All of lore.kernel.org
 help / color / mirror / Atom feed
From: Yao Zi <me@ziyao.cc>
To: Zhouqi Jiang <luojia@rustsbi.com>, Yao Zi <me@ziyao.cc>
Cc: u-boot@lists.u-boot-project.org, trini@konsulko.com,
	tim609@andestech.com, leo.liang@sifive.com, sjg@chromium.org,
	alpernebiyasak@gmail.com, jiangzhouqi2025@iscas.ac.cn,
	plucky@rustsbi.com
Subject: Re: [PATCH] riscv: Add RustSBI firmware support
Date: Thu, 19 Nov 2026 11:41:59 +0000	[thread overview]
Message-ID: <av7gh5SblUm5WYA1@pie> (raw)
In-Reply-To: <1388F640AFD6F7C7+20261116151554.2375371-3-luojia@rustsbi.com>

On Mon, Nov 16, 2026 at 11:15:56PM +0800, Zhouqi Jiang wrote:
> Hi Ziyao,
> 
> > Please quote the context on reply, so the mail is easier to read. I'm
> > adding Tom's comment back here.
> 
> Thanks, and sorry I omitted the context in my previous reply.
> 
> > > > To be clear, I think all of the changes to make things more clearly
> > > > generic here are good. At first read, it looks like aside from display
> > > > strings and a different scratch options value, RustSBI and OpenSBI are
> > > > interchangeable as far as U-Boot is concerned? If so, I'd like to
> > > > explore continuing down that path more.
> >
> > On Mon, Nov 16, 2026 at 01:42:33PM +0800, Zhouqi Jiang wrote:
> > > Thanks. You are right: RustSBI and OpenSBI are interchangeable here because
> > > RustSBI implements OpenSBI's FW_DYNAMIC handoff protocol. This reflects
> > > ecosystem compatibility rather than a formal RISC-V standard for
> > > inter-firmware handoff.
> >
> > So if RustSBI could be supported through OpenSBI's handoff protocol, why
> > not simply keep the current OpenSBI stuff as-is, only replace the
> > fw_dynamic.bin with RustSBI's binary?
> >
> > From the view of this patch, I don't see any improvements by separating
> > RustSBI stuff out for the current situation.
> 
> I agree. For v2, I will remove the separate SPL_RUSTSBI option and use the
> existing FW_DYNAMIC handoff. The current draft keeps RUSTSBI= as an alias
> for OPENSBI=; both use the same FIT entry and handoff. Would that alias be
> useful, or would you prefer using OPENSBI= for both?

I prefer to keep using OPENSBI=.

> > > I initially considered merging the tags, but I now prefer to keep
> > > SPL_OPENSBI and SPL_RUSTSBI separate while moving the shared handoff code into
> > > a common path. This keeps firmware-specific details, such as scratch options,
> > > explicit and leaves room for other SBI implementations.
> >
> > However, scratch options could be modified through Kconfig options
> > easily. Even with a separate Kconfig option introduced, others still
> > need to alternate it to use RustSBI-compatible scratch options.
> >
> > So I think, like what I have said before, a documentation to explain
> > support status of different SBI implementations, and
> > SPL_OPENSBI_SCRATCH_OPTIONS' effects on them, might be more appropriate.
> 
> Agreed. I have documented the supported implementations, firmware input and
> scratch options in the RISC-V architecture documentation.
> 
> In your earlier review of the original patch [1], you also wrote:
> 
> > These changes should be defnitely split into separate commits, since
> > they're related to too many different components (riscv, boot flow, spl
> > loaders, binman, and doc).
> 
> > > @@ -476,8 +476,7 @@ config SBI_V02
> > > -	  Choose this option if OpenSBI release v0.7 or above is used together
> > > -	  with U-Boot.
> > > +	  Choose this option if the SBI firmware implements SBI v0.2 or later.
> >
> > This change itself looks like a good clean-up, and could be accepted
> > independently with RustSBI support. Please split it into a separate
> > commit.
> 
> I plan to split v2 into three patches:
> 
> 1. Make the SBI_V02 Kconfig help text generic.

Makes sense.

> 2. Add the RUSTSBI= firmware input alias, binman test and missing-blob help.

Whoever, the code would be largely identical to the OpenSBI handling,
no? This is one of the reasons I dislike the idea.

> 3. Document SBI firmware support and scratch options in the RISC-V
>    architecture documentation and SPL Kconfig help.

The documentation itself should probably not stay in SPL Kconfig help
text. You could inform the reader to refer to something in doc/ instead.

> > We should probably not cover every possible SBI implementations in a
> > board-specific documentation. A better option might be explaining the
> > support status of different SBI implementations in RISC-V-specific doc
> > doc/arch/riscv.rst, and keep all the board-level ones as-is.
> 
> Agreed. I have kept the board documentation unchanged and moved the
> firmware information to doc/arch/riscv.rst.
> 
> One point I'd appreciate your and Tom's thoughts on: the existing FIT
> description and SPL debug message say OpenSBI even when the binary is
> RustSBI, although the U-Boot "sbi" command can identify the implementation
> at runtime. Would carrying a firmware name for display be useful, or should
> the existing labels remain as they are?

I prefer to keep them as-is.

> [1] https://lore.kernel.org/all/avr68Io04mE4YFs-@pie/
> 
> Thanks,
> Zhouqi

-- 
Thanks,
Yao Zi

Meow-mrrp-mrow, I'm a cat.

  reply	other threads:[~2026-11-19 11:42 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-15 14:38 [PATCH] riscv: Add RustSBI firmware support Zhouqi Jiang
2026-11-15 17:10 ` Tom Rini
2026-11-16  5:42   ` Zhouqi Jiang
2026-11-16 13:02     ` Yao Zi
2026-11-16 15:15       ` Zhouqi Jiang
2026-11-19 11:41         ` Yao Zi [this message]
2026-11-19 12:00           ` Zhouqi Jiang
2026-11-16 12:45 ` Yao Zi
2026-11-19  9:18 ` [PATCH v2 0/3] riscv: Support RustSBI through the FW_DYNAMIC handoff Zhouqi Jiang
2026-11-19 10:31   ` E Shattow
     [not found] ` <20261119091842.3062208-1-luojia@rustsbi.com>
2026-11-19  9:18   ` [PATCH v2 1/3] riscv: Describe SBI_V02 in terms of the SBI version Zhouqi Jiang
2026-11-19 10:43     ` E Shattow
2026-11-19 12:08       ` Yao Zi
2026-11-19  9:18   ` [PATCH v2 2/3] binman: Accept RUSTSBI as an alias for OPENSBI Zhouqi Jiang
2026-11-19 10:47     ` E Shattow
2026-11-19  9:18   ` [PATCH v2 3/3] riscv: Document SBI firmware support in SPL Zhouqi Jiang
2026-11-19 11:14     ` E Shattow
2026-11-19 12:05       ` Zhouqi Jiang

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=av7gh5SblUm5WYA1@pie \
    --to=me@ziyao.cc \
    --cc=alpernebiyasak@gmail.com \
    --cc=jiangzhouqi2025@iscas.ac.cn \
    --cc=leo.liang@sifive.com \
    --cc=luojia@rustsbi.com \
    --cc=plucky@rustsbi.com \
    --cc=sjg@chromium.org \
    --cc=tim609@andestech.com \
    --cc=trini@konsulko.com \
    --cc=u-boot@lists.u-boot-project.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.