Re: [OSDev] How to switch to long mode in x86 CPUs?

On 04/03/2025 18:31, Paul Edwards wrote:

"bart" <bc@freeuk.com> wrote in message news:vq208q$re74$1@dont-email.me...

On 02/03/2025 08:22, Paul Edwards wrote:

"Scott Lurndal" <scott@slp53.sl.home> wrote in message
news:JdFwP.46247$SZca.36276@fx13.iad...

"Paul Edwards" <mutazilah@gmail.com> writes:

"David Brown" <david.brown@hesbynett.no> wrote in message
news:vprtt6$3jah9$1@dont-email.me...

On 27/02/2025 16:57, Ar Rakin wrote:

>

>

Do you consider the concept of a BIOS (as seen as the IBM PC),
"legitimate to use"

>
In the abstract, possibly.  But the last half century has
shown that BIOS as an I/O abstraction layer was a bad idea
from the start.
>

and do you consider MSDOS (which uses that
BIOS) to be an operating system?

>
No, MSDOS was, is, will always be a simple program loader.

>
Plus manages memory.

>
How does it do that? I seem to recall that you got 640KB minus whatever
the resident parts of the OS needed.

 Yes - and? Then you need to manage that memory, so you
need a memory manager, which MSDOS provides. It's not
trivial to write/debug one of those either.
 I'm not sure what your question is exactly, so I'll include code
to call the interrupt (21H AH=48H) you need to obtain memory.

I was questioning whether MSDOS provided means to manage memory, since I don't recall anything like that. My programs handled their own memory as they did most other things.
This 0x48/48H syscall is new to me.
(But it also looks like something it wouldn't use anyway, since it appears to work like malloc in storing the allocated size. That's something I would never need, and would be wasteful with the limited memory available.
Because I will either know the fixed size of the object I want to freed, or would need to keep a record of it anyway if it's more dynamic.)