On Sun, Jan 14, 2024 at 10:43:54AM +0100, Peter J. Philipp wrote:
> Hi!
>
> Yesterday, I was for the last 6 hours or so, trying to get netbsd running in
> vmm.  What  I used was netbsd 10.0 RC2 and here is how I went about it:
>
> 1. upon vmm start press 2 at console this brings one into a boot config
> 2. enter "consdev com0,9600" to initialize com0 console
> 3. boot -cav
> 4. in the kernel config "disable tpm" I found this is helpful continuing on 
> past com0
> 5. after this point the symptoms are similar for NFS diskless booting or 
> letting it boot on the .img file, one enters the root filesystem and 
> eventually the path to init [/sbin/init] and then vmd kills the vmm.  Here is 
> what it says on its debug (vmd -dvvv):
>
> vm/netbsd: vcpu_process_com_lcr: set baudrate = 9600
> vm/netbsd: vcpu_exit_eptviolation: fault already handled
> vm/netbsd: vcpu_exit_eptviolation: fault already handled
> vm/netbsd: vcpu_exit_eptviolation: fault already handled
> vm/netbsd: vcpu_exit_eptviolation: fault already handled
> vm/netbsd: vm/netbsd: vcpu_exit_inout: can't emulate REP prefixed IN(S)/OUT(S)
> vm/netbsd/vionet0: handle_sync_io: pipe dead (EV_READ)
> vm/netbsd/vionet0: dev_dispatch_vm: pipe dead (EV_READ)
> vm/netbsd/vioblk0: handle_sync_io: vioblk pipe dead (EV_READ)
> vm/netbsd/vioblk0: dev_dispatch_vm: pipe dead (EV_READ)
> vmm: vmm_sighdlr: handling signal 20
> vmm: vmm_sighdlr: terminated vm netbsd (id 3)
>
> I think the clue here is the "can't emulate REP prefixed IN(S)/OUT(S)".. I 
> don't know asm too well but is this a limitation with the "rep" opcode?
>

yes. this is not yet supported. dv and I both have diffs to make this work
but they are not yet ready.

> My vmm console looks like this:
>
> --->
>    2.2978862] crypto: driver 0 registers alg 22 flags 0 maxoplen 0
> [   2.2978862] swwdog0: software watchdog initialized
> [   2.3137243] WARNING: 1 error while detecting hardware; check system log.
> [   2.3137243] boot device: ld0
> [   2.3137243] root device (default dk1): vioif0
> [   4.6983828] dump device:
> [   5.2377099] file system (default generic): nfs
> [   6.4505124] root on vioif0
> [   6.4523019] nfs_boot: trying DHCP/BOOTP
> [  68.2946438] nfs_boot: timeout...
> [  85.4568907] nfs_boot: timeout...
> [  87.1923785] nfs_boot: DHCP next-server: 172.16.88.1
> [  87.1923785] nfs_boot: my_domain=mainrechner.de
> [  87.1923785] nfs_boot: my_addr=172.16.88.3
> [  87.1923785] nfs_boot: my_mask=255.255.255.0
> [  87.1923785] nfs_boot: gateway=172.16.88.1
> [ 128.4226631] root on 172.16.88.1:/netbsd
> [ 128.4226631] kern.module.path=/stand/amd64/10.0/modules
> [ 128.4294984] init path (default /sbin/init):
> [ 129.1605695] init: trying /sbin/init
>
> [EOT]
> <---
>
> It always dies on execve'ing /sbin/init...
>
> I tried to investigate this to the best I could, as I'm starting to doing 
> beta testing of my software (delphinusdnsd), but I guess I'll have to recover 
> a hyper-v netbsd image.
>
> Best Regards,
>
> -peter
>
> PS: if this is simply not working for anyone just say it please.
>
> --
> Over thirty years experience on UNIX-like Operating Systems starting with QNX.
>

Reply via email to