|
25 December 2019, 20:28
|
#1 |
|
Registered User
Join Date: Sep 2019
Location: Essen/Germany
Age: 55
Posts: 463
|
Is move.l atomic
When I execute something like this:
Code:
lea NewFunction, a0 move.l a0, VBIService The VBI looks like this (not complete): Code:
... move.l VBIService, a0 jsr(a0) ... rte So can it happen, that the VBI is happening while one half of the ptr is written, so that the jsr will be corrupted? Or is the instruction atomic, so that it is either written, or the old value is still valid? I would expect that it is safe, since we are dealing only with single core here, right? Or are there any other pitfalls involved which might cause a crash with such code? |
|
|
25 December 2019, 20:43
|
#2 |
|
It's coming back!
Join Date: Jul 2018
Location: comp.sys.amiga
Posts: 762
|
Instructions are atomic, yes. Interrupts only happen between instructions.
Plenty of other things can go wrong though. |
|
|
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| non-atomic Copper waits for MrgCop? | Samurai_Crow | Coders. System | 0 | 01 April 2014 20:07 |
| Full versions of Atomic and Evasive Aksion | brolly | request.Old Rare Games | 4 | 05 September 2012 02:50 |
| Move it Move it... (68000) | Gilloo | Coders. Asm / Hardware | 19 | 04 December 2011 17:36 |
| Atomic Cabbage Amiga Page | Enverex | Amiga websites reviews | 10 | 16 May 2003 02:28 |
|
|
