Re: [9fans] vmware installation problems

Re: [9fans] vmware installation problems

[roger peppe]

[-- Attachment #1: Type: text/plain, Size: 732 bytes --]

well, now i've got another problem...
i'm gettting strange time behaviour in the virtual machine.
(as before, this is under VMWare 2.0.5, latest plan 9 ISO)

time goes much faster than it should,
and the VM hangs up for approx 7 seconds
every 40 seconds or so. after a hangup, system
time proceeds even faster for a little while,
before settling down again.

i've attached a pdf graph of the behaviour
(times in seconds, real time on the x-axis, system time on
the y-axis, data points generated by looping alternating
reading /dev/time and sleep(1000)). the
blank spots in the graph are where the system
was hung up unresponsive, not even the mouse cursor
moves.

anyone got any idea what might be going on here?

[-- Attachment #2: plan9-time-bug.pdf --]
[-- Type: application/pdf, Size: 52916 bytes --]

Re: [9fans] vmware installation problems

[Rodolfo (kix)]

Is an strange problem ... problem with interrupts?

Try to disable some hardware stuff, like the CD-ROM, USB, soundcard, etc.

This problem is only with plan9?

On Wed, Jul 29, 2009 at 4:49 PM, roger peppe<rogpeppe@gmail.com> wrote:
> well, now i've got another problem...
> i'm gettting strange time behaviour in the virtual machine.
> (as before, this is under VMWare 2.0.5, latest plan 9 ISO)
>
> time goes much faster than it should,
> and the VM hangs up for approx 7 seconds
> every 40 seconds or so. after a hangup, system
> time proceeds even faster for a little while,
> before settling down again.
>
> i've attached a pdf graph of the behaviour
> (times in seconds, real time on the x-axis, system time on
> the y-axis, data points generated by looping alternating
> reading /dev/time and sleep(1000)). the
> blank spots in the graph are where the system
> was hung up unresponsive, not even the mouse cursor
> moves.
>
> anyone got any idea what might be going on here?
>



-- 
Rodolfo García "kix"
EA4ERH - IN80ER

Re: [9fans] vmware installation problems

[roger peppe]

> This problem is only with plan9?

yup.
although it is plan9-inside-vmware, which could make
a significant difference.

Re: [9fans] vmware installation problems

[Tim Newsham]

>> This problem is only with plan9?
>
> yup.
> although it is plan9-inside-vmware, which could make
> a significant difference.

And you said VMWare 2.x, which is exceedingly old...

Tim Newsham
http://www.thenewsh.com/~newsham/

Re: [9fans] vmware installation problems

[roger peppe]

it's the latest version of VMWare Fusion AFAIK...

2009/7/29 Tim Newsham <newsham@lava.net>:
>>> This problem is only with plan9?
>>
>> yup.
>> although it is plan9-inside-vmware, which could make
>> a significant difference.
>
> And you said VMWare 2.x, which is exceedingly old...
>
> Tim Newsham
> http://www.thenewsh.com/~newsham/
>
>

Re: [9fans] vmware installation problems

[drivers]

Hello Roger,

	I'm currently running latest Plan 9 in vmware fusion Version 2.0 (116369).  This system has been behaving perfectly on every installation (save where i really botched something up trying to learn the system).

What is your host OS?  I'm running this on Mac OS X 10.5.8

hope we can get you a fix for this soon,

respectfully,

++james

Re: [9fans] vmware installation problems

[Tim Newsham]

> it's the latest version of VMWare Fusion AFAIK...

Ahh, my apologies.  I had incorrectly assumed "VMWare Workstation".

>> And you said VMWare 2.x, which is exceedingly old...

Tim Newsham
http://www.thenewsh.com/~newsham/

[9fans] vmware installation problems

[roger peppe]

i'm trying to install plan 9 under vmware (2.0.5 this time, having
failed with 1.0),
and it all worked fine (fossil + venti installation, 8GB disk, all
defaults, rebooted fine,
waited for venti to calm down, fshalt), except that when i came to reboot,
i got i/o errors as it was booting - here are some messages from the console:

init: starting /bin/rc
ndb/dns: can't read my ip address
2009/0728 00:53:27 err 4: write /dev/sdC0/isect offset 0xb0a4000 count
65536 buf 1dc6000 returned -1: i/o error
/boot/venti: part /dev/sdC0/isect addr 0xaf24000: icachewritesect
writepart: write /dev/sdC0/isect offset 0xb0a4000 count 65536 buf
1dc6000 returned -1: i/o error

there was then a long pause and then i got lots and lots more i/o
error messages, and it failed to boot.

initially, i thought that this was a one-off accident, so i
reinstalled everything
from scratch... and got exactly the same errors (with different addr
and offset).

someone else tried almost exactly the same thing (latest iso, md5:
d65178d8d8870c1488c64bc5ad8cdca8),
mac os, vmware fusion 2.0.0, and had no problems.

any ideas, anyone?

Re: [9fans] vmware installation problems

[erik quanstrom]

> there was then a long pause and then i got lots and lots more i/o
> error messages, and it failed to boot.

sounds like missed irqs.

> init: starting /bin/rc
> ndb/dns: can't read my ip address
> 2009/0728 00:53:27 err 4: write /dev/sdC0/isect offset 0xb0a4000 count
> 65536 buf 1dc6000 returned -1: i/o error
> /boot/venti: part /dev/sdC0/isect addr 0xaf24000: icachewritesect
> writepart: write /dev/sdC0/isect offset 0xb0a4000 count 65536 buf
> 1dc6000 returned -1: i/o error

a few things that might differ between the intaller
and the installed image that might cause lost irqs or
other problems, maybe in likelyhood order:
- the mp irq state,
- the dma state,
- the lba settings.

contrib quanstro/sd might have some extra debugging
that may help.  it also has some defense against missed
irqs.  it will log them and soldier on.  i know it's a
gruesome hack, but missed ide interrupts have been
a common disease and the easiest way to debug them
is to have a bootable system.

- erik