Date: Mon, 7 Jun 2004 16:10:12 -0700 From: Marcel Moolenaar <marcel@xcllnt.net> To: Doug Rabson <dfr@nlsystems.com> Cc: freebsd-current@freebsd.org Subject: Re: HEADS UP! KSE needs more attention Message-ID: <20040607231012.GB11313@ns1.xcllnt.net> In-Reply-To: <200406072242.13393.dfr@nlsystems.com> References: <Pine.GSO.4.10.10406061551210.16558-100000@pcnet5.pcnet.com> <1086625355.10911.39.camel@builder02.qubesoft.com> <20040607194237.GA10406@ns1.xcllnt.net> <200406072242.13393.dfr@nlsystems.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, Jun 07, 2004 at 10:42:13PM +0100, Doug Rabson wrote: > On Monday 07 June 2004 20:42, Marcel Moolenaar wrote: > > On Mon, Jun 07, 2004 at 05:22:35PM +0100, Doug Rabson wrote: > > > > > Actually its a bit better than that. It works for most use > > > > > cases right now on i386 but would get confused on dlclose. I'll > > > > > fix that before I move it into current. > > > > > > > > Does it work on static bound executables? > > > > > > Which one is static bound > > > > The executable; you know, no rtld. What I call complete executable to > > distinguish it from static TLS on my page. Does static TLS work? > > > > See also: http://wiki.daemon.li/index.pl?ThreadLocalStorage > > No, this one is not yet supported. I think I can deal with this inside > libc with some small support from the kernel (probably just to provide > details of the TLS segment size etc.) Ok, thanks. BTW, I was thinking along the same lines, although it looks from your description that I probably wanted to put more of the meat in the kernel to avoid making the startup code complex and possibly pessimizing non-TLS processes. Anyway: From my PoV, static TLS is not critical enough to force it in 5.3, but it is important enough to have soon after that. FYI, -- Marcel Moolenaar USPA: A-39004 marcel@xcllnt.net
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040607231012.GB11313>