From: Alan Cox <alan@linux.intel.com>
Date: Sun, 21 May 2017 21:20:18 +0000 (+0100)
Subject: v68-banked: updated README
X-Git-Url: https://git.ndcode.org/public/gitweb.cgi?a=commitdiff_plain;h=455c2f22f3f63597239ecbda9abe1b5d034ecf86;p=FUZIX.git

v68-banked: updated README

Bring a bit more into line with the code status
---

diff --git a/Kernel/platform-v68-banked/README b/Kernel/platform-v68-banked/README
index fcd8aede..431926e1 100644
--- a/Kernel/platform-v68-banked/README
+++ b/Kernel/platform-v68-banked/README
@@ -9,9 +9,11 @@ faster task switch if we kept udata banks all in the kernel and moved a5
 
 We simulate 4 banks of 128K plus swap.
 
-TODO
+The current model this uses only makes sense for small memory sizes. If you
+have lots of memory then we need a much smarter swap algorithm that does real
+allocation and we also need to avoid copying empty zones we can zero.
 
-- Debug swap (swap out works, swap in fails)
+TODO
 
 - Make bankfixed support only writing out used areas of memory