@balaton

I've installed Fedora 12 ppc under windows and I've tried gcc with -pg and gprof and everything is working correctly.
But as I told you, I don't know how it works under linux. If it is using a timer, an interrupt or something else. I can only say that on linux (using windows) it is working.

@balaton

I've just tried Fedora 12 with qemu ppc and profiling is working correctly. I don't know how it works on linux the profiling but at least the CPU should be configured correctly.
So most probably is something related to interrupts

@balaton

I don't know the internal details because I don't have access to the kernel source code. I can only try to explain to you how the code is working now:

https://github.com/afxgroup/clib2/blob ... /library/profile/profil.c

On line 110 you have:
IPM->SetInterruptVector(1, &CounterInt);

That is enabling that interrupt (CounterInt is an Interrupt) and that should call CounterIntFn function.
But while on a real pegasos2 that function is called, on qemu is never called. So I suppose that interrupt is never enabled

If you tell me how, I could try to "grep" something useful for you.

On real peg2 everything is working correctly. It seems that an interrupt is never called and so profiling doesn't work correctly. Or better. Time is always zero. While function count is working correctly. I don't know if you need more informations. However I don't know how performance monitor works internally and which registers are used. But I'm almost sure that interrupt is never executed

I've tried to use gprof with newlib under qemu and pegasos2. But while functions are shown the time is always 0.00. Of course i've used the -Ttext=0x00000000 to compile the exe file to avoid the problem we have on OS4.
Is it possible to use profiling under qemu? Do I have to use specific flag? I've also tried 7447 and apollo7 cpu but with same result

@joerg

https://people.irisa.fr/Anthony.Baire/ ... for-beginners-handout.pdf

@joerg

newlib doesn't use the same clib2 mechanism for constructors. newlib doesn't have CONSTRUCTOR defines that clib2 (original too) has. If you try to change the clib2 crtbegin/end files trying to change .ctors/.dtors to ~0/1 nothing will work with current gcc (don't know the ancient version of 20 years ago). If you think that this can be fixed of course feel free to send a pull request
That's the beauty of open source

@LiveForIt

Libraries are always closed correctly inside clib2 function that call main because exit will jump into longjmp that will call the routine that will close the libraries, call the destructors and so on.
Why setup and use pr_ExitCode and pr_FinalCode?
I'll fix the closing routine that will set the pointers to NULL.

Of course. But who is paying in Amiga Land? Should be better to release sources so any other developer can learn something? We are a small community. Closing the sources is the worst thing can happen

Closed source software is less frequent nowadays. Except maybe for games. That however doesn't have this problem because almost all of them is created through external platforms. I know that on Amiga platform the people think that closed source is still the best solution. But fortunately 99.9% of the world doesn't. And if wasn't for newlib, owb and all other libraries most probably we where stuck on a closed source browser with iff support and nothing else... Same for morphos an all other minor platforms. Android included..

I think that new clib2 has really few bugs compared the current newlib. The only part that is slower (not bugged) it is the I/O because the way was originally written. I've made hundreds of tests against Linux for old and new functions and at moment no bugs are popped out. Not speaking about the cleaning. Linux code (newlib one) is a mess. Clib2 code is way better than that.
Not only, but the most important thing, is that if a bug is found anyone can fix it. And of course this isn't valid for any OS4 version of newlib. Past, present and most probably future.

Do hieronimous use performance.monitor? I remember it was using sampling in the past

We are trying to fix everything. More people help us, faster we can have a working gcc suite..

Thanks mate! Replied!

I'm trying to run an e500v2 machine. Not Mc, not M1. A v2 because I need to test math library under linux
I've also found a powerpc spe freebsd version but same result

Well, don't you think i've tried lot of distros without results?
I've donwloaded ubuntu-14.04-server-powerpc.iso, extracted vmlinux and initrd.gz, downloaded a file called wirtio-win.iso from a fedora web site.. and then nothing happens when I run the command. qemu seems stuck.
Not only. Not all distro has the same kernel file. For example freebsd has a file called KERNEL but not initrd.gz.
That's why I'm asking for something ready that is working for someone. Is it so hard to share two files?

I need your files because I don't have any of them. No kernel, no initrd and so on. I have only the iso files and of course you don't have to give me them. Just the minimum to make the machine work correctly like yours

Well,
Can you please send me everything?
I mean vmlinux and so on?

Thanks

@MartinW

https://gist.github.com/afxgroup/864bf7d87e96ee0e2be25295f52a4a86

Of course it is a dirty toolchain file i'm using. And sometime I play with it but it should work for any project

Did you see -L/opt/code/ogles2/local/newlib/lib?
You are pointing to newlib lib folder and not clib2 one and you have the error:

/opt/code/ogles2/local/newlib/lib/libogles2.a(Clear.o): In function `glClear':

That show you are using newlib version

Great. Of course if you fix problems or add something feel free to push your changes