From mboxrd@z Thu Jan 1 00:00:00 1970 X-Msuck: nntp://news.gmane.io/gmane.comp.sysutils.supervision.general/631 Path: main.gmane.org!not-for-mail From: yoppo@jippii.fi Newsgroups: gmane.comp.sysutils.supervision.general Subject: Re: Cross compiling for VIA C3? Date: Wed, 24 Nov 2004 02:58:46 +0200 (EET) Message-ID: <29331726.1101257926558.JavaMail.yoppo@jippii.fi> NNTP-Posting-Host: deer.gmane.org Mime-Version: 1.0 Content-Type: text/plain; Charset=iso-8859-1; Format=Flowed Content-Transfer-Encoding: quoted-printable X-Trace: sea.gmane.org 1101257943 21442 80.91.229.6 (24 Nov 2004 00:59:03 GMT) X-Complaints-To: usenet@sea.gmane.org NNTP-Posting-Date: Wed, 24 Nov 2004 00:59:03 +0000 (UTC) Original-X-From: supervision-return-870-gcsg-supervision=m.gmane.org@list.skarnet.org Wed Nov 24 01:58:50 2004 Return-path: Original-Received: from antah.skarnet.org ([212.85.147.14] ident=qmailr) by deer.gmane.org with smtp (Exim 3.35 #1 (Debian)) id 1CWlUg-00056B-00 for ; Wed, 24 Nov 2004 01:58:50 +0100 Original-Received: (qmail 14291 invoked by uid 76); 24 Nov 2004 00:59:09 -0000 Mailing-List: contact supervision-help@list.skarnet.org; run by ezmlm List-Post: List-Help: List-Unsubscribe: List-Subscribe: List-Archive: Original-Received: (qmail 14286 invoked from network); 24 Nov 2004 00:59:09 -0000 Original-To: supervision@list.skarnet.org X-Mailer: Saunalahti webmail - http://saunalahti.fi X-Originating-IP: 213.139.166.93 Xref: main.gmane.org gmane.comp.sysutils.supervision.general:631 X-Report-Spam: http://spam.gmane.org/gmane.comp.sysutils.supervision.general:631 >>>> Is it possible to cross-compile runit for VIA C3 system? If yes then h= ow? >>>> If I compile it normally on my athlon machine it doesn't work at all i= n=20 >>> my NAT-machine (VIA). >>> afair via c3 lack of cmov, what produce 'some' problems when=20 >>> crosscompiling, what optimizations are you using? >> The latest try was with -O2 -march=3Di586 -mcpu=3Di586 don't know if Iam= =20 >> missing something, but it seems like it could not be cross compiled just= =20 >> changing the optimizations. >>=20 >> Now after studying the compile process it seems like it makes some kind= =20 >> of testing on compile time but I dont have any clues how to "hack" the= =20 >> compile process so it would work in the VIA C3. And because I have only= =20 > > Yes, the runit package's build process (just as the daemontools' one) > includes some tests to configure system dependencies. This includes > three runtime tests where a test program is compiled, linked, and then > run. These runtime tests obviously won't work when cross-compiling. > > More precisely it's the targets hasshsgr.h:, iopause.h:, and uint64.h:, > in the Makefile which test setgroups(), poll(), and the size of uint64, > to select *.h1 or *.h2 respectively. You can compile and link the test > programs, and then run them on the target device to check the exit code > and choose the correct header. > >> 64Mb Usb drive as a "hard drive" on that VIA I can't even test compiling= =20 >> the runit directly there to see if it works when not cross compiling and= =20 >> what the compile process does when compiling on that platform. > > A nfs mount could help. > > HTH, Gerrit. I compiled the package now on i586 machine and the runit works fine on=20 VIA. Too bad that I don't allways have this possibility so I still need=20 to get the cross compiling to work. I compared the choices which was=20 made during compiling and all of them are the same as on my athlon=20 machine so they can't be causing the problems. When I tried the "cross compiled" binaries only the runit-init seems to=20 fail with illegal instruction other binaries works. Right now I don't=20 have a clue what to try next... __ Hanki uusi t=E4ysin ilmainen netti.fi-s=E4hk=F6posti, jopa 10 Mt tilaa vies= teillesi! http://webmail.saunalahti.fi/