Re: 6.1.4 kernel and Qube2

RĂ©mi Zara <[email protected]> Sat, 1 Nov 2014 11:18:39 +0100
Newsgroups gmane.os.netbsd.ports.cobalt
Message-ID <[email protected]>
> Le 23 oct. 2014 =C3=A0 02:17, Brian M <[email protected]> a =
=C3=A9crit :
>=20
> (did this get out?  I  don't see that gmane updated with the new
> message I sent yesterday).
>=20
> --------
> Hi all;
>=20
> I built the 6.1.4 kernel with SCSI removed, ddb enabled, placed it
> over my 5.2.1 base, and it ran OK for a few days.  I packed/ unpacked
> a few big tar files, no issues.  No need yet for those debug flags I
> enabled!
>=20
> It looks like July 2014 some PR's were resolved related to i/o that
> addressed the stability issues.
>=20
> I'm partway through building 6.1.5 with the restoreCD script, but I
> ran into some issues on my part as the lcd patch directories have been
> moved.  Nothing serious but I didn't get it right the first couple
> times.  I'm traveling for work now and won't be back at this for a
> bit.
>=20
> Anyway- this looks like it could be ready to roll up into a .iso and =
work fine.
>=20
> Brian.


Hi,

thanks for the heads up !
I tried stock 6.1.5 on my qube2, but it was not more stable (could not =
untar pkgsrc)

I then tried current (NetBSD 7.99.1 (GENERIC.201410221550Z)) and this =
one is a winner (up 8 days, /etc/daily not panicing the machine as =
before) !

Except for one thing : I cannot get perl 5.20.1 (the current version in =
pkgsrc) compiled, which is unfortunate since many other package have a =
dependency on perl5 >=3D 5.20

It fails, not always at the same stage, with a perl error of the form =
"panic: memory wrap at lib/Pod/Simple/BlackBox.pm line 213"

Changing dataseg size limit seems to move the failure point.

Has anybody succeeded in compiling this version of perl on a Qube2 ?

Regards,

R=C3=A9mi=