Moal
2020-03-02 5:25 ` Sean Anderson
2020-03-02 5:43 ` Damien Le Moal
2020-03-04 19:44 ` Palmer Dabbelt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=c8197767-c76a-efc2-1fe2-250840bee605@gmail.com \
--to=seanga2@gmail.com \
--cc=Anup.Patel@wdc.com \
--cc=Damien.LeMoal@wdc.com \
--cc=anup@brainfault.org \
--cc=linux-riscv@lists.infradead.org \
--cc=palmer@dabbelt.com \
--cc=paul.walmsley@sifive.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).