Makefile.lib: Lower kernel gzip compression to fastest

You're reading this - so you're trying to understand "JUST WHY OMG".
That's already a good step.

First of all, this is a downstream kernel - always keep that in mind!
Now, this kernel is targeting new *very powerful* Qualcomm platforms
like SM8250 and the Sony Edo platform - which has a very fast UFS card.

Keep in mind that the bootloader sets the CPU at a frequency that is
slightly faster than the "in the middle" ones, which is anyway not
veeeery fast - but that's good, really. I agree.

So.. check this out:   for Image.gz-dtb.....
COMP_LEVEL    SIZE
9             20116171
5      	      20220479
2      	      20940223
1      	      21231290

Remember again that we're loading from a UFS card and that
we are loading ~1.1MB more out of a 20MB file.
If you're smart enough you surely know already about RAM and CPU
overhead of very high compression levels.

If you still disagree with what I just did, read this commit description
another 20 times, or more, until you understand it. :)))

scripts/Makefile.lib

@@ -248,7 +248,7 @@ cmd_objcopy = $(OBJCOPY) $(OBJCOPYFLAGS) $(OBJCOPYFLAGS_$(@F)) $< $@
 # ---------------------------------------------------------------------------
 
 quiet_cmd_gzip = GZIP    $@
-cmd_gzip = (cat $(filter-out FORCE,$^) | gzip -n -f -9 > $@) || \
+cmd_gzip = (cat $(filter-out FORCE,$^) | gzip -n -f -1 > $@) || \
 	(rm -f $@ ; false)
 
 # DTC