fix warning in tlbex.c for CONFIG_32BIT

For 32 bit builds CONFIG_64BIT is not defined. Should be doing #ifdef not #if.

@@ -513,7 +513,7 @@ static __init int rel_lo(long val)
 
 static __init void i_LA_mostly(u32 **buf, unsigned int rs, long addr)
 {
-#if CONFIG_64BIT
+#ifdef CONFIG_64BIT
        if (!in_compat_space_p(addr)) {
                i_lui(buf, rs, rel_highest(addr));
                if (rel_higher(addr))

                                                ^^^^^^^^^^^^^

Applied - BUT: your mailer garbles patches ...

Ralf

Some people on this list are quite adamant that patches be in-line.

I was trying to see how my mailer (Thunderbird) handled this. Obviously
(in hindsight) it screws things up.

Sending as an attachment works well except some mailers (Not Thunderbird) cannot quote attached patches with out jumping through hoops.

I don't really want to change the mailer that I am using, so I am in a bit of a bind WRT submitting patches here.

FWIW other mailing lists (binutils, gcc) don't seem to have the same trouble with attached patches.

David Daney

Maybe a different style of work there. The submission style we're asking people to follow here is exactly the same as on linux-kernel, netdev or other kernel-related lists.

I just asked somebody; this is the answer I got:

<snip>
I have never had any luck getting mailers to send patches in a way that no one complained about. In the end, I used this
http://www.speakeasy.org/~pj99/sgi/sendpatchset

I found the best way is to have a directory with patches like 001_part1.patch, 002_part2.patch etc with matching explainations in 001_part1.mail 002_part2.mail . I have a script that generates the final mails and feeds them to sendpatchset
<snip>

The script may not be what you want but I guess I'll be something like it to deal with the huge patchsets I'm sometimes fiddling with - like the 452 patch monster right now ...

Ralf