Hi!
8-Июл-2004 22:45 [EMAIL PROTECTED] (tom ehlert) wrote to Michael Devore
<[EMAIL PROTECTED]>:
te> WITHOUT PROGRAMMING OR TESTING !!
te> old kernel, NEW a000 EMM386:
te> 10000 - 9ffff available to programs
te> 9ffff - a0000 possibly used as UMB link, but not available
te> a0000 - affff available UMB
Unfortunately, my studies shows, that old kernel (umb_init()) works
incorrectly, when new region connected to start of UMB chain. This is why my
second ask was to make testing emm386 edition, which will use A000, but not
joins it to base memory - in this case I expect wrong old kernel behavior
and wish to prove this.
te> AFAICS, the 9ffff - a0000 range is used as UMB link, but in itself
te> never available. the problems I saw are
te> enable_UMB()
te> dos_alloc() returns 8000-afff
te> disable_UMB()
te> however dos_alloc() returning 8000-afff should never be happen,
te> as the UMB link area is used.
There is bug in (old) umb_init(), which not recreates first umb link at
_real_ RAM_TOP (there was used remember, old, non-updated by EMM386 value).
te> arkady turbo kernel, NEW a000 EMM386:
te> 10000 - affff available to programs
Yes, which MS-EMM386, which joins A000 to base memory.
te> if the a000-afff area is indeed merged into lower memory and
te> UMB area starts at Cxxxx, this might even work.
It _is_ work. Though, I not test what happens (even with my edition) if
EMM386 _not_ joins A000 to base memory.
te> however I still would prefer arkady to recompile emm386
te> and tell us if that works, instead of using mental input bandwidth of
te> all people subscribed to this list.
You have latest EMM386 and compilers/assemblers, which need for
recompilation. Why you can't send me recompiled edition? For you this costs
5 (five) minutes to comment translation, recompile and send me executable.
For me this costs some hours to install all required compilation tools and
study EMM386 sources to understand, how it designed.
te> Finally, an idea that might even be used:
te> I found that using EMM386 /IA000-Afff is dangerous.
? Explain, please.
te> But the Numega SofyIce people had the idea to write a program
te> LOADBIG MyProgram MyArguments
te> that would
te> merge A000-Afff into low memory
te> run MyProgram MyArguments
te> unmerge A000-Afff from low memory
There already exists more flexible VIDRAM from QEMM package, which does
this at any time. Then you may run any programs between merge/unmerge
without adding prefix for each call.
te> so I could run a very normal system; but on request (for make/compile
te> jobs mostly) I could enjoy 64/96 K more memory.
-------------------------------------------------------
This SF.Net email sponsored by Black Hat Briefings & Training.
Attend Black Hat Briefings & Training, Las Vegas July 24-29 -
digital self defense, top technical experts, no vendor pitches,
unmatched networking opportunities. Visit www.blackhat.com
_______________________________________________
Freedos-devel mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/freedos-devel
