Supermicro_arm64_1G_Switch_G3748_Added: Patch of dts file to support …

[SelvaSupermicro]

This patch is to support new platform, Supermicro 1G arm64 switch G3748 which uses marvel AC5X ASIC.
This patch is applied for all arm64 builds and will be effective only for this new platform.

Testing:

1. Tested running M0/PTF32 test suites for this new platform.
2. Verified image partitions
3. Verified uboot and sonic platform functionalities

[saiarcot895]

Has/will this been upstreamed?

[SelvaSupermicro]

Hello Saikrishna Yes, once the PR is reviewed and approved it should automatically get merged to master (upstream) I am sure you know that process.So I really don’t completely understand your question. Please help me understand I would be glad to respond.

[paulmenzel]

Thank you for your contribution.

PS: If you reply via email, please remove the cited part, otherwise the comment in the GitHub Web interface is “ugly” to read.

[saiarcot895]

Yes, once the PR is reviewed and approved it should automatically get merged to master (upstream)
I am sure you know that process.So I really don’t completely understand your question.
Please help me understand I would be glad to respond.

Hi @SelvaSupermicro,

By upstream, I mean submitting this change into the main Linux kernel. See their contribution page for more details.

[SelvaSupermicro]

Yes, once the PR is reviewed and approved it should automatically get merged to master (upstream)
I am sure you know that process.So I really don’t completely understand your question.
Please help me understand I would be glad to respond.

Hi @SelvaSupermicro,

By upstream, I mean submitting this change into the main Linux kernel. See their contribution page for more details.

Thanks for the clarifications. These code changes are applicable only for SONiC. Hence we dont want to upstream to main Linux kernel.

[paulmenzel]

From your commit message:

hence will not be upstreamed to main ONL repo

SONiC and ONL are two separate projects for a network operating system. No idea, why you reference ONL.

To ease maintenance of the SONiC Linux kernel repository, you have to sent these patches upstream. Often upstream is also giving you a high quality review, that the SONiC developers/maintainers are not capable of.

Also the devicetree has a lot of commented out lines, that also need to be cleaned up.

[SelvaSupermicro]

From your commit message:

hence will not be upstreamed to main ONL repo

SONiC and ONL are two separate projects for a network operating system. No idea, why you reference ONL.

To ease maintenance of the SONiC Linux kernel repository, you have to sent these patches upstream. Often upstream is also giving you a high quality review, that the SONiC developers/maintainers are not capable of.

Also the devicetree has a lot of commented out lines, that also need to be cleaned up.

From your commit message:

hence will not be upstreamed to main ONL repo

SONiC and ONL are two separate projects for a network operating system. No idea, why you reference ONL.

To ease maintenance of the SONiC Linux kernel repository, you have to sent these patches upstream. Often upstream is also giving you a high quality review, that the SONiC developers/maintainers are not capable of.

Also the devicetree has a lot of commented out lines, that also need to be cleaned up.

Sorry for the ONL reference. I meant to say Linux. Anyway, I have cleaned up the patch to include file that was already committed instead creating new..
Why I don't want to upstream to Linux repo,

1. As you can see this patch file depends on many other patches. The other patch files all are in SONiC repo and not in main Linux repo. So they cant apply my patch as it is invalid in that repo.
2. Our patch is applicable only for our switch platform running SONiC. So it applicable only for SONiC. It has no value in main Linux.

[prgeor]

@SelvaSupermicro please resolve the conflict

[paulmenzel]

How is it changed?

[ErichYen]

We use different SPI flash partitions from what is used in Marvell RD-AC5X dts file.
Supermicro SSE-G3748 smc_sse-g3748.dts uses:
partition@0 {
label = "spi_flash_part0";
reg = <0x0 0x200000>;
};

            parition@1 {
                    label = "spi_flash_part1";
                    reg = <0x200000 0x10000>;
            };

            parition@2 {
                    label = "spi_flash_part2";
                    reg = <0x210000 0xC00000>;
            };

            parition@3 {
                    label = "spi_flash_part3";
                    reg = <0xE10000 0x1F0000>;
            };

Marvell RD-AC5X ac5-98dx35xx-rd.dts uses:
partition@0 {
label = "spi_flash_part0";
reg = <0x0 0x800000>;
};

            parition@1 {
                    label = "spi_flash_part1";
                    reg = <0x800000 0x700000>;
            };

            parition@2 {
                    label = "spi_flash_part2";
                    reg = <0xF00000 0x100000>;
            };

[saiarcot895]

Sorry for the delay here, could you rebase your changes on top of the master branch, or merge in the current master branch into your branch? This is to resolve the patch conflict in patch/series.

[SelvaSupermicro]

Sorry for the delay here, could you rebase your changes on top of the master branch, or merge in the current master branch into your branch? This is to resolve the patch conflict in patch/series.

Hi Saikrishna,
Thanks for the comment. Now, I have updated patch so that it can be merged with the latest master.
Let me know if that is all good!

[paulmenzel]

Please improve the merge/pull request summary/title.

[ErichYen]

Hello SaiKrishna,
We may need your approve on @SelvaSupermicro new update here?
4aadc91

[saiarcot895]

Looks like there's one conflict somehow?

[saiarcot895]

Please improve the merge/pull request summary/title.

Echoing this, particularly the title.

[saiarcot895]

There's too many changes here given that only one patch file being added in this PR. Please recheck the merge conflict resolution