Re: fault(4)

[Paul Goyette <paul%whooppee.com@localhost>]

On Sat, 8 Feb 2020, Martin Husemann wrote:

> On Sat, Feb 08, 2020 at 06:19:43AM -0800, Paul Goyette wrote:
> > The module should be MODULE_CLASS_DRIVER.  And there should be a
> > sys/module/fault/Makefile to build the module, along with changes to
> > sys/module/Makefile (to descend into the fault directory) and to
> > src/distrib/sets/lists/modules/mi (for new files).  The module init
> > code should also have appropriate config_{init,fini}_component()
> > calls for inserting the device into the autoconfig database.
>
> Since all probes have to be hard-coded this is very intrusive and I am
> not sure a module makes sense.

True.  But in that case it should not be a module at all.

But shouldn't it still be a pseudo-device and not just an option?



+--------------------+--------------------------+-----------------------+
| Paul Goyette       | PGP Key fingerprint:     | E-mail addresses:     |
| (Retired)          | FA29 0E3B 35AF E8AE 6651 | paul%whooppee.com@localhost     |
| Software Developer | 0786 F758 55DE 53BA 7731 | pgoyette%netbsd.org@localhost   |
+--------------------+--------------------------+-----------------------+