Re: [PATCH v3 0/7] x86/time: improvements to wallclock logic

[Jan Beulich <jbeulich@xxxxxxxx>]

On 03.09.2024 15:02, Roger Pau Monne wrote:
> Hello,
> 
> This series started as an attempt to change the default wallclock
> preference from EFI_GET_TIME to CMOS RTC, but has grown quite a lot.
> First 3 patches should be non-functional changes, mostly chopping the
> current logic into smaller functions so that in patch 4 the probing vs
> runtime wallclock logic can be split.
> 
> Patch 5 changes the preference to use CMOS RTC even when booted from EFI
> firmware.
> 
> Finally patches 6 introduces a new command line option to bypass the
> probing an allow specifying which wallclock source to use on the command
> line.  Patch 7 enables CMOS RTC probing by default.
> 
> Thanks, Roger.
> 
> Roger Pau Monne (7):
>   x86/time: introduce helper to fetch Xen wallclock when running as a
>     guest
>   x86/time: move CMOS edge detection into read helper
>   x86/time: split CMOS read and probe logic into function
>   x86/time: introduce probing logic for the wallclock
>   x86/time: prefer CMOS over EFI_GET_TIME
>   x86/time: introduce command line option to select wallclock
>   x86/time: probe the CMOS RTC by default
> 
>  docs/misc/xen-command-line.pandoc |  28 +++-
>  xen/arch/x86/time.c               | 238 +++++++++++++++++++++++-------
>  2 files changed, 208 insertions(+), 58 deletions(-)

Having reached patch 6, it seems pretty clear that somewhere in the series
a CHANGELOG.md entry wants adding.

Jan