Mallocr.c _free_r problem,Really starting to bother me Alot.

[KickinAezz]

which point to

Code: Select all

_free_r
../../../../../newlib/libc/stdlib/mallocr.c:2708

Code: Select all

_free_r
../../../../../newlib/libc/stdlib/mallocr.c:3340

Changing Stack/heap sizes dint help...

There's 14050Kb of usable ram remaining....

The scenario: As I increase the number of statements in some function, EVEN BEFORE the control reaches that function, exception occurs....

[KickinAezz]

Solved ATM,

sceKernelDcacheWritebackAll(); Did the trick...

[jimparis]

Crashes in malloc/free are almost always due to heap corruption. The cache writeback is probably just hiding your problem for now.

[KickinAezz]

Crashes in malloc/free are almost always due to heap corruption. The cache writeback is probably just hiding your problem for now.

Is there any other solution for this?

Coz, this errors make me condense/reduce/limit the number of lines I put in a function...

[J.F.]

The solution is to debug your code. The last time I helped someone find a bug, it was a string overflow (writing too many characters into a fixed size string). Bugs like that will cause all kinds of interesting problems. I'm not saying this is your problem, just giving an example of how one type of bug can lead to issues elsewhere.

[KickinAezz]

The solution is to debug your code. The last time I helped someone find a bug, it was a string overflow (writing too many characters into a fixed size string). Bugs like that will cause all kinds of interesting problems. I'm not saying this is your problem, just giving an example of how one type of bug can lead to issues elsewhere.

Oh, someone! I'll do my best in debugging.

Is it recommended that I invalidate Instruction cache every loop?

sceKernelIcacheInvalidateall(); ????

[Jim]

No, you don't need to do it at all unless you are syncing with some other piece of hardware, such as the GU or ME.

Jim