Date:   Wed, 14 Dec 2022 13:53:54 -0800
From:   Saravana Kannan <saravanak@...gle.com>
To:     Miquel Raynal <miquel.raynal@...tlin.com>
Cc:     Maxim Kiselev <bigunclemax@...il.com>,
        Maxim Kochetkov <fido_max@...ox.ru>,
        Thorsten Leemhuis <regressions@...mhuis.info>,
        linux-mtd@...ts.infradead.org, linux-kernel@...r.kernel.org,
        Vignesh Raghavendra <vigneshr@...com>,
        Richard Weinberger <richard@....at>,
        Rafał Miłecki <rafal@...ecki.pl>,
        "regressions@...ts.linux.dev" <regressions@...ts.linux.dev>,
        Rob Herring <robh+dt@...nel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        "Rafael J. Wysocki" <rafael@...nel.org>
Subject: Re: nvmem-cells regression after adding 'call of_platform_populate()
 for MTD partitions'

On Tue, Dec 13, 2022 at 8:54 AM Miquel Raynal <miquel.raynal@...tlin.com> wrote:
>
> Hi Maxim,
>
> bigunclemax@...il.com wrote on Tue, 13 Dec 2022 14:02:34 +0300:
>
> > I looked closer at commit 658c4448bbbf and bcdf0315a61a, 5db1c2dbc04c16 commits.
> > Looks like we have two different features binded to one property - "compatible".
> >
> > From one side it is the ability to forward the subnode of the mtd
> > partition to the nvmem subsystem (658c4448bbbf and ac42c46f983e).
> > And from another side is the ability to use custom initialization of
> > the mtd partition (bcdf0315a61a and 5db1c2dbc04c16).
> >
> > What I mean:
> > According to ac42c46f983e I can create DT like this:
> >  - |
> >     partitions {
> >         compatible = "fixed-partitions";
> >         #address-cells = <1>;
> >         #size-cells = <1>;
> >
> >         partition@0 {
> >             compatible = "nvmem-cells";
> >             reg = <0x40000 0x10000>;
> >             #address-cells = <1>;
> >             #size-cells = <1>;
> >             macaddr_gmac1: macaddr_gmac1@0 {
> >                 reg = <0x0 0x6>;
> >             };
> >         };
> >     };
> >
> >
> > And according to 5db1c2dbc04c16 I can create DT like this:
> >  - |
> >     partitions {
> >         compatible = "fixed-partitions";
> >         #address-cells = <1>;
> >         #size-cells = <1>;
> >
> >         partition@0 {
> >             compatible = "u-boot,env";
> >             reg = <0x40000 0x10000>;
> >         };
> >     };
> >
> > But I can not use them both, because only one "compatible" property allowed.
> > This will be incorrect:
> >  - |
> >     partitions {
> >         compatible = "fixed-partitions";
> >         #address-cells = <1>;
> >         #size-cells = <1>;
> >
> >         partition@0 {
> >             compatible = "u-boot,env";  # from ac42c46f983e
> >             compatible = "nvmem-cells"; # from 5db1c2dbc04c
>
> What about:
>
>               compatible = "u-boot,env", "nvmem-cells";
>
> instead? that should actually work.
>
> >             reg = <0x40000 0x10000>;
> >             #address-cells = <1>;
> >             #size-cells = <1>;
> >             macaddr_gmac1: macaddr_gmac1@0 {
> >                 reg = <0x0 0x6>;
> >             };
> >         };
> >     };
> >
> > > compatible: Duplicate property name
> >
> > вт, 13 дек. 2022 г. в 12:46, Miquel Raynal <miquel.raynal@...tlin.com>:
> > >
> > > Hi Maxim,
> > >
> > > fido_max@...ox.ru wrote on Mon, 12 Dec 2022 20:57:49 +0300:
> > >
> > > > Hi, Miquel!
> > > >
> > > > On 12.12.2022 19:37, Miquel Raynal wrote:
> > > >
> > > > > Let me try to recap the situation for all the people I just involved:
> > > > >
> > > > > * An Ethernet driver gets its mac address from an nvmem cell. The
> > > > >    Ethernet controller DT node then has an "nvmem-cells" property
> > > > >    pointing towards an nvmem cell.
> > > > > * The nvmem cell comes from an mtd partition.
> > > > > * The mtd partition is flagged with a particular compatible
> > > > >    (which is also named "nvmem-cells") to tell the kernel that the node
> > > > >    produces nvmem cells.
> > > > > * The mtd partition itself has no driver, but is the child node of a
> > > > >    "partitions" container which has one (in this case,
> > > > >    "fixed-partitions", see the snippet below).
> > > > >
> > > > > Because the "nvmem-cells" property of the Ethernet node points at the
> > > > > nvmem-cell node, the core create a device link between the Ethernet
> > > > > controller (consumer) and the mtd partition (producer).
> > > > >
> > > > > The device link in this case will never be satisfied because no driver
> > > > > matches the "nvmem-cells" compatible of the partition node.
> > > > >
> > > > > Reverting commit bcdf0315a61a ("mtd: call of_platform_populate() for MTD
> > > > > partitions") would IMHO not make much sense, the problem comes from the
> > > > > device link side and even there, there is nothing really "wrong",
> > > > > because I really expect the mtd device to be ready before the
> > > > > Ethernet controller probe, the device link is legitimate.
> > > > >
> > > > > So I would like to explore other alternatives. Here are a bunch of
> > > > > ideas, but I'm open:
> > > >
> > > > How about to create simple driver with compatible="nvmem-cell" and to move all the suff from main mtd driver which serves nvmem-cell to the probe function?
> > >
> > > This is probably worth the try but I doubt you can make it work without
> > > regressions because IIRC the nvmem registration happens no matter the
> > > compatible (not mentioning the user-otp and factory-otp cases). You can
> > > definitely try this out if you think you can come up with something
> > > though.
> > >
> > > But I would like to hear from the device-link gurus :) because even if
> > > we fix mtd with a "trick" like above, I guess we'll very likely find
> > > other corner cases like that and I am interested in understanding the
> > > rationale of what could be a proper fix.
> > >

Responding to the whole thread.

I'm going by Miquel's first email in which he cc'ed me and haven't
actually looked at the mtd code. Couple of comments:

Independent of mtd/nvmem-cell, I generally frown on having a
compatible string for a child node that you don't treat as a device.
Even more so if you actually create a struct device for it and then
don't do anything else with it. That's just a waste of memory. So, in
general try to avoid that in the future if you can.

Also, there are flags the parent device's driver can set that'll tell
fw_devlink not to treat a specific DT node as a real device. So, if we
really need that I'll dig up and suggest a fix.

Lastly and more importantly, I've a series[1] that stops depending on
the compatible property for fw_devlink to work. So it should be
smarter than it is today. But that series has known bugs for which I
gave test fixes in that thread. I plan to make a v2 of that series
with that fix and I'm expecting it'll fix a bunch of fw_devlink
issues.

Feel free to give v1 + squashing the fixes a shot if you are excited
to try it. Otherwise, I'll try my best to get around to it this week
(kinda swamped though + holidays coming up, so no promises).

[1] - https://lore.kernel.org/lkml/20220810060040.321697-1-saravanak@google.com/

Thanks,
Saravana

Powered by blists - more mailing lists