Hacking SNEEK v2

[mousex]

Tried setting it equal to zero instead of just deleting it?

God damnit. I should have though of that. Fucking durrr. Yup, that worked.
I gave this tip already an hour ago

Ok, well it compiled, but it did not make it any bit faster. That stinks.
You have to put it into a boot2.bin and place this on your card again. Be sure to remove all compiled es and fs modules before you compile your new one so that you don't put the old one again into your boot2.bin file

 

[giantpune]

ok. i see what it is. there needs to be more #ifdef DEBUG a couple other places in the code. i didn't really try it out till just now. try this http://pastie.org/804570

doing
#DEFINE blabla 0
is not the same as removing the #DEFINE blabla

so you were still using a debug version. the speed increase was due to something else. maybe the placebo effect

 

[FenrirWolf]

ok. i see what it is. there needs to be more #ifdef DEBUG a couple other places in the code. i didn't really try it out till just now. try this http://pastie.org/804570

doing
#DEFINE blabla 0
is not the same as removing the #DEFINE blabla

so you were still using a debug version. the speed increase was due to something else. maybe the placebo effect

In other words, copy-paste that over the matching sections in the code?

 

[giantpune]

yeah. it should be self explanitory. and you have to do the same for both modules. you'll know you did it right when the thing compiles.

 

[FenrirWolf]

Welp, that definitely did the trick. I'd say it loaded around 40% faster than before and things are more responsive. Neato. I should probably set up BootMii Swap with both versions and do a comparison test.

EDIT: Nevermind that last bit. It was just the boot2.bin that changed IIRC, so that wouldn't really work out. lol

 

[giantpune]

find the different file names in the pastie. then find the soction of code that looks a lot like the pastie in the real code. then delete the original code and copy the section of the pastie there. do it for all 3 sections 2 times.

 

[FenrirWolf]

And in the case of vsprintf.c you tack it on at the end since there's no corresponding code to overwrite. Also, I had a hard time trying to do it in notepad before I remembered that devkitPro comes with a developer's notepad that's far easier to use for this sort of thing.

 

[giantpune]

oops. looks like i mixed up the vsprintf and string files then. it should all overwrite existing code. the only differences are the lines starting with #

 

[FenrirWolf]

Interesting in that the speed still improved, though. Mind updating the pastie with the things in the right place? I wanna see if that will make the speed even faster.

 

[giantpune]

it should still work the same. just maybe a extra warning when compiling or something.

 

[FenrirWolf]

So no need to recompile, then? Cool.

EDIT: I went ahead and did it anyway and yup, both versions seemed to run at the same speed. Way better than with the debugger running, though. Sweet. Maybe precompiled ES/FS modules could be released for people who aren't interested in running it with a USB Gecko.

 

[longtom1]

So no need to recompile, then? Cool.

EDIT: I went ahead and did it anyway and yup, both versions seemed to run at the same speed. Way better than with the debugger running, though. Sweet. Maybe precompiled ES/FS modules could be released for people who aren't interested in running it with a USB Gecko.

Can't get it to compile just keep getting errors any chance of precompiled ES/FS modules file upload??

 

[FenrirWolf]

I'm not at my pc anymore so you'll have to hold on for an hour or so.

 

[SanGor]

uhm there is r16 for some time now with important changes for SD stuff for wii apps.

 

[FenrirWolf]

That there is. But what's that got to do with the current discussion?

 

[SanGor]

I doubt disabling will make it much faster, sure it will build the string which will take some time but it is really not that much and if nothing is inserted in Slot B it doesn't send anything or wait for anything.

 

[FenrirWolf]

I doubt disabling will make it much faster, sure it will build the string which will take some time but it is really not that much and if nothing is inserted in Slot B it doesn't send anything or wait for anything.

You know what, you're probably right. I'm doing some better testing and it seems to shave only a few seconds off of the boot time, which is around 17 seconds with the card I'm using.

Heh, that shows me for shouting success before doing more in-depth preparation. But at the same time I'm happy to eek off however many seconds I can, so whatever.

But if anyone wants to play around with the altered modules, here they are:
http://www.megaupload.com/?d=LP4U6ANG

I'll take down the link if I'm not supposed to distribute anything.

 

[leic7]

I am not really sure which part of the following from the es vsprintf.c I should change? It doesn't look like the other vsprintf.c in fs?

static char buffer[1024] ALIGNED(32);
int dbgprintf( const char *fmt, ...)
{
ÂÂÂÂif ( (*(vu32*)(0xd800070) & 1) == 0)
ÂÂÂÂÂÂÂÂreturn 0;

ÂÂÂÂva_list args;
ÂÂÂÂint i;

ÂÂÂÂva_start(args, fmt);
ÂÂÂÂi = vsprintf(buffer, fmt, args);
ÂÂÂÂva_end(args);

ÂÂÂÂ//FIL f;
ÂÂÂÂ//if( f_open( &f, "log.txt", FA_WRITE|FA_OPEN_ALWAYS ) == FR_OK )
ÂÂÂÂ//{
ÂÂÂÂ//ÂÂÂÂf_lseek( &f, f.fsize );
ÂÂÂÂ//ÂÂÂÂu32 read=0;
ÂÂÂÂ//ÂÂÂÂf_write( &f, buffer, strlen(buffer), &read );ÂÂ
ÂÂÂÂ//ÂÂÂÂf_close( &f);
ÂÂÂÂ//}
ÂÂÂÂGeckoSendBuffer( buffer );

ÂÂÂÂ//svc_write(buffer);

ÂÂÂÂreturn i;
}
void fatal(const char *fmt, ...)

 

[FenrirWolf]

I updated my post with precompiled modules that have the changes. I'm pretty sure I did them right, too.

 

[Dmon84]

I don't know if anybody mentioned before, but cIOS rev7 and cMIOS from WiiGator working fine with Sneek. Rev7 doesn't have hdd or double layer support but it plays all my dvd5 backups. So it is not the worst option for now.