All of lore.kernel.org
 help / color / mirror / Atom feed
From: Luca Ceresoli via buildroot <buildroot@buildroot.org>
To: "Frager, Neal" <neal.frager@amd.com>
Cc: Brandon Maier <brandon.maier@collins.com>,
	"thomas.petazzoni@bootlin.com" <thomas.petazzoni@bootlin.com>,
	"buildroot@buildroot.org" <buildroot@buildroot.org>,
	"romain.naour@smile.fr" <romain.naour@smile.fr>,
	"Simek, Michal" <michal.simek@amd.com>,
	"yann.morin.1998@free.fr" <yann.morin.1998@free.fr>
Subject: Re: [Buildroot] [PATCH v2 1/1] board/xilinx: bump to linux 6.6.40
Date: Mon, 19 Aug 2024 12:01:21 +0200	[thread overview]
Message-ID: <20240819120121.50ed89ad@booty> (raw)
In-Reply-To: <CH2PR12MB500412028C07F3C30A5DEDD5F0BA2@CH2PR12MB5004.namprd12.prod.outlook.com>

Hello Nea, Michal, Brandon, Thomas, Peter,

On Fri, 9 Aug 2024 15:19:30 +0000
"Frager, Neal" <neal.frager@amd.com> wrote:

> Hi Brandon,
> 
> > Hi Brandon,
> >  
> > > This patch bumps all Xilinx board defconfigs to Linux 6.6.40.
> > >
> > > Xilinx Evaluation Boards:
> > > configs/versal_vck190_defconfig
> > > configs/versal_vpk180_defconfig
> > > configs/zynq_microzed_defconfig
> > > configs/zynq_zc702_defconfig
> > > configs/zynq_zc706_defconfig
> > > configs/zynq_zed_defconfig
> > > configs/zynqmp_kria_kd240_defconfig
> > > configs/zynqmp_kria_kr260_defconfig
> > > configs/zynqmp_kria_kv260_defconfig
> > > configs/zynqmp_zcu102_defconfig
> > > configs/zynqmp_zcu104_defconfig
> > > configs/zynqmp_zcu106_defconfig  
> >  
> > > I attempted to boot this patch on a zcu106, but I get no output over the
> > > uart port from anything, U-Boot SPL, ATF, etc. So this board may be
> > > broken on master.  
> >  
> > > Can you verify if it works on your end?  
> >
> > I can help you with this problem.  It is because there are two different
> > versions of the zcu106 having different DDR memories.  This is because the
> > DDR that came on the revA version of the board became obsolete.
> >
> > I left the revA device tree in buildroot as I know this is the version of the
> > zcu106 that Peter Korsgaard has.  Based on your comment, I believe you have
> > the newer version of the zcu106.
> >
> > To fix this, only the u-boot psu_init_gpl.c needs to be changed, so the
> > correct spl psu_init routine will run.
> >
> > Please implement the following in your zcu106 defconfig:
> >
> > - BR2_TARGET_UBOOT_CUSTOM_MAKEOPTS="DEVICE_TREE=zynqmp-zcu106-revA
> > + BR2_TARGET_UBOOT_CUSTOM_MAKEOPTS="DEVICE_TREE=zynqmp-zcu106-rev1.0
> >
> > This will build in the correct psu_init_gpl.c from the two choices below:
> > https://github.com/Xilinx/u-boot-xlnx/tree/master/board/xilinx/zynqmp/zynqmp-zcu106-rev1.0
> > https://github.com/Xilinx/u-boot-xlnx/tree/master/board/xilinx/zynqmp/zynqmp-zcu106-revA
> >
> > You can read about this issue for the zcu102 and zcu106 here:
> > https://support.xilinx.com/s/article/71961?language=en_US  
> 
> > Thank you! Switching to the rev1.0 devicetree does let my board boot.  
> 
> Great!
> 
> >
> > Do you have any opinions for how this can be better handled in buildroot to
> > support both versions of the zcu102 and zcu106?  
> 
> > My preference would be to having two defconfigs, as they are effectively
> > different boards. But given Thomas would prefer not adding new
> > defconfigs, documenting it in the defconfig as you did is a good idea.  
> 
> As I mentioned in my previous response, I will move this documentation to the
> readme.txt file.  It makes more sense there.

I agree the readme file seems the most appropriate.

> > At the moment, the older version of the zcu106 is supported and the newer
> > version of the zcu102 is supported in buildroot natively.  
> 
> > Would it be prudent to default the zcu106 to the newest revision? Is
> > there any idea of if the zcu106 revA or rev1.0 is more widely used?  
> 
> Now that I think about it, you are absolutely right.  Not only are all new
> boards using the rev1.0 DDR memories, but I have to believe the vast majority
> of boards out there are on the new memory at this point.
> 
> The only reason I kept the zcu106 on revA was because I knew that Peter's
> zcu106 has revA DDR memory.  That is probably not the best reason to keep the
> zcu106 defconfig on revA.  Sorry Peter!
> 
> So unless anyone objects, I plan to do the following.
> 
> 1. Migrate the zynqmp_zcu106_defconfig to rev1.0 DDR memory.
> 
> 2. Document the revA/rev1.0 issue in the board/zynqmp/readme.txt, so that
> users of the old boards have at least a chance of finding that they need to
> change the u-boot device_tree config for revA DDR memory.

Would it be possible to have a single psu_init_gpl.c that works for
both revisions, albeit with non-optimal performance for both? I don't
know what the differences are between the two revisions, but if it is a
matter of timings it should be possible.

Custom designs will use a custom psu_init_gpl.c anyway, so that
wouldn't affect performance of industrial products.

Luca

-- 
Luca Ceresoli, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

  reply	other threads:[~2024-08-19 10:01 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-08 13:23 [Buildroot] [PATCH v2 1/1] board/xilinx: bump to linux 6.6.40 Neal Frager via buildroot
2024-08-08 19:44 ` Brandon Maier via buildroot
2024-08-09  3:52   ` Frager, Neal via buildroot
2024-08-09  6:30     ` Thomas Petazzoni via buildroot
2024-08-09  6:39       ` Frager, Neal via buildroot
     [not found]         ` <3036afae-2027-4993-bb1f-acd3b312759e@amd.com>
2024-08-09  8:22           ` Frager, Neal via buildroot
2024-08-09 15:02     ` Brandon Maier via buildroot
2024-08-09 15:19       ` Frager, Neal via buildroot
2024-08-19 10:01         ` Luca Ceresoli via buildroot [this message]
2024-08-09 15:51 ` Brandon Maier via buildroot

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=20240819120121.50ed89ad@booty \
    --to=buildroot@buildroot.org \
    --cc=brandon.maier@collins.com \
    --cc=luca.ceresoli@bootlin.com \
    --cc=michal.simek@amd.com \
    --cc=neal.frager@amd.com \
    --cc=romain.naour@smile.fr \
    --cc=thomas.petazzoni@bootlin.com \
    --cc=yann.morin.1998@free.fr \
    /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.