Re: [Soekris] FreeBSD/amd64 on Net6501 working!

2011-12-10 Thread Nicholas Esborn
On 12/10/2011 11:04 PM, Denis Fortin wrote:
> On 2011-12-11 01:54, Nicholas Esborn wrote:
>> I was able to boot FreeBSD/amd64 9.0-RC3 just now. The key was to
>> enable mptable and atpic support in the kernel. device atpic #
>> Optional legacy pic support device mptable # Optional MPSPEC mptable
>> support I also removed "device acpi", but that may not have been
>> necessary. -nick
> 
> This is excellent news.
> 
> Based on your input and what I had found earlier, I made a couple of
> FreeBSD 9.0-RC3 kernels to play with the options, and I found that
> "device mptable" is needed for the FreeBSD/amd64 kernel to boot on the
> net6501.
> 
> Anyway, it's good to know that we'll be able to build and use 64-bit
> kernels... but I doubt we can get "device mptable" enabled in GENERIC at
> this late stage of the 9.0 release cycle (we are at RC3).

From what I have seen, mptable and atpic support are both designated as
legacy capabilities on FreeBSD.  The FreeBSD developers may not want to
re-enable them in GENERIC.  Support for non-ACPI configurations could
become increasingly tenuous.

With luck, the required options will remain available for some time to
come, but it seems like an ACPI BIOS would be a very positive
development for the longevity of the Net6501 with modern operating systems.

-nick

> Cheers,
> 
> Denis, for...@acm.org
> 
> PS. Now onwards to find out why my USB 3.0 PCIe card is not working as
> expected.


-- 
n...@desert.net - all messages cryptographically signed



signature.asc
Description: OpenPGP digital signature
___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] FreeBSD/amd64 on Net6501 working!

2011-12-10 Thread Denis Fortin
On 2011-12-11 01:54, Nicholas Esborn wrote:
> I was able to boot FreeBSD/amd64 9.0-RC3 just now. The key was to 
> enable mptable and atpic support in the kernel. device atpic # 
> Optional legacy pic support device mptable # Optional MPSPEC mptable 
> support I also removed "device acpi", but that may not have been 
> necessary. -nick

This is excellent news.

Based on your input and what I had found earlier, I made a couple of 
FreeBSD 9.0-RC3 kernels to play with the options, and I found that 
"device mptable" is needed for the FreeBSD/amd64 kernel to boot on the 
net6501.

Anyway, it's good to know that we'll be able to build and use 64-bit 
kernels... but I doubt we can get "device mptable" enabled in GENERIC at 
this late stage of the 9.0 release cycle (we are at RC3).

Cheers,

Denis, for...@acm.org

PS. Now onwards to find out why my USB 3.0 PCIe card is not working as 
expected.
___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] FreeBSD/amd64 on Net6501 working!

2011-12-10 Thread Nicholas Esborn
On 12/10/2011 06:07 AM, Denis Fortin wrote:
> On 2011-12-08 08:55, Aragon Gouveia wrote:
>> On 12/07/11 23:00, Soren Kristensen wrote:
>>> I don't have the knowledge to find out where and why FreeBSD stops, but
>>> if some FreeBSD hacker can go down and tell me why, I will do what I can
>>> do fix it if the BIOS need a little adjustment
>>
>> Considering all the ACPI error messages, I guess it's a problem there
>> somewhere.  The freebsd-acpi@ mailing list is a good place to seek ACPI
>> wizards. :)
> 
> If I try to boot FreeBSD 9.0-RC2 in verbose mode with amd64 and i386 and 
> compare the boot messages side-by-side, there are as many error messages 
> about ACPI with the (successful) i386 boot than with the (failed) amd64 
> boot.
> 
> The amd64 boot ends with the following lines:
>  > panic: No usable event timer found!
>  > cpuid = 0
>  > KDB: stack backtrace:
>  > #0 0x80867e4e at kdb_backtrace+0x5e
>  > #1 0x80832ab7 at panic+0x187
>  > #2 0x80b5250b at cpu_initclocks_bsp+0x3cb
>  > #3 0x807e9ec0 at initclocks+0x20
>  > #4 0x807e7507 at mi_startup+0x77
>  > #5 0x8029f6ec at btext+0x2c
>  > Uptime: 1s
>  > Automatic reboot in 15 seconds - press a key on the console to abort
> 
> Whereas the i386 boot goes on at this same spot with:
>  > lapic: Divisor 2, Frequency 50001280 Hz
>  > Timecounters tick every 1.000 msec
> 
> Somehow, the "LAPIC" timer is reported with the i386 boot thus:
>  > MPTable: 
>  > Event timer "LAPIC" quality 400
>  > FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
>  > FreeBSD/SMP: 1 package(s) x 1 core(s) x 2 HTT threads
>  >  cpu0 (BSP): APIC ID:  0
>  >  cpu1 (AP/HT): APIC ID:  1
> 
> but not on the amd64 boot.
> 
> Also, the i386 reports early on:
>  > MP Configuration Table version 1.4 found at 0xc00fe110
> 
> and *not* the amd64 boot.
> 
> Anyway, just some preliminary information in case it triggers an Ah-ha! 
> with some of the list members.

I was able to boot FreeBSD/amd64 9.0-RC3 just now.

The key was to enable mptable and atpic support in the kernel.

device  atpic  # Optional legacy pic support
device  mptable# Optional MPSPEC mptable support

I also removed "device acpi", but that may not have been necessary.

-nick

> Cheers,
> 
> Denis F.
> ___
> Soekris-tech mailing list
> Soekris-tech@lists.soekris.com
> http://lists.soekris.com/mailman/listinfo/soekris-tech


-- 
n...@desert.net - all messages cryptographically signed



signature.asc
Description: OpenPGP digital signature
___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] vr0 using OpenBSD stops responding.

2011-12-10 Thread Stuart Henderson
On 2011-12-10, Frank Schuhmann  wrote:
> Hi Brandan,
>
> on the Windows 7 side all significant updates are done?
> You?ve checked the Windows 7 internal firewall, to be sure that the false is 
> not
> pointed or placed in their?
>
> Because you wrote that the Win XP was not producing the problem, but Win 7
> should do it, it is perhaps the TCP window scaling, as declared in the RFC 
> 1323.
> In some similar other cases it was sometimes fixing the problem (pfsense) and 
> in
> sometimes also not, but to quick test it out the spended time will to be of
> value.   

Yes this is what I meant about the improvements in the Windows network stack.
This sort of thing used to cause a problem with stateful firewall rules in 
places
(because we do sequence number tracking) until we started doing "flags s/sa keep
state" by default. However it wouldn't result into the interface totally locking
up until it's reset. (btw Brandon, Note that "ifconfig vrX down; ifconfig vrX 
up"
would probably also clear it without a full reboot).

> On the BSD side perhaps you try to turn of the scaling, I?m not an OpenBSD
> professional but it must be turn out by setting up a shell order likes  -w net.inet.tcp.rfc1323=0> 
> If not or it must be typed in, in other direction it will be super that 
> perhaps
> an OpenBSD familiar list member would correct this please.

This is the correct command but it only affects the local tcp stack,
not forwarded packets, so I think this will have no effect here.

>
> And at the other point I consider what Stuart Henderson was mentoring to you.
> Perhaps the best at all,   
> but keeping an eye on the net6501 is also a nice option ;)
>
>
> I hope this is helping you out. 
>
> 
> Kind regards
>
> Frank
>
> -Original Message-
> From: soekris-tech-boun...@lists.soekris.com
> [mailto:soekris-tech-boun...@lists.soekris.com] On Behalf Of Brandan Rowley
> Sent: Friday, December 09, 2011 11:11 PM
> To: 'soekris-tech@lists.soekris.com'
> Subject: [Soekris] vr0 using OpenBSD stops responding.
>
>>Hi,
>>I am new to the list so be gentle if this has been posted already.  I am using
> two net5501 (with the VPN chip) running OpenBSD 4.9 to setup a VPN tunnel.  
>>The tunnel has been up and running for a while.  We've recently added Windows 
>>7
> PC to the network.  Performing file transfers from the Windows 7 PC's 
>>across the VPN tunnel causes the internal interface of the net5501 to stop
> responding.  A reboot is needed to get the interface communicating again.  
>>This repeatable.  Windows XP clients have no issues.  Is this a fix or
> workaround for this?  I've tried OpenBSD 5.0 and read of similar issues on
> OpenBSD, 
>>but have not found a resolution.
>
>>Regards,
>>Brandan
>
>
>  
>
> __ Hinweis von ESET NOD32 Antivirus, Signaturdatenbank-Version 6699
> (20111210) __
>
> E-Mail wurde gepr?ft mit ESET NOD32 Antivirus.
>
> http://www.eset.com
> =

___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] FreeBSD/amd64 on Net6501

2011-12-10 Thread Denis Fortin
On 2011-12-08 08:55, Aragon Gouveia wrote:
> On 12/07/11 23:00, Soren Kristensen wrote:
>> I don't have the knowledge to find out where and why FreeBSD stops, but
>> if some FreeBSD hacker can go down and tell me why, I will do what I can
>> do fix it if the BIOS need a little adjustment
>
> Considering all the ACPI error messages, I guess it's a problem there
> somewhere.  The freebsd-acpi@ mailing list is a good place to seek ACPI
> wizards. :)

If I try to boot FreeBSD 9.0-RC2 in verbose mode with amd64 and i386 and 
compare the boot messages side-by-side, there are as many error messages 
about ACPI with the (successful) i386 boot than with the (failed) amd64 
boot.

The amd64 boot ends with the following lines:
 > panic: No usable event timer found!
 > cpuid = 0
 > KDB: stack backtrace:
 > #0 0x80867e4e at kdb_backtrace+0x5e
 > #1 0x80832ab7 at panic+0x187
 > #2 0x80b5250b at cpu_initclocks_bsp+0x3cb
 > #3 0x807e9ec0 at initclocks+0x20
 > #4 0x807e7507 at mi_startup+0x77
 > #5 0x8029f6ec at btext+0x2c
 > Uptime: 1s
 > Automatic reboot in 15 seconds - press a key on the console to abort

Whereas the i386 boot goes on at this same spot with:
 > lapic: Divisor 2, Frequency 50001280 Hz
 > Timecounters tick every 1.000 msec

Somehow, the "LAPIC" timer is reported with the i386 boot thus:
 > MPTable: 
 > Event timer "LAPIC" quality 400
 > FreeBSD/SMP: Multiprocessor System Detected: 2 CPUs
 > FreeBSD/SMP: 1 package(s) x 1 core(s) x 2 HTT threads
 >  cpu0 (BSP): APIC ID:  0
 >  cpu1 (AP/HT): APIC ID:  1

but not on the amd64 boot.

Also, the i386 reports early on:
 > MP Configuration Table version 1.4 found at 0xc00fe110

and *not* the amd64 boot.

Anyway, just some preliminary information in case it triggers an Ah-ha! 
with some of the list members.

Cheers,

Denis F.
___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] vr0 using OpenBSD stops responding.

2011-12-10 Thread Frank Schuhmann
Hi Brandan,

on the Windows 7 side all significant updates are done?
You´ve checked the Windows 7 internal firewall, to be sure that the false is not
pointed or placed in their?

Because you wrote that the Win XP was not producing the problem, but Win 7
should do it, it is perhaps the TCP window scaling, as declared in the RFC 1323.
In some similar other cases it was sometimes fixing the problem (pfsense) and in
sometimes also not, but to quick test it out the spended time will to be of
value.   
On the BSD side perhaps you try to turn of the scaling, I´m not an OpenBSD
professional but it must be turn out by setting up a shell order likes  
If not or it must be typed in, in other direction it will be super that perhaps
an OpenBSD familiar list member would correct this please.

And at the other point I consider what Stuart Henderson was mentoring to you.
Perhaps the best at all,   
but keeping an eye on the net6501 is also a nice option ;)


I hope this is helping you out. 


Kind regards

Frank

-Original Message-
From: soekris-tech-boun...@lists.soekris.com
[mailto:soekris-tech-boun...@lists.soekris.com] On Behalf Of Brandan Rowley
Sent: Friday, December 09, 2011 11:11 PM
To: 'soekris-tech@lists.soekris.com'
Subject: [Soekris] vr0 using OpenBSD stops responding.

>Hi,
>I am new to the list so be gentle if this has been posted already.  I am using
two net5501 (with the VPN chip) running OpenBSD 4.9 to setup a VPN tunnel.  
>The tunnel has been up and running for a while.  We've recently added Windows 7
PC to the network.  Performing file transfers from the Windows 7 PC's 
>across the VPN tunnel causes the internal interface of the net5501 to stop
responding.  A reboot is needed to get the interface communicating again.  
>This repeatable.  Windows XP clients have no issues.  Is this a fix or
workaround for this?  I've tried OpenBSD 5.0 and read of similar issues on
OpenBSD, 
>but have not found a resolution.

>Regards,
>Brandan


 

__ Hinweis von ESET NOD32 Antivirus, Signaturdatenbank-Version 6699
(20111210) __

E-Mail wurde geprüft mit ESET NOD32 Antivirus.

http://www.eset.com
 

___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] vr0 using OpenBSD stops responding.

2011-12-10 Thread Stuart Henderson
Might be worth looking at ciphers/hashes and see if switching to
a different type improves things.

Don't automatically assume that hw acceleration is going to help;
it reduces cpu use but has high overheads so it can make things
worse in some cases.


On 2011-12-09, Brandan Rowley  wrote:
> Hi Aric,
>
> Thanks for the reply.  That sounds about right, I'm getting about 7Mbit of 
> IPsec traffic.  I have tried turning on/off ipcomp with no real improvement.  
> For this particular tunnel I'm now needing more throughput.  I've been 
> looking at the net6501 and am considering giving it a try.
>
> Regards,
> Brandan
>
> -Original Message-
> From: Aric Warsaw [mailto:awar...@gmail.com] 
> Sent: Friday, December 09, 2011 3:43 PM
> To: Brandan Rowley
> Cc: soekris-tech@lists.soekris.com
> Subject: Re: [Soekris] vr0 using OpenBSD stops responding.
>
> Hi Brandan
>
> I've experienced just this even with the crypto accelerator card.
> That was back on OpenBSD 4.8.  I've found that about 8-10Mbit of IPsec 
> traffic is all you're going to get out of these guys...your mileage may vary. 
>  Taking IPsec full throttle for 30 minutes or longer was causing my 5501 to 
> fall off the networklost layer 2 and serial to the box entirely.
>
> My work around was to QoS the IPsec traffic down.  Another option that I 
> hadn't done personally and assuming you've enabled ipcomp, you can turn it 
> off to save CPU resources.  Again hadn't done it.  Most IPsec documentation 
> for OpenBSD tells you to enable this in /etc/sysctl.conf, so it might break 
> your flows and SAs.
>
> This is the setting:  net.inet.ipcomp.enable=1  # 0 to disable or just 
> comment out the line
>
> Let us know if you try it and notice a difference.  Hope that helps.
>
> -Aric
>
>
>
>
>
> 2011/12/9 Brandan Rowley :
>> Hi,
>>
>> I am new to the list so be gentle if this has been posted already.? I 
>> am using two net5501 (with the VPN chip) running OpenBSD 4.9 to setup 
>> a VPN tunnel.? The tunnel has been up and running for a while.? We've 
>> recently added Windows 7 PC to the network.? Performing file transfers 
>> from the Windows 7 PC's across the VPN tunnel causes the internal 
>> interface of the
>> net5501 to stop responding.? A reboot is needed to get the interface 
>> communicating again.? This repeatable.? Windows XP clients have no issues.
>> Is this a fix or workaround for this?? I've tried OpenBSD 5.0 and read 
>> of similar issues on OpenBSD, but have not found a resolution.
>>
>> Regards,
>> Brandan
>>
>>
>> ___
>> Soekris-tech mailing list
>> Soekris-tech@lists.soekris.com
>> http://lists.soekris.com/mailman/listinfo/soekris-tech
>>

___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech


Re: [Soekris] vr0 using OpenBSD stops responding.

2011-12-10 Thread Stuart Henderson
On 2011-12-09, Brandan Rowley  wrote:
> Windows 7 PC to the network.  Performing file transfers from the Windows 7 =
> PC's across the VPN tunnel causes the internal interface of the net5501 to =
> stop responding.  A reboot is needed to get the interface communicating aga=
> in.  This repeatable.  Windows XP clients have no issues.  Is this a fix or=
>  workaround for this?  I've tried OpenBSD 5.0 and read of similar issues on=
>  OpenBSD, but have not found a resolution.

There was a very recent fix for a vr(4) problem committed to -current,
I would try updating the OS to 5.0 then updating sys/dev/pci/if_vr*
to the latest version and building a new kernel.

http://www.openbsd.org/cgi-bin/cvsweb/src/sys/dev/pci/

This is probably a candidaatte for going into 5.0-stable after it's
had more testing.

As to why Windows 7 but not XP triggers it, this is probably due to
improvements in their network stack making it likely traffic is sent
more rapidly over a medium/high-latency link.


___
Soekris-tech mailing list
Soekris-tech@lists.soekris.com
http://lists.soekris.com/mailman/listinfo/soekris-tech