Hi all,

>      Of course, I remember discussions about handling A20 behavior. Only
> main reason to support switching A20 off is bugs in MS exepacker (which was
> used by MS link in many programs) and old PKLITE - they both behave
> wrongly, when runs in first 64k of memory and there is HMA enabled.

This reminds me that on SOME systems, older versions of HTMLHELP crashed
when DOS was using the HMA. Either caused by CATS problems or by ZLIB
problems with the memory model. Somebody recompiled with KITTEN and a
fresh compile of ZLIB, but I do not know if the new binary is online.

In addition, I would like to know if OTHER programs which still use CATS
have that kind of problem. In either case, the crash only happened when
using HTMLHELP more than once without rebooting... Bernd can describe
this better than me. We found out that filling the first 64k stopped the
problem, but we also found that FREECOM compilation is configured to
DISABLE LOADFIX if you ENABLE XMS SWAP - does anybody know if this is
actually required? Looks like we really do need loadfix sometimes!
Having loadfix would allow us to use A20control:off option of HIMEM,
or whatever the "only ENABLE A20 and never disable it again" option is
called at the moment.

Even a kernel config sys SWITCH to disable disabling of the A20 would
be nice: You could save memory by removing the stubs in low memory and
by moving more data structures, maybe even List of Lists, into HMA, as
soon as that "A20 never disabled" switch is activated. Well, LoL might
be asking a bit too much - drivers could already be using it before HMA
becomes accessible, and then you can no longer move LoL around.

Eric



-------------------------------------------------------
The SF.Net email is sponsored by: Beat the post-holiday blues
Get a FREE limited edition SourceForge.net t-shirt from ThinkGeek.
It's fun and FREE -- well, almost....http://www.thinkgeek.com/sfshirt
_______________________________________________
Freedos-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/freedos-devel

Reply via email to