From: Geert Uytterhoeven <geert@linux-m68k.org>
To: Eugeniu Rosca <erosca@de.adit-jv.com>
Cc: Marek Vasut <marek.vasut+renesas@gmail.com>,
Linux-Renesas <linux-renesas-soc@vger.kernel.org>,
Takeshi Kihara <takeshi.kihara.df@renesas.com>,
Michael Dege <michael.dege@renesas.com>,
Andrew_Gabbasov@mentor.com,
"George G. Davis" <george_davis@mentor.com>,
Tobias Franzen <tfranzen@de.adit-jv.com>,
Eugeniu Rosca <roscaeugeniu@gmail.com>
Subject: Re: [RFC DO-NOT-MERGE PATCH] arm64: dts: renesas: R8A77961: Add Renesas M3-W+ (M3 ES3.0) SoC support
Date: Fri, 23 Aug 2019 16:18:09 +0200 [thread overview]
Message-ID: <CAMuHMdWdObHAesUvF1BLwnEFJ6dsdpwM2yPRdUFW4D1Rp6d-tQ@mail.gmail.com> (raw)
In-Reply-To: <20190821124441.22319-1-erosca@de.adit-jv.com>
Hi Eugeniu,
Thanks for bringing this up!
On Wed, Aug 21, 2019 at 2:45 PM Eugeniu Rosca <erosca@de.adit-jv.com> wrote:
> Similar to the revision update from H3-ES1.x to H3-ES2.0, the update
> from M3-ES1.x to M3-ES3.0, in addition to fixing HW bugs/erratas, drops
> entire silicon IPs [1-2] (for cost efficiency and other reasons).
>
> However, unlike in the H3 ES1.x->ES2.0 revision update, the M3-ES3.0
> came with a new SoC id, i.e. r8a77961 (according to both [2] and
Actually R-Car H3 ES2.0 is r8a77951, while ES1.x is r8a77950.
But we ignored the fifth digit (see below).
> the updated SoC HW manual Rev.2.00 Jul 2019). The choice to allocate a
> new identifier seems to strengthen the HW differences between M3-ES1.x
> and M3-ES3.0 (as it is the case for M3N/r8a77965).
While H3 ES2.0 was an evolutionary step, obsoleting H3 ES1.x, it looks
like M3-W and M3-W+ may exist as two separate products, next to each
other.
> Given the above, there are several ways to differentiate between
> M3-ES1.x and M3-ES3.0:
>
> A. The BSP way [1]. Move/rename r8a7796.dtsi to r8a7796-es1.dtsi and
> keep using r8a7796.dtsi for M3-ES3.x.
>
> Pros:
> * Resembles commit 291e0c4994d081 ("arm64: dts: r8a7795: Add support
> for R-Car H3 ES2.0")
> * Reuses the r8a7796 (e.g. sysc, cpg-mssr) drivers for r8a77961 (i.e.
> minimizes the bring-up effort)
> Cons:
> * Deliberately diverges from the vendor documentation [2] by
> ignoring the new SoC identifier r8a77961, i.e. leading to
> inconsistencies in the names of the drivers and DTS
>
> B. The approach taken in this patch, i.e. create a brand new
> r8a77961.dtsi (similar to r8a77965.dtsi).
>
> Pros:
> * Reflects the reality documented by HW designers [2]
> * Maintains drivers/DTS naming consistency and avoids mismatch between
> documentation and code
> Cons:
> * higher bring-up effort than (A)
> * more discussion is needed on whether it makes sense to separate:
> - DTS only
> - DTS + Kconfig (ARCH_R8A77961)
> - DTS + Kconfig (ARCH_R8A77961) + drivers (sysc, cpg-mssr, other?)
>
> Comments appreciated!
When we started work on H3 ES2.0, it was considered an evolutionary step
from ES1.x, not a different SoC. We also were used to 4-digit IDs in
compatible values, as before the 5th digit was typically used to
indicate a minor difference, like a different package, or a different
ROM option. Hence we ignored the 5th digit, reused the compatible
values for H3 ES1.x, and went with soc_device_match() to differentiate,
where needed.
However, it turned out H3 ES2.0 was more like a different SoC in the
same family: it has more similarities with R-Car M3-W ES1.0 than with
R-Car H3 ES1.x. In the mean time, with the advent of R-Car D3 and M3-N,
we also got used to 5 digits. Hence in hindsight, it might have been
better if we had considered H3 ES1.x and ES2.0 to be two different
SoCs.
Given R-Car M3-W and M3-W+ may co-exist as separate SoCs, I think
approach B is the best approach, using separate DTS, compatible values,
Kconfig, and drivers, like we did for e.g. R-Car M3-N.
What do you think?
Thanks!
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
next prev parent reply other threads:[~2019-08-23 14:18 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-21 12:44 [RFC DO-NOT-MERGE PATCH] arm64: dts: renesas: R8A77961: Add Renesas M3-W+ (M3 ES3.0) SoC support Eugeniu Rosca
2019-08-23 14:18 ` Geert Uytterhoeven [this message]
2019-08-28 17:09 ` Eugeniu Rosca
2019-09-27 8:13 ` Geert Uytterhoeven
2019-09-27 8:35 ` Eugeniu Rosca
2019-08-31 8:01 ` Simon Horman
2019-09-27 8:39 ` Eugeniu Rosca
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=CAMuHMdWdObHAesUvF1BLwnEFJ6dsdpwM2yPRdUFW4D1Rp6d-tQ@mail.gmail.com \
--to=geert@linux-m68k.org \
--cc=Andrew_Gabbasov@mentor.com \
--cc=erosca@de.adit-jv.com \
--cc=george_davis@mentor.com \
--cc=linux-renesas-soc@vger.kernel.org \
--cc=marek.vasut+renesas@gmail.com \
--cc=michael.dege@renesas.com \
--cc=roscaeugeniu@gmail.com \
--cc=takeshi.kihara.df@renesas.com \
--cc=tfranzen@de.adit-jv.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).