This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: [PATCH] PPC linuxthreads tls.h fix
- From: Alan Modra <amodra at bigpond dot net dot au>
- To: Roland McGrath <roland at redhat dot com>
- Cc: Paul Mackerras <paulus at samba dot org>, sjmunroe at vnet dot ibm dot com, libc-alpha <libc-alpha at sources dot redhat dot com>, Franz-Sirl <Franz-Sirl at lauterbach dot com>
- Date: Fri, 28 Feb 2003 12:56:32 +1030
- Subject: Re: [PATCH] PPC linuxthreads tls.h fix
- References: <20030228015520.GH31111@bubble.sa.bigpond.net.au> <200302280206.h1S26ca01917@magilla.sf.frob.com>
On Thu, Feb 27, 2003 at 06:06:38PM -0800, Roland McGrath wrote:
> > I'm not at all fussed over what you do with the TCB. Since our tls
> > abi has the thread pointer relative to the end of the TCB, you can
> > change what's in TCB and its size, and binutils and gcc won't be
> > affected. Perhaps our ABI document shouldn't even mention the size
> > of the TCB.
>
> He is talking about the words the thread register points to, i.e. the
> word(s) after the DTV pointer. This is what the TLS document calls the TCB
> in "variant I" (TLS_DTV_AT_TP). Statically laid-out TLS offsets begin
> immediately after this.
Sure, I realize Paul is talking about the words after the DTV pointer.
And our thread register _doesn't_ point to the TCB, rather to the end
of TCB + 0x7000. Which is why the TCB size and contents are entirely
a glibc issue for powerpc.
--
Alan Modra
IBM OzLabs - Linux Technology Centre