Date: Tue, 22 Jun 2010 00:40:48 -0600 (MDT) From: "M. Warner Losh" <imp@bsdimp.com> To: c.jayachandran@gmail.com Cc: freebsd-mips@freebsd.org Subject: Re: Merging 64 bit changes to -HEAD - part 3 Message-ID: <20100622.004048.624519315654756672.imp@bsdimp.com> In-Reply-To: <AANLkTikv__dQnu3iEMl-Of2xzJc6xoPSCrEpLz5GVg6R@mail.gmail.com> References: <AANLkTimQbw57AUGXd2A9Ql98VvtkInsdHX7jDor1rLwe@mail.gmail.com> <20100621.164056.10150326092187581.imp@bsdimp.com> <AANLkTikv__dQnu3iEMl-Of2xzJc6xoPSCrEpLz5GVg6R@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
In message: <AANLkTikv__dQnu3iEMl-Of2xzJc6xoPSCrEpLz5GVg6R@mail.gmail.c= om> "Jayachandran C." <c.jayachandran@gmail.com> writes: : On Tue, Jun 22, 2010 at 4:10 AM, M. Warner Losh <imp@bsdimp.com> wrot= e: : > The changes to asm.h change the ABI for all ABIs, no? =A0Does that : > matter? =A0The jumpbuf is a user-visible thing... =A0I'm guessing t= hat we : > can say it is OK, but it sure would be nice if we could MFC this : > change before 8.1-RELEASE (but given the locked-down nature of the : > tree, I'm not hopeful). : = : Currently, it should not disturb the o32 ABI, since it is inside an i= fdef. : In n32/n64 the jmpbuf entries are 64bit and we will use one more : register for GP. Good point. I missed that when I was looking at it. My bad.a : Slightly off-topic, _JBLEN is 95 for mips, I cannot see a reason for= : this value, any hints? Hmmm. 31 normal registers (we don't save $0 :). 32 floating point registers that are double size. 31 + 2 * 32 =3D=3D 95. Maybe that's why... There's at least one Mips that has CP1 registers for its specialized hardware. If we ever are going to support saving/restoring that state along with the core and FP registers, we will need to expand this... Warner
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100622.004048.624519315654756672.imp>