ACPI panic on boot with new Lua loader and other minor issues

Mon Feb 19 22:13:41 UTC 2018

Le 19/02/2018 à 21:21, Kyle Evans a écrit :
> Hello!
> 
> On Mon, Feb 19, 2018 at 8:21 AM, Juan Ramón Molina Menor <listjm at club.fr> wrote:
>> I have done a full build of r329555 to test the new Lua boot loader.
>>
>> Both the new and the old kernels panic after being loaded with:
>>
>> panic: running without device atpic requires a local APIC
>>
>> For reasons unknown, ACPI is off, as shown by David Wolfskill in a previous
>> message:
>> https://lists.freebsd.org/pipermail/freebsd-current/2018-February/068497.html
>>
>> OK show hint.acpi.0.disabled
>> 1
>>
>> Setting ACPI to On resolves the issue.
> 

Hi Kyle.

> As David noted, this should actually Just Work (TM) now. Can you break
> into a loader prompt with just the forth loader and tell me what "show
> hint.acpi.0.rsdp" looks like?

OK show hint.acpi.0.rsdp
Command error

I tested both with hint.acpi.0.disabled= 1 and 0.

> 
>> Also, I can not stop boot2 to try to use the copy of the Forth loader: the
>> keyboard only becomes responsive at the loader stage.
> 
> Hmm...

In fact, I don’t think this has ever worked here… I’ve found a very old 
(July 2016) FreeBSD 12 memstick and neither can I stop the boot2 stage.

>> There is an error during this stage:
>>
>> Loading /boot/defaults/loader.conf
>> Failed to open config: ’/boot/loader.conf.local’
> 
> David's diagnosis of this is right- this is more of an informational
> message that you don't need to worry about.

Thanks.

>> Moreover, the "boot [kernel]" loader command does not work:
>>
>> OK ls /boot/kernel.old/kernel
>>      /boot/kernel.old/kernel
>> OK boot kernel.old
>> Command failed
>> OK boot /boot/kernel.old/kernel
>> Command failed
>> OK boot kernel
>> Command failed
>>
>> On the other hand, just "boot" works.
> 
> It seems that the Forth loader might be doing something sneaky and
> replacing the standard common "boot" with a Forth boot that handles
> this a lot better. CC'ing dteske@ so they can confirm.
> 
>> Finally, the double lines drawing a frame around the loader menu do not work
>> with the new loader and are replaced by ? characters in a box.
> 
> Interesting, I'll look into that... anything interesting/unique about
> your setup? r329387 should have addressed one potential cause of this,
> but I see you're past that.

I’m using a memory stick to boot a Lenovo ThinkPad S440 (i3-4030U 
processor, 4GB RAM). The only thing I can think of is that the ACPI of 
this model is not well supported, but the errors I have are related to 
thermal zones…:
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=201678

To build the memstick I’m using a 11.1-RELEASE VM under Hyper-V, with 
ccache and WITH_META_MODE, but this build process has been working 
nicely for months.

The kernel is based on GENERIC-NODEBUG and has been also working reliably:

juan at Server ~ % cat /root/kernels/MEMSTICK
include GENERIC-NODEBUG

ident   MEMSTICK

nodevice        fdc

nodevice        ch
nodevice        sa
nodevice        ses

nodevice        amr
nodevice        arcmsr
nodevice        ciss
nodevice        dpt
nodevice        hptmv
nodevice        hptnr
nodevice        hptrr
nodevice        hpt27xx
nodevice        iir
nodevice        ips
nodevice        mly
nodevice        twa
nodevice        tws

nodevice        aac
nodevice        aacp
nodevice        aacraid
nodevice        ida
nodevice        mfi
nodevice        mlx
nodevice        mrsas
nodevice        pmspcv
nodevice        twe

nodevice        nvme
nodevice        nvd

nodevice        virtio
nodevice        virtio_pci
nodevice        vtnet
nodevice        virtio_blk
nodevice        virtio_scsi
nodevice        virtio_balloon

nooptions       HYPERV
nodevice        hyperv

nooptions       XENHVM
nodevice        xenpci

nodevice        vmx

There is maybe something fishy in my src.conf, where I disable a lot of 
things to slim down the memstick, but still, it has been stable till now:

juan at Server ~ % cat /etc/src.conf
# For memory sticks

WITH_CCACHE_BUILD=

WITHOUT_ACCT=
WITHOUT_AMD=
WITHOUT_ATM=
WITHOUT_AUTHPF=
WITHOUT_AUTOFS=
WITHOUT_BHYVE=
WITHOUT_BLACKLIST=
# iwm does not support Bluetooth
WITHOUT_BLUETOOTH=
WITHOUT_BOOTPARAMD=
WITHOUT_BOOTPD=
# WITHOUT_BSDINSTALL enforced by WITHOUT_DIALOG
WITHOUT_BSNMP=
WITHOUT_CALENDAR=
# Don't set this when building HEAD from RELENG
# WITHOUT_CROSS_COMPILER=
WITHOUT_CTM=
WITHOUT_DEBUG_FILES=
#WITHOUT_DIALOG=
WITHOUT_DICT=
WITHOUT_EE=
WITHOUT_EXAMPLES=
WITHOUT_FDT=
WITHOUT_FINGER=
WITHOUT_FLOPPY=
# For testing the Lua loader (WITH_LOADER_LUA)
WITHOUT_FORTH=
WITHOUT_FREEBSD_UPDATE=
WITHOUT_GAMES=
WITHOUT_GCOV=
WITHOUT_GPIO=
# You disable Kerberos later, but try to keep GSSAPI for curl > pkg
# But this does not work, base Kerberos is required
#WITH_GSSAPI=
WITHOUT_GSSAPI=
WITHOUT_HAST=
WITHOUT_HESIOD=
WITHOUT_HTML=
WITHOUT_HYPERV=
WITHOUT_IPFILTER=
WITHOUT_IPFW=
WITHOUT_ISCSI=
WITHOUT_JAIL=
WITHOUT_KERBEROS=
WITHOUT_KERNEL_SYMBOLS=
WITHOUT_KVM=
WITHOUT_LDNS=
# This disables moused
#WITHOUT_LEGACY_CONSOLE=
WITHOUT_LLDB=
# This requires WITHOUT_FORTH
WITH_LOADER_LUA=
# This breaks setting locale and thus tmux
#WITHOUT_LOCALES=
WITHOUT_LPR=
WITHOUT_MAIL=
WITHOUT_NETCAT=
WITHOUT_PC_SYSINSTALL=
WITHOUT_PF=
WITHOUT_PORTSNAP=
WITHOUT_PPP=
WITHOUT_PROFILE=
WITHOUT_QUOTAS=
WITHOUT_RADIUS_SUPPORT=
WITHOUT_RBOOTD=
WITHOUT_RCS=
WITHOUT_SHAREDOCS=
WITH_SVN=
WITHOUT_SYSCONS=
WITHOUT_TALK=
WITHOUT_TCP_WRAPPERS=
WITHOUT_TELNET=
WITHOUT_TESTS=
WITHOUT_TFPT=
WITHOUT_TIMED=
WITHOUT_UNBOUND=
WITHOUT_UTMPX=
WITHOUT_ZFS=
WITHOUT_ZONEINFO=

Thanks for your attention.
Juan