Brian McEwen | 1 May 04:34 2006
Picon
Picon

libgpg-error build issue


Hi; anyone have a workaround for this?

from -current pkgsrc as of late April  (how do I check when last I  
did a cvs update, anyway-- short of a paper log ;)

thanks,

Brian
-------------

===> Extracting for libgcrypt-1.2.2nb1
===> Required package libgpg-error>=1.0nb1: NOT found
===> Verifying reinstall for ../../security/libgpg-error
make: "../../mk/tools/../../mk/tools/../../mk/tools/../../mk/tools/ 
replace.mk" line 154: Malformed conditional (($ 
{_TOOLS_DEPMETHOD.gawk} == "BUILD_DEPENDS") &&  defined 
(_TOOLS_DEPMETHOD.awk))
make: "../../mk/tools/../../mk/tools/../../mk/tools/../../mk/tools/ 
replace.mk" line 154: Missing dependency operator
make: "../../mk/tools/../../mk/tools/bsd.tools.mk" line 219: if-less  
endif
make: "../../mk/tools/../../mk/tools/bsd.tools.mk" line 219: Need an  
operator
make: Fatal errors encountered -- cannot continue

Izumi Tsutsui | 8 May 15:50 2006
Picon

Re: Qube2 Network Performance & Forwarding

In article <17491.5821.507807.34738 <at> cochise.kilbi.de>
mk <at> kilbi.de wrote:

> Maybe I can help here with my qube2 (which shows identical ifconfig -m
> tlp{0,1} results):
 :
> Does it only contain the mac addresses?

Yes, and I think if_tlp_pci.c has already handled it
by vendor ID in the MAC address.
(Maybe Linux tulip driver implicitly refers SROM contents
 specified by DEC regardless of its vendor)
If the performance problem is caused by inappropriate
MII settings it would cause more awful results,
like <100KB/s xfer rate.

Anyway, basically tulip chips can't receive packets at unaligned
word address via DMA directly, so RX function in tulip driver
must copy RX data from DMA buffer to mbufs, so xfer rate will
be limitted by bandwidth of memory-to-memory-copy.
(see sys/dev/ic/tulip.c around #ifdef __NO_STRICT_ALIGNMENT)
ex(4), epic(4), and pcn(4) etc. don't have such restriction
(i.e. they can do DMA to unaligned buffer) so we could get
a bit better performance with them, but maybe another problem
is that NetBSD/mips pmap is not so optimized for MIPS3 CPUs
(especially ones which don't have VCE support) and there are
too many cache flush ops to avoid virtual cache aliases,
so it might cause less performance than Linux.
(though I haven't checked Linux VM sources...)
---
(Continue reading)

JEFF SCOTT | 8 May 18:47 2006
Picon
Picon

perl on netbsd 3.99

hello,

I'm new to this group, and just build my first raq2 w/netbsd.  I'm a long time
cobalt and linux user, so I'm not to savvy on netbsd.  The build went find,
but slow (2-3 hours on 60gb disk).  However, I don't seem to have perl and can
not find a perl package for 3.99.  

Any reason perl is not included?  Box is worthless to me without it.

Thanks.

Andy Ruhl | 8 May 20:34 2006
Picon

Re: perl on netbsd 3.99

On 5/8/06, JEFF SCOTT <jeff.scott <at> usa.net> wrote:
> hello,
>
> I'm new to this group, and just build my first raq2 w/netbsd.  I'm a long time
> cobalt and linux user, so I'm not to savvy on netbsd.  The build went find,
> but slow (2-3 hours on 60gb disk).  However, I don't seem to have perl and can
> not find a perl package for 3.99.
>
> Any reason perl is not included?  Box is worthless to me without it.

If you're asking why Perl is not included with NetBSD, it never has
been as far as I know. It's not up to the NetBSD core developers to
maintain it and ship it.

If you want it, you have to get it through pkgsrc.

Get the pkgsrc tarball from the NetBSD ftp site (packages directory)
untar it in the /usr directory, and then go to the directory for perl
(don't remember what it is off hand) and build it through pkgsrc.

Andy

Brian McEwen | 8 May 20:41 2006
Picon
Picon

Re: perl on netbsd 3.99


On May 8, 2006, at 12:47 PM, JEFF SCOTT wrote:

> hello,
>
> I'm new to this group, and just build my first raq2 w/netbsd.  I'm  
> a long time
> cobalt and linux user, so I'm not to savvy on netbsd.  The build  
> went find,
> but slow (2-3 hours on 60gb disk).  However, I don't seem to have  
> perl and can
> not find a perl package for 3.99.
>
> Any reason perl is not included?  Box is worthless to me without it.

perl is part of pkgsrc,

/usr/pkgsrc/lang/perl5/work/perl-5.8.6/
I guess is what you want?

to get pkgsrc,

setenv CVSROOT anoncvs <at> anoncvs.NetBSD.org:/cvsroot
setenv CVS_RSH ssh
cd /usr
cvs checkout -P pkgsrc

and wait a while.
This will give you the -current pksrc to go with your -current install.

(Continue reading)

Francis Koclas | 8 May 21:46 2006

Re: Qube2 Network Performance & Forwarding

Izumi Tsutsui wrote:
> ex(4), epic(4), and pcn(4) etc. don't have such restriction
> (i.e. they can do DMA to unaligned buffer) so we could get
> a bit better performance with them

Don't know whats up with ex but I never got a decent performance out of 
that driver on the Qube2.. the 3c905C have excessively poor performance 
on the Qubes. They seem to be able to use Tx properly, but Rx is quite a 
bit slower than 100 KB/s...

So far a cheap D-Link DFE-530TX (via-rhine) performs much better than 
the 3Coms on these... it should be the other way around :-D

Francis

Izumi Tsutsui | 8 May 22:05 2006
Picon

Re: Qube2 Network Performance & Forwarding

> Don't know whats up with ex but I never got a decent performance out of 
> that driver on the Qube2.. the 3c905C have excessively poor performance 
> on the Qubes. They seem to be able to use Tx properly, but Rx is quite a 
> bit slower than 100 KB/s...

Do you have proper MII devices in your kernel config?
I think 3C905C requires exphy(4) or bmtphy(4), and
a wrong MII device (generic ukphy(4) is used in that case)
could cause such awful performance, as I wrote.

"dmesg might help."

> So far a cheap D-Link DFE-530TX (via-rhine) performs much better than 
> the 3Coms on these... it should be the other way around :-D

BTW, vr(4) can't DMA to/from unaligned buffers both on TX and RX,
so it would have less performance even on i386.
---
Izumi Tsutsui

Francis Koclas | 8 May 22:12 2006

Re: Qube2 Network Performance & Forwarding

Izumi Tsutsui wrote:

> Do you have proper MII devices in your kernel config?
> I think 3C905C requires exphy(4) or bmtphy(4), and
> a wrong MII device (generic ukphy(4) is used in that case)
> could cause such awful performance, as I wrote.

Yes, I'm using bmtphy

3Com 3c905C-TX 10/100 Ethernet with mngmt (rev. 0x74)
Broadcom 3c905C internal PHY, rev. 7

> 
>> So far a cheap D-Link DFE-530TX (via-rhine) performs much better than 
>> the 3Coms on these... it should be the other way around :-D
> 
> BTW, vr(4) can't DMA to/from unaligned buffers both on TX and RX,
> so it would have less performance even on i386.

It does have less performance on an x86 machine (than the 3Com), but in 
the case of the Qube, its the other way around by quite a huge margin :S

Francis

sensia | 8 May 20:37 2006

Re: perl on netbsd 3.99

hi,

you can install perl with Andreas Fuchs website :

http://www.aironaut.ch/Cobalt/NetBSD3.0/
read this before : http://www.aironaut.ch/Cobalt/NetBSD3.0/README

ciao

JEFF SCOTT a écrit :
> hello,
> 
> I'm new to this group, and just build my first raq2 w/netbsd.  I'm a long time
> cobalt and linux user, so I'm not to savvy on netbsd.  The build went find,
> but slow (2-3 hours on 60gb disk).  However, I don't seem to have perl and can
> not find a perl package for 3.99.  
> 
> Any reason perl is not included?  Box is worthless to me without it.
> 
> Thanks.
> 
> 
> 

	

	
		
___________________________________________________________________________ 
Faites de Yahoo! votre page d'accueil sur le web pour retrouver directement vos services préférés :
(Continue reading)

Izumi Tsutsui | 8 May 22:25 2006
Picon

Re: Qube2 Network Performance & Forwarding

> It does have less performance on an x86 machine (than the 3Com), but in 
> the case of the Qube, its the other way around by quite a huge margin :S

Hmm. At least 3c905B worked charm on my R5000 O2,
but I don't have Qube or 3c905C so there is few thing
I can help without real hardware (or detailed info). sorry.
---
Izumi Tsutsui


Gmane