Who should I talk to about making kernel changes that could potential break CI

Who should I talk to about making kernel changes that could potential break CI

[Naveen Nathan]

Hi,

A few months ago I submitted a patch [1] to make /dev/urandom block until
sufficient entropy is gathered; this is blocked by a kernel config
option CONFIG_SECURE_URANDOM=[yn].

If set, this will break systemd systems.

If set, this will probably break Kernel CI (afaik tests are run on
Debian which use systemd) - specifically when randconfig is used
and the config option is set to y.

There is no longer a 'CONFIG_EXPERIMENTAL', to depend/block on, so
I cannot rely on that going forward.

I was hoping to work with the Kernel CI peeps to see if we can prepare
the Kernel CI system to support this config option.

Thanks,
Naveen

[1]: https://lore.kernel.org/patchwork/patch/1079511/

Re: Who should I talk to about making kernel changes that could potential break CI

[Mark Brown]

> On 25 Jun 2019, at 23:21, Naveen Nathan <naveen@lastninja.net> wrote:
> 
> Hi,
> 
> A few months ago I submitted a patch [1] to make /dev/urandom block until
> sufficient entropy is gathered; this is blocked by a kernel config
> option CONFIG_SECURE_URANDOM=[yn].
> 
> If set, this will break systemd systems.
> 
> If set, this will probably break Kernel CI (afaik tests are run on
> Debian which use systemd) - specifically when randconfig is used
> and the config option is set to y.

KernelCI doesn’t do randconfigs so it won’t be affected. 

> 
> There is no longer a 'CONFIG_EXPERIMENTAL', to depend/block on, so
> I cannot rely on that going forward.
> 
> I was hoping to work with the Kernel CI peeps to see if we can prepare
> the Kernel CI system to support this config option.
> 
> Thanks,
> Naveen
> 
> [1]: https://lore.kernel.org/patchwork/patch/1079511/
> 
> 
>