no 5.7 arm packages

classic Classic list List threaded Threaded
49 messages Options
123
Reply | Threaded
Open this post in threaded view
|

no 5.7 arm packages

akita1000
Hello

5.7/packages/arm is not available even on ftp.openbsd.org.
Is that lost beyond recall?

Jake

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Stuart Henderson-6
On 2015/05/21 20:07, [hidden email] wrote:
> Hello
>
> 5.7/packages/arm is not available even on ftp.openbsd.org.
> Is that lost beyond recall?
>
> Jake
>

We haven't been building arm packages since 5.6.

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Philip Cheney
That's unfortunate, since arm is one of the few architectures where building ports is often infeasible due to limitations of the installed machine.

Why aren't arm packages built anymore? Is a hardware donation needed for the build farm?


> On 21 May 15, at 13:05, Stuart Henderson <[hidden email]> wrote:
>
> On 2015/05/21 20:07, [hidden email] wrote:
>> Hello
>>
>> 5.7/packages/arm is not available even on ftp.openbsd.org.
>> Is that lost beyond recall?
>>
>> Jake
>>
>
> We haven't been building arm packages since 5.6.
>
>


Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Stuart Henderson-6
> > On 2015/05/21 20:07, [hidden email] wrote:
> >> 5.7/packages/arm is not available even on ftp.openbsd.org.
> >> Is that lost beyond recall?
> >
> > On 21 May 15, at 13:05, Stuart Henderson <[hidden email]> wrote:
> > We haven't been building arm packages since 5.6.
>
On 2015/05/21 18:54, Philip Cheney wrote:
> That's unfortunate, since arm is one of the few architectures where building ports is often infeasible due to limitations of the installed machine.

That applies even more so to building thousands of packages in a bulk
build ;)

> Why aren't arm packages built anymore? Is a hardware donation needed for the build farm?

I used to build them but the N2100 died (and in any event it was quite
labour intensive babysitting the build, and took about 3 months to
complete). Other people stepped in for the last couple of releases,
though not for the last one and knowing how long it takes I didn't feel
like asking for volunteers. It might make more sense when there are
supported machines that are more suited to running bulk builds.

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Philip Cheney
> On 22 May 15, at 01:34, Stuart Henderson <[hidden email]> wrote:
>
>>> On 2015/05/21 20:07, [hidden email] wrote:
>>>> 5.7/packages/arm is not available even on ftp.openbsd.org.
>>>> Is that lost beyond recall?
>>>
>>> On 21 May 15, at 13:05, Stuart Henderson <[hidden email]> wrote:
>>> We haven't been building arm packages since 5.6.
>>
> On 2015/05/21 18:54, Philip Cheney wrote:
>> That's unfortunate, since arm is one of the few architectures where building ports is often infeasible due to limitations of the installed machine.
>
> That applies even more so to building thousands of packages in a bulk
> build ;)

I completely understand!

>> Why aren't arm packages built anymore? Is a hardware donation needed for the build farm?
>
> I used to build them but the N2100 died (and in any event it was quite
> labour intensive babysitting the build, and took about 3 months to
> complete). Other people stepped in for the last couple of releases,
> though not for the last one and knowing how long it takes I didn't feel
> like asking for volunteers. It might make more sense when there are
> supported machines that are more suited to running bulk builds.

I've been building packages as I need them on a Beaglebone Black and I can go ahead and start a bulk build.

While it's building, I'll pick up something a little more suitable for the task, too. It looks like the Wandboard is fairly well supported and the Quad model has a bit more punch for the cost. Does anyone know of any caveats to using that system?

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Henrik Lund Kramshøj-2

> On 22 May 2015, at 16:40, Philip Cheney <[hidden email]> wrote:
>
> ...
> I've been building packages as I need them on a Beaglebone Black and I can go ahead and start a bulk build.
>
> While it's building, I'll pick up something a little more suitable for the task, too. It looks like the Wandboard is fairly well supported and the Quad model has a bit more punch for the cost. Does anyone know of any caveats to using that system?
>

I am a happy BBB user and will be happy add some EUR/$ for donation to someone willing to build packages.

I will prefer to have something permanent setup, but if we start out getting some hardware and estimating the "cost" time, hardware etc. for doing this it would be great.


Mvh/Best regards

Henrik
--
Henrik Lund Kramshøj, Follower of the Great Way of Unix
internet samurai cand.scient CISSP
[hidden email] [hidden email] +45 2026 6000
http://solidonetworks.com/ Network Security is a business enabler


signature.asc (858 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Warner Losh

> On May 22, 2015, at 8:50 AM, Henrik Lund Kramshøj <[hidden email]> wrote:
>
>
>> On 22 May 2015, at 16:40, Philip Cheney <[hidden email]> wrote:
>>
>> ...
>> I've been building packages as I need them on a Beaglebone Black and I can go ahead and start a bulk build.
>>
>> While it's building, I'll pick up something a little more suitable for the task, too. It looks like the Wandboard is fairly well supported and the Quad model has a bit more punch for the cost. Does anyone know of any caveats to using that system?
>>
>
> I am a happy BBB user and will be happy add some EUR/$ for donation to someone willing to build packages.
>
> I will prefer to have something permanent setup, but if we start out getting some hardware and estimating the "cost" time, hardware etc. for doing this it would be great.
FreeBSD has had good luck with using userland qemu emulation to build packages. More stable than random hardware that was never intended to have such a load on it, and quite a bit faster than these little boards.

Warner


signature.asc (859 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Edwin Amsler
That's my thinking too. I know building is supposed to put a platform through its paces, but virtualizing the more exotic platforms would ease a lot of package building.

--
Edwin (on the move)

> On May 22, 2015, at 12:26 PM, Warner Losh <[hidden email]> wrote:
>
>
>> On May 22, 2015, at 8:50 AM, Henrik Lund Kramshøj <[hidden email]> wrote:
>>
>>
>>> On 22 May 2015, at 16:40, Philip Cheney <[hidden email]> wrote:
>>>
>>> ...
>>> I've been building packages as I need them on a Beaglebone Black and I can go ahead and start a bulk build.
>>>
>>> While it's building, I'll pick up something a little more suitable for the task, too. It looks like the Wandboard is fairly well supported and the Quad model has a bit more punch for the cost. Does anyone know of any caveats to using that system?
>>
>> I am a happy BBB user and will be happy add some EUR/$ for donation to someone willing to build packages.
>>
>> I will prefer to have something permanent setup, but if we start out getting some hardware and estimating the "cost" time, hardware etc. for doing this it would be great.
>
> FreeBSD has had good luck with using userland qemu emulation to build packages. More stable than random hardware that was never intended to have such a load on it, and quite a bit faster than these little boards.
>
> Warner
>

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Patrick Wildt-2
qemu system emulation (= full emulated system) on my x86 is a third slower than on my usual build machine.

Not sure how fast userland emulation is.

> Am 23.05.2015 um 02:24 schrieb Edwin Amsler <[hidden email]>:
>
> That's my thinking too. I know building is supposed to put a platform through its paces, but virtualizing the more exotic platforms would ease a lot of package building.
>
> --
> Edwin (on the move)
>
>> On May 22, 2015, at 12:26 PM, Warner Losh <[hidden email]> wrote:
>>
>>
>>> On May 22, 2015, at 8:50 AM, Henrik Lund Kramshøj <[hidden email]> wrote:
>>>
>>>
>>>> On 22 May 2015, at 16:40, Philip Cheney <[hidden email]> wrote:
>>>>
>>>> ...
>>>> I've been building packages as I need them on a Beaglebone Black and I can go ahead and start a bulk build.
>>>>
>>>> While it's building, I'll pick up something a little more suitable for the task, too. It looks like the Wandboard is fairly well supported and the Quad model has a bit more punch for the cost. Does anyone know of any caveats to using that system?
>>>
>>> I am a happy BBB user and will be happy add some EUR/$ for donation to someone willing to build packages.
>>>
>>> I will prefer to have something permanent setup, but if we start out getting some hardware and estimating the "cost" time, hardware etc. for doing this it would be great.
>>
>> FreeBSD has had good luck with using userland qemu emulation to build packages. More stable than random hardware that was never intended to have such a load on it, and quite a bit faster than these little boards.
>>
>> Warner
>>
>


Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Nick Holland
In reply to this post by Warner Losh
On 05/22/15 13:26, Warner Losh wrote:
...
> FreeBSD has had good luck with using userland qemu emulation to build
> packages. More stable than random hardware that was never intended to
> have such a load on it, and quite a bit faster than these little
> boards.

Think about that a bit.  Emulated hw is more stable than the real thing?
 That means one of two things:
1) The real HW sucks, and thus, should not be used.
2) They code to the emulator, not real HW, and thus the SW sucks, and
should not be used.

There's a reason the OpenBSD project does only native builds.  You
pretty well nailed it right there.

The ARM platform needs work, not more hardware, this is why it was
demoted from a "supported" platform to a "current porting effort" around
5.6.  The people working on the armv7 platform are brilliant people, but
like most brilliant people, there is no shortage of things for them to
work on, both for pay and for fun.

Last I was building packages for it (5.6 -- and was told to stop), the
platform would just crash, trying to keep a dpb cluster running was
basically reverse wack-a-mole (trying to quickly restart systems that
crashed).  That's not supposed to happen.  This isn't a case of "never
intended to have such load on it", it just needs some work.

Nick.

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Edward Fochler
I agree that improved stability would not justify emulation, because of the code targeting problem you cited.  But would speed help?  

Warner suggested that emulated hardware could be faster than the native hardware being used.  If emulated hardware were twice as fast, would that justify using it in addition to native builds?

        ED.


> On 2015, May 22, at 11:10 PM, Nick Holland <[hidden email]> wrote:
>
> On 05/22/15 13:26, Warner Losh wrote:
> ...
>> FreeBSD has had good luck with using userland qemu emulation to build
>> packages. More stable than random hardware that was never intended to
>> have such a load on it, and quite a bit faster than these little
>> boards.
>
> Think about that a bit.  Emulated hw is more stable than the real thing?
> That means one of two things:
> 1) The real HW sucks, and thus, should not be used.
> 2) They code to the emulator, not real HW, and thus the SW sucks, and
> should not be used.
>
> There's a reason the OpenBSD project does only native builds.  You
> pretty well nailed it right there.
>
> The ARM platform needs work, not more hardware, this is why it was
> demoted from a "supported" platform to a "current porting effort" around
> 5.6.  The people working on the armv7 platform are brilliant people, but
> like most brilliant people, there is no shortage of things for them to
> work on, both for pay and for fun.
>
> Last I was building packages for it (5.6 -- and was told to stop), the
> platform would just crash, trying to keep a dpb cluster running was
> basically reverse wack-a-mole (trying to quickly restart systems that
> crashed).  That's not supposed to happen.  This isn't a case of "never
> intended to have such load on it", it just needs some work.
>
> Nick.
>


Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Stuart Henderson-6
On 2015/05/22 23:54, Edward Fochler wrote:
> I agree that improved stability would not justify emulation, because
> of the code targeting problem you cited. But would speed help?

Not particularly. If the machines and/or OS arch are reliable in the
first place it's not a problem to spread builds across more machines.
But in reality, it's not very reliable, so 1) adding machines just
means babysitting additional crashes, and 2) if emulation is used
instead of real machines, then you're dealing with both OS bugs and
emulator bugs.

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Stuart Henderson-6
In reply to this post by Philip Cheney
On 2015/05/22 08:40, Philip Cheney wrote:
> While it's building, I'll pick up something a little more suitable for
> the task, too. It looks like the Wandboard is fairly well supported and
> the Quad model has a bit more punch for the cost. Does anyone know of
> any caveats to using that system?

Of the "recent" bulk builds, the one on the pandaboard worked best.
I have no idea about the newer machines though, last time I asked about
what was best, IIRC there were problems mentioned with everything...

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Hrishikesh Murukkathampoondi

> Of the "recent" bulk builds, the one on the pandaboard worked best.
> I have no idea about the newer machines though, last time I asked about
> what was best, IIRC there were problems mentioned with everything…
>

Do you think the the problem is hardware stability of ARM based boards or if OpenBSD support for ARM has not matured yet?

Thanks
Hrishi



Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Stuart Henderson-6
On 2015/05/23 15:43, Hrishikesh Murukkathampoondi wrote:

>
> > Of the "recent" bulk builds, the one on the pandaboard worked best.
> > I have no idea about the newer machines though, last time I asked about
> > what was best, IIRC there were problems mentioned with everything…
> >
>
> Do you think the the problem is hardware stability of ARM based boards or if OpenBSD support for ARM has not matured yet?
>
> Thanks
> Hrishi
>
>

The latter is a problem.

There have been some ARM boards with crappy hardware stability for sure,
but given how they are used in various situations (particularly thinking
of things like rpi, beaglebone, etc) and the general lack of complaints
about hw problems, they can't all be bad...


Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Joel Rees-2
In reply to this post by Edward Fochler
2015/05/23 15:56 "Edward Fochler" <[hidden email]>:
>
> I agree that improved stability would not justify emulation, because of
the code targeting problem you cited.  But would speed help?
>

Crash at higher speed?

I think the subtext is that they need more people, people who don't have
important other things going on that nursing the arm platforms would drag
them away from.

Although I am wondering whether building on a device where the disks are
attached via USB is ever going to be stable enough. I see arm devices with
SATA interfaces, but the only SAS/ARM devices I see are arm controllers in
SAS inteligent interfaces. The market definitely shows a bit of being under
Intel's illusion of speed.

> Warner suggested that emulated hardware could be faster than the native
hardware being used.  If emulated hardware were twice as fast, would that
justify using it in addition to native builds?
>

If only such theoretical speed advantage were real, it might be worth
talking about. Maybe. On the other hand, the openbsd project never used
emulation on the 68k platforms, where the clockspeed difference was very
real.

I myself have a couple of dead computers and a dead printer to fix or
replace before I can buy arm hardware and volunteer on the arm platforms.
We need people who can get past that roadblock called "if". (And I do feel
the irony when I say so.)

-- Joel Rees

>         ED.
>
>
> > On 2015, May 22, at 11:10 PM, Nick Holland <[hidden email]>
wro

> >
> > On 05/22/15 13:26, Warner Losh wrote:
> > ...
> >> FreeBSD has had good luck with using userland qemu emulation to build
> >> packages. More stable than random hardware that was never intended to
> >> have such a load on it, and quite a bit faster than these little
> >> boards.
> >
> > Think about that a bit.  Emulated hw is more stable than the real thing?
> > That means one of two things:
> > 1) The real HW sucks, and thus, should not be used.
> > 2) They code to the emulator, not real HW, and thus the SW sucks, and
> > should not be used.
> >
> > There's a reason the OpenBSD project does only native builds.  You
> > pretty well nailed it right there.
> >
> > The ARM platform needs work, not more hardware, this is why it was
> > demoted from a "supported" platform to a "current porting effort" around
> > 5.6.  The people working on the armv7 platform are brilliant people, but
> > like most brilliant people, there is no shortage of things for them to
> > work on, both for pay and for fun.
> >
> > Last I was building packages for it (5.6 -- and was told to stop), the
> > platform would just crash, trying to keep a dpb cluster running was
> > basically reverse wack-a-mole (trying to quickly restart systems that
> > crashed).  That's not supposed to happen.  This isn't a case of "never
> > intended to have such load on it", it just needs some work.
> >
> > Nick.
> >
>
>
Reply | Threaded
Open this post in threaded view
|

Odroid-XU OpenBSD cluster / Re: no 5.7 arm packages

ms-2
I have some ARM based boards from Hardkernel. In case OpenBSD 5.7
supports them I would love to use a bunch of them to create OpenBSD 5.7
packages. Hopefully it will be possible to build a cluster (OpenBSD) and
use it for such tasks like creating packages.

Odroid-XU
http://odroid.com/dokuwiki/doku.php?id=en:odroid-xu
Processor Samsung Exynos5 Octa ARM Cortex™-A15 Quad 1.6Ghz and
Cortex™-A7 Quad 1.2GHz CPUs
Memory 2Gbyte LPDDR3 RAM PoP
<http://odroid.com/dokuwiki/doku.php?id=en:pop> (800Mhz, 1600Mbps/pin, 2
x 32bit Bus)


Odroid-U3
http://www.hardkernel.com/main/products/prdt_info.php?g_code=G138745696275
http://odroid.com/dokuwiki/doku.php?id=en:odroid-u3

Processor *CPU* 1.7GHz *Exynos4412 Prime*
<http://www.samsung.com/global/business/semiconductor/minisite/Exynos/index.html> Cortex-A9
Quad-core processor
Memory PoP (Package on Package) 2Gbyte LPDDR2 880Mega Data Rate



Till know I run Ubuntu on them and tried Android. Due to limited free
time a year ago I gave up my inverstigations to get OpenBSD running on them.

Pleae let me know in case I could help in this matter.

Kind regards, Mark

--
[hidden email]

http://rsync.it-infrastrukturen.org

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Douglas Beattie
In reply to this post by Stuart Henderson-6
I think there is a difference between building OpenBSD "on real hardware"
and building port binaries for ARMv7 which are intended to run across all
of the ARMv7 target ports.

And I think the concern should be less about “OS bugs and emulator bugs”
since (1.) you’re not building the OS, and (2.) ARM emulation is quite mature
with several standard platform configurations which may be targeted (including
ARMv7).

A dual-Xeon-quad-core server (8 cores, 16 threads) with 48G of RAM can
easily host 10 or more VMs, and do a distributed build across QEMU/ARM
running on each of the VMs. The resulting compiled code for a port ‘binary’
distribution should be no different regardless of where it was actually built.

So, what then is the reason (other than philosophical) that an OpenBSD port
does not yet exist for QEMU/ARM ?  If you can no longer blame the hardware
to such a degree, then either your port builds should become more reliable, or
there is still a problem with the OS arch.

The behavior of emulated hardware is repeatable and fairly predictable, especially
if it’s a mature emulation platform (i.e. having been pounded on and considered
stable already for multiple OS). Taking the hardware (excuse) out of the equation,
you could move forward with efforts to continuously build and validate ARMv7 ports
with every iteration of OpenBSD.

And yes, the native building of the OS itself is still a great stress test for these up-and-
coming ARMv7 hardware platforms. But the building of port packages -- or the choice
of making excuses to not build or provide them -- is really a very separate issue.


On 2015-05-23, Stuart Henderson wrote:
> Not particularly. If the machines and/or OS arch are reliable in the
> first place it's not a problem to spread builds across more machines.
> But in reality, it's not very reliable, so 1) adding machines just
> means babysitting additional crashes, and 2) if emulation is used
> instead of real machines, then you're dealing with both OS bugs and
> emulator bugs.

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Scarlett-2
On 23/05/2015 22:19, Douglas Beattie wrote:

> I think there is a difference between building OpenBSD "on real hardware"
> and building port binaries for ARMv7 which are intended to run across all
> of the ARMv7 target ports.
>
> And I think the concern should be less about “OS bugs and emulator bugs”
> since (1.) you’re not building the OS, and (2.) ARM emulation is quite mature
> with several standard platform configurations which may be targeted (including
> ARMv7).
>
> A dual-Xeon-quad-core server (8 cores, 16 threads) with 48G of RAM can
> easily host 10 or more VMs, and do a distributed build across QEMU/ARM
> running on each of the VMs. The resulting compiled code for a port ‘binary’
> distribution should be no different regardless of where it was actually built.
>
> So, what then is the reason (other than philosophical) that an OpenBSD port
> does not yet exist for QEMU/ARM ?  If you can no longer blame the hardware
> to such a degree, then either your port builds should become more reliable, or
> there is still a problem with the OS arch.
>
> The behavior of emulated hardware is repeatable and fairly predictable, especially
> if it’s a mature emulation platform (i.e. having been pounded on and considered
> stable already for multiple OS). Taking the hardware (excuse) out of the equation,
> you could move forward with efforts to continuously build and validate ARMv7 ports
> with every iteration of OpenBSD.
>
> And yes, the native building of the OS itself is still a great stress test for these up-and-
> coming ARMv7 hardware platforms. But the building of port packages -- or the choice
> of making excuses to not build or provide them -- is really a very separate issue.
>
>
> On 2015-05-23, Stuart Henderson wrote:
>> Not particularly. If the machines and/or OS arch are reliable in the
>> first place it's not a problem to spread builds across more machines.
>> But in reality, it's not very reliable, so 1) adding machines just
>> means babysitting additional crashes, and 2) if emulation is used
>> instead of real machines, then you're dealing with both OS bugs and
>> emulator bugs.
>

Why are OS bugs not a concern when doing ports builds? The OS must be
suitably stable to complete the build, unless you're advocating for
cross-compilation. I could point to NetBSD's history with toolchain bugs
and cross-built executables not actually running on the target platform
as being valid reasons for avoiding it. It's probably unfair to single
out NetBSD here, plenty of other projects have had these headaches.

I think the reason for there being no arm-on-qemu support is the usual
supect: it's not a priority for any developer.

Maybe it's easy to accuse others of making excuses when you aren't
responsible for making sure the builds work and aren't hacking on the port.

(whoops, I forgot to CC the list).

Reply | Threaded
Open this post in threaded view
|

Re: no 5.7 arm packages

Edwin Amsler
In reply to this post by Joel Rees-2
> If only such theoretical speed advantage were real

Last I checked, Qemu could show an arbitrary number of cores. I think OpenBSD current has issues on ARM with multi-core but if it didn't, I could run a non-BigLittle 8 core VM with huge amounts of RAM and SSDs. I think I could build a pretty cheap build machine to cover AMD64, PPC and ARM at the same time.

Plus there's the administrative and other advantages with VMs I've grown to love in our rack at the office.

123