[9fans] random question

[9fans] random question

[Nick Owens]

[-- Attachment #1: Type: text/plain, Size: 113 bytes --]

why does /sys/src/9/port/random.c:/^randomclock use rb.next to only run
1/4th of the time it normally would?


[-- Attachment #2: Type: application/pgp-signature, Size: 836 bytes --]

Re: [9fans] Random Question

[Don]

> Doubtful; a documented 3 byte signature isn't much of a hurdle.
Agreed
> Smells more like a buggy attempt to guess at what partitioning scheme
> might be in place.
Probably. Sure would have been a cool trick though :)
Don

RE: [9fans] Random Question

[greg andruk]

Don wrote:

> Heyas,
> 	I just finished writing most of the MBR
> for my Autumn OS and began reading through FS
> documentation to help me develop my FS. Now.. an
> intereting notion crossed my mind while reading
> the FAT specification by Microsoft. If byte zero
> of a BPB jmp_Boot is 0xEB (the preference is this
> short jmp according to the specs) the third byte
> _must_ be a NOP (0x90).

Right.

> Now, I remember reading
> in several other operating system's comments
> that some BIOS actually bail out if there isn't
> a jmp short followed by a nop in the first three
> bytes of code. Noone seemed to know why the
> BIOS needed to see that nop there.

It's historical identification cruft.

2 bytes of Short jump + Nop -> DOS 3 or later
3 bytes of near jump = DOS 2
DOS 1 didn't have a BPB structure.

> Could this
> have been a tactic used by Microsoft to attempt
> to thwart other Operating Systems from occupying
> an IBM PC?

Doubtful; a documented 3 byte signature isn't much of a hurdle.

> I mean, if the BIOS is designed to check
> for a documented FAT specification, wouldn't that
> be a clue that Microsoft coerced the manufacturers
> to implement the check? Anyone out there have a
> comment on this issue?

Smells more like a buggy attempt to guess at what partitioning scheme
might be in place.

> Don

[9fans] Random Question

[Don]

Heyas,
	I just finished writing most of the MBR
for my Autumn OS and began reading through FS
documentation to help me develop my FS. Now.. an
intereting notion crossed my mind while reading
the FAT specification by Microsoft. If byte zero
of a BPB jmp_Boot is 0xEB (the preference is this
short jmp according to the specs) the third byte
_must_ be a NOP (0x90). Now, I remember reading
in several other operating system's comments
that some BIOS actually bail out if there isn't
a jmp short followed by a nop in the first three
bytes of code. Noone seemed to know why the
BIOS needed to see that nop there. Could this
have been a tactic used by Microsoft to attempt
to thwart other Operating Systems from occupying
an IBM PC? I mean, if the BIOS is designed to check
for a documented FAT specification, wouldn't that
be a clue that Microsoft coerced the manufacturers
to implement the check? Anyone out there have a
comment on this issue?
Don