Help! JMP out of range?

[Nicolae Nasturica]

Conditional jumps such as JZ or JNC, etc are limited to 127/128 bytes. You cant
jump over 127 or 128 bytes backward or forward.
------8<-------
JZ label1
------8<-------
------8<-------
JNZ label2
JMP label1
------8<-------
this solves your problem.
Good Luck
Alexei A. Frounze
-----------------------------------------
Homepage: http://alexfru.chat.ru
Mirror: http://members.xoom.com/alexfru

Can someone please explain why MASM is reporting "Jump out of range by
130(bytes)
What does this means?
Any help would be very much appreciated..
Vassone

I like your games :)

[wolfgang kern]

On 22/11/2021 20:16, Nicolae Nasturica posted:
---

Can someone please explain why MASM is reporting "Jump out of range by
130(bytes)

because the range is a signed byte (+127..-128) and because MASM don't
know about long CC-instructions (perhaps it knows with hint "long"):

0F 8x 43 21 jmp cc rel+4+2143

versus short:
7x 12 jmp cc rel+2+12

both variants use signed values relative to the address after the jump.
__
wolfgang

[Kerr-Mudd, John]

On Tue, 23 Nov 2021 07:40:41 +0100

Post by wolfgang kern
---

Can someone please explain why MASM is reporting "Jump out of
range by 130(bytes)

because the range is a signed byte (+127..-128) and because MASM
don't know about long CC-instructions (perhaps it knows with hint
0F 8x 43 21 jmp cc rel+4+2143

There are also the unconditional jmps
E9 xx xx jmp
EB xx jmp near[/short]
has the same +127-128 range, but means a slightly increased jmp
forward, ability as the jmp is from the address of the next op code.

Post by wolfgang kern
7x 12 jmp cc rel+2+12
both variants use signed values relative to the address after the
jump. __
wolfgang

It might be useful to someone here; but Alexei answered Vassone's
question back in 2000.

[Terje Mathisen]

Post by Kerr-Mudd, John
On Tue, 23 Nov 2021 07:40:41 +0100

Post by wolfgang kern
---

Can someone please explain why MASM is reporting "Jump out of
range by 130(bytes)

because the range is a signed byte (+127..-128) and because MASM
don't know about long CC-instructions (perhaps it knows with hint
0F 8x 43 21 jmp cc rel+4+2143

There are also the unconditional jmps
E9 xx xx jmp
EB xx jmp near[/short]
has the same +127-128 range, but means a slightly increased jmp
forward, ability as the jmp is from the address of the next op code.

Post by wolfgang kern
7x 12 jmp cc rel+2+12
both variants use signed values relative to the address after the
jump. __
wolfgang

It might be useful to someone here; but Alexei answered Vassone's
question back in 2000.

2000?

You young wippersnappers! :-)

This particular issue came up regularly even before the 808x was
invented, the solution was always the same, i.e. use some other method
for the far jump and a short branch past this code to avoid it.

TASM (maybe around 1986?) had options to handle this automatically: By
starting with the long jump macro at every branch location and then make
one or a couple of passes, nearly all of them could be converted to the
short form.

By starting with the long form

Long branch on Greater to target
jle $+3
jmp target

and then for each target checking the offset and if in the -128 to +127
range, convert to the two-byte form, no previously converted branch can
go out of range, so the process is stable. After 2 or 3 iterations
you'll probably find that either all of them are now short, or that no
more can be converted.

I think there were other assemblers which did the same trick.

Terje

[Kerr-Mudd, John]

On Wed, 24 Nov 2021 16:22:57 +0100

Post by Terje Mathisen

Post by Kerr-Mudd, John
On Tue, 23 Nov 2021 07:40:41 +0100

Post by wolfgang kern
---

Can someone please explain why MASM is reporting "Jump out of
range by 130(bytes)

because the range is a signed byte (+127..-128) and because MASM
don't know about long CC-instructions (perhaps it knows with hint
0F 8x 43 21 jmp cc rel+4+2143

There are also the unconditional jmps
E9 xx xx jmp
EB xx jmp near[/short]
has the same +127-128 range, but means a slightly increased jmp
forward, ability as the jmp is from the address of the next op code.

Post by wolfgang kern
7x 12 jmp cc rel+2+12
both variants use signed values relative to the address after the
jump. __
wolfgang

It might be useful to someone here; but Alexei answered Vassone's
question back in 2000.

2000?
You young wippersnappers! :-)
This particular issue came up regularly even before the 808x was
invented, the solution was always the same, i.e. use some other
method for the far jump and a short branch past this code to avoid it.
By starting with the long jump macro at every branch location and
then make one or a couple of passes, nearly all of them could be
converted to the short form.
By starting with the long form
Long branch on Greater to target
jle $+3
jmp target

I presume the 2nd jmp is the 3 byte version.

Post by Terje Mathisen
and then for each target checking the offset and if in the -128 to
+127 range, convert to the two-byte form, no previously converted
branch can go out of range, so the process is stable. After 2 or 3
iterations you'll probably find that either all of them are now
short, or that no more can be converted.
I think there were other assemblers which did the same trick.

Nasm does; if I spot this I try re-ordering my code to remove them.

Post by Terje Mathisen
--
- <Terje.Mathisen at tmsw.no>
"almost all programming can be viewed as an exercise in caching"