DistroKit Mailinglist
 help / color / mirror / Atom feed
From: Ahmad Fatoum <a.fatoum@pengutronix.de>
To: Roland Hieber <rhi@pengutronix.de>
Cc: distrokit@pengutronix.de
Subject: Re: [DistroKit]  [PATCH v3 3/6] v7a: barebox: pass root=mmcblkXpN instead of root=PARTUUID=… to kernel
Date: Wed, 22 Nov 2023 15:24:07 +0100	[thread overview]
Message-ID: <9afb1d8a-f90d-7d74-781d-c33f6b7a2efb@pengutronix.de> (raw)
In-Reply-To: <20231122141711.iudghqftvkpp7wf2@pengutronix.de>

Hello Roland,

On 22.11.23 15:17, Roland Hieber wrote:
> On Wed, Nov 22, 2023 at 02:27:00PM +0100, Ahmad Fatoum wrote:
>> Hello Roland,
>>
>> On 20.10.23 11:05, Roland Hieber wrote:
>>> Kernels prior to v5.10-rc1 could not ensure deterministic ordering for
>>> MMC devices. For this reason, the default 'append-root' behaviour of
>>> barebox for MMC block devices is to append a 'root=PARTUUID=XYZ' option
>>> to the kernel command line. The PARTUUID is derived from the partition
>>> table and is expected to be unique system-wide. This is however
>>> susceptible to breakage if the exact same image is written to different
>>> boot media in the system, e.g. SD and eMMC, so the PARTUUID will no
>>> longer uniquely refer to a single partition.
>>>
>>> Newer kernel versions honor the numbering used for 'mmc' aliases,
>>> which makes 'root=mmcblkXpN' deterministic as long as barebox and
>>> kernel aliases are identical. Since we're using kernel 6.4, enable this
>>> option in barebox.
>>
>> As you mention this feature requires that barebox and kernel aliases
>> are identical. At least for the STM32MP157C-EV1, this is not the case
>> as the kernel device tree has no aliases at all.
> 
> Oh. How come? Can we patch the kernel DT to make them the same?

Having aliases for everything is uncommon outside i.MX. I am in favor
of adding aliases upstream to all DTs we support, but that would take
a while.

> 
>> Can you send a revert?
> 
> I can send a revert, but I don't like the implications. Incidentally,
> this topic was discussed in length last week on a systemd pull request
> of mine [1], with the TL;DR that filesystem and partition UUIDs should
> not generally be considered unique. While that might be the case in the
> DistroKit test runs, we also have some boards with eMMCs that can hold
> the same image as the SD cards that we boot from. I suspect this might
> come back to bite us some time in the future.

I don't mind restoring CONFIG_MMCBLKDEV_ROOTARG for devices whose aliases
have been vetted to be identical between barebox and Linux, but right now,
this breaks boot on at least one board.

Cheers,
Ahmad



> 
> [1]: https://github.com/systemd/systemd/pull/29219#issuecomment-1812226959
> 
>  - Roland
> 

-- 
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:[~2023-11-22 14:24 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-20  9:05 [DistroKit] [PATCH v3 1/6] v7a: image-stm32mp: use 'partition-table-type' option instead of 'gpt' Roland Hieber
2023-10-20  9:05 ` [DistroKit] [PATCH v3 2/6] images: use 'partition-table-type' option instead of 'partition-table' Roland Hieber
2023-10-20  9:05 ` [DistroKit] [PATCH v3 3/6] v7a: barebox: pass root=mmcblkXpN instead of root=PARTUUID=… to kernel Roland Hieber
2023-11-22 13:27   ` Ahmad Fatoum
2023-11-22 14:17     ` Roland Hieber
2023-11-22 14:24       ` Ahmad Fatoum [this message]
2023-12-11  9:37         ` [DistroKit] [PATCH] v7a: barebox-stm32mp1: pass " Roland Hieber
2023-12-12 17:58           ` Robert Schwebel
2023-10-20  9:05 ` [DistroKit] [PATCH v3 4/6] v7a: barebox: enable GUID Partition Table support Roland Hieber
2023-10-20  9:05 ` [DistroKit] [PATCH v3 5/6] images: remove qemu size hack Roland Hieber
2023-10-20  9:05 ` [DistroKit] [PATCH v3 6/6] images: convert to GPT format Roland Hieber
2023-10-20 10:56 ` [DistroKit] [PATCH v3 1/6] v7a: image-stm32mp: use 'partition-table-type' option instead of 'gpt' Robert Schwebel

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=9afb1d8a-f90d-7d74-781d-c33f6b7a2efb@pengutronix.de \
    --to=a.fatoum@pengutronix.de \
    --cc=distrokit@pengutronix.de \
    --cc=rhi@pengutronix.de \
    /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