Richard,
I got your stealth comment below: However...... I believe you said
there times when devfs_register could return NULL (e.g. out of memory??).
In some cases, this would be an indicator to back out/unload. I have a
module that matches that criteria. No use having the module loaded (and
using up some variety of resources) if an application cannot access it.
Depending on the envinonment, this may or may not be highly theoretical.
As a workaround, I did just as you suggested, ignoring the return
code. Looking at the drivers in the kernel source, they all seem to do
this anyway. I'll just comment this caveat in the drivers' code.
Regards,
Jon Grimm
jgrimm@xxxxxxxxxx
Sent by: owner-devfs@xxxxxxxxxxx
To: Paul Hedderly <paul@xxxxxxx>
cc: devfs <devfs@xxxxxxxxxxx>
Subject: Re: devfs for 2.2.18 - any chance?
Paul Hedderly writes:
> The subject says it all. I couldn't get devfs working with 2218.
Wait. I'll be producing a patch.
NOTE to John Grimm (?). The email I sent to you was lost. Just ignore
error codes. That should work for both cases.
Regards,
Richard....
Permanent: rgooch@xxxxxxxxxxxxx
Current: rgooch@xxxxxxxxxxxxxxx
|