RST #10, command

I suddenly saw on 05-12-97, at 02:24, Denis Parinov wrote to All: Hi Denis! DP> Maybe someone was involved in calling subroutines via RST #10: DP> RST #10 DP> db DP> I'm interested in a program that does this as quickly as possible. DP> 48 BASIC does not need to be processed, we will get to RST DP> this program right away. I’m not ashamed to admit that I’m working on it... :) There's no way to get anything quick here. the gist is this: First we initialize the interceptions... where we write the thread: LD HL,rst16_ LD (#5C51),HL and on the rst16_ label there should be this is: rst16_DW rst16 that is, this address contains the real interceptor address. Well, this is the processing procedure itself, there were no optimizations! this is an example from raw source. rst16 LD HL,0 ADD HL,SP LD BC,4 ADD HL,BC LD (ADR_ST+1),HL LD E,(HL) INC HL LD D,(HL); found the address LD A,(DE); which follows INC DE; just after rst 16 ; saved ix ; in register A with us ; next byte immediately ; for rst The DE register is vital, it is now points to the second byte after rst16.after the ADR_ST label it must indicate per byte after all data! that is already to the next command. this is of course for of this example. here by type we switch to ours procedures. AND A JR Z,proc1 DEC A JR Z,proc2 exit ADR_ST LD HL,0; replace the address ; return on stack LD (HL), E to the one we need, INC HL contained in DE LD(HL),D RET that's all, in a nutshell say... try, optimize, I hope you've already seen what's going on in the ROM for this about... trace any rst 16... ps. This procedure does not allow recursions! that is, it cannot be called if we have already went via rst16. only one occurrence, but You can do anything you want. ▌▌║▌█▐│▌▌▐▐ WiTh The BeST wIsheS fROM ▌▌║▌█▐│▌▌▐▐ *C*R*E*A*T*O*R* ▌812/03.00▐ -+- SMM version 1.05

Greetings, Paul! In the echo of CODE.ZX, you wrote a letter for Denis Parinov. It was 06-12-97, at 00:00. DP>> Maybe someone was involved in calling subroutines via RST #10: DP>> RST #10 DP>> db DP>> I'm interested in a program that does this as quickly as possible. DP>> 48 BASIC does not need to be processed, we will get to RST DP>> this program right away. [skip] The point is that the handler must be in ROM and save registers. I wrote it in principle, but it has a drawback he ruins the flags. rst_16 PUSH HL PUSH DE LD HL,4 ADD HL,SP LD E,(HL) ;Take the lowest byte of the return address INC (HL) ;Adjust the low byte of the address. on the stack INC HL LD D,(HL) ;Take the high byte of the return address JP NZ,LP INC (HL); If necessary, corr. high byte return address LP EX DE,HL ;B HL - return address LD L,(HL) ;Take the code LD H,PROG/256 LD E,(HL) ;Remove the smallest byte of the subroutine address from ;tables INC H ;Go to the table of high bytes LD D,(HL);Remove high byte EX DE,HL POP DE EX (SP),HL ;Place the address of the required subroutine on the stackRET ; and go to it PROG DEFS 256 ;256 ml bytes of subroutine addresses DEFS 256 ;256 st.bytes ---------//-------- Maybe it can be done faster somehow. PF> PF> that's all, in a nutshell PF> say... try, optimize, I hope Well, of course I’ll try it :) PF> you've already seen what's going on in the ROM for this PF> about... trace any rst 16... Yes, there is nothing interesting in the choice of channels, but for RST 8, In general, it looks like they decided not to bother :) PF> ▌▌║▌█▐│▌▌▐▐ WiTh The BeST wIsheS fROM PF> ▌▌║▌█▐│▌▌▐▐ *C*R*E*A*T*O*R* PF> ▌812/03.00▐ All the best, Mr. Paul Falcon That's actually all I wanted to say on the topic "RST #10, command". Best regards, Denis. -+- SMM version 1.05

Suddenly I saw on 12/12/97, at 05:08, Denis Parinov wrote to Paul Falcon: Hi Denis! giving two and a half hours of his great life to this problem, that's what I got it... ;) rst_16 PUSH HL; 11 POP HL; 10th place X POP HL; 10 INC HL; 6 PUSH HL; 11 DEC HL; 6 LD L,(HL) ; 7 LD H,PROG/256; 7 DEC SP; 6 DEC SP ; 6 PUSH AF ; 11th place Y LD A,(HL) ; 7 INC H; 4 LD H,(HL) ; 7 LD L,A ; 4 POP AF ; 10 EX (SP),HL ; 19 RET ; 10 after all, 152 bars... PROG DEFS 256 ;256 ml bytes of subroutine addresses DEFS 256 ;256 st.bytes ---------//-------- OOO! it looks like something happened, it works at first glance, but even if not, I think you understand. 152 clock cycles... in my opinion this is not the limit ;) but... if there is an interruption between place X and Y then HL will be lost. ▌▌║▌█▐│▌▌▐▐ WiTh The BeST wIsheS fROM ▌▌║▌█▐│▌▌▐▐ *C*R*E*A*T*O*R* ▌812/03.00▐ -+- asm...