request for i.MX8MM Venice test

[Adam Ford]

On Mon, Mar 8, 2021 at 2:35 PM Tim Harvey <tharvey at gateworks.com> wrote:
>
> On Mon, Mar 8, 2021 at 8:47 AM Tim Harvey <tharvey at gateworks.com> wrote:
> >
> > On Sat, Mar 6, 2021 at 6:10 PM Marek Behun <marek.behun at nic.cz> wrote:
> > >
> > > [sorry for the spam, I accidentally sent this e-mail from my personal
> > > address]
> > >
> > > Hello Tim,
> > >
> > > you are listed as maintainer of i.MX8MM Venice board in U-Boot.
> > >
> > > I am currently working on LTO support for U-Boot, and I have
> > > encountered a problem with i.MX8MM Venice board:
> > > when LTO is enabled, the linking process for SPL does not throw away
> > > relocation information, making the resulting SPL image too big for that
> > > board.
> > >
> > > I have added a patch that discards symbols from .rela* section to my
> > > patch series, but I would like you to test whether the patch series
> > > works for your board and does not break anything.
> > >
> > > Could you please clone https://github.com/elkablo/u-boot branch lto,
> > > build for imx8mm_venice_defconfig and test whether it boots on your
> > > board and maybe test some U-Boot commands (disk reads, kernel
> > > booting, ...)? If it does not work, could you also please check with
> > > current U-Boot master, to see if it got broken with my patches or with
> > > something different?
> > >
> >
> > Marek,
> >
> > Your lto branch does not boot - I don't get anything after the SPL
> > banner. U-Boot master shows the same problem so some regression has
> > taken place.
> >
> > There should be several other IMX8 boards using SPL that show the same
> > issue I would think?
> >
> > I will have to bisect to find the breakage.
> >
>
> Marek,
>
> My testing may be flawed as I'm having issues with even the branches I
> submitted to add my board (which I previously tested without issues).
> I'm not sure if I'm having a compiler related issue or something else.
>
> Perhaps there is another IMX8 SPL user out there that can test with
> your patches while I figure out what's wrong on my end?

I have tested and booted an i.MX8M Nano with a slightly older revision
of Marek's git repo. I wrongly assumed that the Mini and Nano were
close enough to not warrant testing on the mini, but based on what I'm
hearing, I will test the Beacon i.mx8mm board.   I can try to do that
sometime later this week.

adam

>
> Tim