DMRtest failures on bigendian hardware (3.14.0)

[alastair....@gmail.com]

Hi,

I'm the libdap maintainer in Debian, and have recently been testing the 3.14.0 release which is in Debian unstable now (+ Ubuntu, etc).

It works fine on Intel, amd64, etc. but DMRtest breaks on big-endian architectures: mips, powerpc, sparc, s390x, etc.

Is this a known issue?

regards

Alastair McKinstry

[James Gallagher]

On Jun 28, 2015, at 11:57 AM, alastair....@gmail.com wrote:

Hi,

I'm the libdap maintainer in Debian, and have recently been testing the 3.14.0 release which is in Debian unstable now (+ Ubuntu, etc).

Nice!

Do you have any plans to package the BES software (which forms the bulk of our Hyrax data server)? We have streamlined the software so that the build is essentially "configure; make; make check; make install” for a complete working BES given that the dependencies are available.

It works fine on Intel, amd64, etc. but DMRtest breaks on big-endian architectures: mips, powerpc, sparc, s390x, etc.

Is this a known issue?

No, it is not. Any more info you can give would be great.

regards

Alastair McKinstry

--
You received this message because you are subscribed to the Google Groups "OPeNDAP Tech" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opendap-tech...@opendap.org.
To post to this group, send email to openda...@opendap.org.
Visit this group at http://groups.google.com/a/opendap.org/group/opendap-tech/.
For more options, visit https://groups.google.com/a/opendap.org/d/optout.

--

James Gallagher

jgallagher at opendap.org

406.723.8663

[or...@cora.nwra.com]

On Monday, June 29, 2015 at 9:54:34 AM UTC-6, James Gallagher wrote:

On Jun 28, 2015, at 11:57 AM, alastair....@gmail.com wrote:

It works fine on Intel, amd64, etc. but DMRtest breaks on big-endian architectures: mips, powerpc, sparc, s390x, etc.

Is this a known issue?

No, it is not. Any more info you can give would be great.

Perhaps this helps a little:

http://ppc.koji.fedoraproject.org/kojifiles/work/tasks/3913/2633913/build.log

[or...@cora.nwra.com]

On Monday, June 29, 2015 at 9:54:34 AM UTC-6, James Gallagher wrote:

On Jun 28, 2015, at 11:57 AM, alastair....@gmail.com wrote:

It works fine on Intel, amd64, etc. but DMRtest breaks on big-endian architectures: mips, powerpc, sparc, s390x, etc.

Is this a known issue?

No, it is not. Any more info you can give would be great.

Here's one with DMRTest.log as well: http://ppc.koji.fedoraproject.org/kojifiles/work/tasks/3955/2633955/build.log

[Rafael Fonseca]

On Monday, 29 June 2015 17:54:34 UTC+2, James Gallagher wrote:

It works fine on Intel, amd64, etc. but DMRtest breaks on big-endian architectures: mips, powerpc, sparc, s390x, etc.

Is this a known issue?

No, it is not. Any more info you can give would be great.

I'm also having problems with endianness in ppc64 and s390[x] on Fedora. Attached are the logs I was able to collect.

Let me know if you need any more info and/or help to test on these architectures.

Att.

--

Rafael Fonseca

[James Gallagher]

Thanks for the info - I will look into this ASAP, but I’m on travel next week and then out again until 3 Aug. 

The problem is likely a combination of our switching to ‘reader-makes-right’ meaning the byte-order used for the response is based on the server’s native byte order and either a testing glitch or a bug in the code that computes the received data’s checksum. Thanks very much for finding and reporting this!

James

Att.

--

Rafael Fonseca

--
You received this message because you are subscribed to the Google Groups "OPeNDAP Tech" group.
To unsubscribe from this group and stop receiving emails from it, send an email to opendap-tech...@opendap.org.
To post to this group, send email to openda...@opendap.org.
Visit this group at http://groups.google.com/a/opendap.org/group/opendap-tech/.
For more options, visit https://groups.google.com/a/opendap.org/d/optout.

<DMRTest.log><test-suite.log>

[rdos...@redhat.com]

Hi,

Thanks for the info - I will look into this ASAP, but I’m on travel next week and then out again until 3 Aug. 

The problem is likely a combination of our switching to ‘reader-makes-right’ meaning the byte-order used for the response is based on the server’s native byte order and either a testing glitch or a bug in the code that computes the received data’s checksum. Thanks very much for finding and reporting this!

Any update on this? Is there anything I can do to help?

Att.

--

Rafael Fonseca 

[James Gallagher]

No progress. Yes, I could use access to a machine to test on. Since we’ve just finished a big release, now is a good time to look into this.

Thanks,

James

Att.

--

Rafael Fonseca 

--

[Rafael dos Santos]

James,

> > Any update on this? Is there anything I can do to help?
>
> No progress. Yes, I could use access to a machine to test on. Since we’ve
> just finished a big release, now is a good time to look into this.

Sure thing. Can you please fill in this form [1]?

Let me know when you're done and I'll make sure the access is granted asap.


[1] https://fit-rhlab.rhcloud.com/powerlinux-openpower-development-hosting/


Att
--
Rafael Fonseca

[d...@danny.cz]

 Hi,

is there any news about the big endian arches support?

New URL for the machine access request form is http://research.redhat.com/powerlinux-openpower-development-hosting/


Thanks, Dan

[James Gallagher]

On November 27, 2015 at 2:38:34 AM, d...@danny.cz (d...@danny.cz) wrote:

On Wednesday, September 23, 2015 at 4:47:05 PM UTC+2, Rafael dos Santos wrote:

James,

> > Any update on this? Is there anything I can do to help?
>
> No progress. Yes, I could use access to a machine to test on. Since we’ve
> just finished a big release, now is a good time to look into this.

Sure thing. Can you please fill in this form [1]?

Let me know when you're done and I'll make sure the access is granted asap.


[1] https://fit-rhlab.rhcloud.com/powerlinux-openpower-development-hosting/

 Hi,

is there any news about the big endian arches support?

Fixes have been pushed and merged to the master branch. 

We no longer need the VM - My apologies if that message never made it back to you.

The issue with big endian support had to do with the checksums - they are dependent on word order. I fixed the tests so they work for each byte order.

James

New URL for the machine access request form is http://research.redhat.com/powerlinux-openpower-development-hosting/


Thanks, Dan

-- 
James Gallagher
jgall...@opendap.org

[Dan Horák]

On Fri, 27 Nov 2015 14:32:25 -0700
James Gallagher <jgall...@opendap.org> wrote:

>
>
> On November 27, 2015 at 2:38:34 AM, d...@danny.cz (d...@danny.cz) wrote:
>
> On Wednesday, September 23, 2015 at 4:47:05 PM UTC+2, Rafael dos
> Santos wrote: James,
>
> > > Any update on this? Is there anything I can do to help?
> >
> > No progress. Yes, I could use access to a machine to test on. Since
> > we’ve just finished a big release, now is a good time to look into
> > this.
>
> Sure thing. Can you please fill in this form [1]?
>
> Let me know when you're done and I'll make sure the access is granted
> asap.
>
>
> [1] https://fit-rhlab.rhcloud.com/powerlinux-openpower-development-hosting/
>
>
>  Hi,
>
> is there any news about the big endian arches support?
> Fixes have been pushed and merged to the master branch. 
>
> We no longer need the VM - My apologies if that message never made it
> back to you.
>
> The issue with big endian support had to do with the checksums - they
> are dependent on word order. I fixed the tests so they work for each
> byte order.

ah, thanks for the work, I can confirm we could build the 3.16.0
release successfully on ppc64 platform (and the test-suite passed) [1].
Unfortunately it fails on other big-endian arches - 64-bit s390x [2]
and 32-bit s390 [3] (see the build.logs there). I'll try to get more
information.

[1] http://ppc.koji.fedoraproject.org/koji/taskinfo?taskID=2938873
[2] http://s390.koji.fedoraproject.org/koji/taskinfo?taskID=1995702
[3] http://s390.koji.fedoraproject.org/koji/taskinfo?taskID=1995699


Dan