On Tue, Apr 27, 2010 at 19:05, Martin Pitt <martin.pitt@xxxxxxxxxx> wrote:
> harald@xxxxxxxxxx [2010-04-27 12:26 +0200]:
>> diff --git a/extras/cdrom_id/60-cdrom_id.rules b/extras/cdrom_id/60-cdrom_id.rules
>> index 16b3af9..b3109f1 100644
>> --- a/extras/cdrom_id/60-cdrom_id.rules
>> +++ b/extras/cdrom_id/60-cdrom_id.rules
>> @@ -5,7 +5,7 @@ SUBSYSTEM!="block", GOTO="cdrom_end"
>> KERNEL!="sr[0-9]*|xvd*", GOTO="cdrom_end"
>> ENV{DEVTYPE}!="disk", GOTO="cdrom_end"
>>
>> -ENV{ID_CDROM}="1"
>> +KERNEL=="sr[0-9]*", ENV{ID_CDROM}="1"
>
> This would require the distros/platforms (such as some powerpc) which
> still use the IDE CD driver to adapt this as well (which shouldn't be
> a big deal.
Yeah, and they need to patch moer than this rule. All the persistent
namning stuff does no longer match hd* too.
> I wonder why we need this at all. cdrom_id already sets ID_CDROM once
> it figured out that it is a drive. Was this added for cases where
> cdrom_id failed (we recently fixed a ton of bugs in it)? Or for the
> case where you build without extras?
We do O_ECXL now, which means we could validly fail to open the device
while handling an event, and would never set ID_CDROM, even when we
can be sure we have one.
The result is that we remove all /dev/cdrom, ... links with the same
event, which isn't what we want, and therefore just set it when the
kernel device matches and we know we can expect one.
Kay
--
To unsubscribe from this list: send the line "unsubscribe linux-hotplug" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
