From: Sebastian A. <sp...@sy...> - 2007-01-18 18:49:59
|
Kristian, Thanks for the feedback and links. I have switched to i586 and Generic i586 device for now and it seems to be going OK, except that I ran out of disk space at one point and I was missing uudecode. I had to basically start again after the disk space problem: cp -r dl dl-bak ; make clean ; make distclean ; rm -f -r toolchain_build_i586 ; svn update ; make If I didn't delete toolchain_build_i586 it would just immediately bomb again. ftp://ftp.samba.org/ isn't working today but I found the file at http://samba.org/ftp/ppp/ppp-2.4.3.tar.gz (btw, there is a newer version available too). I copied the file into the dl directory manually, but it is still retrying via FTP (try: 4) and the timeouts are extremely long. Is there anyway to cancel the getter_script.sh without cancelling the make? The server has spent so long getting to this point I don't want to have to start it again. Having said that, I would be willing to try VIA with i586. Is this likely to work then? What are the differences between Generic i586 and VIA builds? Presumably it will overwrite build_i586/root so I'd want to back that up... I saw your svn changes. :) Is trunk nearly ready for a release? Or more importantly for me right now, should it be stable? Thanks for your help. This is my first time trying to compile AstLinux. I need to add support, i.e. drivers, for a device that is not supported in the standard build. How do I go about doing this? Presumably I can do it after the main make has finished... Sebastian -----Original Message----- From: ast...@li... [mailto:ast...@li...] On Behalf Of Kristian Kielhofner Sent: 18 January 2007 15:52 To: AstLinux Users Mailing List Subject: Re: [Astlinux-users] Compile problems with i686 as the TargetArchitecture Variant On 1/18/07, Sebastian Auriol <sp...@sy...> wrote: > > > Well trying to compile with i386/i686/Generic i586 doesn't work either (on > 0.4 and trunk), with the same problems, so it looks like the i686 variant > may be the problem. I'm going to try again now using i586 as the variant > for 0.4 branch and trunk... > > Sebastian > > Sebastian, Thank you for your comments. A couple of things: 1) Almost all VIA processors have known problems when running with "i686" compiled code: http://lists.digium.com/pipermail/asterisk-dev/2003-July/001194.html http://www.voip-info.org/wiki/view/Asterisk+hardware+mini-itx There are other examples all over the place. On the VIA, stick with i586. 2) In general, if you select a specific device under AstLinux target devices, you really shouldn't just change the processor architecture. If you want to do this, create a new target device that will select your intended target CPU. Those devices are there (and work with 586) because they are a specific piece of hardware with a known configuration. 3) "make clean" is broken all over the place. Myself and the other devs know this. I have submitted a patch to both 0.4 and trunk to fix the few problems that you pointed out but there are certainly other problems with make clean. 4) These days, almost all development is done in trunk. 0.4 isn't updated and fixed nearly as often (your mISDN CVS problem, etc). -- Kristian Kielhofner |
From: olivier.taylor <oli...@gm...> - 2007-01-18 19:22:19
|
<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Sebastian,<br> <br> I have an HP 5500 thin client (Via chip) running perfectly here with the I586 build, just Fyi.<br> <br> Olivier<br> <br> Sebastian Auriol a écrit : <blockquote cite="mid!~!UENERkVCMDkAAQACAAAAAAAAAAAAAAAAAC4AAAAAAAAAvrTCuIsz1xGxNWiNbQAAAAEAUR1BVCzq0xGwsQCgzFqsgwAAAAGTZwAAEAAAADO2lXjA+91IvP7aWfG%2FAnIBAAAAAA==@syntec.co.uk" type="cite"> <pre wrap="">Kristian, Thanks for the feedback and links. I have switched to i586 and Generic i586 device for now and it seems to be going OK, except that I ran out of disk space at one point and I was missing uudecode. I had to basically start again after the disk space problem: cp -r dl dl-bak ; make clean ; make distclean ; rm -f -r toolchain_build_i586 ; svn update ; make If I didn't delete toolchain_build_i586 it would just immediately bomb again. <a class="moz-txt-link-freetext" href="ftp://ftp.samba.org/">ftp://ftp.samba.org/</a> isn't working today but I found the file at <a class="moz-txt-link-freetext" href="http://samba.org/ftp/ppp/ppp-2.4.3.tar.gz">http://samba.org/ftp/ppp/ppp-2.4.3.tar.gz</a> (btw, there is a newer version available too). I copied the file into the dl directory manually, but it is still retrying via FTP (try: 4) and the timeouts are extremely long. Is there anyway to cancel the getter_script.sh without cancelling the make? The server has spent so long getting to this point I don't want to have to start it again. Having said that, I would be willing to try VIA with i586. Is this likely to work then? What are the differences between Generic i586 and VIA builds? Presumably it will overwrite build_i586/root so I'd want to back that up... I saw your svn changes. :) Is trunk nearly ready for a release? Or more importantly for me right now, should it be stable? Thanks for your help. This is my first time trying to compile AstLinux. I need to add support, i.e. drivers, for a device that is not supported in the standard build. How do I go about doing this? Presumably I can do it after the main make has finished... Sebastian -----Original Message----- From: <a class="moz-txt-link-abbreviated" href="mailto:ast...@li...">ast...@li...</a> [<a class="moz-txt-link-freetext" href="mailto:ast...@li...">mailto:ast...@li...</a>] On Behalf Of Kristian Kielhofner Sent: 18 January 2007 15:52 To: AstLinux Users Mailing List Subject: Re: [Astlinux-users] Compile problems with i686 as the TargetArchitecture Variant On 1/18/07, Sebastian Auriol <a class="moz-txt-link-rfc2396E" href="mailto:sp...@sy..."><sp...@sy...></a> wrote: </pre> <blockquote type="cite"> <pre wrap=""> Well trying to compile with i386/i686/Generic i586 doesn't work either (on 0.4 and trunk), with the same problems, so it looks like the i686 variant may be the problem. I'm going to try again now using i586 as the variant for 0.4 branch and trunk... Sebastian </pre> </blockquote> <pre wrap=""><!----> Sebastian, Thank you for your comments. A couple of things: 1) Almost all VIA processors have known problems when running with "i686" compiled code: <a class="moz-txt-link-freetext" href="http://lists.digium.com/pipermail/asterisk-dev/2003-July/001194.html">http://lists.digium.com/pipermail/asterisk-dev/2003-July/001194.html</a> <a class="moz-txt-link-freetext" href="http://www.voip-info.org/wiki/view/Asterisk+hardware+mini-itx">http://www.voip-info.org/wiki/view/Asterisk+hardware+mini-itx</a> There are other examples all over the place. On the VIA, stick with i586. 2) In general, if you select a specific device under AstLinux target devices, you really shouldn't just change the processor architecture. If you want to do this, create a new target device that will select your intended target CPU. Those devices are there (and work with 586) because they are a specific piece of hardware with a known configuration. 3) "make clean" is broken all over the place. Myself and the other devs know this. I have submitted a patch to both 0.4 and trunk to fix the few problems that you pointed out but there are certainly other problems with make clean. 4) These days, almost all development is done in trunk. 0.4 isn't updated and fixed nearly as often (your mISDN CVS problem, etc). </pre> </blockquote> </body> </html> |
From: Gary G. H. <GHe...@co...> - 2007-01-18 20:07:44
|
Most of the ones I am dealing with here are VIA C3 1GHz machines on a m'boards from LEX ... I use the standard i586 build and it seems to work fine ... I think the last time the question came up, I was told that there is no difference between the VIA and standard i586 build ... That was a while ago so I am not sure if anything has changed since ... I am not sure that the kernel in Astlinux takes advantage of the fancy random number generator in the VIA chip ... I understand that this function might improve performance on VIA machines when doing codec conversion or when doing encryption as in VPN ... But the machines have performed well enough for my scenario that I just don't worry about it ... I would be more concerned about making modifications to the build that might version updates difficult ... There are many on this list who report good results using 233MHz and 266MHz CPU's ... I would assume that the VIA machines with 800MHz to 1.5GHz CPU's should be able to handle most anything you might throw at them even without using any unique VIA features ... All that said, if you get SAMBA working under Astlinus, please take time to put together a recipe of how you got it to work ... I would be interested in playing with that ... Not sure I would want to expose a SAMBA mahine to the Internet though ... That scares me a bit ... Would have to tweak up the Arno firewall script to compensate for the security issues I suspect ... G.Hendershot |