On Tue, Sep 26, 2017 at 01:19:04PM -0400, Daniel Wilkins wrote:
> On Tue, Sep 26, 2017 at 02:21:34AM -0700, Mike Larkin wrote:
> > On Tue, Sep 26, 2017 at 01:26:47AM -0400, Daniel Wilkins wrote:
> > > On Mon, Sep 25, 2017 at 10:16:21PM -0700, Mike Larkin wrote:
> > > > Please define "enough times".
> > > > 
> > > > -ml
> > > 
> > > Sadly it's not consistent. I think I got to 7 boots this time today or so;
> > > 5-10 seems like a good estimate to me. Pretty sure I've had it happen 
> > > after 3
> > > boots, think I got to 11 or 12 once. Since I don't recall whether I 
> > > brought
> > > it up on the original email the OS is irrelevant as far as I can tell, 
> > > I've
> > > had Linux trigger it, had OpenBSD trigger it, had Plan9 trigger it, I 
> > > think
> > > I triggered it once when I was seeing if I could get a DOS running in 
> > > there.
> > > 
> > 
> > Please provide your vm.conf or equivalent vmctl start command line, and the
> > vmd -dv log output (pkill -9 vmd && vmd -dv   and then vmctl log verbose)
> > when the issue occurs.
> > 
> > -ml
> > 
> 
> I've attached vm.conf (whether or not I have it is irrelevant though) and a 
> script
> session with the commands asked to make sure I didn't miss anything. The 
> vmctl start
> was the most minimal that triggers it, which is literally just "put it on an 
> image that
> wants a bios". vmd.log looks like the vm stuff somehow gets into a state 
> where an interrupt
> is messed up that seabios triggers?

> startup
> vm_opentty: vm lottie tty /dev/ttyp5 uid 0 gid 4 mode 620
> lottie: started vm 12 successfully, tty /dev/ttyp5
> loadfile_bios: loaded BIOS image
> run_vm: initializing hardware for vm lottie
> run_vm: starting vcpu threads for vm lottie
> vcpu_reset: resetting vcpu 0 for vm 12
> run_vm: waiting on events for VM lottie
> i8259_write_datareg: master pic, reset IRQ vector to 0x8
> i8259_write_datareg: slave pic, reset IRQ vector to 0x70
> virtio_blk_io: device reset
> i8259_nonspecific_eoi: master pic nonspecific eoi not supported

> Script started on Tue Sep 26 13:11:04 2017
> # pkill -9 vmd
> # cd lottie
> # 
# vmd -dv 2> vmd.log  &
> [1] 98067
> # vmctl log verbose
> # vmctl start lottie -d lottie.img                                            
>                                                                               
>                                                                               
>                               
> 
>  
# vmctl start lottie -d lottie.img  -c
> vmctl: starting without network interfaces
> Connected to /dev/ttyp5 (speed 9600)
> Changing serial settings was 0/0 now 3/0
> SeaBIOS (version 1.10.2p2-OpenBSD-vmm)

This is slightly downlevel. Probably not the issue but you might want to
fw_update anyway.

IIRC your machine is a Westmere. I have one of those too and I ran through
about 20 or 30 vmm+seabios boots just now without any problem. I'm wondering
if it's a leak of some sort, since my machine has more memory than yours.

Do you get more boots if you reduce the memory on each vmm boot to 128MB or
256MB?

-ml

> BUILD: gcc: (GCC) 4.2.1 20070719  binutils: 2.17
> enabling shadow ram
> Unable to unlock ram - bridge not found
> RamSize: 0x20000000 [cmos]
> malloc preinit
> malloc init
> init ivt
> init bda
> init bios32
> init keyboard
> init pic
> math cp init
> pci setup
> === PCI bus & bridge init ===
> PCI: pci_bios_init_bus_rec bus = 0x0
> === PCI device probing ===
> PCI probe
> Found 4 PCI devices (max PCI bus is 00)
> === PCI new allocation pass #1 ===
> PCI: check devices
> === PCI new allocation pass #2 ===
> PCI: IO: c000 - efff
> PCI: 32: 0000000020000000 - 00000000fec00000
> PCI: map device bdf=00:01.0  bar 0, addr 0000c000, size 00001000 [io]
> PCI: map device bdf=00:02.0  bar 0, addr 0000d000, size 00001000 [io]
> PCI: map device bdf=00:03.0  bar 0, addr 0000e000, size 00001000 [io]
> PCI: map device bdf=00:00.0  bar 6, addr febfc000, size 00001000 [mem]
> PCI: map device bdf=00:01.0  bar 6, addr febfd000, size 00001000 [mem]
> PCI: map device bdf=00:02.0  bar 6, addr febfe000, size 00001000 [mem]
> PCI: map device bdf=00:03.0  bar 6, addr febff000, size 00001000 [mem]
> PCI: init bdf=00:00.0 id=0b5d:0666
> PCI: init bdf=00:01.0 id=1af4:1005
> PCI: init bdf=00:02.0 id=1af4:1001
> PCI: init bdf=00:03.0 id=0b5d:0777
> PCI: No VGA devices found
> No apic - only the main cpu is present.
> init timer
> init virtio-blk
> found virtio-blk at 00:02.0
> pci dev 00:02.0 using legacy (0.9.5) virtio mode
> virtio-blk 00:02.0 blksize=512 sectors=104857600
> Registering bootable: Virtio disk PCI:00:02.0 (type:2 prio:9999 data:f1f70)
> init serial
> Found 1 serial ports
> Searching bootorder for: HALT
> Mapping hd drive 0x000f1f70 to 0
> drive 0x000f1f70: PCHS=0/0/0 translation=lba LCHS=1024/255/63 s=104857600
> malloc finalize
> Space available for UMB: c0000-ef000, f0000-f1f70
> Returned 253952 bytes of ZoneHigh
> e820 map has 6 items:
>   0: 0000000000000000 - 000000000009fc00 = 1 RAM
>   1: 000000000009fc00 - 00000000000a0000 = 2 RESERVED
>   2: 00000000000f0000 - 0000000000100000 = 2 RESERVED
>   3: 0000000000100000 - 000000001fffe000 = 1 RAM
>   4: 000000001fffe000 - 0000000020000000 = 2 RESERVED
>   5: 00000000fffc0000 - 0000000100000000 = 2 RESERVED
> locking shadow ram
> Unable to lock ram - bridge not found
> Jump to int19
> enter handle_19:
>   NULL
> enter handle_18:
>   NULL
> invalid handle_legacy_disk:729:
>    a=00000201  b=00000000  c=00000001  d=00000000 ds=0000 es=07c0 ss=0000
>   si=00000000 di=00000000 bp=00000000 sp=000066dc cs=f000 ip=d9bf  f=0202
> enter handle_18:
>   NULL
> 
> [EOT]
> # exit
> 
> Script done on Tue Sep 26 13:12:19 2017

Reply via email to