Re: CVS commit: [matt-nb5-mips64] src/sys/arch/mips/mips

[Matt Thomas <jabbathespud%gmail.com@localhost>]

On Aug 24, 2009, at 9:09 AM, Jason Thorpe wrote:

> On Aug 23, 2009, at 9:04 PM, Matt Thomas wrote:
>
> > Module Name:    src
> > Committed By:   matt
> > Date:           Mon Aug 24 04:04:53 UTC 2009
> >
> > Modified Files:
> >         src/sys/arch/mips/mips [matt-nb5-mips64]: mipsX_subr.S
> >
> > Log Message:
> > SyscallCall exception didn't save the temporaries which libc no  
> > longer likes.
> > So now it saves them.
>
> This seems wrong... the syscall trap is like making a function call,  
> so the temporaries should be expected to be clobbered.  If the  
> syscall stubs don't like this, they should make use of callee-saved  
> registers.
>
> What am I missing?

That the syscall needs the arguments and stack pointer unchanged from  
the start of syscall stub. Thus if the syscall needs to store things,  
it's limited t using the temporary registers.  And  MIPS NewABI makes  
the global pointer callee-saved so it has to be saved somewhere.  And  
that place is a temporary register.  If the trap nukes the  
temporaries, that would be bad.