Message-ID: <002501be0d0f$0d1fa200$c5223182@marst96.m.resnet.pitt.edu>
From: "mark reed" <marst96+@pitt.edu>
To: <djgpp AT delorie DOT com>
Subject: Re: Question regarding the use of __djgpp_nearptr_enable()
Date: Tue, 10 Nov 1998 20:03:09 -0500
MIME-Version: 1.0
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
X-Priority: 3
X-MSMail-Priority: Normal
X-Mailer: Microsoft Outlook Express 4.72.3110.1
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3110.3
Reply-To: djgpp AT delorie DOT com

Something is wrong with that tutorial, isnt fill dma buffer called by an
interrupt? and near ptr enable/disable  is called in it ?? if you want to
use nearptrs, call enable and disable once.  I think someone just didnt want
to bother learning how to do it without using nearptrs.  What you can do, is
go to

http://www.soundblaster.com/wwwnew/tech/devcnr/sbhwfaq.html

, and look for the question how do i get the sbhwpg (Sound Blaster HardWare
Programming Guide)
then download it, and it is a big manual in pdf i think, and i am sure it
comes with source for
watcom c++.  look under samples/dmaw32  ,  thats the 32 bit protected mode
dos watcom c++ source

Watcom c++ is just like djgpp. Just replace their dos memory functions with
the respeective djgpp dos memory functions. in watcom they call
int86(0x31,whatver); with ax set to function number like 100
with djgpp you just call __dpmi_allocate_dos_memory

if you need help just email me. I have source code and stuff, and havnet
looked over it in a long time, so i wouldnt mine helping you and going back
and remembering what i did :)

i have never found a reason to use nearptrs, i have checked the speed and
stuff, and there is really no diff
(from what i remember i may be wrong, if i am let me know )

doing it this way is a lot better than following tutorials. you learn a lot
more.

Mark



>Hello I am currently programming a sound player for soundblaster cards and
I
>have a question regarding the
>__djgpp_nearptr_enable() function.
>
>I have seen tutorials on the web for programming the DMA but all them use
>the __djgpp_nearptr_enable function.
>I don't want to use that because it disable the memory protection.
>
>Here's some code from the tutorials:
>--------------------------------------------------
>


nearptrs then MKFP?? weird wheres the disable?
>void *MK_FP(DWORD seg, DWORD ofs)
>{
> if (!(_crt0_startup_flags & _CRT0_FLAG_NEARPTR))
>  if
>(!__djgpp_nearptr_enable()) -----------------------------------------------
-
>> Oh no ! no memory protection!
>   return (void*)0;
>  return (void *) (seg*16+ofs+__djgpp_conventional_base);
>}
>
>
>


>void Set_DMA()
>{
> SegInfo.size = BUFFSIZE*2/16;
> _go32_dpmi_allocate_dos_memory(&SegInfo); ------------> Allocate some dos
>memory
> phys = SegInfo.rm_segment<<4;
> if ((phys >> 16) != ((phys+BUFFSIZE)>>16))
> {
>  phys+=BUFFSIZE;
>  printf(" Warning: Hit page division!\n");
> }
> memset((BYTE *)MK_FP(SegInfo.rm_segment,0),0,BUFFSIZE); ----------> Set
>buffer to 0 using nearpointer!
>
> /* Re-enable memory protection */
> __djgpp_nearptr_disable();            -------------------> This is a line
>that I have put to return to the protected mem.
>}
>
>
>/* This function is called often!*/
>void Fill_DMA_buffer()
>{
>


this is in an interrupt right?? dont do this. dont need nearptrs or memset.
>/* if I put: */
>(A)  __djgpp_nearptr_enable();     ------------------> Does this line will
>slow down my real time app.??
>
>....
>/* COPY using DOS memory using Nearpointers */
> memset(...[DMAbuffer -> in DOS memory])
>....
>(B)  __djgpp_nearptr_disable();   ------------------> Does this line will
>slow down my real time app.??
>
>}
>
>
>If I want to have memory protection can I add the two lines (A) and (B)
with
>no major speed down??
>Or is there a better method to allocate DOS memory for the DMA transfert
and
>for Copying sound sample to DMA buffer (This mean copying dpmi memory to
DOS
>memory) without disabling memory protection.
>
>
>Any suggestions or piece of code for the allocation and copy function will
>be greatly appreciated!
>
>Thank you!
>
>Jonathan Villani
>jonathan DOT villani AT videotron DOT ca
>
>
>
>