Re: RFC: fullfs — disk-full simulation filesystem

[Roland Illig <roland.illig%gmx.de@localhost>]

Am 18.03.2026 um 05:46 schrieb Hudson C. Dalpra:
> Hello tech-kern,
> 
> Before I proceed with the remaining of the project, it would be great to
> have some guidance on the following:
> 
>> I am not familiar with patches through email, so I've created a PR 
>> in Github: https://github.com/NetBSD/src/pull/76
> 
> Is this an acceptable approach or should I stop at once?
> 
> On the fullfsctl subject, I was planning to add support to:
> - turning failling on/off
> - changing the error
> - selecting which operations fails
> - setting number of operations before failing
> - setting number of bytes written before failing

While these ideas are interesting, they stretch the name "fullfs". Maybe
a more appropriate name would be "errfs".

The SQLite test suite (https://sqlite.org/testing.html) takes a similar
approach, with the failure modes "fail once" and "fail from now on" that
kick in after a specified number of operations. Maybe you find further
ideas there.

Roland