SWI PSR corruption

[Stewart Brodie]

Christian Ludlam <chris at recoil.org> wrote:

> Hi,
> 
> When you build an application targeting 32 bit OSs, it seems you MUST use
> the -zpq512 compiler switch to tell the compiler that inline SWI calls
> corrupt the PSR (why this isn't the default I don't know).

I thought that was the one that stopped multiple calls to malloc being
optimised away, no?

> Of course, the -zpq command line isn't documented and is subject to
> change. What's the recommended workaround?

The -zpq option is an internal development feature of the compiler - it
doesn't require documentation - it shouldn't really be compiled in to public
releases, but it actually has been for quite a while now (the 32-bit betas
that Acorn released had it enabled)


-- 
Stewart Brodie