--------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2H00011Date: 02/11/98 From: SCOTT MCNAY Time: 10:50pm \/To: DARRYL GREGORASH (Read 1 times) Subj: System timer tick counter *** Darryl Gregorash wrote in a message to Robert Fortune: DG> This may or may not have a partial basis in the memory DG> architecture of the 16-bit CPUs.. on those processors, it is DG> better to align one's data on a word boundary than on a byte DG> boundary. Similarly, on the 32-bit CPUs, it is better to DG> align on DWord boundaries.. it all has to do with timings DG> when reading from or writing to memory. I wouldn't bet on this, personally, since the original IBM PC used the 8088, which had an 8-bit data path, so that alignment didn't affect the speed. DG> But the singlemost important factor is probably that DW DG> (long int) is already a data type, whereas a 3-byte DG> structure has to be declared during compilation/assembly. Most likely, not to mention it allows the clock speed to be fiddled with, without overflowing the data structure. The Tandy 2000, for example, has 20 ticks per second, I believe (going by memory from Ralf Brown's list). --Scott. --- timEd 1.01 * Origin: Wizard's PC Services, BBS=254-554-2146, Pager=903-3097 (1:395/11) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00000Date: 02/11/98 From: TOM WASSON Time: 04:44pm \/To: SYLVAIN LAUZON (Read 1 times) Subj: Re: i need to allocate some memory TW> Was the first thing that your *.COM program did was free all but TW> the memory used by the program? SL> So i free all but 64k? You can free up all but what your program requires. I have, for example, freed up all but 12K. But remember that you must leave room (at least 1k to be safe) for the Stack and you must change the SS and SP registers accordingly. --- Maximus 3.01 * Origin: Castrovalva BBS 610-917-0380 (1:2626/102) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00001Date: 02/12/98 From: ROBERT KOHL Time: 08:53pm \/To: RICH VERAA (Read 1 times) Subj: pure Hex Programming * Copied from: SAVEAREA Hi Rich, Tuesday, February 10, 1998 18:14:34, Rich Veraa wrote to Robert Kohl Subject: pure Hex Programming RV> In a message to Rich Veraa, Robert Kohl wrote: RV>> The PS/2 Model 30 gave purchasers the choice of a 8088 or RV>> 80286 processor. I believe there was also a Model 25 offered only RV>> with the 8088 RK>> This is the only area of discussion I know a lot about. I RK>> had a model 30. There was no choice in the model 30. You got a RK>> 8086. RV> Sorry, your memory's at fault. I've still got my Model 30-286 RV> right here, and I assure you it came with an 80286 processor, 512 RV> kB RAM (expandable to 2 MB), and a 20 MB hard disk RV> Cheers, Rich Yup, you're right. They added the 286 version of the model 30 when they rolled out the model 55. The original was a 8086 with choice for 8088 or other. RV> http://www.netside.net/~rveraa/ Regards and *Terminate-ing*, Robert (Bob) Kohl Rio Rancho, New Mexico Home Page: http://www.geocities.com/SiliconValley/Way/7854 http://members.tripod.com/~Bob_Kohl/index.html Internet: bobakohl@abq.com bobakohl1@juno.com barbarianh@aol.com .!. Why is it newbies have a Pentium and I still have a 486dx2 50? --- Terminate 5.00/Pro >> Almost All my software is named Bob < * Origin: The Barbarian Hitman FIDONET POINT [private bbs] (1:109/921.66) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00002Date: 02/15/98 From: PETER MAGNUSSON Time: 11:10pm \/To: BRYAN SCHWARTZ (Read 2 times) Subj: flat.xx BS> How do i decode your code? Where do you get this 'xx3402' My code? ;-) Really, I dunno. Maybe someone else knows. I just copied something I found on my disk somewhere... Maybe you should search for help on The Other Net... (internet) [peter] --- FMail 1.02 * Origin: Server*7GB* D0S.0S2.WiN USR33600 0300-13564 (2:203/253) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00003Date: 02/15/98 From: PETER MAGNUSSON Time: 11:15pm \/To: JASEN BETTS (Read 1 times) Subj: REAL BIG! JB> So I guess 16 bit code won't run in flat mode then.... It JB> sounds JB> to me like protected mode with memory management turned off. 16 bit works fine... But you have to keep code bellow the 640k limit. Data above 640k must of course be read by 32bit instructions. Whatever it sounds like, it's still realmode and interrupts works as easily as in normal realmode. Well, I just "know" the theroy, haven't programmed in bigreal, so I guess I can't answer much more questions. [peter] --- FMail 1.02 * Origin: Server*7GB* D0S.0S2.WiN USR33600 0300-13564 (2:203/253) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00004Date: 02/12/98 From: JASEN BETTS Time: 04:51pm \/To: SYLVAIN LAUZON (Read 1 times) Subj: i need to allocate se SL> Okay but how come a system can runs two COM programs at the same time SL> if one COM take all memory available (below 640k). the com file must have given some memory back to the system... --- EzyQwk V1.20 * Origin: CSS Brisbane, Qld, Australia. (61-7-3367-3890) (3:640/350) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00005Date: 02/12/98 From: BRIAN MCCLOUD Time: 04:43am \/To: FERNANDO ARIEL GONT (Read 1 times) Subj: micro-ops DG>But you will never find the micro-code anywhere legal, unless you have ots DG>of money to give to Intel :) I found a large PDF doc. file from Intel that might have it... However it'll take me some time to d/l it, then sift through it (over 1.5 MB). ((Cloud)) MauveCloud@juno.com * OLX 2.2 * hAS ANYONE SEEN MY cAPSLOCK KEY? --- PCBoard (R) v15.3/M 10 * Origin: Next time, Dial The Wrong Number! (209) 943-1880 (1:208/205) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00006Date: 02/11/98 From: SERGUEI SHTYLIOV Time: 11:57am \/To: JEFF QUINDLEN (Read 1 times) Subj: Directly writing to memory Hail! Once you wrote to All: JQ> I consider myself to be fairly experienced assembly programmer, For what kind of CPU? ;) JQ> however there is a simple task in which I've tried and failed numerous JQ> times over. That task is moving data into a specific address. If somebody JQ> could help me I would deeply apreciate it, as it has given me headaches JQ> for days. Would code like this work (to place the value 1h into memory at JQ> 0000:0000)? mov ax,01h mov [00000],ax No, you must first load 0 into some segment register (like DS), and the use it as a segment prefix for move instruction. BTW haven't you tried to simply: xor ax,ax mov ds,ax > mov [ds:0],1 ; clobber int 0 vector ;) All this is for real mode. Selector 0 is illegal in protected mode, and will cause the General Protection fault if my sclerosis doesn't fail. :) JQ> think it is funny that people mistake this for XXX stuff. That one who named the echo should have considered the consequencies. :) Farewell! --- GoldED 2.50+ * Origin: The Wake Of The Medusa (FidoNet 2:5020/157.59) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00007Date: 02/11/98 From: SERGUEI SHTYLIOV Time: 12:17pm \/To: DARRYL GREGORASH (Read 1 times) Subj: PIC/ISRs Hail! Once you wrote to Sam Izzo: SI>> on Creative's Web site that in their ISR, if the IRQ was SI>> greater than 8, they signalled an EOI by outputting 20h to SI>> PIC 2 and also did the usual 20h to PIC 1 Is this SI>> necessary? Absolutely. SI>> Here's some pseudocode of what they did: DG> Yes, it is necessary, because PIC2 actually signals to PIC1 that an DG> interrupt has occurred, and it is actually PIC1 that informs the CPU hat DG> an interrupt has occurred. This is the heart of "INT02 redirection" that DG> confuses many people. My memory is a bit vague, Yeah, looks like this. :) DG> but I think this is how it works: DG> INT02 is not actually redirected anywhere; No, IRQ2 was (or at least could be) used by EGA on XTs, but on ISA bus (286+) this pin (located on XT legacy "8-bit" part of the ISA connector) is called IRQ9 instead of IRQ2, so actually any card using IRQ2 will raise IRQ9 instead. BIOS INT 71h handler looks like this: push ax mov al,20h out 0A0h,al ; send EOI to the "slave" PIC pop ax int 0Ah iret So, as you can see IRQ9 is routed to IRQ2 handler. DG> rather the INT09 bit on PIC2 is latched onto the INT02 line on the DG> system No, there's no IRQn lines on the system bus (or "host bus", or "CPU bus"), just on the ISA bus. DG> bus, while the cascade bit What bit? Do you think that PICs use some dedicated pin for cascading? You're wrong then -- they use INTR/IRQn pins (INTR output is used by a "slave", IRQn input is used by the "master"). DG> on PIC2 is latched onto the INT02 bit of PIC1. Yes, INTR pin of PIC2 ("slave") is connected (not latched) to IRQ2 pin of PIC1 ("master"). DG> When PIC2 issues an interrupt, it raises the cascade bit, INTR pin. DG> which in turn raises an INT02 on PIC1; it is actually PIC1 that raises DG> the INT signal to the CPU by raising its cascade bit. Yes. All right. But I still haven't understood how could IRQ9 input of PIC2 be "latched onto" the IRQ2. Maybe my English simply loses. :) Farewell! --- GoldED 2.50+ * Origin: Billy's Bones (FidoNet 2:5020/157.59) --------------- FIDO MESSAGE AREA==> TOPIC: 145 ASSEMBLY LANG. Ref: F2K00008Date: 02/14/98 From: ROBERT KOHL Time: 01:00am \/To: CRAIG HART (Read 1 times) Subj: pure Hex Programming * Copied from: SAVEAREA Hi Craig, Wednesday, February 11, 1998 17:10:00, Craig Hart wrote to Robert Kohl Subject: pure Hex Programming CH> Hi.. >> This is the only area of discussion I know a lot about. I had >> a model 30. There was no choice in the model 30. You got a 8086. CH> Incorrect. There are several 'model 30' PS/2's. The later CH> model 30's came with a 286 CPU and a 1.44 FDD as opposed to the CH> eariler ones which were 8086's and 720k FDD's CH> You probably owned a model 30 8530-021 which is the 8086/720k CH> version. I am kicking one as I type :- CH> Craig I think I already covered this with the mail which is on it's way out. But here's what I did say. The model with the 286 came out about the same time the model 55 came out. I had the the 8530-021 one anemic 20 meg HD and a single 720k A: drive. When I had mine, I added a 5.25" drive via external cable. The drive ran without the need of an external power supply. Regards and *Terminate-ing*, Robert (Bob) Kohl Rio Rancho, New Mexico Home Page: http://www.geocities.com/SiliconValley/Way/7854 http://members.tripod.com/~Bob_Kohl/index.html Internet: bobakohl@abq.com bobakohl1@juno.com barbarianh@aol.com .!. PS/2 an idea.....somewhere. Ahead, behind and such, of it's time. --- Terminate 5.00/Pro >> Almost All my software is named Bob < * Origin: The Barbarian Hitman FIDONET POINT [private bbs] (1:109/921.66)