Re: More toolchain issues

["Michael L. Hitch" <mhitch%lightning.msu.montana.edu@localhost>]

On Wed, 16 Sep 2009, John Klos wrote:

> After about two and a half minutes of CPU time and after growing to more than 
> 60 megs, cc1plus stops and complains. Between this and the last problem, I'm 
> confused - isn't our unlimited datasize 128 megs? Why are these dying after 
> reaching 64 megs?
>
> I've tried Frank's libc with phk-allocator in place of jemalloc, and the 
> compilation (at least of scan_engine.o) is still running after two hours (60 
> MHz m68060), and the size went above 64 megs (up to 69, at least).
>
> So what's broken with jemalloc that doesn't allow more than 64 megs to be 
> used? Is this an m68k problem, or are different allocators used by different 
> ports?

  Can you tell (perhaps from pmap(1)) where the data memory is being 
allocated.  I'm guess that perhaps at least jemalloc allocates memory
by mmaping address space outside the 'reserved' space for a process.  On
the amiga, the user address space is limited to 224MB.  Out of that space,
there's 6MB for text, 128MB for data, and 32MB for stack.  The remainder
is used for shared libraries.  There's much less than 128MB in that space
that could be allocated via mmap.  I don't have my amigas running at the
momement to be able to check this.

  If this is the case, you can make more space available by building a 
kernel with USRSTACK (in the kernel config file or 
sys/arch/amiga/include/vmparam.h) set to 0x1c000000.  [You can go higher,
but that runs the risk of kernel crashes on 680[46]0 systems due to 
current limitations of the pmap.]

--
Michael L. Hitch                        mhitch%montana.edu@localhost
Computer Consultant
Information Technology Center
Montana State University        Bozeman, MT     USA