Heads up: today's u-boot breaks raspberry pi 4 boot
Hi all, today I did the usual "zypper -v dup --no-recommends" and the following reboot stuck in u-boot, it looks like it did no longer like to execute the GRUB2 EFI binary and instead proceeded to try tftp booting. Sorry, no serial console on that machine :-( The machine is right now needed to build stuff in PMBS, so I installed a raspi-os via network onto an usb stick and copied back just u-boot.bin from the second, not-updated raspberrypi4 machine, now it is booting again. No bug report yet, in case nobody else can reproduce it, I'll create one later tomorrow once the builds finished :-) have fun, seife working u-boot: u-boot-rpiarm64-2022.10-2.1.aarch64 borken u-boot: u-boot-rpiarm64-2023.01-1.1.aarch64 -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
Hi,
-----Original Message----- From: Stefan Seyfried <stefan.seyfried@googlemail.com> Sent: 26 January 2023 18:14 To: arm@lists.opensuse.org Subject: Heads up: today's u-boot breaks raspberry pi 4 boot
Hi all,
today I did the usual "zypper -v dup --no-recommends" and the following reboot stuck in u-boot, it looks like it did no longer like to execute the GRUB2 EFI binary and instead proceeded to try tftp booting. Sorry, no serial console on that machine :-(
The machine is right now needed to build stuff in PMBS, so I installed a raspi-os via network onto an usb stick and copied back just u-boot.bin from the second, not- updated raspberrypi4 machine, now it is booting again.
No bug report yet, in case nobody else can reproduce it, I'll create one later tomorrow once the builds finished :-)
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679 Cheers, Guillaume
have fun, seife
working u-boot: u-boot-rpiarm64-2022.10-2.1.aarch64 borken u-boot: u-boot-rpiarm64-2023.01-1.1.aarch64 -- Stefan Seyfried
"For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
IMPORTANT NOTICE: The contents of this email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please notify the sender immediately and do not disclose the contents to any other person, use it for any purpose, or store or copy the information in any medium. Thank you.
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
Same for me on my rpi4s
have fun, seife
working u-boot: u-boot-rpiarm64-2022.10-2.1.aarch64 borken u-boot: u-boot-rpiarm64-2023.01-1.1.aarch64
The "borken" ;) version is installed on my rpi4s and they booted fine. Thorsten -- Thorsten Kukuk, Distinguished Engineer, Senior Architect, Future Technologies SUSE Software Solutions Germany GmbH, Frankenstraße 146, 90461 Nuernberg, Germany Managing Director: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman (HRB 36809, AG Nürnberg)
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562 On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
Same for me on my rpi4s
have fun, seife
working u-boot: u-boot-rpiarm64-2022.10-2.1.aarch64 borken u-boot: u-boot-rpiarm64-2023.01-1.1.aarch64
The "borken" ;) version is installed on my rpi4s and they booted fine.
Thorsten
-- Thorsten Kukuk, Distinguished Engineer, Senior Architect, Future Technologies SUSE Software Solutions Germany GmbH, Frankenstraße 146, 90461 Nuernberg, Germany Managing Director: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman (HRB 36809, AG Nürnberg)
-- Thorsten Kukuk, Distinguished Engineer, Senior Architect, Future Technologies SUSE Software Solutions Germany GmbH, Frankenstraße 146, 90461 Nuernberg, Germany Managing Director: Ivo Totev, Andrew Myers, Andrew McDonald, Martje Boudien Moerman (HRB 36809, AG Nürnberg)
On 26.01.23 18:52, Thorsten Kukuk wrote:
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562
Yes, that's exactly the problem (same error message) I was seeing.
On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
...but probably not from USB :-) I need to boot from USB3 SSD because the SD card performance / endurance is not good enough to run an OBS worker (I could boot from SD and just use the SSD for the worker, but then why? It even boots much faster from the SSD ;-) Have fun, seife -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
I am also having this problem (booting from USB3 SSD). So, where/how does one set CONFIG_LMB_MAX_REGIONS=64, as mentioned in https://bugzilla.suse.com/show_bug.cgi?id=1207562? David On 1/26/23 10:25 AM, Stefan Seyfried wrote:
On 26.01.23 18:52, Thorsten Kukuk wrote:
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562
Yes, that's exactly the problem (same error message) I was seeing.
On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
...but probably not from USB :-)
I need to boot from USB3 SSD because the SD card performance / endurance is not good enough to run an OBS worker (I could boot from SD and just use the SSD for the worker, but then why? It even boots much faster from the SSD ;-)
Have fun, seife
Hello, On Thu, Jan 26, 2023 at 12:31:20PM -0800, David Walker wrote:
I am also having this problem (booting from USB3 SSD). So, where/how does one set CONFIG_LMB_MAX_REGIONS=64, as mentioned in https://bugzilla.suse.com/show_bug.cgi?id=1207562?
Something like this: https://build.opensuse.org/package/rdiff/home:michals/u-boot?opackage=u-boot&oproject=hardware%3Aboot&rev=29 HTH Michal
David
On 1/26/23 10:25 AM, Stefan Seyfried wrote:
On 26.01.23 18:52, Thorsten Kukuk wrote:
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562
Yes, that's exactly the problem (same error message) I was seeing.
On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
...but probably not from USB :-)
I need to boot from USB3 SSD because the SD card performance / endurance is not good enough to run an OBS worker (I could boot from SD and just use the SSD for the worker, but then why? It even boots much faster from the SSD ;-)
Have fun, seife
Ah! No wonder I didn't find it. It looks like it's on its way to Tumbleweed, so I can wait. Thanks! David On 1/26/23 1:15 PM, Michal Suchánek wrote:
Hello,
On Thu, Jan 26, 2023 at 12:31:20PM -0800, David Walker wrote:
I am also having this problem (booting from USB3 SSD). So, where/how does one set CONFIG_LMB_MAX_REGIONS=64, as mentioned in https://bugzilla.suse.com/show_bug.cgi?id=1207562? Something like this:
HTH
Michal
David
On 1/26/23 10:25 AM, Stefan Seyfried wrote:
On 26.01.23 18:52, Thorsten Kukuk wrote:
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562 Yes, that's exactly the problem (same error message) I was seeing.
On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
...but probably not from USB :-)
I need to boot from USB3 SSD because the SD card performance / endurance is not good enough to run an OBS worker (I could boot from SD and just use the SSD for the worker, but then why? It even boots much faster from the SSD ;-)
Have fun, seife
Hi,
-----Original Message----- From: Michal Suchánek <msuchanek@suse.de> Sent: 26 January 2023 22:16 To: David Walker <David@walkerstreet.info> Cc: arm@lists.opensuse.org Subject: Re: Heads up: today's u-boot breaks raspberry pi 4 boot
Hello,
On Thu, Jan 26, 2023 at 12:31:20PM -0800, David Walker wrote:
I am also having this problem (booting from USB3 SSD). So, where/how does one set CONFIG_LMB_MAX_REGIONS=64, as mentioned in https://bugzilla.suse.com/show_bug.cgi?id=1207562?
Something like this:
https://build.opensuse.org/package/rdiff/home:michals/u-boot?opackage=u- boot&oproject=hardware%3Aboot&rev=29
Could you try this u-boot package (from https://build.opensuse.org/package/show/home:michals/u-boot ) and report the result here, please? It sets CONFIG_LMB_MAX_REGIONS to 64. Cheers, Guillaume
HTH
Michal
David
On 1/26/23 10:25 AM, Stefan Seyfried wrote:
On 26.01.23 18:52, Thorsten Kukuk wrote:
Looks like somebody else hit this, too and found the problem: https://bugzilla.suse.com/show_bug.cgi?id=1207562
Yes, that's exactly the problem (same error message) I was seeing.
On Thu, Jan 26, Thorsten Kukuk wrote:
On Thu, Jan 26, Guillaume Gardet wrote:
That's weird, since it boots properly in openQA: https://openqa.opensuse.org/tests/3061679
...but probably not from USB :-)
I need to boot from USB3 SSD because the SD card performance / endurance is not good enough to run an OBS worker (I could boot from SD and just use the SSD for the worker, but then why? It even boots much faster from the SSD ;-)
Have fun, ����seife
So, I sat down a while ago to try this u-boot package from Michal's repository, and now I have some questions... 1. I have an Rpi4b; should I use the u-boot:rpi4 openSUSE_Factory_ARM or u-boot:rpiarm64? Is there a difference? 2. Do I just install the package, or do I need to do something more? 3. The SSD doesn't boot, what's the most effective way to rescue it? Is it as simple as copying /boot/efi/u-boot.bin (and maybe also /boot/efi/uboot* ?) from a working Rpi4b running the same Tumbleweed snapshot (20230116) as the non-booting Rpi4b was running before its breaking update to 20230121? Thanks! David On 1/26/23 11:04 PM, Guillaume Gardet wrote: ...
On Thu, Jan 26, 2023 at 12:31:20PM -0800, David Walker wrote:
I am also having this problem (booting from USB3 SSD). So, where/how does one set CONFIG_LMB_MAX_REGIONS=64, as mentioned in https://bugzilla.suse.com/show_bug.cgi?id=1207562? Something like this:
https://build.opensuse.org/package/rdiff/home:michals/u-boot?opackage=u- boot&oproject=hardware%3Aboot&rev=29 Could you try this u-boot package (fromhttps://build.opensuse.org/package/show/home:michals/u-boot ) and report the result here, please?
It sets CONFIG_LMB_MAX_REGIONS to 64.
Cheers, Guillaume
On 28.01.23 05:11, David Walker wrote:
So, I sat down a while ago to try this u-boot package from Michal's repository, and now I have some questions...
1. I have an Rpi4b; should I use the u-boot:rpi4 openSUSE_Factory_ARM or u-boot:rpiarm64? Is there a difference?
u-boot-rpiarm64, that's at least what is installed on all my raspberry pi 4 and 400 machines. There probably is a difference, but I don't know what it is offhand ;-)
2. Do I just install the package, or do I need to do something more?
Just installing the package is enough.
3. The SSD doesn't boot, what's the most effective way to rescue it? Is it as simple as copying /boot/efi/u-boot.bin (and maybe also /boot/efi/uboot* ?) from a working Rpi4b running the same Tumbleweed snapshot (20230116) as the non-booting Rpi4b was running before its breaking update to 20230121?
Yes, that's what I did to rescue the machine. As I had no easy way of connecting the SSD to another machine (it's in an argon 40 case with an USB3 A-to-A connector), I disconnected it, adden an USB stick, installed raspi-OS via the in-ROM net-installer and then reconnected the SSD and copied over u-boot.bin. Disconnected the USB stick, rebooted, everything was fine again. Then updated the u-boot-rpiarm64 package and still everything is fine :-) Good luck, -- Stefan Seyfried "For a successful technology, reality must take precedence over public relations, for nature cannot be fooled." -- Richard Feynman
Thanks, Stefan! I was able to rescue my system SSD, but installing u-boot-rpiarm64-2023.01-246.1.aarch64.rpm from Michal's repository broke booting again with the "** Reading file would overwrite reserved memory **" error message. I notice that there's a new build scheduled in OBS for that package, so I'll try again when it's finished. David On 1/28/23 12:41 AM, Stefan Seyfried wrote:
On 28.01.23 05:11, David Walker wrote:
So, I sat down a while ago to try this u-boot package from Michal's repository, and now I have some questions...
1. I have an Rpi4b; should I use the u-boot:rpi4 openSUSE_Factory_ARM or u-boot:rpiarm64? Is there a difference?
u-boot-rpiarm64, that's at least what is installed on all my raspberry pi 4 and 400 machines.
There probably is a difference, but I don't know what it is offhand ;-)
2. Do I just install the package, or do I need to do something more?
Just installing the package is enough.
3. The SSD doesn't boot, what's the most effective way to rescue it? Is it as simple as copying /boot/efi/u-boot.bin (and maybe also /boot/efi/uboot* ?) from a working Rpi4b running the same Tumbleweed snapshot (20230116) as the non-booting Rpi4b was running before its breaking update to 20230121?
Yes, that's what I did to rescue the machine. As I had no easy way of connecting the SSD to another machine (it's in an argon 40 case with an USB3 A-to-A connector), I disconnected it, adden an USB stick, installed raspi-OS via the in-ROM net-installer and then reconnected the SSD and copied over u-boot.bin. Disconnected the USB stick, rebooted, everything was fine again. Then updated the u-boot-rpiarm64 package and still everything is fine :-)
Good luck,
I haven't had any luck with u-boot-rpiarm64-2023.01-248.1.aarch64.rpm from Michal's repository, or with the version released in the 20230129 Tumbleweed Snapshot. Any ideas of where I can look next? David On 1/28/23 7:51 PM, David Walker wrote:
Thanks, Stefan! I was able to rescue my system SSD, but installing u-boot-rpiarm64-2023.01-246.1.aarch64.rpm from Michal's repository broke booting again with the "** Reading file would overwrite reserved memory **" error message. I notice that there's a new build scheduled in OBS for that package, so I'll try again when it's finished.
David
On 1/28/23 12:41 AM, Stefan Seyfried wrote:
On 28.01.23 05:11, David Walker wrote:
So, I sat down a while ago to try this u-boot package from Michal's repository, and now I have some questions...
1. I have an Rpi4b; should I use the u-boot:rpi4 openSUSE_Factory_ARM or u-boot:rpiarm64? Is there a difference?
u-boot-rpiarm64, that's at least what is installed on all my raspberry pi 4 and 400 machines.
There probably is a difference, but I don't know what it is offhand ;-)
2. Do I just install the package, or do I need to do something more?
Just installing the package is enough.
3. The SSD doesn't boot, what's the most effective way to rescue it? Is it as simple as copying /boot/efi/u-boot.bin (and maybe also /boot/efi/uboot* ?) from a working Rpi4b running the same Tumbleweed snapshot (20230116) as the non-booting Rpi4b was running before its breaking update to 20230121?
Yes, that's what I did to rescue the machine. As I had no easy way of connecting the SSD to another machine (it's in an argon 40 case with an USB3 A-to-A connector), I disconnected it, adden an USB stick, installed raspi-OS via the in-ROM net-installer and then reconnected the SSD and copied over u-boot.bin. Disconnected the USB stick, rebooted, everything was fine again. Then updated the u-boot-rpiarm64 package and still everything is fine :-)
Good luck,
participants (5)
-
David Walker -
Guillaume Gardet -
Michal Suchánek -
Stefan Seyfried -
Thorsten Kukuk