I think my commit to md yesterday torpedoed your patch, because one
part of it rej'ected. I applied the .rej by hand and found a
few nits here and there.
I change the list function extensively, because the "512" size of
the array were not picked from where it should come, ,
because it is under
> The "md" problem should really be solved by adding
> #define MD_NAME "md"
> to and using this in both the driver and the mdconfig
> program.
>
> Can I get you to incorporate that in your patch ?
Certainly! The updated patch is at
http://www.unixfreak.org/~dima/home/md-list4.diff (it's
In message <[EMAIL PROTECTED]>, Dima Dorfman write
s:
>> 2) I'm not sure I like the strncmp(.., "md", 2) stuff, as that means that
>> it would also match any other device name that might begin with md, which
>> potentially might not be provided by the md driver. This is currently (I
>> suspect) h
> 2) I'm not sure I like the strncmp(.., "md", 2) stuff, as that means that
> it would also match any other device name that might begin with md, which
> potentially might not be provided by the md driver. This is currently (I
> suspect) hypothetical as we don't have any other drivers beginning w
The only comments I really had on this were:
1) Your comment about vnode->name mapping is correct, there is no way to
turn a vnode pointer into an absolute path meaningful to the caller. There
are countless reasons why believing this is possible is a bad idea, and we
keep having to hit people wi
5 matches
Mail list logo
