Re: aarch64 test systems implementation
"Richard W.M. Jones" writes: > On Thu, Sep 19, 2019 at 11:27:21AM +0100, Dave Love wrote: >> (Is there a way to tell?) > > virt-what ! I guess if I'd read the man page I'd have seen that the "kvm" I saw was distinct from "qemu"; apologies for the noise. ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
On Thu, Sep 19, 2019 at 11:27:21AM +0100, Dave Love wrote: > (Is there a way to tell?) virt-what ! Rich. -- Richard Jones, Virtualization Group, Red Hat http://people.redhat.com/~rjones Read my programming and virtualization blog: http://rwmj.wordpress.com virt-p2v converts physical machines to virtual machines. Boot with a live CD or over the network (PXE) and turn machines into KVM guests. http://libguestfs.org/virt-v2v ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
On Fri, 20 Sep 2019 11:10:41 +0100 Dave Love wrote: > Stephen John Smoogen writes: > > > It depends on what level of emulation you mean. Are you meaning "is > > it an x86_64 acting like a aarch64?" then the answer is no. > > Yes, that's what I meant, thanks. > > > If you are > > meaning "is it an aarch64 guest system on an aarch64" then probably > > yes... we don't have enough hardware to not use virtualization which > > does have some emulation at some level. > > Sure, they're documented as VMs (and virt-what shows so). I may > misunderstand the virtualization, but I thought that a straight KVM VM > runs the instructions directly on the ThunderX hardware it advertises, > which would pin the failure either on target specifics in the library > or gcc. I'm not sure I have the enthusiasm to debug it anyway... generally user space runs without any emulation in KVM and only privileged instructions (used in kernel) are emulated or have some special treatment Dan ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
Kevin Fenzi writes: > On 9/19/19 3:27 AM, Dave Love wrote: >> Are the aarch64 test systems running on real or emulated hardware? (Is >> there a way to tell?) > > You mean the ones listed at > https://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers > ? Yes, apologies for not being explicit. Thanks. ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
Stephen John Smoogen writes: > It depends on what level of emulation you mean. Are you meaning "is it > an x86_64 acting like a aarch64?" then the answer is no. Yes, that's what I meant, thanks. > If you are > meaning "is it an aarch64 guest system on an aarch64" then probably > yes... we don't have enough hardware to not use virtualization which > does have some emulation at some level. Sure, they're documented as VMs (and virt-what shows so). I may misunderstand the virtualization, but I thought that a straight KVM VM runs the instructions directly on the ThunderX hardware it advertises, which would pin the failure either on target specifics in the library or gcc. I'm not sure I have the enthusiasm to debug it anyway... ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
On 9/19/19 3:27 AM, Dave Love wrote: > Are the aarch64 test systems running on real or emulated hardware? (Is > there a way to tell?) You mean the ones listed at https://fedoraproject.org/wiki/Test_Machine_Resources_For_Package_Maintainers ? If so, they are vm's running on real hardware with kvm. kevin -- > > I ask because I'm seeing bad numerical results from a test and would > like to know if I can eliminate qemu as a possible cause -- not that I > think that's likely. > ___ > devel mailing list -- devel@lists.fedoraproject.org > To unsubscribe send an email to devel-le...@lists.fedoraproject.org > Fedora Code of Conduct: > https://docs.fedoraproject.org/en-US/project/code-of-conduct/ > List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines > List Archives: > https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org > signature.asc Description: OpenPGP digital signature ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
On Thu, 19 Sep 2019 at 06:28, Dave Love wrote: > > Are the aarch64 test systems running on real or emulated hardware? (Is > there a way to tell?) > > I ask because I'm seeing bad numerical results from a test and would > like to know if I can eliminate qemu as a possible cause -- not that I > think that's likely. It depends on what level of emulation you mean. Are you meaning "is it an x86_64 acting like a aarch64?" then the answer is no. If you are meaning "is it an aarch64 guest system on an aarch64" then probably yes... we don't have enough hardware to not use virtualization which does have some emulation at some level. -- Stephen J Smoogen. ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
> > Do you mean a system to test SRPM on aarch64? or a system to test a > > open source code? > > Do you mean CI testing system or an aarch64 server for adhoc test? > > I don't know why it matters, but I'm testing dynamic micro-architecture > selection added to a library I've packaged. Because there are some proper choices depending on the situation. * Test by mock command (aarch64 qemu environment). (SRPM, adhoc) https://github.com/rpm-software-management/mock/wiki/Feature-forcearch * Test by Koji (aarch64 native) (SRPM, a kind of CI) * Test by Copr (aarch64 native or qemu. I am not sure) * Test by Packit service (aarch64 native, maybe) (SRPM, CI) * Test qemu-user-static RPM and aarch64 container * on local (Source, adhoc) * with CI services (Source, CI) * Test with Travis CI with qemu (aarch64 qemu environment) (Source, CI) * Test with native aarch64 supporting CI (Source CI) * only CI: Shippable CI, Drone CI, * Works on ARM (free aarch64 server for open source) (Source, adhoc or CI) For 2nd half, below documents might be helpful for you. * https://github.com/junaruga/fedora-workshop-multiarch * https://github.com/junaruga/ci-multi-arch-test Jun ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
Jun Aruga writes: > On Thu, Sep 19, 2019 at 12:28 PM Dave Love > wrote: >> >> Are the aarch64 test systems running on real or emulated hardware? (Is >> there a way to tell?) >> >> I ask because I'm seeing bad numerical results from a test and would >> like to know if I can eliminate qemu as a possible cause -- not that I >> think that's likely. > > Do you mean a system to test SRPM on aarch64? or a system to test a > open source code? > Do you mean CI testing system or an aarch64 server for adhoc test? I don't know why it matters, but I'm testing dynamic micro-architecture selection added to a library I've packaged. ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
Re: aarch64 test systems implementation
On Thu, Sep 19, 2019 at 12:28 PM Dave Love wrote: > > Are the aarch64 test systems running on real or emulated hardware? (Is > there a way to tell?) > > I ask because I'm seeing bad numerical results from a test and would > like to know if I can eliminate qemu as a possible cause -- not that I > think that's likely. Do you mean a system to test SRPM on aarch64? or a system to test a open source code? Do you mean CI testing system or an aarch64 server for adhoc test? ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
aarch64 test systems implementation
Are the aarch64 test systems running on real or emulated hardware? (Is there a way to tell?) I ask because I'm seeing bad numerical results from a test and would like to know if I can eliminate qemu as a possible cause -- not that I think that's likely. ___ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org