Re: firefox instability may be fixed now in HEAD

2010-10-04 Thread Matthew Dillon

:I'm still using 2010Q1. Should I switch to the 2.8 build of packages once I 
:recompile the kernel and world?
:
:Pierre

We won't have 2.8 packages until about a week after netbsd rolls Q3.
In fact, right now we are waiting for netbsd to roll Q3 so we can roll
the 2.8 release branch and start building packages.

I think the 2.6 and 2.7 packages are Q2.  The firefox in Q2 is
firefox-3.6.9.  I've had good stability with the recent patches except
for one annoying firefox lockup which occurs w/ slashdot when entering
text into a text field.  It's possible that some of the dependencies
aren't compatible but I won't know until I can reload my /usr/pkg with
a fresh set of packages from scratch.

-Matt
Matthew Dillon 



Re: What does "HAMMER UPGRADE" flag means during hammer cleanup

2010-10-04 Thread Matthew Dillon

:Hi,
:
:While running hammer cleanup a new pfs i created and mounted showed
:
:cleanup /vms/vm-archlinux- HAMMER UPGRADE: Creating snapshots
:Creating snapshots in /var/hammer/vms/vm-archlinux
: handle PFS #8 using /var/hammer/vms/vm-archlinux
:   snapshots - run
:   prune - run
:   rebalance - run..
: reblock - run
:  recopy - run
:
:where as the old pfses does not show the  "HAMMER UPGRADE" flag

It says that when it needs to create snapshot meta-data for
hammer v4.  I'm not sure why it decided it needed to do that if
the filesystem was already v4 (and assuming the hammer utility
wasn't an old copy either).  It is a one-time process when it
does happen.

-Matt
Matthew Dillon 



Re: firefox instability may be fixed now in HEAD

2010-10-04 Thread Pierre Abbat
On Monday 04 October 2010 18:32:14 Matthew Dillon wrote:
> A double free can occur in the nmalloc code when a pthreads thread
> is torn down (which firefox does quite a bit).  There is also a
> library call issue with libgcc that can cause firefox to exit on
> a dynamic loader symbol not found when it triesto run certain
> operations.
>
> Both libc and libthread_xu have to be rebuilt, but it is best to
> just rebuild the whole world (people trying to upgrade from bases
> which are too old will get caught up with syscall changes and that
> requires a full buildworld/installworld).

I'm still using 2010Q1. Should I switch to the 2.8 build of packages once I 
recompile the kernel and world?

Pierre
-- 
Don't buy a French car in Holland. It may be a citroen.


Talks about DFBSD in BSDday-AR

2010-10-04 Thread Sdävtaker
Hi,
Alex Hornung confirmed 2 talks about DFBSD in next BSDday Argentina.
Also I ll give an introductory to HammerFS
BSDday-AR is coming November 5th and 6th in Buenos Aires
Damian

-- 
http://dfbsd.trackbsd.org.ar


Re: Is the DRM stuff done?

2010-10-04 Thread elekktretterr
> I don't think it is committable at this stage.  After the release
> I will have to spend a few days and see how much can be salvaged.
> A good chunk of the work was syncing up code from FreeBSD and we
> should be able to get that in but the actual linux GEM API and
> mmap()ing...  that doesn't look committable to me.


Hmm that's sad. I was under the impression that all the GSoC projects went
really well.

What seems to be the problem with the GEM API?

Petr



Re: firefox instability may be fixed now in HEAD

2010-10-04 Thread Matthew Dillon

:I've had Firefox crash on me. What were the issues?
:
:Pierre

A double free can occur in the nmalloc code when a pthreads thread
is torn down (which firefox does quite a bit).  There is also a
library call issue with libgcc that can cause firefox to exit on
a dynamic loader symbol not found when it triesto run certain
operations.

Both libc and libthread_xu have to be rebuilt, but it is best to
just rebuild the whole world (people trying to upgrade from bases
which are too old will get caught up with syscall changes and that
requires a full buildworld/installworld).

-Matt
Matthew Dillon 



Re: firefox instability may be fixed now in HEAD

2010-10-04 Thread Pierre Abbat
On Monday 04 October 2010 19:12:45 Matthew Dillon wrote:
> Firefox became a bit unstable due to two issues which are now fixed
> in HEAD (in the world build, not the kernel).  People running HEAD who
> are experiencing unexpected firefox window closures can re-test w/
> the latest and see if they go away.

I've had Firefox crash on me. What were the issues?

Pierre
-- 
I believe in Yellow when I'm in Sweden and in Black when I'm in Wales.


firefox instability may be fixed now in HEAD

2010-10-04 Thread Matthew Dillon
Firefox became a bit unstable due to two issues which are now fixed
in HEAD (in the world build, not the kernel).  People running HEAD who
are experiencing unexpected firefox window closures can re-test w/
the latest and see if they go away.

-Matt
Matthew Dillon 



What does "HAMMER UPGRADE" flag means during hammer cleanup

2010-10-04 Thread Siju George
Hi,

While running hammer cleanup a new pfs i created and mounted showed

cleanup /vms/vm-archlinux- HAMMER UPGRADE: Creating snapshots
Creating snapshots in /var/hammer/vms/vm-archlinux
 handle PFS #8 using /var/hammer/vms/vm-archlinux
   snapshots - run
   prune - run
   rebalance - run..
 reblock - run
  recopy - run

where as the old pfses does not show the  "HAMMER UPGRADE" flag

cleanup /usr - handle PFS #3 using /var/hammer/usr
   snapshots - run
   prune - run
   rebalance - run..
 reblock - run
  recopy - run

There is no need to upgrade the pfs because it is actuall on hammer version 4.
So is there any thing I need to know about this falg? :-)

thanks

--Siju