[arch-general] /etc/modules-load.d | /etc/modprobe.d -- options determinative?

[Matthew Monaco]

On 08/17/2012 09:10 AM, David C. Rankin wrote:
> On 08/17/2012 10:49 AM, David C. Rankin wrote:
>> Guys,
>> 
>> Reading the wiki pages regarding /etc/modules-load.d and /etc/modprobe.d, 
>> is the only difference that makes /etc/modules-load.d to correct place for
>> the <moduleconfname>.conf file a lack of options required on module load?
>> 
>> Or put another way, if options need to be provided along with the module 
>> name on load, then use
>> 
>> /etc/modprobe.d/<moduleconfname>.conf
>> 
>> otherwise if just the module name is required, use:
>> 
>> /etc/modules-load.d/<moduleconfname>.conf?
>> 
>> That was the only difference I saw in their use in the wiki. Anything
>> else?
>> 
> 
> Also,
> 
> for dmraid, the old MODULES=(dm_mod dm_mirror) would be replace by say 
> '/etc/modules-load.d/dmraid.conf' with the contents:
> 
> # load dm_mod dm_mirror for dmraid dm_mod dm_mirror
> 
> But what of the 'USEDMRAID="yes"' directive presently in the HARDWARE
> section of rc.conf? Is the call to 'dmraid -i -ay' automatically done if
> dmraid is specified by HOOKS= in mkinitcpio.conf?
> 

Doesn't look like it, from rc.sysinit:

----------------------------8<------------------------------
# FakeRAID devices detection
[[ $USEDMRAID = [Yy][Ee][Ss] && -x $(type -P dmraid) ]] &&
	status "Activating FakeRAID arrays" dmraid -i -ay
----------------------------8<------------------------------

> Lastly, will the new '/etc/modules-load.d/dmraid.conf' work with the current 
> init before the switch to systemd?
> 

Yes, they are identical, initscripts leverages /usr/lib/systemd/systemd-modules-load