[9fans] plan9port openbsd 4.4

[9fans] plan9port openbsd 4.4

[Fernan Bolando]

Hi all

I recently installed plan9port on an openbsd pc in order gain access
to acme and other tools.

rio seems to work fine, but I have been getting a few 9serve.core dumps.

I also get the following when running acme
_procstart rfork_thread: Resource temporarily unavailable
Abort trap
....and few broken pipes message.

any openbsd experts here?

fernan
--
http://www.fernski.com

Re: [9fans] plan9port openbsd 4.4

[Iruata Souza]

could you send me the core file? are you linking with rthreads?

iru

On Sun, Jan 25, 2009 at 11:34 PM, Fernan Bolando
<fernanbolando@mailc.net> wrote:
> Hi all
>
> I recently installed plan9port on an openbsd pc in order gain access
> to acme and other tools.
>
> rio seems to work fine, but I have been getting a few 9serve.core dumps.
>
> I also get the following when running acme
> _procstart rfork_thread: Resource temporarily unavailable
> Abort trap
> ....and few broken pipes message.
>
> any openbsd experts here?
>
> fernan
> --
> http://www.fernski.com
>
>



--
iru

Re: [9fans] plan9port openbsd 4.4

[Fernan Bolando]

On 1/26/09, Iruata Souza <iru.muzgo@gmail.com> wrote:
> could you send me the core file? are you linking with rthreads?
>
> iru
>
> On Sun, Jan 25, 2009 at 11:34 PM, Fernan Bolando
> <fernanbolando@mailc.net> wrote:
>> Hi all
>>
>> I recently installed plan9port on an openbsd pc in order gain access
>> to acme and other tools.
>>
>> rio seems to work fine, but I have been getting a few 9serve.core dumps.
>>
>> I also get the following when running acme
>> _procstart rfork_thread: Resource temporarily unavailable
>> Abort trap
>> ....and few broken pipes message.
>>
>> any openbsd experts here?
>>
>> fernan
>> --
>> http://www.fernski.com
>>
Hi Iru

Applogies for the multiple replies, having problems sending with
attachments. Core files was sent directly to your gmail account.

I just followed the same install process I do with linux...type ./INSTAKK

again sorry about the multiple replies. and thanks for your help

regards
fernan
--
http://www.fernski.com

Re: [9fans] plan9port openbsd 4.4

[andrey mirtchovski]

> rio seems to work fine, but I have been getting a few 9serve.core dumps.

that makes sense: p9p's rio (used to be 9wm) is written with X-windows
in mind. it's not a Plan 9 app.

Re: [9fans] plan9port openbsd 4.4

[sqweek]

On Mon, Jan 26, 2009 at 10:34 AM, Fernan Bolando
<fernanbolando@mailc.net> wrote:
> rio seems to work fine, but I have been getting a few 9serve.core dumps.
>
> I also get the following when running acme
> _procstart rfork_thread: Resource temporarily unavailable
> Abort trap
> ....and few broken pipes message.

 Hm, does this work?

9p -n -a sources.cs.bell-labs.com ls /

 I run netbsd and after recently updating p9p I get a Memory Fault
trying to do this, and a couple of process hanging around. I had a
quick go at debugging it but it was fairly inscrutible. I got the
impression it was libthread related but haven't found time to look at
it in more detail.
-sqweek

Re: [9fans] plan9port openbsd 4.4

[Russ Cox]

The plan9port code depends on the operating system's pthreads
being real kernel-level threads, not a fake user-level simulation.
The user-level simulations are not good enough, because
on the x86 they cut corners and use the stack pointer
to locate the thread-local state.  The Plan 9 threaded
programs manage their own stacks, making it impossible
for the user-space simulations to find their thread-local state.

Most Linux distributions switched to real threads (i.e., dropped
LinuxThreads in favor of NPTL) around the time they switched
to the 2.6 kernel.  FreeBSD switched in the FreeBSD 5 release.

Last I had heard, OpenBSD was still plodding along with
user-level threads.  Until they fix that, programs like acme
will not run.

Russ

Re: [9fans] plan9port openbsd 4.4

[Anthony Sorace]

if you want to work on it some, this message talks about getting
inferno working on OpenBSD using the rthreads library (the pending
replacement for the userland threads russ talked about):

http://www.xs4all.nl/~mechiel/inferno/openbsd43-old.html

you could likely follow a similar path for p9p if you're so inclined.

inferno has since gained an explicit OpenBSD port (for those
directions, remove "-old" from the above URL), but i'm not sure
whether that involves working around the userland thread model or
explicitly relying on rthreads.

Re: [9fans] plan9port openbsd 4.4

[Mechiel Lukkien]

On Mon, Jan 26, 2009 at 01:01:55PM -0500, Anthony Sorace wrote:
> if you want to work on it some, this message talks about getting
> inferno working on OpenBSD using the rthreads library (the pending
> replacement for the userland threads russ talked about):
>
> http://www.xs4all.nl/~mechiel/inferno/openbsd43-old.html
>
> you could likely follow a similar path for p9p if you're so inclined.

i don't recall the details, but i think those instructions let you do
too much.  i.e. i think the rfork-like function has been in openbsd for
longer than rthreads, and the instructions & patch at the url don't
really use the rthreads library.  btw, rthreads on openbsd had been
stalled, there does seem to be a dev woroking on that now.

as for the error of the original poster:  perhaps your process limit (or
other limit) is too low?  inferno spawns quite a few procs, which can
make me hit the default process limit (defined in /etc/login.conf).

mjl

>
> inferno has since gained an explicit OpenBSD port (for those
> directions, remove "-old" from the above URL), but i'm not sure
> whether that involves working around the userland thread model or
> explicitly relying on rthreads.

Re: [9fans] plan9port openbsd 4.4

[Iruata Souza]

can't remember on 4.4, but 4.3 did run acme fine.

On Mon, Jan 26, 2009 at 2:04 PM, Russ Cox <rsc@swtch.com> wrote:
> The plan9port code depends on the operating system's pthreads
> being real kernel-level threads, not a fake user-level simulation.
> The user-level simulations are not good enough, because
> on the x86 they cut corners and use the stack pointer
> to locate the thread-local state.  The Plan 9 threaded
> programs manage their own stacks, making it impossible
> for the user-space simulations to find their thread-local state.
>
> Most Linux distributions switched to real threads (i.e., dropped
> LinuxThreads in favor of NPTL) around the time they switched
> to the 2.6 kernel.  FreeBSD switched in the FreeBSD 5 release.
>
> Last I had heard, OpenBSD was still plodding along with
> user-level threads.  Until they fix that, programs like acme
> will not run.
>
> Russ
>
>



--
iru

Re: [9fans] plan9port openbsd 4.4

[Fernan Bolando]

> On Mon, Jan 26, 2009 at 2:04 PM, Russ Cox <rsc@swtch.com> wrote:
>> The plan9port code depends on the operating system's pthreads
>> being real kernel-level threads, not a fake user-level simulation.
>> The user-level simulations are not good enough, because
>> on the x86 they cut corners and use the stack pointer
>> to locate the thread-local state.  The Plan 9 threaded
>> programs manage their own stacks, making it impossible
>> for the user-space simulations to find their thread-local state.
>>
>> Most Linux distributions switched to real threads (i.e., dropped
>> LinuxThreads in favor of NPTL) around the time they switched
>> to the 2.6 kernel.  FreeBSD switched in the FreeBSD 5 release.
>>
>> Last I had heard, OpenBSD was still plodding along with
>> user-level threads.  Until they fix that, programs like acme
>> will not run.
>>
>> Russ
>>
>>
>
On 1/27/09, Iruata Souza <iru.muzgo@gmail.com> wrote:
> can't remember on 4.4, but 4.3 did run acme fine.
>
>
> --
> iru
>
>

Yes, actually acme runs fine, I just get an occasional burp.

--
http://www.fernski.com

Re: [9fans] plan9port openbsd 4.4

[Fernan Bolando]

Hi all

I just everybody to know that...As usual Russ was right.
The occasional burp happens when acme tries to fork a new thread.

now I am back to using drawterm+qemu on openbsd

fernan

On 1/27/09, Fernan Bolando <fernanbolando@mailc.net> wrote:
>> On Mon, Jan 26, 2009 at 2:04 PM, Russ Cox <rsc@swtch.com> wrote:
>>> The plan9port code depends on the operating system's pthreads
>>> being real kernel-level threads, not a fake user-level simulation.
>>> The user-level simulations are not good enough, because
>>> on the x86 they cut corners and use the stack pointer
>>> to locate the thread-local state.  The Plan 9 threaded
>>> programs manage their own stacks, making it impossible
>>> for the user-space simulations to find their thread-local state.
>>>
>>> Most Linux distributions switched to real threads (i.e., dropped
>>> LinuxThreads in favor of NPTL) around the time they switched
>>> to the 2.6 kernel.  FreeBSD switched in the FreeBSD 5 release.
>>>
>>> Last I had heard, OpenBSD was still plodding along with
>>> user-level threads.  Until they fix that, programs like acme
>>> will not run.
>>>
>>> Russ
>>>
>>>
>>
> On 1/27/09, Iruata Souza <iru.muzgo@gmail.com> wrote:
>> can't remember on 4.4, but 4.3 did run acme fine.
>>
>>
>> --
>> iru
>>
>>
>
> Yes, actually acme runs fine, I just get an occasional burp.
>
> --
> http://www.fernski.com
>


--
http://www.fernski.com