| |
| |
| |
|
Page: 1 2 3
<BobtheAvenger> hey
<skuggi> ho
<BobtheAvenger> sup?
<skuggi> just about to go to bed.
<BobtheAvenger> hmm, bed, i should do that too
<BobtheAvenger> but i'm outa a mood so i got energy
<skuggi> well, i've got an exam tomorrow morning, so i'd like to sleep some..
<skuggi> night.
<BobtheAvenger> ooo, thats what i forgot
<BobtheAvenger> peace
<looth> hum... when I get an interrupt... do I need to provide an interrupt stack? I mean... else, it will use whatever thread stack was active?
<prolific> What's the syntax to jmp ahead X number of bytes?
<Robert> jmp $+123 in NASM/FASM/YASM, I think.
<prolific> Ah nice. What about Gas?
<Robert> No idea.
<Robert> rtfm? :)
<prolific> Yeah.. :p
<imaginator> prolific: you can do: jmp (1f + 123) \n :1
<prolific> imaginator, I don't get it :/
<prolific> 1f and 123 are arbitrary?
<imaginator> :1 is a local label. 1f means the address of local label 1 that is forward
<prolific> Oh interesting
<prolific> But since when does jmp care if a label is forward or backward?
<imaginator> prolific: 1f and 1b are just special gas instructions for local labels.
<imaginator> you can also do things like: 2f and so on
<prolific> Hm. Placing ':1' on a line by itself doesn't work. Nor 1: if that's what you meant
<prolific> 2f = two label's forward?
<prolific> Or a label named '2' that is forward?
<imaginator> oh I meant 1: sorry
<imaginator> here's an example: 1: \n movl $1b,%eax \n 2: \n movl $2b,%ecx
<prolific> So it wouldn't work if I did: jmp here+number_of_bytes?
<prolific> here = the label just following the jmp
<imaginator> it should work. I think the jmp is relative to the next instruction.
<prolific> I think so also
<imaginator> jmp (1f + (2f - 1f))
<imaginator> 1:
<imaginator> ret
<imaginator> call abort
<imaginator> 2:
<imaginator> ret
<imaginator> that seems to work
<imaginator> at least it doesn't abort and the code returns
<prolific> So, that example wouldn't work if the first line were: jmp (one + (two - one) )?
<prolific> Sorry, I'm slow
<imaginator> are one and two labels?
<prolific> Yes; and the same as in that example
<prolific> Just 1 = one. etc
<imaginator> it should work
<imaginator> try it :)
<prolific> Ok. Thanks for you help
<prolific> I'm :p
<imaginator> you're welcome. have fun
<prolific> I just have a bunch of string comparisons and call'ing procedures based on which is inputted. So instead of saying: if this!="command1" / je command2
<prolific> I want to simply jmp ahead 9 bytes to the next string comparison
<prolific> But now it looks like I'm going to have labels all over the place with this method :)
<prolific> M way would require a new COMMAND_X label for each string comparison I add
<prolific> My
<prolific> Wish I could just jmp $9 :/ oh well
<imaginator> jmp (label + 9) ; but why do that? labels are meant to be used. I mostly use local labels for anything that isn't a procedure.
<imaginator> one thing I find especially frustrating in asm in unrolling loops. I'm still practicing this a lot
<prolific> Well, my basically boils down to the switch/case concept, although I'm trying to implement it myself of course. In psuedo code I have something like: if buf != this_string: jmp do_next_comparison / do_next_comparison2: if buf != this_string2: jmp do_next_comparison3
<prolific> But each time I add a new command (string comparison), I have to add another "do_next_comparisonX" and replace the old...bleh
<prolific> And each jmp do_next_comparison is a constant distance
<prolific> Ah , you mean like detecting compiler-rolled loops?
<imaginator> I think switch/case uses a jump table usually., so there is like an array of pointers in memory for code to jump to, and maybe a range cmp before that.
<imaginator> prolific: no. I mean unrolling loops in pure asm code.
<prolific> Right
<prolific> imaginator, For what purpose?
<imaginator> more speed
<prolific> Hm can you give me an example?
<imaginator> sure, just a sec. I'll show you something I want to unroll
<prolific> Ok
<imaginator> http://rafb.net/paste/results/GcWQDf59.html
<prolific> So you wouldn't totally get rid of the loop in that case, right?
<imaginator> what I want to do is more bts patterns, but also not set bits that shouldn't be set.
<imaginator> I suppose I should do something like Duff's device using the modulus as an index into some labels
<prolific> Ah, I see how that can be _a lot_ quicker
<prolific> Later
<|isubx|> woohoo almost got it http://www.rafb.net/paste/results/fAwGOE46.html
<|isubx|> its big and its ugly but it compiles and it works outside of the fact it is skipping the first 3 fibonacci numbers 1,1,2 starting at 3 :)
<daleXXgribbel> hey all
<|isubx|> hello Mr. Gribbel
<daleXXgribbel> does this room include the masm ***embler?
<imaginator> daleXXgribbel: yes
<|isubx|> who has cursed you with masm :P
<daleXXgribbel> my instructor ;D
<|isubx|> sounds like were both in the same leaky boat
<daleXXgribbel> i'm working on a recursive fibonacci procedure that is just not working
<|isubx|> lemme guess your also reading the holy grail of kip r. irvine :P
<daleXXgribbel> lol, yes, which i have been told is not a great source
<|isubx|> the book is actually fairly good.. its his damn library that is garbage
<daleXXgribbel> haha
<|isubx|> what school you going too?
<daleXXgribbel> well my instructor tells me that the book isn't great, but its been a decent resource
<daleXXgribbel> austin community college
<daleXXgribbel> you?
<|isubx|> rofl! mr black?
<imaginator> you're both in the same cl*** aren't you?
<|isubx|> sounds like it! rofl
<daleXXgribbel> hahaha
<daleXXgribbel> thats awesome
<|isubx|> and who in my cl*** are you?? lol
<daleXXgribbel> ben, and who are you?
<|isubx|> gabriel
<daleXXgribbel> which program are you working on?
<|isubx|> same one lol finally got it working for the most part
<daleXXgribbel> oh man
<daleXXgribbel> i have been slaving over this thing for every night over the past week and i feel like i'm still not making much headway
<|isubx|> i have been bugging these guys all weekend
<daleXXgribbel> haha
<|isubx|> ohh yea ive been bangin my head against the desk
<|isubx|> my last problem at this point is my fib sequence is starting at 3 and skipping 1,1,2
<daleXXgribbel> so, does it work for 4?
<daleXXgribbel> or above?
<|isubx|> ya works fine
<daleXXgribbel> oh thats cool
<daleXXgribbel> mine works great for 0,1, or 2
<daleXXgribbel> lol
<|isubx|> i read lots of other peoples source on fib sequences posted on the net and then my biggest problem was doing it different from thiers
<daleXXgribbel> yeah i've gone over lots as well, none seem to be helping me much
<|isubx|> and its pretty ugly but better than just living in a debugger
<daleXXgribbel> sure
<|isubx|> evening xark
<daleXXgribbel> did you find one example that was particularly helpful?
<|isubx|> here is sheng jiang's source.. he took mr black last semester http://spaces.msn.com/members/jiangsheng/Blog/cns!1pVUmC6mid-3zVpLZNQTgtuw!187.entry
<|isubx|> met him in the computer lab acouple weeks ago
<daleXXgribbel> wow
<|isubx|> i believe he is the student who mr black was saying came back to test his microcode on the pic
<daleXXgribbel> oh thats awesome
<|isubx|> funny how small a world we live in!
<daleXXgribbel> yeah i enjoy the cl***, i think he is a good instructor
<daleXXgribbel> haha definitely
<daleXXgribbel> sorry, i don't think we've actually met in cl***, i have longer hair and gl***es
<|isubx|> ya he is, im just not that great a student lol
<daleXXgribbel> haha
<|isubx|> ya last tuesday you came into the computer lab before cl*** and you sat down next to me
<|isubx|> usually have a black cap on, think a guy sits between us in cl***
<daleXXgribbel> oh cool
<daleXXgribbel> how did you do on the test?
<|isubx|> made an 88
<daleXXgribbel> thats awesome
<|isubx|> ya i should have done better, everything i got wrong was stupid stuff that i knew better
<daleXXgribbel> yeah i studied a ton for it
<|isubx|> same here, i had 2 other tests that same day tho was a day from hell
<daleXXgribbel> yikes
<daleXXgribbel> omg it is finally giving me something back
<|isubx|> lol it always feels so good when things start to work
<daleXXgribbel> i lost so much sleep last week over this
Return to asm
or
Go to some related
logs:
beginner
moo.dll np script now playing
java
|
|