F32 UEFI reboot complete OK if 'cold/hard'; but 'warm/soft' reboot hangs after Grub menu entry is selected ?

[PGNet Dev]

I've built up an older box (ASRock J1900D2Y mobo), on Fedora32 + kernel 5.8.18 + grub 2.04 (from F32 pkgs), booting UEFI.

I can cold boot -- or hard reboot -- the box, and it boots to desktop with no problems.
Afaict, everything's up & running without issue.

But,

on SOFT reboot -- via either `shutdown -r now` or `systemctl start reboot.target`,
the system DOES start the reboot.
UEFI boot passes to grub menu as usual, latest kernel is selected, and then ...
... nothing.

Just sits there.  No further error logging, or progress -- even at serial console with kernel/systemd debug logging turned up.

A cold, power-cycling reboot immediately fixes the problem -- e.g., after an IPMI 'Reset', the system again boots fine.

Since cold boot is OK, seems kernel & initrd are ok.

Since the issue arises only AFTER grub menu's in the picture, it doesn't seem that BIOS/UEFI should be involved in any way.

Any hints as to where to dig around next? Or suggestions how to workaround?

[Roger Heflin]

I am going to guess a "warm" reboot does not leave some piece of
critical hardware in a correct state (either to work, or that some
driver and/or grub needs/expects) and so something fails early.

Does the kernel post any messages? you might try removing quiet and
see if anything shows up at all, if nothing at all does then something
must be broken really early which would likely make it a firmware
issue.

There are options you can specify on the kernel line that will force a
hard/cold boot when you do a reboot, and setting that option is
probably the easiest solution.  I think it is boot=cold but at the
moment the machine I have to set that on to get it to work on reboots
is off so I cannot double check.

A bios update might fix it, but if the boot=cold does it, then I would
just set that.  The option is I believe designed for this sort of
issue.

On Sat, Nov 14, 2020 at 3:00 PM PGNet Dev  wrote:
> I've built up an older box (ASRock J1900D2Y mobo), on Fedora32 + kernel 5.8.18 + grub 2.04 (from F32 pkgs), booting UEFI.
>
> I can cold boot -- or hard reboot -- the box, and it boots to desktop with no problems.
> Afaict, everything's up & running without issue.
>
> But,
>
> on SOFT reboot -- via either `shutdown -r now` or `systemctl start reboot.target`,
> the system DOES start the reboot.
> UEFI boot passes to grub menu as usual, latest kernel is selected, and then ...
> ... nothing.
>
> Just sits there.  No further error logging, or progress -- even at serial console with kernel/systemd debug logging turned up.
>
> A cold, power-cycling reboot immediately fixes the problem -- e.g., after an IPMI 'Reset', the system again boots fine.
>
> Since cold boot is OK, seems kernel & initrd are ok.
>
> Since the issue arises only AFTER grub menu's in the picture, it doesn't seem that BIOS/UEFI should be involved in any way.
>
> Any hints as to where to dig around next? Or suggestions how to workaround?
> _______________________________________________
> users mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
> Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
> List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
> List Archives: https://lists.fedoraproject.org/archives/list/[e-mail]

[PGNet Dev]

On 11/15/20 5:23 AM, Roger Heflin wrote:
> Does the kernel post any messages?

nope
> you might try removing quiet and see if anything shows up at all,

those^ results are withOUT 'quiet' ...
specifically, WITH 'try-some-massively-overkill-debugging-first',

rd.systemd.show_status=1 rd.shell=1 rd.auto=1 rd.udev.log_priority=debug rd.debug=1 rd.break systemd.log_target=kmsg log_buf_len=1M printk.devkmsg=on systemd.log_level=debug debug showopts noquiet print_fatal_signals=1 loglevel=8 initcall_debug

and serial-console enabled/monitored.
> if nothing at all does

nope. complete 'radio silence'.  not a peep after selecting the grub menu entry after a soft-reboot.
> then something
> must be broken really early which would likely make it a firmware
> issue.

likely.  tho, i'm not entirely clear on what could be getting borked in a soft-reboot case that's NOT an issue in the hard-reboot case. and doesn't seem to appear as a problem in the running environment.
> There are options you can specify on the kernel line that will force a
> hard/cold boot when you do a reboot, and setting that option is
> probably the easiest solution.  I think it is boot=cold but at the
> moment the machine I have to set that on to get it to work on reboots
> is off so I cannot double check.

here,

https://www.kernel.org/doc/html/v5.8/admin-guide/kernel-parameters.html

i see only

reboot=

not

boot=

I've already tried any/all combinations of 'reboot=' options.

I've also tried the usual assortment of acpi= options ; to no avail.

I can't get it to FAIL to RE-boot; in all soft cases, e.g. 'shutdown -r now', I _do_ get a reboot, back to the grub menu ... and it's there, and only there, that it freezes / goes silent on soft-reboot.
> A bios update might fix it, but if the boot=cold does it, then I would
> just set that.  The option is I believe designed for this sort of
> issue.

BIOS is latest available; which _is_ a bit old.

I've cleared/reflashed ... no effect.

(Re)booting the machine with different OSs (Win10, Ubu20, OpensuseTW) all work without problem; both in the soft- and hard-reboot cases.

So, here, it's Fedora-specific; It's 'local' to my F32 install on this machine; elusive, atm :-/

If I can just manage to get _some_ sort of useful logging at that stage ...