All of lore.kernel.org
 help / color / mirror / Atom feed
From: Tom Rini <trini@konsulko.com>
To: Andrew Davis <afd@ti.com>
Cc: Bryan Brattlof <bb@ti.com>, Denys Dmytriyenko <denis@denix.org>,
	Ryan Eatmon <reatmon@ti.com>,
	Meta-TI <meta-ti@lists.yoctoproject.org>,
	Nishanth Menon <nm@ti.com>, Res Sapp <rs@ti.com>,
	Anand Gadiyar <gadiyar@ti.com>,
	Praneeth Bajjuri <praneeth@ti.com>
Subject: Re: [meta-ti] [dunfell PATCH] wic: move uboot/spls mountpoint to /boot/firmware
Date: Wed, 27 Sep 2023 15:02:39 -0400	[thread overview]
Message-ID: <20230927190239.GC305624@bill-the-cat> (raw)
In-Reply-To: <ba8d9b4c-a500-5775-bfaa-88f6bb2ef881@ti.com>

On Wed, Sep 27, 2023 at 02:00:26PM -0500, Andrew Davis wrote:
> On 7/28/22 6:31 AM, Tom Rini wrote:
> > On Mon, Jul 25, 2022 at 06:29:04PM -0500, Bryan Brattlof wrote:
> > > On July 22, 2022 thus sayeth Tom Rini:
> > > > On Thu, Jul 21, 2022 at 09:31:30PM -0500, Bryan Brattlof wrote:
> > > > > On July 21, 2022 thus sayeth Denys Dmytriyenko:
> > > > > > I don't think this is a correct solution, or maybe I'm not understanding the
> > > > > > problem. Can you please elaborate a bit more on the problem?
> > > > > > 
> > > > > 
> > > > > I'm fairly new to yocto so I'm sure I've found the entirely wrong way to
> > > > > get what I wanted :)
> > > > > 
> > > > > After boot it appears /etc/fstab is setup to mount our vfat boot
> > > > > partition to the /boot directory currently holding the Image and dtbs.
> > > > > I guess because uboot has already thrown the kernel and dtb into ddr at
> > > > > this point only Anand was able to notice this.
> > > > > 
> > > > > My change should mount the vfat partition to /boot/firmware so we can
> > > > > access to both our spls and kernel binaries in /boot.
> > > > > 
> > > > > I'm also not sure what folder name to use here. It seem like there may
> > > > > be a standard to this? Would /boot/uboot be more correct?
> > > > 
> > > > So which SoCs are we talking about here?  For the 64bit parts, the
> > > > intention should be that the FAT partition on p1 be functional as the
> > > > ESP.  So that means mounting it to /boot/efi.  But I'm not sure off-hand
> > > > how that's being treated in upstream OE around the notion of making an
> > > > image that's what a UEFI system expects.
> > > 
> > > Ideally this should be for all TI's K3 devices. I'm unfamiliar with the
> > > UEFI format and don't know if we currently follow it, however I like the
> > > idea of /boot/efi from what I googled.  I also found raspian is using
> > > /boot/firmware. Could this be less of a standard than I think?
> > 
> > Keep in mind this is a generic OE on 64bit Arm problem and not a TI
> > specific problem.  Where to mount the ESP should have a consistent
> > default.  And it should look like it does on an off the shelf distro.
> > 
> 
> While not an exact ESP partition, we are working on making our boot
> partition as close as we can given ROM constraints.
> 
> I'm running into the problem now of updating kernel on a live system,
> would fail as we have mounted our boot partition over the rootfs boot/
> directory, so new images are not installed to the right spot/partition.
> 
> Let's go with /boot/efi if that works for everyone.
> 
> Bryan, could you re-send this with that change and the same for
> sdimage-2part-efi.wks.in?

The only feedback I would provide here is to work with the ROM team to
fix whatever constraints prevent an "ESP" itself from being used as this
will hinder the long term viability of the processors being widely
supported.

-- 
Tom


  reply	other threads:[~2023-09-27 19:02 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-21 21:01 [dunfell PATCH] wic: move uboot/spls mountpoint to /boot/firmware Bryan Brattlof
2022-07-21 22:57 ` Denys Dmytriyenko
2022-07-22  2:31   ` Bryan Brattlof
2022-07-22 12:29     ` Tom Rini
2022-07-25 23:29       ` Bryan Brattlof
2022-07-28 11:31         ` Tom Rini
2023-09-27 19:00           ` [meta-ti] " Andrew Davis
2023-09-27 19:02             ` Tom Rini [this message]
2023-09-28  5:44               ` Res Sapp
2023-09-28 13:39                 ` Bryan Brattlof

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=20230927190239.GC305624@bill-the-cat \
    --to=trini@konsulko.com \
    --cc=afd@ti.com \
    --cc=bb@ti.com \
    --cc=denis@denix.org \
    --cc=gadiyar@ti.com \
    --cc=meta-ti@lists.yoctoproject.org \
    --cc=nm@ti.com \
    --cc=praneeth@ti.com \
    --cc=reatmon@ti.com \
    --cc=rs@ti.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 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.