On 19.09.16 13:21, Dirk Müller wrote:
> Hi Alex,
> 
> 2016-09-19 10:32 GMT+02:00 Alexander Graf <ag...@suse.de>:
> 
>> Yes, but it's a can of worms that we have to open anyway unfortunately.
>> Systems are growing and I doubt 42 bits are going to get us very far.
> 
> sure, I still prefer to have tests, and even tests that are passing
> over just enablign something and then not fixing it.

So can we install with btrfs and 64k yet?

>>> Are we talking about 42.2 or tumbleweed ?  I'm sort of okay with doing
>>> that for 42.2 when we document it properly and mayb ealso fix the swap
>>> issue. Because I seriously hope not that many people are using btrfs.
>> You mean if you're on an ext4 rootfs and upgrade to 42.2, swap should
>> stay working? Any constructive ideas?
> 
> Well, the problem is that boot completely fails since systemd
> considers a non-mountable swap a fatal error condition, and you can
> not easily recover from that if you don't have serial console.
> swapon --fixpgsize is to the rescue though. Alternatively we could try

So I crawled through systemd and util-linux sources a bit. Systemd does
not know anything about a "-f" option to swapon. The only thing it can
do is pass -o parameters from fstab to swapon.

Swapon on the other hand has -o mappings for a good chunk of its
parameters, but not for fixpgsize. So the parameter is not reachable
with -o.

That means we can

  a) Patch systemd to pass -f on aarch64
  b) Patch util-linux to allow passing -f in -o and modify fstab (how?)

Preferences?

> to convince it to not fail to boot just because swap isn't there.

There is a -o option called "nofail". But reading the source I don't
think it actually doesn't fail. It only doesn't fail if the target path
doesn't exist, not if it's unusable.

>>> For tumbleweed, I would only want to switch when we have sorted out
>>> the remaining issues. And only because I am going to be hit by this
>>> badly
>> What are "the remaining issues"?
> 
> 48 bit, 

What is missing apart from qt5 which has a patch available?

> yast2-bootloader seems broken again.

What exactly?

> I fixed ovmf over the weekend finally.

Cool. What did you do?


Alex
-- 
To unsubscribe, e-mail: opensuse-arm+unsubscr...@opensuse.org
To contact the owner, e-mail: opensuse-arm+ow...@opensuse.org

Reply via email to