Attempt to update Rock64 to head -r355976 failed to boot afterwards, anyone have a recent FreeBSD booting a Rock64?

Mon Dec 23 01:14:09 UTC 2019

I’m using a ’special’  2019.10 u-boot - version ,developed  by a  BSD-colleague,
where the dtb-clock-settings were backported to 2019.10 u-boot  for Rock64.
Before using that special version I’ve booted FreeBSD on the Rock64 by
typing  'boot disk0s2:/boot/kernel/kernel‘ at the prompt with a standard linux-uboot-version 
by Ayufan .

> Am 23.12.2019 um 01:24 schrieb Mark Millard via freebsd-arm <freebsd-arm at freebsd.org>:
> 
> [The start_init: message was a dumb thing to use as a
> reference point.]
> 
> On 2019-Dec-22, at 15:06, Mark Millard <marklmi at yahoo.com> wrote:
> 
>> [It has taken explicitly controlling the DTB used and
>> use of "boot -v" together to be able to boot to
>> completion . . .]
>> 
>> On 2019-Dec-22, at 12:47, Mark Millard <marklmi at yahoo.com> wrote:
>> 
>>> On 2019-Dec-22, at 11:16, Mark Millard <marklmi at yahoo.com> wrote:
>>> 
>>>> On 2019-Dec-22, at 02:38, Emmanuel Vadot <manu at bidouilliste.com> wrote:
>>>> 
>>>>> On Sun, 22 Dec 2019 00:22:16 -0800
>>>>> Mark Millard via freebsd-arm <freebsd-arm at freebsd.org> wrote:
>>>>> 
>>>>>> [OverDrive 1000 and MACCHIATObin Doubleshot updates went fine.
>>>>>> The code has Peter Jeremy's rk_tsadc.c patch.]
>>>>>> 
>>>>>> 
>>>>>> The console shows for boot -v . . .
>>>>>> 
>>>>>> 
>>>>>> Loading kernel...
>>>>>> /boot/kernel/kernel text=0x98af14 data=0x18e618 data=0x0+0x6fc8e8 syms=[0x8+0x142020+0x8+0x12d3fd]
>>>>>> Loading configured modules...
>>>>>> /boot/kernel/umodem.ko text=0x2120 text=0x13e0 data=0x6e8+0x10 syms=[0x8+0xf60+0x8+0xb7f]
>>>>>> /boot/kernel/ucom.ko text=0x217f text=0x3340 data=0x880+0x858 syms=[0x8+0x1170+0x8+0xb0d]
>>>>>> /boot/entropy size=0x1000
>>>>>> 
>>>>>> Hit [Enter] to boot immediately, or any other key for command prompt.
>>>>>> Booting [/boot/kernel/kernel] in 8 seconds... 
>>>>>> 
>>>>>> Type '?' for a list of commands, 'help' for more detailed help.
>>>>>> OK boot -v
>>>>>> Using DTB provided by EFI at 0x80f3000.
>>>>>> ---<<BOOT>>---
>>>>>> . . .
>>>>> 
>>>>> 
>>>>> I don't have the same clocks from the dtb, make sure that you are
>>>>> using the latest one.
>>>>> rk3328_cru0: <Rockchip RK3328 Clock and Reset Unit> mem 0xff440000-0xff440fff on ofwbus0
>>>>> . . .
>>>>> sha256 /boot/dtb/rockchip/rk3328-rock64.dtb 
>>>>> SHA256 (/boot/dtb/rockchip/rk3328-rock64.dtb) = 50a180fed37f1d5dbfda60a6c55261c7b87b5b2bc97e428042481c94877da317
>>>> 
>>>> Thanks.
>>>> 
>>>> # sha256 /mnt/boot/dtb/rockchip/rk3328-rock64.dtb
>>>> SHA256 (/mnt/boot/dtb/rockchip/rk3328-rock64.dtb) = 50a180fed37f1d5dbfda60a6c55261c7b87b5b2bc97e428042481c94877da317
>>>> 
>>>> Looks like a match to me. So I need to look elsewhere than
>>>> the contents of that file . . .
>>>> 
>>>> My getting: "Using DTB provided by EFI at 0x80f3000" suggests
>>>> that the file is not being used for some reason: Instead some
>>>> sort of nonFreeBSD/internal-to-something-else DTB seems to be
>>>> in use?
>>>> 
>>>> So my current guess is that I need to figure out how to control
>>>> which DTB source is used so that /boot/dtb/rockchip/rk3328-rock64.dtb
>>>> is used in my context. (Although I've no clue why I'd need a
>>>> different configuration for controlling such things now.)
>>>> 
>>>> Note: I tried putting back the prior EFI/BOOT/bootaa64.efi but
>>>> it made no difference to the failed-boot behavior.
>>> 
>>> Well, using load -t explicitly got farther:
>>> (So once I figure out an equivalent in /boot/loader.conf
>>> it should automatically get farther.)
>> . . .
>> 
>> The following forced the desired .dtb to be used:
>> 
>> # more /boot/loader.conf 
>> . . .
>> rk3328_rock64_load="YES"
>> rk3328_rock64_type="dtb"
>> rk3328_rock64_name="/boot/dtb/rockchip/rk3328-rock64.dtb"
>> . . .
>> 
>> Interestingly, so far, boot -v works for power up booting,
>> but default booting does not, even for when "boot" is
>> typed to the loader prompt.
>> 
>> The default usually just hangs instead of showing all
>> the information that I reported previously. The hangup
>> (possibly with some text) is just before the start_init
>> line in:
>> 
>> Warning: no time-of-day clock registered, system time will not be set accurately
>> start_init: trying /sbin/init
> 
> This was a dumb choice of reference point:
> 
>        while ((path = strsep(&tmp_init_path, ":")) != NULL) {
>                if (bootverbose)
>                        printf("start_init: trying %s\n", path);
> 
> The message only shows up for bootverbose in the first place.
> 
>> So far, no hang-up has shown part of the "start_init"
>> message line.
> 
> Again, a dumb reference point to have used.
> 
> A normal boot shows:
> 
> . . .
> Root mount waiting for: CAM
> Warning: no time-of-day clock registered, system time will not be set accurately
> Setting hostuuid: . . .
> Setting hostid: . . .
> Starting file system checks:
> . . .
> 
> So the real range for failure is between:
> 
> Warning: no time-of-day clock registered, system time will not be set accurately
> Setting hostuuid: . . .
> 
> 
>> But I've not had troubles (so far) with "shutdown -r now"
>> reboots, defaults or otherwise: problems Just for going
>> from power-off to power-on and trying to boot.
>> 
>> (Someday, I also want to figure out how to set up the
>> Rock64 to get a stable DHCP binding: a fixed MAC
>> address, I guess.)
>> 
> 
> So far, trying a debug kernel is like trying "boot -v": the
> problem has not occurred (and there are no interesting
> messages).
> 
> 
> ===
> Mark Millard
> marklmi at yahoo.com
> ( dsl-only.net went
> away in early 2018-Mar)
> 
> _______________________________________________
> freebsd-arm at freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-arm
> To unsubscribe, send any mail to "freebsd-arm-unsubscribe at freebsd.org"