![]() |
MINIX NEO Z83-4 running Fedora 27 with Rawhide kernel (showing working WiFi, Bluetooth and audio including headphone jack) |
Three recent events have made me question whether to keep using Ubuntu as my preferred Linux distribution:
- Six weeks have passed since v4.15-rc1 was released and the Ubuntu config for the 'unstable' kernel still hasn't been updated to reflect the patches around the Serial Device Bus managing serial devices declared as attached to an UART in ACPI table. This prevents Bluetooth working on some mini PCs.
- The Ubuntu 17.10 ISO is still not available following the scramble to address the corrupted BIOS due to "Intel SPI bug in kernel" arguably actually caused by incorrectly enabling the Intel SPI drivers in the config.
- Canonical's slow response coupled with what appears to be a justification to adhere to previously agreed timescales resulting in their inability to immediately address the 'Meltdown' exploit. This has created a 'Window of Vulnerability' for existing Ubuntu users unaware they could temporarily upgrade to a the most recent Canonical mainline build.
The core of the problem appears to lie in the architectural approach of how kernels are adopted and released as part of the overall Ubuntu release structure. Whilst implementing Ubuntu LTS enablement (known as HWE or Hardware Enablement) stacks which provide newer kernel support for existing Ubuntu LTS releases, there is no equivalent for regular releases which are only guaranteed to receive security updates for the duration of their release life (normally 9 months for desktop and server releases). As a result the Ubuntu kernel support per se is not aligned with the mainline kernel development strategy of supporting the current release together with nominated LTS kernels.
Aside from the above points there is also the somewhat emotive issue of dropping Unity in preference for the GNOME desktop. Through dressing up GNOME to look similar to Unity by including the 'dock-to-dash' extension it inadvertently shows that changing the back-end distribution might not be so visible if GNOME is to be used especially for users reliant on a GUI.
Furthermore as I run Ubuntu 17.04 on a number of devices I have to make a 'release' decision now that Ubuntu 17.04 reaches 'end of life' on 13th January. Canonical will not be providing updated kernel packages for Ubuntu 17.04 as they will not be patching the 4.10 HWE kernel to address the Meltdown and Spectre vulnerabilities resulting in a '4.10 HWE EARLY END OF LIFE'. The consequence is that the rolling HWE kernel for Ubuntu 16.04 will go to 4.13 early.
To continue using Ubuntu on those devices I will need to either:
- upgrade to Ubuntu 17.10 and use the GNOME desktop or gamble with 'Ubuntu Unity' as the Unity desktop in the long term, or
- replace existing Ubuntu 17.04 instances with fresh installations of Ubuntu 16.04 and continue using the Unity desktop in the short term
or, and given my initial concerns, look for an alternative to Ubuntu.
Choosing a distributions is somewhat complex and ultimately personal however one candidate does immediately comes to mind: Fedora. Arguably not so user-friendly it is focused on leading-edge software. The principle difference between Ubuntu and Fedora is package management due to the origins of each namely Debian vs Red Hat respectively.
Becuase the ISO for Ubuntu 17.10 is unavailable and installing Ubuntu 16.04 is effectively downgrading I have made the decision to look at transitioning to Fedora. I'll start by using Fedora as a desktop for everyday use and gradually increase usage. There will be challenges ahead as Fedora 'OOTB' doesn't work on a number of mini PCs both old and new. For inspiration I've combined several well know quotes:
If things don’t change
they’ll stay the way they are,
and if they stay the way they are
they won't get any better,
but the more things change
the more they stay the same.
as making things better need not necessarily make everything different. We shall see.
Please donate if you find my work useful using the following link http://goo.gl/nXWSGf.