Hi Stefan,
I think even without these ifdefs you suggested, the patch should not
change semantics:
+++ linux-2.4.18ac2-stefan/net/core/dev.c Wed Mar 27 00:32:17 2002
- dev->flags &= ~IFF_UP;
+ dev->flags &= ~(IFF_UP | IFF_RUNNING);
That should not harm. Currently, 2.4+ does not use the in kernel
IFF_RUNNING flag, and those broken drivers that still do from 2.2 should
be aware that somebody else can change the flags, because dev_close in
2.2 just did this. But may be I'm overseeing something.
I was referring to that. You and me know that but the poor people
downloading the kernel and having one of that borken card. But summa
summarum I agree with the patch as is and I think you don't need to
change anything then.
Hmm, isn't that what the obj-$(OPTION) construct is for: Evaluate to
obj-1 if the option is selected, and to obj- if not?
Yes, you're right, consider me having some brain rush sometimes. Cheers,
Roberto Nibali, ratz
|