mail archive of the barebox mailing list
 help / color / mirror / Atom feed
From: anis chali <chalianis1@gmail.com>
To: Ahmad Fatoum <a.fatoum@pengutronix.de>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH 1/2] drivers: misc: external_state: add a barebox external state.
Date: Mon, 3 Nov 2025 09:09:20 -0500	[thread overview]
Message-ID: <CAL+1fyC5bD3k6rufJx6eYXjL4Fo00QOWRRC3JzgfcU-3maROhA@mail.gmail.com> (raw)
In-Reply-To: <e6fe90ea-3ba4-402b-9b9e-dbe2f386dc63@pengutronix.de>

>> strlen can't return negative numbers.
ack, right it is size_t (unsigned)

>> The original function had a number of safety checks that are missing here:
>>
>> - It only ran the code when in EFI payload mode, as the same barebox
>>  binary could be chainloaded without EFI
>>
>> - It checked that /boot is mounted and gave a useful error message to hint
>>  at e.g., EFIFS missing, which we are missing out on here.
>>
>> I am not yet convinced this is strictly an improvement.

Yes you are right but I didn't reintegrate the code in generic driver
since it will just use Kconfigs to reference a state dtb file.
In case this code was accepted there is no context to do the check,
but I ensured that barebox gives a behaves in way that the user
will quickly see that there is no /boot directory with the message
telling No such file or directory, anyway you suggestion to add -f to
state command
is more elegant, I will do some experimentations. the aim of this
patch is to support a state.dtb given in runtime which permits
userspace interaction with
native board because I didn't figure out how barebox state command
could communicate with a barebox intree state.

>> If you have the need to customize where your state comes from, how about adding
>> a state -f /path/to/state.dtb command option and calling it from your init script?
right, more elegant suggestion, I will try it.

Thank's for feedback

Anis.

Le lun. 3 nov. 2025 à 01:46, Ahmad Fatoum <a.fatoum@pengutronix.de> a écrit :
>
> Hi,
>
> On 03.11.25 05:38, chalianis1@gmail.com wrote:
> > From: Chali Anis <chalianis1@gmail.com>
> >
> > Add a driver to use an external state dtb, gives the ability to define
> > an external state at compile time. useful for yocto or buildroot defining
> > a state.dtb that will be passed to barebox at compile time vi a defconfig
> > fragment.
>
> We already had code that did this unconditionally for EFI.
> Can you explain why this needs to be customizable?
>
> > +config EXTERNAL_STATE
> > +     tristate "Use external barebox state dtb"
> > +     depends on OFDEVICE
> > +     depends on STATE
> > +     help
> > +       This permits the use of an extranl dtb state blob
>
> external
>
> > +       which permits to dynamicly at compile time specify
>
> dynamically
>
> > +       an external blob vi EXTERNAL_STATE_DTB_PATH
> > +
> > +config EXTERNAL_STATE_DTB_PATH
> > +     string "the external barebox state dtb path"
> > +     depends on EXTERNAL_STATE
>
> Really needs help text as EXTERNAL makes me think of a file on the
> build host.
>
> > +static int state_external_init(void)
> > +{
> > +     const char *dt_path = CONFIG_EXTERNAL_STATE_DTB_PATH;
> > +     struct device_node *state_root = NULL;
> > +     size_t size;
> > +     void *fdt;
> > +     int ret;
> > +
> > +     if (strlen(dt_path) <= 0)
> > +             return -EINVAL;
>
> strlen can't return negative numbers.
>
> The original function had a number of safety checks that are missing here:
>
> - It only ran the code when in EFI payload mode, as the same barebox
>   binary could be chainloaded without EFI
>
> - It checked that /boot is mounted and gave a useful error message to hint
>   at e.g., EFIFS missing, which we are missing out on here.
>
>
> I am not yet convinced this is strictly an improvement.
>
> If you have the need to customize where your state comes from, how about adding
> a state -f /path/to/state.dtb command option and calling it from your init script?
>
> Cheers,
> Ahmad
>
> > +
> > +     fdt = read_file(dt_path, &size);
> > +     if (!fdt) {
> > +             pr_info("unable to read %s: %m\n", dt_path);
> > +             return 0;
> > +     }
> > +
> > +     state_root = of_unflatten_dtb(fdt, size);
> > +     if (!IS_ERR(state_root)) {
> > +             struct device_node *np = NULL;
> > +             struct state *state;
> > +
> > +             ret = barebox_register_of(state_root);
> > +             if (ret)
> > +                     pr_warn("Failed to register device-tree: %pe\n", ERR_PTR(ret));
> > +
> > +             np = of_find_node_by_alias(state_root, "state");
> > +
> > +             state = state_new_from_node(np, false);
> > +             if (IS_ERR(state))
> > +                     return PTR_ERR(state);
> > +
> > +             ret = state_load(state);
> > +             if (ret != -ENOMEDIUM)
> > +                     pr_warn("Failed to load persistent state, continuing with defaults, %d\n",
> > +                             ret);
> > +
> > +             return 0;
> > +     }
> > +
> > +     return -EINVAL;
> > +}
> > +
> > +late_initcall(state_external_init);
>
>
> --
> Pengutronix e.K.                           |                             |
> Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
> 31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
> Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |



  reply	other threads:[~2025-11-03 14:10 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-11-03  4:38 chalianis1
2025-11-03  4:38 ` [PATCH 2/2] efi: payload: refactor to use the external barebox state driver chalianis1
2025-11-03  6:40   ` Ahmad Fatoum
2025-11-03 14:10     ` anis chali
2025-11-03 14:24       ` Ahmad Fatoum
2025-11-03 15:28         ` anis chali
2025-11-03  6:46 ` [PATCH 1/2] drivers: misc: external_state: add a barebox external state Ahmad Fatoum
2025-11-03 14:09   ` anis chali [this message]
2025-11-03 14:22     ` Ahmad Fatoum
2025-11-03 15:26       ` anis chali

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=CAL+1fyC5bD3k6rufJx6eYXjL4Fo00QOWRRC3JzgfcU-3maROhA@mail.gmail.com \
    --to=chalianis1@gmail.com \
    --cc=a.fatoum@pengutronix.de \
    --cc=barebox@lists.infradead.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox