MS free space problem

Art · Post by **Art** » Mon Jul 20, 2009 9:12 pm

Hi Guys,
I have always used this snippet to retrieve Memory Stick space remaining
I don't know how it works:

void free_ms_memory&#40;&#41; &#123;
int memd;

sceIoDevctl&#40;"ms0&#58;", 0x02425818, &pbuf, sizeof&#40;pbuf&#41;, 0, 0&#41;;
for &#40;memd=0; memd<5; memd++&#41;;

freemem = buf&#91;2&#93; * buf&#91;3&#93; * buf&#91;4&#93;;
freememx = freemem;
freemem = freemem/1024;
freemem = freemem/1024;

sprintf&#40;filler, "    MS&#58; %d Mb",freemem&#41;;
printTextScreen&#40;3, 250, filler, Blue&#41;;
sceDisplayWaitVblankStart&#40;&#41;;
flipScreen&#40;&#41;;
&#125;

Problem is it doesn't work with Memory Sticks 8Gb and over.
Is there a commonly used replacement used for this purpose?

Post by **TyRaNiD** » Wed Jul 22, 2009 3:05 am

It works fine, you are just hitting a overflowing the integer arithmetic. You need to convert your 3 values to 64bit integers.

Art · Post by **Art** » Wed Jul 22, 2009 8:23 am

thankyou :)

PsPfReAK · Post by **PsPfReAK** » Tue Jul 28, 2009 7:28 am

Would this work for the flash0 aswell?

m0skit0 · Post by **m0skit0** » Tue Jul 28, 2009 7:17 pm

Why would use that on flash0? It's fixed size :P

slasher2661996 · Post by **slasher2661996** » Tue Jul 28, 2009 9:12 pm

no it's not. You can partition it from DC

m0skit0 · Post by **m0skit0** » Tue Jul 28, 2009 10:09 pm

Ok, didnt know, sorry.

That code should work on flash partitions too as it uses sceIoDevctl().

Art · Post by **Art** » Wed Jul 29, 2009 9:22 am

I still didn't get the thing to work.
I assume the 64 bit integers are long long.
I got a 16Gb MS now.

cory1492 · Post by **cory1492** » Wed Jul 29, 2009 9:40 am

I've always assumed the actual io command (0x02425818) would be different for different devices, at least it seems to me it is for flushes at any rate.

Going by types.h a 64 bit is either u64 or s64, though signed 64bit should be long long. Are you sure you retyped all the buf/pbuf etc? Is there any particular reason to have that for loop on memd in there as it braces no code? I'd try my own ideas, but I lack anything over 4G.

Post by **TyRaNiD** » Wed Jul 29, 2009 3:11 pm

Urm, what I meant when I said use 64bit ints is to keep the buf as is (leave as 32bit ints) then when doing the calculation cast each element to a 64bit int.

Torch · Post by **Torch** » Wed Jul 29, 2009 11:15 pm

TyRaNiD wrote:Urm, what I meant when I said use 64bit ints is to keep the buf as is (leave as 32bit ints) then when doing the calculation cast each element to a 64bit int.

When I divide u64 by a smaller value like (u64)1024 it crashes.

Post by **TyRaNiD** » Thu Jul 30, 2009 2:29 am

That is unfortunate ;)

Torch · Post by **Torch** » Thu Jul 30, 2009 3:14 am

TyRaNiD wrote:That is unfortunate ;)

...
Even if I assign the result to a u64.
However if I have something like (u64var1 - u64var2)/small then it works. Presumably since the numerator equates to a 32bit value, like what usually occurs when using the RTC timer values in calculations.