>>> On 9/11/2015 at 09:59 AM, Cole Robinson <crobinso@xxxxxxxxxx> wrote:
> On 09/11/2015 11:41 AM, Charles Arnold wrote:
>> Allow virt-install to correctly find the SLES kernel/initrd on the media
>> when installing an s390x VM.
>>
>> Signed-off-by: Charles Arnold <carnold@xxxxxxxx>
>>
>
> Thanks for the patch. Is there a public URL/media to test this against?
I don't think so or at least I'm not aware of a public URL. s390 being the big
expensive piece of hardware that it is I doubt anyone makes it public.
>
> There's some minor pep8 errors:
>
> virtinst/urlfetcher.py:940: [E201] whitespace after '['
> virtinst/urlfetcher.py:941: [E202] whitespace before ']'
> virtinst/urlfetcher.py:986: [E201] whitespace after '['
> virtinst/urlfetcher.py:987: [E202] whitespace before ']'
I'll clean these up and resubmit unless you have more comments
for my comments below :)
>
> One comment below:
>
>> diff --git a/virtinst/urlfetcher.py b/virtinst/urlfetcher.py
>> index c48e9d5..3e62c2b 100644
>> --- a/virtinst/urlfetcher.py
>> +++ b/virtinst/urlfetcher.py
>> @@ -369,6 +369,8 @@ def _distroFromSUSEContent(fetcher, arch, vmtype=None):
>> arch = "x86_64"
>> elif cbuf.find("i586") != -1:
>> arch = "i586"
>> + elif cbuf.find("s390x") != -1:
>> + arch = "s390x"
>>
>> dclass = GenericDistro
>> if distribution:
>> @@ -934,16 +936,22 @@ class SuseDistro(Distro):
>> oldkern += "64"
>> oldinit += "64"
>>
>> - # Tested with Opensuse >= 10.2, 11, and sles 10
>> - self._hvm_kernel_paths = [("boot/%s/loader/linux" % self.arch,
>> - "boot/%s/loader/initrd" % self.arch)]
>> - # Tested with Opensuse 10.0
>> - self._hvm_kernel_paths.append(("boot/loader/%s" % oldkern,
>> - "boot/loader/%s" % oldinit))
>> + if self.arch == "s390x":
>> + self._hvm_kernel_paths = [ ("boot/%s/linux" % self.arch,
>> + "boot/%s/initrd" % self.arch) ]
>> + # No Xen on s390x
>> + self._xen_kernel_paths = []
>> + else:
>> + # Tested with Opensuse >= 10.2, 11, and sles 10
>> + self._hvm_kernel_paths = [("boot/%s/loader/linux" % self.arch,
>> + "boot/%s/loader/initrd" %
> self.arch)]
>> + # Tested with Opensuse 10.0
>> + self._hvm_kernel_paths.append(("boot/loader/%s" % oldkern,
>> + "boot/loader/%s" % oldinit))
>>
>> - # Matches Opensuse > 10.2 and sles 10
>> - self._xen_kernel_paths = [("boot/%s/vmlinuz-xen" % self.arch,
>> - "boot/%s/initrd-xen" % self.arch)]
>> + # Matches Opensuse > 10.2 and sles 10
>> + self._xen_kernel_paths = [("boot/%s/vmlinuz-xen" % self.arch,
>> + "boot/%s/initrd-xen" % self.arch)]
>>
>> def _variantFromVersion(self):
>> distro_version = self.version_from_content[1].strip()
>> @@ -971,6 +979,13 @@ class SuseDistro(Distro):
>> self._variantFromVersion()
>>
>> self.os_variant = self._detect_osdict_from_url()
>> +
>> + # Reset kernel name for sle11 source on s390x
>> + if self.arch == "s390x":
>> + if self.os_variant == "sles11" or self.os_variant == "sled11":
>> + self._hvm_kernel_paths = [ ("boot/%s/vmrdr.ikr" %
> self.arch,
>> + "boot/%s/initrd" % self.arch) ]
>> +
>
> From my reading of _detect_osdict_from_url it will only return opensuse* OS
> names, not sles* or sled*, so I'm not sure how this path is triggered with
> virt-manager.git
Not exactly. It will only return an 'opensuse' variant if the url has an opensuse in
the path which it won't find for sles and so instead returns what is already in
self.os_variant. In this case, self.os_variant was previously set to some sles version
in the call to self._variantFromVersion.
Thanks,
- Charles
>
> Thanks,
> Cole
_______________________________________________
virt-tools-list mailing list
virt-tools-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/virt-tools-list
