u-boot branches

u-boot branches

[Joel Stanley]

As Aspeed develop their ast2600 u-boot features, we've had a need to
integrate changes from their tree into the openbmc tree. In the past
we lined this up with a major version bump (as with the kernel),
meaning we could create a new branch with our "out of tree" changes
applied on top. However, in this case their SDK is still based on the
2019.04 release, so there's not a clear way forward.

I've chosen to rebase the openbmc changes on top of their branch. The
old branch is still accessible at archive/v2019.04-aspeed-openbmc-1.
The bitbake recipe will continue to work as it uses the nobranch=1
flag, and specifies the version based on the SHA.

The new branch contains their aspeed-master-v2019.04 tree as of today,
with the OpenBMC changes applied on top.

I've boot tested it running from SPI NOR on AST2600 A1 hardawre, as
well as QEMU, and it reached userspace. I will bump the bitbake recipe
to integrate the changes into openbmc.

Eddie, can you see what changes are required on top of this branch to
support eMMC boot, and rebase your patch series? Please include an
emmc boot defconfig based on ast2600_openbmc_spl_defconfig.

Cheers,

Joel

Re: u-boot branches

[Michael Richardson]

[-- Attachment #1.1: Type: text/plain, Size: 512 bytes --]


Joel Stanley <joel@jms.id.au> wrote:
    > I've chosen to rebase the openbmc changes on top of their branch. The
    > old branch is still accessible at archive/v2019.04-aspeed-openbmc-1.
    > The bitbake recipe will continue to work as it uses the nobranch=1
    > flag, and specifies the version based on the SHA.

    > The new branch contains their aspeed-master-v2019.04 tree as of today,
    > with the OpenBMC changes applied on top.

Is this any closer to upstream u-boot?
Or is it diverging further?


[-- Attachment #1.2: Signature --]
[-- Type: text/plain, Size: 241 bytes --]

--
]               Never tell me the odds!                 | ipv6 mesh networks [
]   Michael Richardson, Sandelman Software Works        |    IoT architect   [
]     mcr@sandelman.ca  http://www.sandelman.ca/        |   ruby on rails    [


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 487 bytes --]

Re: u-boot branches

[Dan Zhang]

[-- Attachment #1: Type: text/plain, Size: 2017 bytes --]

We have two options to pick up changes from Aspeed and upstream.
1. Be closer to Aspeed temporary:
 Rebase OpenBMC u-boot changes to aspeed-master-v2019.04, with CIT ( daily
or at least weekly). After the aspeed-master-v2019.04 activity slowed down,
which means the u-boot for AST2600 from Aspeed SDK is getting stable, we
start the effort to upstream the changes.

2. Be closer to upstream:
OpenBMC takes yocto u-boot branch, submit to OpenBMC u-boot branch,
will target upstream immediately.


option-1: get changes from Aspeed is easy, but in this case we need some
important feature / patch from upstream, we need back ports.

option-2: get changes from Aspeed will need more effort than just a simple
cherry pick.

Considering Aspeed is still developing both the AST2600 chip and SDK, I
would like to vote for option-1, as now, there will be much more changes
needed from Aspeed instead of from upstream.

BRs
Dan Zhang

>
>
> ---------- Forwarded message ----------
> From: Michael Richardson <mcr@sandelman.ca>
> To: Joel Stanley <joel@jms.id.au>, OpenBMC Maillist <
> openbmc@lists.ozlabs.org>, Eddie James <eajames@linux.ibm.com>
> Cc:
> Bcc:
> Date: Thu, 24 Sep 2020 14:02:29 -0400
> Subject: Re: u-boot branches
>
> Joel Stanley <joel@jms.id.au> wrote:
>     > I've chosen to rebase the openbmc changes on top of their branch. The
>     > old branch is still accessible at archive/v2019.04-aspeed-openbmc-1.
>     > The bitbake recipe will continue to work as it uses the nobranch=1
>     > flag, and specifies the version based on the SHA.
>
>     > The new branch contains their aspeed-master-v2019.04 tree as of
> today,
>     > with the OpenBMC changes applied on top.
>
> Is this any closer to upstream u-boot?
> Or is it diverging further?
>
> --
> ]               Never tell me the odds!                 | ipv6 mesh
> networks [
> ]   Michael Richardson, Sandelman Software Works        |    IoT
> architect   [
> ]     mcr@sandelman.ca  http://www.sandelman.ca/        |   ruby on
> rails    [
>
>

[-- Attachment #2: Type: text/html, Size: 3233 bytes --]