Xen project Mailing List

I think I have found the problem.

when I output the jiffies value when exec wait_for_devices at drivers/xen/xenbus/xen_probe_frontend.c

When I can boot my guest successfully, jiffies's value increases, and it is correct.

But when I can't boot it. the jiffies stay the same.

correct:

tc@box:~$ dmesg|grep jiffies

function wait_for_devices at jiffies 4294877484

function wait_for_devices at jiffies 4294877485

function wait_for_devices at jiffies 4294877486

function wait_for_devices at jiffies 4294877544

function wait_for_devices at jiffies 4294877850

function wait_for_devices at jiffies 4294877910<6>Setting capacity to 102400

wrong:

Initialising Xen virtual ethernet driver.

function wait_for_devices at jiffies 4294877302

Initializing USB Mass Storage driver...

i8042: PNP: No PS/2 controller found. Probing ports directly.

i8042: No controller found

mousedev: PS/2 mouse device common for all mice

function wait_for_devices at jiffies 4294877302

rtc_cmos rtc_cmos: rtc core: registered rtc_cmos as rtc0

rtc_cmos: probe of rtc_cmos failed with error -38

xen_wdt: Xen WatchDog Timer Driver v0.01

xen_wdt: initialized (timeout=60s, nowayout=0)

watchdog: Software Watchdog: cannot register miscdev on minor=130 (err=-16).

watchdog: Software Watchdog: a legacy watchdog module is probably present.

softdog: Software Watchdog Timer: 0.08 initialized. soft_noboot=0 soft_margin=60 sec soft_panic=0 (nowayout=0)

ledtrig-cpu: registered to indicate activity on CPUs

EFI Variables Facility v0.08 2004-May-17

zram: num_devices not specified. Using default: 1

zram: Creating 1 devices ...

Netfilter messages via NETLINK v0.30.

nfnl_acct: registering with nfnetlink.

IPv4 over IPv4 tunneling driver

IPv4 over IPSec tunneling driver

TCP: cubic registered

Initializing XFRM netlink socket

NET: Registered protocol family 17

NET: Registered protocol family 15

Key type dns_resolver registered

registered taskstats version 1

dpf test sleep

function wait_for_devices at jiffies 4294877302

Does anyone know how to get a stable clock?

On Tue, Apr 22, 2014 at 8:55 AM, Tuan Pengfei <pengfeituan@xxxxxxxxx> wrote:

My hypervisor is Fedora 20 withÂ3.11.10-301, and my guest is 3.8.13 of tiny core linux.

On Tue, Apr 22, 2014 at 8:47 AM, Wei Liu <wei.liu2@xxxxxxxxxx> wrote:

On Tue, Apr 22, 2014 at 08:40:47AM -0400, Tuan Pengfei wrote:
[...]

> now, i think the problems is around wait_for_devices.
> When I enable all the debug infomation, the guest boots successfully, but
> when I disable them, the guest fails with a probability. It freezes at
> stdout "registered taskstats version 1" Â.
> And I can't output kernel debug information, even though I set both
> hypervisor boot(log_lvl=all guest_loglvl=all) and guest boot
> parameter(earlyprintk=xen debug loglevel=8)
>
>

What kernel version do you use? If it is too old you might have better
luck trying a latest stable kernel.

Otherwise you need to have kernel debug message compiled in, capture log
with sensible backtrace and file a proper bug report.

Wei.

Re: [Xen-users] Fwd: xl can't create and boot vm when with on hard disk assignment