From mboxrd@z Thu Jan  1 00:00:00 1970
From: mathieu.desnoyers@polymtl.ca (Mathieu Desnoyers)
Date: Tue, 18 Aug 2009 23:40:46 -0400
Subject: [ltt-dev] sparc: sys32.S incorrect compat-layer splice() system
	call
In-Reply-To: <20090818.201613.15171258.davem@davemloft.net>
References: <20090819025623.GA11677@Krystal>
	<20090818.201613.15171258.davem@davemloft.net>
Message-ID: <20090819034046.GA20959@Krystal>

* David Miller (davem at davemloft.net) wrote:
> From: Mathieu Desnoyers <mathieu.desnoyers@polymtl.ca>
> Date: Tue, 18 Aug 2009 22:56:23 -0400
> 
> > I think arch/sparc/kernel/sys32.S has an incorrect splice definition:
> > 
> > SIGN2(sys32_splice, sys_splice, %o0, %o1)
> > 
> > The splice() prototype looks like :
> > 
> >        long splice(int fd_in, loff_t *off_in, int fd_out,
> >                    loff_t *off_out, size_t len, unsigned int flags);
> > 
> > So I think we should have :
> > 
> > SIGN2(sys32_splice, sys_splice, %o0, %o2)
> > 
> > instead, am I correct ?
> 
> Indeed, that's correct, thanks for your fix.  I'll apply it.
> 
> > 
> > BTW, I can't figure out why we have %o5 in :
> > 
> > SIGN2(sys32_sync_file_range, compat_sync_file_range, %o0, %o5)
> > 
> > which takes only 4 arguments:
> > 
> >        int sync_file_range(int fd, off64_t offset, off64_t nbytes,
> >                   unsigned int flags);
> > 
> > maybe it has something to do with the return value ? Anyway it should
> > not hurt if it is unused.
> 
> It takes 4 arguments, but they are passed in 6 registers.  Each
> off64_t is passed in two 32-bit register parts.
> 

Thanks for the clarification. So the %o5 is there to sign-extend
"unsigned int flags" ?

Mathieu

-- 
Mathieu Desnoyers
OpenPGP key fingerprint: 8CD5 52C3 8E3C 4140 715F  BA06 3F25 A8FE 3BAE 9A68