Re: Skip LinuxInitrd.test_with_2gib_file_should_exit_error_msg_with_linux_v3_6 on 32-bit hosts?

Re: Skip LinuxInitrd.test_with_2gib_file_should_exit_error_msg_with_linux_v3_6 on 32-bit hosts?

[Daniel P. Berrangé]

Re-adding qemu-devel since it seems to have been Bcc'd.

On Fri, Oct 04, 2019 at 03:29:06PM +1000, David Gibson wrote:
> When I run make check-acceptance in a 32-bit (i686) container, this
> test fails, because it tries to start a guest with 4G of RAM, which
> can't fit in the userspace address space on a 32-bit host, obviously.
> 
>  (16/44) /home/dwg/src/qemu/tests/acceptance/linux_initrd.py:LinuxInitrd.test_with_2gib_file_should_exit_error_msg_with_linux_v3_6: FAIL: Regex didn't match: '.*initrd is too large.*max: \\d+, need 2147483648.*' not found in 'qemu-system-x86_64: -m 4096: ram size too large\n' (0.12 s)
> 
> Should we simply skip this test on 32-bit hosts?

There's a thread right now asking if we should drop support for 32-bit
hosts entirely

  https://lists.gnu.org/archive/html/qemu-devel/2019-09/msg06168.html

I'm curious why you were testing with 32-bit host ? Do you have some
use case that needs it, or was this just responding to some user bug
report

As long as we do support 32-bit hosts though, this test definitely needs
fixing. It is attempting to boot with a 2 GB initrd, so I guess RAM
needs to be big enough to fit that & boot the guest OS image. Possibly
it might still work in 2.5-3.0 GB of RAM in guest if that's sufficiently
small for 32-bit host testing. If not, then skipping seems reasonable,
as no one will ever use such huge initrds on 32-bit hosts in real world.

Regards,
Daniel
-- 
|: https://berrange.com      -o-    https://www.flickr.com/photos/dberrange :|
|: https://libvirt.org         -o-            https://fstop138.berrange.com :|
|: https://entangle-photo.org    -o-    https://www.instagram.com/dberrange :|

Re: Skip LinuxInitrd.test_with_2gib_file_should_exit_error_msg_with_linux_v3_6 on 32-bit hosts?

[David Gibson]

[-- Attachment #1: Type: text/plain, Size: 1924 bytes --]

On Fri, Oct 04, 2019 at 09:33:38AM +0100, Daniel P. Berrangé wrote:
> Re-adding qemu-devel since it seems to have been Bcc'd.
> 
> On Fri, Oct 04, 2019 at 03:29:06PM +1000, David Gibson wrote:
> > When I run make check-acceptance in a 32-bit (i686) container, this
> > test fails, because it tries to start a guest with 4G of RAM, which
> > can't fit in the userspace address space on a 32-bit host, obviously.
> > 
> >  (16/44) /home/dwg/src/qemu/tests/acceptance/linux_initrd.py:LinuxInitrd.test_with_2gib_file_should_exit_error_msg_with_linux_v3_6: FAIL: Regex didn't match: '.*initrd is too large.*max: \\d+, need 2147483648.*' not found in 'qemu-system-x86_64: -m 4096: ram size too large\n' (0.12 s)
> > 
> > Should we simply skip this test on 32-bit hosts?
> 
> There's a thread right now asking if we should drop support for 32-bit
> hosts entirely
> 
>   https://lists.gnu.org/archive/html/qemu-devel/2019-09/msg06168.html

Huh.

> I'm curious why you were testing with 32-bit host ? Do you have some
> use case that needs it, or was this just responding to some user bug
> report

A 32-bit build (actually, a couple) is part of my routine pre-pull
tests, precisely because I've accidentally included some things that
broken them.

> As long as we do support 32-bit hosts though, this test definitely needs
> fixing. It is attempting to boot with a 2 GB initrd, so I guess RAM
> needs to be big enough to fit that & boot the guest OS image. Possibly
> it might still work in 2.5-3.0 GB of RAM in guest if that's sufficiently
> small for 32-bit host testing. If not, then skipping seems reasonable,
> as no one will ever use such huge initrds on 32-bit hosts in real
> world.

Right.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]