Hi!
> Does anything use those two interrupts in an unbuffered state for UMB space?
NOT LBAcache, but TICKLE. The point is, only TICKLE has local buffers.
So you should not load TICKLE to UMB if you have UMB DMA problems.
LBAcache itself does not alter the buffer pointers - it just uses the
buffe
Hi!
9-Дек-2004 13:36 [EMAIL PROTECTED] (Michael Devore) wrote to
[EMAIL PROTECTED]:
>>So if you can force buffering in the int21-calling software by
>>reading/writing <= 511 bytes or using an UMB or HMA transfer address but
>>not globally.
MD> Okay, stick with me on this for a bit longer. Since,
At 08:53 PM 12/9/2004 +1300, Bart Oldeman wrote:
The low level disk device driver also contains code that forces buffering
if the segment of the address is >= 0xa000, using a different
(internal to the device driver) buffer.
So if you can force buffering in the int21-calling software by
reading/wri
Hi!
9-Дек-2004 01:22 [EMAIL PROTECTED] (Michael Devore) wrote to
[EMAIL PROTECTED]:
MD> Do FreeDOS disk accesses via INT 21h, INT 13h, INT whatever,
INT13 only (INT21 and INT25/26 are DOS internal entries).
MD> read and write
MD> directly from/to a passed address or do they always buffer t
On Thu, 9 Dec 2004, Michael Devore wrote:
> Do FreeDOS disk accesses via INT 21h, INT 13h, INT whatever, read and write
> directly from/to a passed address or do they always buffer the disk
> access? Never? Sometimes?
Sometimes.
> If never/sometimes, is there is a way to make it buffer always
Simple questions, hopefully simple answers. I could probably figure it out
myself eventually by digging through a few tons of source code, but
somebody here probably knows the answers a whole lot quicker.
Do FreeDOS disk accesses via INT 21h, INT 13h, INT whatever, read and write
directly from
