Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On Tuesday 09 June 2009, Jason Carson wrote: Greetings, I am trying to setup a wireless access point using the atheros kernel driver (Built into the kernel, not as a module). I am using Vanilla-Sources 2.6.29.4. I need my wireless network card to start up in master mode but for some reason it is starting up in managed mode. When wlan0 starts up I get this error message... * Bringing up interface wlan0 *configuring wireless network for wlan0 Error for wireless request Set Mode (8B06) : SET failed on device wlan0 ; invalid argument. *wlan0 connected to SSID MyNetwork *in managed mode (WEP Disabled) * null...[ ok ] then when hostapd starts up I get this error message... * Starting hostapd... Configuration file: /etc/hostapd/hostapd.conf Failed to set interface wlan0 to master mode. nl80211 driver initialization failed. ELOOP: remaining socket: sock=5 eloop_data=0x80f5a38 user_data=(nil) handler=0x8094b70 * start-stop-daemon: failed to start `/usr/sbin/hostapd' [ !! ] * ERROR: hostapd failed to start Here is my /etc/conf.d/net config_eth0=69.196.152.151 netmask 255.255.255.0 broadcast 69.196.152.255 config_eth1=null config_wlan0=null bridge_br0=eth1 wlan0 config_br0=192.168.0.1 netmask 255.255.255.0 broadcast 192.168.0.255 channel_wlan0=1 mode_wlan0=master essid_wlan0=MyNetwork Here is my hostapd.conf interface=wlan0 bridge=br0 driver=nl80211 ssid=MyNetwork hw_mode=g channel=1 macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 country_code=CA wpa=1 wpa_passphrase=passphrase wpa_key_mgmt=WPA-PSK wpa_pairwise=TKIP rsn_pairwise=CCMP Anyone know how to change my wireless card to master mode and make it so I don't get those error messages? Thanks P.S. I have everything working with madwifi and an older kernel so worst case scenario I stay with that configuration until I get this problem figured out. Should you also define in your /etc/conf.d/net the driver? modules=( wpa_supplicant ) wpa_supplicant_wlan0=-Dwext (but I'm not sure of the options because I have never set up the card as master with hostapd). HTH. -- Regards, Mick Maybe, I don't know. I thought wpa_supplicant was for clients not wireless access points. I am currently running the madwifi driver and an older kernel. I am not using wpa_supplicant and it works fine as a access point.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson wrote: Hey all, I think I might know the problem. After doing some reading apparently many distrutions, including Gentoo, don't ship hostapd with support for the nl80211 driver. I checked the Gentoo tarball and this is the case. I tried editing the defconfig file in hostapd.tar.gz (enabling CONFIG_DRIVER_NL80211=y)but when I opened the tarball, edited it, and created a new tarball then tried to emerge it it gave me this... penguin distfiles # emerge hostapd Calculating dependencies... done! Verifying ebuild manifests Emerging (1 of 1) net-wireless/hostapd-0.6.9 Refetching... File renamed to '/usr/portage/distfiles/hostapd-0.6.9.tar.gz._checksum_failure_.70f9W1' ...and then wanted to download hostapd again. Anyone know how I edit the source of the hostapd tarball so it will work? What it is doing is correct. The package has been changed and portage doesn't know if you did it or some random hacker. That said, man ebuild and check out the manifest section. I have never done this but maybe it will make sense to you. Also, I think emerge used to have the --digest option. Basically it tells emerge to skip checking the digest. I didn't see it in the man page so it may not be there anymore. Dale :-) :-)
Re: [gentoo-user] Finding orphaned libs
On Tuesday 09 June 2009 02:27:39 Dale wrote: That's sort of what I was thinking. It was generated when it was started up the first time. I also noticed some things that I installed in the Seamonkey directory too. Adblock was one of those. Tho there was a good many of them, I suspect most all of them belongs to something here. I may just leave well enough alone, may poke around a little but nothing risky. Yes, you do want to be very careful, especially any required files that ebuilds generate in post_install() - portage usually doesn't know about those. I'd stick to .so libs initially (revdep-rebuild can find your mistakes!). In my case, I had: liblcms.so.1.0.13 liblcms.so.1.0.18 and eic showed that there is no.13 version. So apparently, this comes from a time when I had a portage that didn't cleanly remove stuff. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson wrote: Hey all, I think I might know the problem. After doing some reading apparently many distrutions, including Gentoo, don't ship hostapd with support for the nl80211 driver. I checked the Gentoo tarball and this is the case. I tried editing the defconfig file in hostapd.tar.gz (enabling CONFIG_DRIVER_NL80211=y)but when I opened the tarball, edited it, and created a new tarball then tried to emerge it it gave me this... penguin distfiles # emerge hostapd Calculating dependencies... done! Verifying ebuild manifests Emerging (1 of 1) net-wireless/hostapd-0.6.9 Refetching... File renamed to '/usr/portage/distfiles/hostapd-0.6.9.tar.gz._checksum_failure_.70f9W1' ...and then wanted to download hostapd again. Anyone know how I edit the source of the hostapd tarball so it will work? What it is doing is correct. The package has been changed and portage doesn't know if you did it or some random hacker. That said, man ebuild and check out the manifest section. I have never done this but maybe it will make sense to you. Also, I think emerge used to have the --digest option. Basically it tells emerge to skip checking the digest. I didn't see it in the man page so it may not be there anymore. Dale :-) :-) H, I tried emerging hostapd normally and I noticed this... * Enabling Wi-Fi Protected Setup support * Enabling drivers: * HostAP driver enabled * Wired driver enabled * Prism54 driver enabled * Madwifi driver enabled * nl80211 driver enabled ...which says that nl80211 is enabled, yet the defconfig file in the tar ball shows this line commented out so I am confused as to how I fix my problem... #CONFIG_DRIVER_NL80211=y
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tue, 09 Jun 2009 07:16:32 +0300, Nikos Chantziaras wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? You can use RESTRICT=nomirror to prevent portage trying the mirrors at all. I don't know of a way of reversing the order in which things are tried. -- Neil Bothwick Those are my principles. If you don't like them I have others. signature.asc Description: PGP signature
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On Tue, 09 Jun 2009 01:33:35 -0500, Dale wrote: Also, I think emerge used to have the --digest option. Basically it tells emerge to skip checking the digest. I didn't see it in the man page so it may not be there anymore. It was dangerous, because it ignored digests for all packages emerged, not just the one. Instead use ebuild /path/to/ebuild manifest You should copy the ebuild to an overlay before doing this, or this will be undone at the next sync. -- Neil Bothwick God created the world in six days. On the seventh day he also decided to create England... just to try out his Practical Joke Weather Machine. signature.asc Description: PGP signature
Re: [gentoo-user] slim login manager issues
Sebastian Günther wrote: * Laurent Lejeune (olo...@gmail.com) [08.06.09 19:19]: Hi all, I'm having some issues with the slim login manager, which suits my very basic needs very well, except from those two things: * When starting, the keyboard doesn't respond quite well. The characters don't show on screen imediately, and sometimes, it gets even worse: it freezes for maybe 1-2 seconds, then the first key i hit gets stuck, which basically means it repeats that caracter a fine dozen of times. I have that problem, too. But upstream has an even bigger problem: Maintainer needed. * when i have to restart x, slim doesn't restarts itself up, i'm thrown back to basic console login prompt. https://bugs.gentoo.org/show_bug.cgi?id=261359 Any help greatly appreciated. cya. I think until there is a new maintainer, nothing of this will get fixed anytime soon. :-( Sebastian I noticed that by changing the theme, the keyboard freeze issue disappeared. As for the x restart sending me back to console login prompt, i still have to investigate that.
Re: [gentoo-user] Finding orphaned libs
Willie Wong ww...@math.princeton.edu wrote: qfile --orphans $(find -H /usr/lib /lib -type f) which avoids checking all the symlinks. In case that you find implementation works correctly, this will not change anything unless either /usr/lib or /lib are symlinks. If you however use find -L, -type f will find even all symlinks in case they are pointing to a file. In negative logic: find -L /usr/lib -type l will find all symlinks in /usr/lib that do not point anywhere. BTW, though, $(find -H ...) will have too big a list. Perhaps better to do find -H /usr/lib /lib -type f | xargs qfile -o This is why find has -exec ... {} + since 19 years. xargs will have problems in case that a file name contains spaces, tabs or newlines. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson schrieb: Greetings, I am trying to setup a wireless access point using the atheros kernel driver (Built into the kernel, not as a module). I am using Vanilla-Sources 2.6.29.4. I need my wireless network card to start up in master mode but for some reason it is starting up in managed mode. When wlan0 starts up I get this error message... * Bringing up interface wlan0 *configuring wireless network for wlan0 Error for wireless request Set Mode (8B06) : SET failed on device wlan0 ; invalid argument. *wlan0 connected to SSID MyNetwork *in managed mode (WEP Disabled) * null...[ ok ] then when hostapd starts up I get this error message... * Starting hostapd... Configuration file: /etc/hostapd/hostapd.conf Failed to set interface wlan0 to master mode. nl80211 driver initialization failed. ELOOP: remaining socket: sock=5 eloop_data=0x80f5a38 user_data=(nil) handler=0x8094b70 * start-stop-daemon: failed to start `/usr/sbin/hostapd' [ !! ] * ERROR: hostapd failed to start Here is my /etc/conf.d/net config_eth0=69.196.152.151 netmask 255.255.255.0 broadcast 69.196.152.255 config_eth1=null config_wlan0=null bridge_br0=eth1 wlan0 config_br0=192.168.0.1 netmask 255.255.255.0 broadcast 192.168.0.255 channel_wlan0=1 mode_wlan0=master essid_wlan0=MyNetwork Here is my hostapd.conf interface=wlan0 bridge=br0 driver=nl80211 ssid=MyNetwork hw_mode=g channel=1 macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 country_code=CA wpa=1 wpa_passphrase=passphrase wpa_key_mgmt=WPA-PSK wpa_pairwise=TKIP rsn_pairwise=CCMP Anyone know how to change my wireless card to master mode and make it so I don't get those error messages? Thanks P.S. I have everything working with madwifi and an older kernel so worst case scenario I stay with that configuration until I get this problem figured out. Hello, do _NOT_ initialize the master mode of your nic with the rc-script. Let hostapd do that. rc-script will fail! So your /etc/conf.d/net would look like this: config_eth0=69.196.152.151 netmask 255.255.255.0 broadcast 69.196.152.255 config_eth1=null config_wlan0=null bridge_br0=eth1 wlan0 config_br0=192.168.0.1 netmask 255.255.255.0 broadcast 192.168.0.255 Regards Norman
Re: [gentoo-user] KDE menu missing: solved
090608 Philip Webb wrote: 090608 Mick wrote: I seem to have installed kde-base/mimelib, I guess as dependency: eix kde-base/mimelib Installed versions: 3.5.10(3.5)(20:11:53 06/07/09)(-debug -elibc_FreeBSD) I also tried changing fluxbox for kdm (in case this is a fluxbox WM issue) and restarted X with kdm. Ran kbuildsycoca, but same error again. I always 'startx' from a raw terminal. I recompiled kcminit kfind kpersonalizer kreadconfig rebooted, but there is no change in Krusader's behaviour. After the usual night's sleep to clear the cerebral tmp files I thought of starting the new day in KDE to try fixing things there. I swapped to my KDE version of ~/.xinitrc then did 'startx' yes the desktop menus the panel menu had vanished. Then I started Krusader, got the 'open with' box for a text file told it to open the file with Leafpad: it did some updating obeyed. After that, the other default 'open with' apps worked as before the KDE menus had been restored. To get Apwal back with L-mouse I had to use the KDE Control Centre 'desktop behaviour' re-apply it. So everything is back to normal Krusader behaves properly on Fluxbox too. HTH anyone else. -- ,, SUPPORT ___//___, Philip Webb ELECTRIC /] [] [] [] [] []| Cities Centre, University of Toronto TRANSIT`-O--O---' purslowatchassdotutorontodotca
Re: [gentoo-user] Finding orphaned libs
On Tue, 09 Jun 2009 12:05:05 +0200, Joerg Schilling wrote: qfile --orphans $(find -H /usr/lib /lib -type f) which avoids checking all the symlinks. In case that you find implementation works correctly, this will not change anything unless either /usr/lib or /lib are symlinks. If you however use find -L, -type f will find even all symlinks in case they are pointing to a file. The idea is not have find NOT follow symlinks, except when the given path is a symlink (which it often is on a 64 bit machine). Checking the symlinks is a waste of time, it's faster to remove only real files that are orphaned, then remove any dangling symlinks after. Otherwise qfile has to check about three times as many objects. -- Neil Bothwick Handy Guide to Modern Science: 1. If it's green or it wiggles, it's biology. 2. If it stinks, it's chemistry. 3. If it doesn't work, it's physics. signature.asc Description: PGP signature
Re: [gentoo-user] KDE menu missing: solved
2009/6/9 Philip Webb purs...@ca.inter.net: After the usual night's sleep to clear the cerebral tmp files I thought of starting the new day in KDE to try fixing things there. I swapped to my KDE version of ~/.xinitrc then did 'startx' yes the desktop menus the panel menu had vanished. Then I started Krusader, got the 'open with' box for a text file told it to open the file with Leafpad: it did some updating obeyed. After that, the other default 'open with' apps worked as before the KDE menus had been restored. To get Apwal back with L-mouse I had to use the KDE Control Centre 'desktop behaviour' re-apply it. So everything is back to normal Krusader behaves properly on Fluxbox too. Thanks Phillip. Did you get any apps in the 'Open With' menu list? Mine is empty. To get any application to open a file I have to type it manually (or choose one from the cached drop down entries in the Open With panel). I ran kate %U, but nothing much happened. :( Will try to repeat all this in a KDE session just in case. Meanwhile, what does your /home/user/.config/menus/applications.menu look like. Mine is empty and comes up with the error shown below: $ kbuildsycoca --menutest Warning: kbuildsycoca is unable to register with DCOP. kbuildsycoca running... kbuildsycoca running... Reusing existing ksycoca kbuildsycoca: WARNING: Parse error in /home/michael/.config/menus/applications.menu, line 1, col 1: unexpected end of file -- Regards, Mick
Re: [gentoo-user] KDE menu missing: solved
2009/6/9 Mick michaelkintz...@gmail.com: 2009/6/9 Philip Webb purs...@ca.inter.net: After the usual night's sleep to clear the cerebral tmp files I thought of starting the new day in KDE to try fixing things there. I swapped to my KDE version of ~/.xinitrc then did 'startx' yes the desktop menus the panel menu had vanished. Then I started Krusader, got the 'open with' box for a text file told it to open the file with Leafpad: it did some updating obeyed. After that, the other default 'open with' apps worked as before the KDE menus had been restored. To get Apwal back with L-mouse I had to use the KDE Control Centre 'desktop behaviour' re-apply it. So everything is back to normal Krusader behaves properly on Fluxbox too. Thanks Phillip. Did you get any apps in the 'Open With' menu list? Mine is empty. To get any application to open a file I have to type it manually (or choose one from the cached drop down entries in the Open With panel). I ran kate %U, but nothing much happened. :( Will try to repeat all this in a KDE session just in case. Meanwhile, what does your /home/user/.config/menus/applications.menu look like. Mine is empty and comes up with the error shown below: $ kbuildsycoca --menutest Warning: kbuildsycoca is unable to register with DCOP. kbuildsycoca running... kbuildsycoca running... Reusing existing ksycoca kbuildsycoca: WARNING: Parse error in /home/michael/.config/menus/applications.menu, line 1, col 1: unexpected end of file Fixed it! I had to delete /home/michael/.config/menus/applications.menu, then change my setup to login using kdm/KDE instead xdm/fluxbox and this time kbuildsycoca worked. Phew! All my menus are back, Kleopatra works fine, and kcontrol has a list on the LH side. Getting back to fluxbox works as normal again. Thanks for the pointer! :) It seems that the elog message was implying that one must be using KDE as a DE rather than individual applications. -- Regards, Mick
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 9 Jun 2009, at 08:15, Jason Carson wrote: H, I tried emerging hostapd normally and I noticed this... * Enabling Wi-Fi Protected Setup support * Enabling drivers: * HostAP driver enabled * Wired driver enabled * Prism54 driver enabled * Madwifi driver enabled * nl80211 driver enabled ...which says that nl80211 is enabled, yet the defconfig file in the tar ball shows this line commented out so I am confused as to how I fix my problem... #CONFIG_DRIVER_NL80211=y In my experience this notation will be used to suggest the default. I.E. if there's no entry for CONFIG_DRIVER_NL80211 (or it's commented out) then the option will be enabled by default. If the config file said #CONFIG_DRIVER_NL80211=n then I would tend to assume that the option will be disabled by default, and that this line needs uncommenting and changing in order to enable it. I can't say that everyone uses this convention, or even that I've ever read it stated, but 1) if I look at /etc/ssh/sshd_config, for example, all the options that I've set myself are lines I've had to uncomment AND change. 2) the compiler / build tool says it's building your module, anyway. Who cares what the config file says? Stroller.
Re: [gentoo-user] Finding orphaned libs
On Mon, 08 Jun 2009 19:27:39 -0500 Dale rdalek1...@gmail.com wrote: Also, be careful when you parse the output of the command. Most of the .pyc and .pyo files in the python2.5 directories are byte-compiled version that python generated dynamically the first time they are used. For example: /Numeric/numeric_version.py was installed by the ebuild and thuse qfile tells me it belongs to dev-python/numeric, but .../Numeric/numeric_version.pyc is listed as an orphan. While it is safe to delete, it will just be regenerated again later, wasting computing cycles. That's sort of what I was thinking. It was generated when it was started up the first time. I also noticed some things that I installed in the Seamonkey directory too. Adblock was one of those. That's not quite correct: .py[co] are generated by emerge right after package installaton and these won't come back unless you use these libs as root, since python won't have write access to these paths and will be byte-compiling each script on-the-fly. -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
Re: [gentoo-user] Trying SRC_URI first during fetching
On 9 Jun 2009, at 05:16, Nikos Chantziaras wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? This is undesirable behaviour - te mirrors exist because SRC_URI may often have limited bandwidth. The package's hosting may be donated to the software's author, for instance, by a 3rd party, so when you go directly to SRC_URI, avoiding mirrors, you wear out the author's welcome. Mirrors are hosted by people with gallons gallons of bandwidth to spare, who expect you to use it. It makes sense to use the mirrors FIRST. Stroller.
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tue, 9 Jun 2009 14:04:54 +0100, Stroller wrote: The package's hosting may be donated to the software's author, for instance, by a 3rd party, so when you go directly to SRC_URI, avoiding mirrors, you wear out the author's welcome. There's another factor that some projects will change the contents of a tarball without bumping the name. If you get the altered file from SRC_URI, emerge will bail out with a checksum failure. -- Neil Bothwick WITLAG: The delay between delivery and comprehension of a joke. signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
On Montag 08 Juni 2009, Neil Bothwick wrote: qfile --orphans $(find -H /usr/lib /lib -type f) qfile --orphans $(find -H /usr/lib /lib -type f) zsh: argument list too long: qfile bash energy ~ # qfile --orphans $(find -H /usr/lib /lib -type f) bash: /usr/bin/qfile: Argument list too long bash --version GNU bash, version 4.0.24(2)-release (x86_64-pc-linux-gnu) Copyright (C) 2009 Free Software Foundation, Inc. License GPLv3+: GNU GPL version 3 or later http://gnu.org/licenses/gpl.html This is free software; you are free to change and redistribute it. There is NO WARRANTY, to the extent permitted by law. qfile -V portage-utils-0.2: compiled on May 10 2009 $Id: qfile.c,v 1.45 2007/05/24 14:47:18 solar Exp $ file written for Gentoo by solar and vapier @ gentoo.org
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tuesday 09 June 2009 15:04:54 Stroller wrote: On 9 Jun 2009, at 05:16, Nikos Chantziaras wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? This is undesirable behaviour - te mirrors exist because SRC_URI may often have limited bandwidth. The package's hosting may be donated to the software's author, for instance, by a 3rd party, so when you go directly to SRC_URI, avoiding mirrors, you wear out the author's welcome. Mirrors are hosted by people with gallons gallons of bandwidth to spare, who expect you to use it. It makes sense to use the mirrors FIRST. Definitely. I have to beat users over the head (metaphorically) with a stick to get them to use my mirror. They somehow have the idea that SRC_URI has better quality bits than my ftp server... By going to SRC_URI every time, they use up precious international bandwidth instead of local (of which there is heaps). Every six months, when Fedora or Ubuntu does a release, those users can saturate the entire pipe into this *country* - just to get isos that I already have publicly available and am begging them to use. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Trying SRC_URI first during fetching
On Dienstag 09 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 15:04:54 Stroller wrote: On 9 Jun 2009, at 05:16, Nikos Chantziaras wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? This is undesirable behaviour - te mirrors exist because SRC_URI may often have limited bandwidth. The package's hosting may be donated to the software's author, for instance, by a 3rd party, so when you go directly to SRC_URI, avoiding mirrors, you wear out the author's welcome. Mirrors are hosted by people with gallons gallons of bandwidth to spare, who expect you to use it. It makes sense to use the mirrors FIRST. Definitely. I have to beat users over the head (metaphorically) with a stick to get them to use my mirror. They somehow have the idea that SRC_URI has better quality bits than my ftp server... By going to SRC_URI every time, they use up precious international bandwidth instead of local (of which there is heaps). Every six months, when Fedora or Ubuntu does a release, those users can saturate the entire pipe into this *country* - just to get isos that I already have publicly available and am begging them to use. my university has a nice volume cap for all students. But everything downloaded from its own network - including the ftp servers is 'free' - only outside traffic counts. Luckily, my university hosts a major gentoo mirror. Not rsync, but distfiles. They also have ubuntu, suse, fedora stuff. Windows updates.. and still people don't use it. Annoying.
Re: [gentoo-user] Finding orphaned libs
Volker Armin Hemmann volkerar...@googlemail.com wrote: On Montag 08 Juni 2009, Neil Bothwick wrote: qfile --orphans $(find -H /usr/lib /lib -type f) qfile --orphans $(find -H /usr/lib /lib -type f) zsh: argument list too long: qfile This is not a zsh limitation but a Linux limitation. You may have longer arg lists if you use Solaris ;-) You may avoid the problem with find . -exec prog args {} + Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tuesday 09 June 2009 15:48:26 Volker Armin Hemmann wrote: On Dienstag 09 Juni 2009, Alan McKinnon wrote: Definitely. I have to beat users over the head (metaphorically) with a stick to get them to use my mirror. They somehow have the idea that SRC_URI has better quality bits than my ftp server... By going to SRC_URI every time, they use up precious international bandwidth instead of local (of which there is heaps). Every six months, when Fedora or Ubuntu does a release, those users can saturate the entire pipe into this *country* - just to get isos that I already have publicly available and am begging them to use. my university has a nice volume cap for all students. But everything downloaded from its own network - including the ftp servers is 'free' - only outside traffic counts. Luckily, my university hosts a major gentoo mirror. Not rsync, but distfiles. They also have ubuntu, suse, fedora stuff. Windows updates.. and still people don't use it. Annoying. I'm six hours behind every mirror I sync - all major distros, every free BSD I can find and every major project out there; the only thing that lags is Ubuntu and Fedora at release time. And give it away at local prices over ftp, http, rsync Why? The company has 2000 employees. The international bandwidth bill is larger than the salary bill - including bonuses, expense claims, subsidies... So management is very very happy that I have a way to reduce that, and rather unhappy that users don't use it more -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Finding orphaned libs
On Dienstag 09 Juni 2009, Joerg Schilling wrote: Volker Armin Hemmann volkerar...@googlemail.com wrote: On Montag 08 Juni 2009, Neil Bothwick wrote: qfile --orphans $(find -H /usr/lib /lib -type f) qfile --orphans $(find -H /usr/lib /lib -type f) zsh: argument list too long: qfile This is not a zsh limitation but a Linux limitation. You may have longer arg lists if you use Solaris ;-) You may avoid the problem with find . -exec prog args {} + ah, thank you - I just copied from Neil (and had a look into the manpage which uses the same line more or less).
Re: [gentoo-user] Finding orphaned libs
Mickaël Bucas mbu...@gmail.com wrote: You may avoid the problem with find . -exec prog args {} + The right way to handle any size for the list returned by find, is by using xargs : find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans No, this is definitely wrong: the right way to handle this is execplus (since 19 years). Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] Finding orphaned libs
On Tue, 9 Jun 2009 16:04:41 +0200, Volker Armin Hemmann wrote: This is not a zsh limitation but a Linux limitation. You may have longer arg lists if you use Solaris ;-) You may avoid the problem with find . -exec prog args {} + ah, thank you - I just copied from Neil (and had a look into the manpage which uses the same line more or less). I've only used that command with smaller directories, where the limitation doesn't occur. find -exec does the job but is rather slow. Piping through xargs is faster and the limitations Joerg mentions, like spaces in filenames, don't apply in the case of */lib. -- Neil Bothwick All general statements are false. signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
2009/6/9 Joerg Schilling joerg.schill...@fokus.fraunhofer.de: Mickaël Bucas mbu...@gmail.com wrote: You may avoid the problem with find . -exec prog args {} + The right way to handle any size for the list returned by find, is by using xargs : find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans No, this is definitely wrong: the right way to handle this is execplus (since 19 years). Jörg I think I was too assertive. I prefer xargs, and it's still handy for more elaborate command lines Mickaël Bucas
Re: [gentoo-user] Finding orphaned libs
On Tue, Jun 9, 2009 at 9:36 AM, Neil Bothwickn...@digimed.co.uk wrote: If it's been around 19 years, why doesn't Google know anything about it? What is it? from the manpage of find, it even compares it to the xargs method: -exec command {} + This variant of the -exec action runs the specified command on the selected files, but the command line is built by appending each selected file name at the end; the total number of invocations of the command will be much less than the number of matched files. The command line is built in much the same way that xargs builds its command lines. Only one instance of `{}' is allowed within the command. The command is executed in the starting directory.
Re: [gentoo-user] Finding orphaned libs
2009/6/9 Joerg Schilling joerg.schill...@fokus.fraunhofer.de Volker Armin Hemmann volkerar...@googlemail.com wrote: On Montag 08 Juni 2009, Neil Bothwick wrote: qfile --orphans $(find -H /usr/lib /lib -type f) qfile --orphans $(find -H /usr/lib /lib -type f) zsh: argument list too long: qfile This is not a zsh limitation but a Linux limitation. You may have longer arg lists if you use Solaris ;-) You may avoid the problem with find . -exec prog args {} + Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de (uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily The right way to handle any size for the list returned by find, is by using xargs : find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans xargs fills the command line up to the maximum size, and then create another command line. Mickaël Bucas
Re: [gentoo-user] Finding orphaned libs
On Tue, 09 Jun 2009 16:15:21 +0200, Joerg Schilling wrote: find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans No, this is definitely wrong: the right way to handle this is execplus (since 19 years). If it's been around 19 years, why doesn't Google know anything about it? What is it? -- Neil Bothwick The facts, although interesting, are irrelevant. signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
Neil Bothwick n...@digimed.co.uk wrote: On Tue, 9 Jun 2009 16:04:41 +0200, Volker Armin Hemmann wrote: This is not a zsh limitation but a Linux limitation. You may have longer arg lists if you use Solaris ;-) You may avoid the problem with find . -exec prog args {} + ah, thank you - I just copied from Neil (and had a look into the manpage which uses the same line more or less). I've only used that command with smaller directories, where the limitation doesn't occur. find -exec does the job but is rather slow. Piping through xargs is faster and the limitations Joerg mentions, like spaces in filenames, don't apply in the case of */lib. Please try to inform yourself about execplus in order to learn that this is the only unlimited was to do the job. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] Finding orphaned libs
On Tue, 9 Jun 2009 09:42:34 -0500, Paul Hartman wrote: -exec command {} + This variant of the -exec action runs the specified command on the selected files, but the command line is built by appending each selected file name at the end; the total number of invocations of the command will be much less than the number of matched files. The command line is built in much the same way that xargs builds its command lines. Only one instance of `{}' is allowed within the command. The command is executed in the starting directory. How does this handle quoted filenames? Does -exec command {} + do command file 1 file 2 file 3 or command file 1 file 2 file 3 -- Neil Bothwick All generalizations are false, including this one. signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
Neil Bothwick n...@digimed.co.uk wrote: On Tue, 9 Jun 2009 09:42:34 -0500, Paul Hartman wrote: -exec command {} + This variant of the -exec action runs the specified command on the selected files, but the command line is built by appending each selected file name at the end; the total number of invocations of the command will be much less than the number of matched files. The command line is built in much the same way that xargs builds its command lines. Only one instance of `{}' is allowed within the command. The command is executed in the starting directory. How does this handle quoted filenames? correctly. Jörg -- EMail:jo...@schily.isdn.cs.tu-berlin.de (home) Jörg Schilling D-13353 Berlin j...@cs.tu-berlin.de(uni) joerg.schill...@fokus.fraunhofer.de (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/private/ ftp://ftp.berlios.de/pub/schily
Re: [gentoo-user] Finding orphaned libs
On Tuesday 9 June 2009, 16:36, Neil Bothwick wrote: On Tue, 09 Jun 2009 16:15:21 +0200, Joerg Schilling wrote: find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans No, this is definitely wrong: the right way to handle this is execplus (since 19 years). If it's been around 19 years, why doesn't Google know anything about it? What is it? Well, google does not know everything :) Basically, using + instead of ; after -exec allows to run the specified command less times, each time with the highest possible number of arguments (instead of running it once per file, which is what happens with ;). And yes, that's been in POSIX for a long time now. Example: $ ls file1 file2 file3 file4 file5 $ find . -type f -exec sh -c 'echo number of arguments: $#' sh {} \; number of arguments: 1 number of arguments: 1 number of arguments: 1 number of arguments: 1 number of arguments: 1 $ find . -type f -exec sh -c 'echo number of arguments: $#' sh {} + number of arguments: 5 So when you have to run a command on a very big number of files, say 1000 or more, with ; you spawn 1000 processes, with + you span just one or two (well, depending on the maximum command line length on the system anyway). This is of course much less resource intensive. Basically, using -exec with + does what xargs does, but without the need to care for strange characters in file names (well, a bit simplified, but you get the idea).
Re: [gentoo-user] Finding orphaned libs
On Tuesday 9 June 2009, 16:52, Neil Bothwick wrote: On Tue, 9 Jun 2009 09:42:34 -0500, Paul Hartman wrote: -exec command {} + This variant of the -exec action runs the specified command on the selected files, but the command line is built by appending each selected file name at the end; the total number of invocations of the command will be much less than the number of matched files. The command line is built in much the same way that xargs builds its command lines. Only one instance of `{}' is allowed within the command. The command is executed in the starting directory. How does this handle quoted filenames? Does -exec command {} + do command file 1 file 2 file 3 or command file 1 file 2 file 3 The latter.
Re: [gentoo-user] Finding orphaned libs
On Tue, 9 Jun 2009 17:21:32 +0200, Etaoin Shrdlu wrote: command file 1 file 2 file 3 or command file 1 file 2 file 3 The latter. Thanks, and bye-bye xargs (not that I used it very often). -- Neil Bothwick Cross-country skiing is great in small countries. signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
On Tue, 09 Jun 2009 16:59:39 +0200, Joerg Schilling wrote: How does this handle quoted filenames? correctly. Are you part Vorlon ;-) -- Neil Bothwick The world is a tragedy to those who feel, but a comedy to those who think.(Horace Walpole) signature.asc Description: PGP signature
Re: [gentoo-user] Finding orphaned libs
On Tuesday 09 June 2009 17:20:49 Etaoin Shrdlu wrote: On Tuesday 9 June 2009, 16:36, Neil Bothwick wrote: On Tue, 09 Jun 2009 16:15:21 +0200, Joerg Schilling wrote: find -H /usr/lib /lib -type f | xargs -d'\n' qfile --orphans No, this is definitely wrong: the right way to handle this is execplus (since 19 years). If it's been around 19 years, why doesn't Google know anything about it? What is it? Well, google does not know everything :) Basically, using + instead of ; after -exec allows to run the specified command less times, each time with the highest possible number of arguments (instead of running it once per file, which is what happens with ;). And yes, that's been in POSIX for a long time now. Example: $ ls file1 file2 file3 file4 file5 $ find . -type f -exec sh -c 'echo number of arguments: $#' sh {} \; number of arguments: 1 number of arguments: 1 number of arguments: 1 number of arguments: 1 number of arguments: 1 $ find . -type f -exec sh -c 'echo number of arguments: $#' sh {} + number of arguments: 5 So when you have to run a command on a very big number of files, say 1000 or more, with ; you spawn 1000 processes, with + you span just one or two (well, depending on the maximum command line length on the system anyway). This is of course much less resource intensive. Some numbers. These are from memory, I ran these commands today while preparing a machine for an installation of a package. The directory tree at the starting point had about 5000 files, more than 80% with a UID not attached to an account: chown -R user:group * about 2 seconds find . -nouser -o -nogroup -exec chown user:group {} + about 30 seconds (wild guess) find . -nouser -o -nogroup -exec chown user:group {} \; I killed this one after 5 minutes and it was nowhere near complete Admittedly, this was on a vmware guest with a rather poor disk configuration, but it does illustrate that the naive find \; performs extremely poorly. chown on it's own is foolish as the whole point of the exercise was to find files meeting certain criteria, and there was definitely some that didn't. execplus is a fine middle ground giving the best possible bang for buck. -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Trying SRC_URI first during fetching
On 6/9/09, Nikos Chantziaras rea...@arcor.de wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? As others have pointed out, caches, proxies et cetera are often Your Friends. Thus, I wonder if there is something specific you are trying to achieve here? Maybe these wise men from odd countries (like Mr McKinnon's weird country on an unknown planet, with companies with actual clueful management! ;) ) can help you even more if you reveal a bit of the *why* behind the question. -- Arttu V.
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tuesday 09 June 2009 17:56:41 Arttu V. wrote: On 6/9/09, Nikos Chantziaras rea...@arcor.de wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? As others have pointed out, caches, proxies et cetera are often Your Friends. Thus, I wonder if there is something specific you are trying to achieve here? Maybe these wise men from odd countries (like Mr McKinnon's weird country on an unknown planet, with companies with actual clueful management! ;) ) We're a carrier-grade Telco/ISP in a third world country that pretends to be first world. Techies rule here :-) -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Trying SRC_URI first during fetching
On Dienstag 09 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 17:56:41 Arttu V. wrote: On 6/9/09, Nikos Chantziaras rea...@arcor.de wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? As others have pointed out, caches, proxies et cetera are often Your Friends. Thus, I wonder if there is something specific you are trying to achieve here? Maybe these wise men from odd countries (like Mr McKinnon's weird country on an unknown planet, with companies with actual clueful management! ;) ) We're a carrier-grade Telco/ISP in a third world country that pretends to be first world. Techies rule here :-) let me guess: South Africa?
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tuesday 09 June 2009 20:26:29 Volker Armin Hemmann wrote: On Dienstag 09 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 17:56:41 Arttu V. wrote: On 6/9/09, Nikos Chantziaras rea...@arcor.de wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? As others have pointed out, caches, proxies et cetera are often Your Friends. Thus, I wonder if there is something specific you are trying to achieve here? Maybe these wise men from odd countries (like Mr McKinnon's weird country on an unknown planet, with companies with actual clueful management! ;) ) We're a carrier-grade Telco/ISP in a third world country that pretends to be first world. Techies rule here :-) let me guess: South Africa? Correct first time :-) -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson schrieb: Greetings, I am trying to setup a wireless access point using the atheros kernel driver (Built into the kernel, not as a module). I am using Vanilla-Sources 2.6.29.4. I need my wireless network card to start up in master mode but for some reason it is starting up in managed mode. When wlan0 starts up I get this error message... * Bringing up interface wlan0 *configuring wireless network for wlan0 Error for wireless request Set Mode (8B06) : SET failed on device wlan0 ; invalid argument. *wlan0 connected to SSID MyNetwork *in managed mode (WEP Disabled) * null...[ ok ] then when hostapd starts up I get this error message... * Starting hostapd... Configuration file: /etc/hostapd/hostapd.conf Failed to set interface wlan0 to master mode. nl80211 driver initialization failed. ELOOP: remaining socket: sock=5 eloop_data=0x80f5a38 user_data=(nil) handler=0x8094b70 * start-stop-daemon: failed to start `/usr/sbin/hostapd' [ !! ] * ERROR: hostapd failed to start Here is my /etc/conf.d/net config_eth0=69.196.152.151 netmask 255.255.255.0 broadcast 69.196.152.255 config_eth1=null config_wlan0=null bridge_br0=eth1 wlan0 config_br0=192.168.0.1 netmask 255.255.255.0 broadcast 192.168.0.255 channel_wlan0=1 mode_wlan0=master essid_wlan0=MyNetwork Here is my hostapd.conf interface=wlan0 bridge=br0 driver=nl80211 ssid=MyNetwork hw_mode=g channel=1 macaddr_acl=0 auth_algs=1 ignore_broadcast_ssid=0 country_code=CA wpa=1 wpa_passphrase=passphrase wpa_key_mgmt=WPA-PSK wpa_pairwise=TKIP rsn_pairwise=CCMP Anyone know how to change my wireless card to master mode and make it so I don't get those error messages? Thanks P.S. I have everything working with madwifi and an older kernel so worst case scenario I stay with that configuration until I get this problem figured out. Hello, do _NOT_ initialize the master mode of your nic with the rc-script. Let hostapd do that. rc-script will fail! So your /etc/conf.d/net would look like this: config_eth0=69.196.152.151 netmask 255.255.255.0 broadcast 69.196.152.255 config_eth1=null config_wlan0=null bridge_br0=eth1 wlan0 config_br0=192.168.0.1 netmask 255.255.255.0 broadcast 192.168.0.255 Regards Norman Hey Norman, I tried changing my /etc/conf.d/net to what you suggested but it resulted in the following errors... *Bringing up interface wlan0 * Configuring wireless network for wlan0 * Scanning for access points *no access points found * Failed to configure wireless for wlan0 Then because wlan0 wouldn't start I got a bunch of errors like... * Error: cannot start named as net.wlan0 would not start * Error: cannot start sshd as net.wlan0 would not start * Error: cannot start apache2 as net.wlan0 would not start etc...
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 9 Jun 2009, at 08:15, Jason Carson wrote: H, I tried emerging hostapd normally and I noticed this... * Enabling Wi-Fi Protected Setup support * Enabling drivers: * HostAP driver enabled * Wired driver enabled * Prism54 driver enabled * Madwifi driver enabled * nl80211 driver enabled ...which says that nl80211 is enabled, yet the defconfig file in the tar ball shows this line commented out so I am confused as to how I fix my problem... #CONFIG_DRIVER_NL80211=y In my experience this notation will be used to suggest the default. I.E. if there's no entry for CONFIG_DRIVER_NL80211 (or it's commented out) then the option will be enabled by default. If the config file said #CONFIG_DRIVER_NL80211=n then I would tend to assume that the option will be disabled by default, and that this line needs uncommenting and changing in order to enable it. I can't say that everyone uses this convention, or even that I've ever read it stated, but 1) if I look at /etc/ssh/sshd_config, for example, all the options that I've set myself are lines I've had to uncomment AND change. 2) the compiler / build tool says it's building your module, anyway. Who cares what the config file says? Stroller. ok, I guess it is enabled then. I wonder why its not working.
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
On 9 Jun 2009, at 08:15, Jason Carson wrote: H, I tried emerging hostapd normally and I noticed this... * Enabling Wi-Fi Protected Setup support * Enabling drivers: * HostAP driver enabled * Wired driver enabled * Prism54 driver enabled * Madwifi driver enabled * nl80211 driver enabled ...which says that nl80211 is enabled, yet the defconfig file in the tar ball shows this line commented out so I am confused as to how I fix my problem... #CONFIG_DRIVER_NL80211=y In my experience this notation will be used to suggest the default. I.E. if there's no entry for CONFIG_DRIVER_NL80211 (or it's commented out) then the option will be enabled by default. If the config file said #CONFIG_DRIVER_NL80211=n then I would tend to assume that the option will be disabled by default, and that this line needs uncommenting and changing in order to enable it. I can't say that everyone uses this convention, or even that I've ever read it stated, but 1) if I look at /etc/ssh/sshd_config, for example, all the options that I've set myself are lines I've had to uncomment AND change. 2) the compiler / build tool says it's building your module, anyway. Who cares what the config file says? Stroller. ok, I guess it is enabled then. I wonder why its not working. I came across this... http://wireless.kernel.org/en/users/Documentation/hostapd?action=showredirect=RTFM-AP and it says... Now find this line: #CONFIG_DRIVER_NL80211=y and uncomment it by removing the '#' sign. Repeat for other settings that you may be interested in. The basic configuration, with only this line uncommented is enough to get hostapd up and running with WPA/WPA2 authentication and encryption. ...so I don't know what to believe anymore lol :-)
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tue, 9 Jun 2009 20:49:43 +0200, Alan McKinnon wrote: let me guess: South Africa? Correct first time :-) It's not hard to work out ;-) Received: from nazgul.localnet (196-210-153-123-rrdg-esr-2.dynamic.isadsl.co.za -- Neil Bothwick Remember the good old days, when CPU was singular? signature.asc Description: PGP signature
Re: [gentoo-user] Trying SRC_URI first during fetching
On Tuesday 09 June 2009 23:57:54 Neil Bothwick wrote: On Tue, 9 Jun 2009 20:49:43 +0200, Alan McKinnon wrote: let me guess: South Africa? Correct first time :-) It's not hard to work out ;-) Received: from nazgul.localnet (196-210-153-123-rrdg-esr-2.dynamic.isadsl.co.za Drat. Horrors. Now my secret is out :-) And on a completely different but related topic, herewith a puzzle: How long does it take to sync the brand new Fedora 11 release? How long should it take? -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Trying SRC_URI first during fetching
On Mittwoch 10 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 23:57:54 Neil Bothwick wrote: On Tue, 9 Jun 2009 20:49:43 +0200, Alan McKinnon wrote: let me guess: South Africa? Correct first time :-) It's not hard to work out ;-) Received: from nazgul.localnet (196-210-153-123-rrdg-esr-2.dynamic.isadsl.co.za Drat. Horrors. Now my secret is out :-) And on a completely different but related topic, herewith a puzzle: How long does it take to sync the brand new Fedora 11 release? two hours? How long should it take? 1 second?
Re: [gentoo-user] Trying SRC_URI first during fetching
On Wednesday 10 June 2009 00:17:43 Volker Armin Hemmann wrote: On Mittwoch 10 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 23:57:54 Neil Bothwick wrote: On Tue, 9 Jun 2009 20:49:43 +0200, Alan McKinnon wrote: let me guess: South Africa? Correct first time :-) It's not hard to work out ;-) Received: from nazgul.localnet (196-210-153-123-rrdg-esr-2.dynamic.isadsl.co.za Drat. Horrors. Now my secret is out :-) And on a completely different but related topic, herewith a puzzle: How long does it take to sync the brand new Fedora 11 release? two hours? 5 days, 9 hours, 27 minutes and counting How long should it take? 1 second? 200G shouldn't take more than a day. Part of that is a booboo on the Fedora master mirror (content was available, it went away, it came back). That's bandwidth constraints for you. Into Africa it gets even worse. Total bandwidth to Kenya is not even 1M. International companies get their mail over dialup with fetchmail. And let's not even mention Zimbabwe... -- alan dot mckinnon at gmail dot com
Re: [gentoo-user] Trying SRC_URI first during fetching
On Mittwoch 10 Juni 2009, Alan McKinnon wrote: On Wednesday 10 June 2009 00:17:43 Volker Armin Hemmann wrote: On Mittwoch 10 Juni 2009, Alan McKinnon wrote: On Tuesday 09 June 2009 23:57:54 Neil Bothwick wrote: On Tue, 9 Jun 2009 20:49:43 +0200, Alan McKinnon wrote: let me guess: South Africa? Correct first time :-) It's not hard to work out ;-) Received: from nazgul.localnet (196-210-153-123-rrdg-esr-2.dynamic.isadsl.co.za Drat. Horrors. Now my secret is out :-) And on a completely different but related topic, herewith a puzzle: How long does it take to sync the brand new Fedora 11 release? two hours? 5 days, 9 hours, 27 minutes and counting sweet. That is even worse than I imagined. 1 second? 200G shouldn't take more than a day. I was joking, but yeah, more than a day starts to stink. Small pipes and bad upstream are no a loveable, cuddly cute situation. Part of that is a booboo on the Fedora master mirror (content was available, it went away, it came back). sounds like real fun ... the fun you wish your enemy to have. That's bandwidth constraints for you. Into Africa it gets even worse. Total bandwidth to Kenya is not even 1M. International companies get their mail over dialup with fetchmail. that is indeed horrible. Nobody should be forced to use fetchmail. And let's not even mention Zimbabwe... I am surprised that Zimbabwe still exists to be honest. But for some reason that trainwreck still jerks around. Just like a headless chicken.
[gentoo-user] KVM and no keyboard at start of boot
Summary: What can I do to get my keyboard recognized (through a kvm switch) right at bootup. I mean like when the grub prompt comes up. Details: I've had this curious problem for some time now. Over at least several kernels but I think beginning with changing from one KVM to another a few mnths ago. Now using an IOMEGA 4prt `symphany'. All usb. I guess `Symphany' because you can switch speaker connections too. I rarely use sound so that isn't even a factor here. My keyboard (through kvm) is not recognized until bootup gets to the login prompt. Once there... no problems with keyboard. If I want to do anything early in boot process, like at grub prompt, I must keep a keyboard plugged in direct to machine. Its been a while since I've messed with it, but in the process of a new install and decided to take some time with that problem now. I've forgotten all the options I've tried enabling in the kernel but most recently just used the massive enablement of `genkernel all' thinking surely with all that junk enabled what ever is missing would be there. But it turned out not to be true. So how can I figure out what gets enabled by the time bootup reaches login prompt that is not enabled when grub screen comes up? Something finally allows the kvm pass through to be recognized but not until I've reached the login prompt.
Re: [gentoo-user] Trying SRC_URI first during fetching
On Wed, 10 Jun 2009 00:10:09 +0200, Alan McKinnon wrote: How long does it take to sync the brand new Fedora 11 release? Don't know. How long should it take? Don't care :) -- Neil Bothwick NOTICE: -- THE ELEVATORS WILL BE OUT OF ORDER TODAY -- (The nearest working elevators are in the building across the street.) signature.asc Description: PGP signature
[gentoo-user] Re: Trying SRC_URI first during fetching
Stroller wrote: On 9 Jun 2009, at 05:16, Nikos Chantziaras wrote: Normally, portage will try fetching from GENTOO_MIRRORS during an emerge and SRC_URI comes last. I want to reverse this; try SRC_URI first and if that fails, only then proceed to fetch from GENTOO_MIRRORS. Doable? This is undesirable behaviour - te mirrors exist because SRC_URI may often have limited bandwidth. Understandable, but my problem lies in ebuild digest of my own ebuilds and ebuilds of third-party overlays. Along the URLs tried by portage is one that seems to need over a minute to reply with 401 not found. This is highly annoying. I would want to try SRC_URI in those cases first. Of course that doesn't mean SRC_URI should be tried by default first. I never suggested that and look what happened to this thread :P
[gentoo-user] 64-bit kernel on a 32-bit installation
Hi, I lurk on the LKML, say hi once in awhile, ask a question once in awhile, and try to read at least the interesting to a non-programmer posts. I was curious about this one that came up today. Seems like this is a natural for Gentoo. I have a Gentoo 64-bit setup but have had lots of troubles over the years (far less now though) with web media and other things that need to be more Windows compatible. (I do audio work with my Gentoo boxes - interface to studios and a few bands, etc) I've found that my 32-bit Gentoo installations have been more compatible than 64-bit. Outside stuff like Java is better. In general when I have a problem I wonder if it's because I'm running 64-bit. How would one go about building a 64-bit kernel on a 32-bit machine with Gentoo? I presume that's mostly just how I configure the kernel, along with maybe some cross-compile options? Are there any projects going on in this area where I might become a test case? Wiki? Docs? Do others see value - getting 64-bit memory management, new CPU flags, etc., but keeping the apps 32-bit for compatibility? Take care, Mark SNIP On Tue, 9 Jun 2009, H. Peter Anvin wrote: A major problem is that distros don't seem to be willing to push 64-bit kernels for 32-bit distros. There are a number of good (and not-so-good) reasons why users may want to run a 32-bit userspace, but not running a 64-bit kernel on capable hardware is just problematic. Yeah, that's just stupid. A 64-bit kernel should work well with 32-bit tools, and while we've occasionally had compat issues (the intel gfx people used to claim that they needed to work with a 32-bit kernel because they cared about 32-bit tools), they aren't unfixable or even all _that_ common. And they'd be even less common if the whole 64-bit kernel even if you do a 32-bit distro was more common. The nice thing about a 64-bit kernel is that you should be able to build one even if you don't in general have all the 64-bit libraries. So you don't need a full 64-bit development environment, you just need a compiler that can generate code for both (and that should be the default on x86 these days). Linus SNIP
Re: [gentoo-user] Latest X on G3 PPC?
On Wed, 27 May 2009, Saphirus Sage wrote: I've attached my xorg from my G4 eMac, hopefully it will be of some help. Actually, that config works without any changes! I merged in the monitor and screens config from my old config and I have my desktop back again. The only difference as far as I can tell is that my old config specified at ATI device and yours specifies a Radeon device. Anyway, thanks very much! -- A
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson schrieb: Hey Norman, I tried changing my /etc/conf.d/net to what you suggested but it resulted in the following errors... *Bringing up interface wlan0 * Configuring wireless network for wlan0 * Scanning for access points *no access points found * Failed to configure wireless for wlan0 Then because wlan0 wouldn't start I got a bunch of errors like... * Error: cannot start named as net.wlan0 would not start * Error: cannot start sshd as net.wlan0 would not start * Error: cannot start apache2 as net.wlan0 would not start etc... This is still rc-script output... do not use the rc-script. Do not start it! Your errors would be, because you have to detach these services from wlan0 and let them use the bridge. Here is some output from my box: mimir ~ # rc-update -s -v | grep net local | default nonetwork net.br0 | default net.eth0 | default net.lo | boot netmount | default mimir ~ # ifconfig wlan0 wlan0 Protokoll:Ethernet Hardware Adresse 00:80:48:5e:57:3d inet6 Adresse: fe80::280:48ff:fe5e:573d/64 Gültigkeitsbereich:Verbindung UP BROADCAST RUNNING PROMISC MULTICAST MTU:1500 Metric:1 RX packets:48 errors:0 dropped:0 overruns:0 frame:0 TX packets:80 errors:0 dropped:0 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlänge:1000 RX bytes:7344 (7.1 KiB) TX bytes:14190 (13.8 KiB) mimir ~ # ls /etc/init.d/net.* /etc/init.d/net.br0 /etc/init.d/net.eth0 /etc/init.d/net.lo As you can see, there is no rc-script on wlan, but it is initialized an running in AP mode. Regards Norman
Re: [gentoo-user] Latest X on G3 PPC?
On Tue, Jun 9, 2009 at 5:35 PM, Ajai Khattria...@bway.net wrote: On Wed, 27 May 2009, Saphirus Sage wrote: I've attached my xorg from my G4 eMac, hopefully it will be of some help. Actually, that config works without any changes! I merged in the monitor and screens config from my old config and I have my desktop back again. The only difference as far as I can tell is that my old config specified at ATI device and yours specifies a Radeon device. Anyway, thanks very much! -- A Humm...if it's that easy maybe I should try it just for test purposes? My need for X sort of went away on that machine, but I'd be interested to see if it fixed my Mac Mini problem. - Mark
[gentoo-user] Fresh install on reboot endless scrolling text
I've just completed first part of a fresh install. Mounted the new root and emerged a few things while still chrooted. It'd be difficult to list quickly since I can't boot it, but the highlights are: eix genkernel gentoolkit gentoolkit-dev gentoo-sources grub lynx ntp reiserfsprogs rsyslog vim vixie-cron And a number of dependancies pulled in from my use flags. I didn't want to start from scratch dinking around with kernel params so elected to go the `genkernel all' route, hoping that with that much junk being compiled... surely I'd get whatever I needed to start the process of building up this installation. I realize there are many who have lots of rants against genkernel but unless you think that is the core of the problem... please hold off on advice to build my own kenel. kernel = 2.6.29-r5 What I see on reboot: Normal booting appears to be going along then moments after dev is mounted followed by filesystems... a massive screen full of text begins scrolling by and never stops. Impossible to read any of the text and neither pause/break or scroll/lock keys have any effect. Appears to be a continuous wrapping line repeated. But it completely kills the boot process, and no further progress is possible. I really have no idea what to change when I boot off the install media.
Re: [gentoo-user] Fresh install on reboot endless scrolling text
On Tue, Jun 9, 2009 at 5:50 PM, Harry Putnamrea...@newsguy.com wrote: I've just completed first part of a fresh install. Mounted the new root and emerged a few things while still chrooted. It'd be difficult to list quickly since I can't boot it, but the highlights are: eix genkernel gentoolkit gentoolkit-dev gentoo-sources grub lynx ntp reiserfsprogs rsyslog vim vixie-cron And a number of dependancies pulled in from my use flags. I didn't want to start from scratch dinking around with kernel params so elected to go the `genkernel all' route, hoping that with that much junk being compiled... surely I'd get whatever I needed to start the process of building up this installation. I realize there are many who have lots of rants against genkernel but unless you think that is the core of the problem... please hold off on advice to build my own kenel. kernel = 2.6.29-r5 What I see on reboot: Normal booting appears to be going along then moments after dev is mounted followed by filesystems... a massive screen full of text begins scrolling by and never stops. Impossible to read any of the text and neither pause/break or scroll/lock keys have any effect. Appears to be a continuous wrapping line repeated. But it completely kills the boot process, and no further progress is possible. I really have no idea what to change when I boot off the install media. You emerged reiserfsprogs before the machine was up and running. Does this somehow imply that you are trying to boot from a Reiser partition and could that be part of the problem? Does the machine have a serial port and do you have access to a second machine that you could capture and save the early boot messages? That might be helpful. I know many machine now don't support that. There may be ways to do something similar over the network or USB if the boot gets that far. - Mark
Re: [gentoo-user] Fresh install on reboot endless scrolling text
On Jun 9, 2009, at 7:50 PM, Harry Putnam wrote: I've just completed first part of a fresh install. [snip] What I see on reboot: Normal booting appears to be going along then moments after dev is mounted followed by filesystems... a massive screen full of text begins scrolling by and never stops. Impossible to read any of the text and neither pause/break or scroll/lock keys have any effect. Appears to be a continuous wrapping line repeated. But it completely kills the boot process, and no further progress is possible. I really have no idea what to change when I boot off the install media. So you have grub installed, the grub menu appears, you select a boot option, boot starts, the usual kernel messages flash by, you get to the second part of the boot where the lines look like: blah... [OK] Then something goes crazy. Assuming that's the scenario, boot with your install media, enter the chroot, then take a peek at the logs in /var/log. Hopefully you can find a hint. HTH, Roy
[gentoo-user] Firefox 3.5 and Thunderbird 3 - Preferences dialog broken
I decided to check out Firefox 3.5_beta4-r1 and Thunderbird 3.0_beta2 in one of my machines (from the mozilla overlay). They both emerged fine without problems. They even run without problems (I'm sending this from TB 3) except that I can't change their configuration. In both Firefox and Thunderbird, Edit-Preferences brings up a big but totally empty dialog. I did rm -rf ~/.mozilla ~/.thunderbird so everything is clean, but still no go. Anyone else encountered this? The machine in question is ACCEPT_KEYWORDS=~amd64 so everything is up to date.
[gentoo-user] Re: Fresh install on reboot endless scrolling text
Mark Knecht markkne...@gmail.com writes: I really have no idea what to change when I boot off the install media. You emerged reiserfsprogs before the machine was up and running. Does this somehow imply that you are trying to boot from a Reiser partition and could that be part of the problem? Yes, as advised in the quick install guide. And yes I am booting off of a reiserfs and have been doing so for at least 2 yrs. I haven't seen this problem in that time. Does the machine have a serial port and do you have access to a second machine that you could capture and save the early boot messages? That might be helpful. I know many machine now don't support that. There may be ways to do something similar over the network or USB if the boot gets that far. No serial port
[gentoo-user] Re: Fresh install on reboot endless scrolling text
Roy Wright r...@wright.org writes: [...] I really have no idea what to change when I boot off the install media. So you have grub installed, the grub menu appears, you select a boot option, boot starts, the usual kernel messages flash by, you get to the second part of the boot where the lines look like: blah... [OK] Then something goes crazy. Yes, just the first few `blah [ok]' lines... maybe 3. Assuming that's the scenario, boot with your install media, enter the chroot, then take a peek at the logs in /var/log. Hopefully you can find a hint. In progress on you suggestion now. But no syslog is running in the chroot.(is it?) But still looking around. And just incase its something related to the newest kernel I've downgraded to 2.6.29-r2 and building from an old .config of 2.6.28. (not genkernel on this one)
Re: [gentoo-user] Re: Fresh install on reboot endless scrolling text
On Tue, Jun 9, 2009 at 6:48 PM, Harry Putnamrea...@newsguy.com wrote: Mark Knecht markkne...@gmail.com writes: I really have no idea what to change when I boot off the install media. You emerged reiserfsprogs before the machine was up and running. Does this somehow imply that you are trying to boot from a Reiser partition and could that be part of the problem? Yes, as advised in the quick install guide. And yes I am booting off of a reiserfs and have been doing so for at least 2 yrs. I haven't seen this problem in that time. I understand, and I'm not suggesting it was the problem. Just asking questions. Keep in mind that I ran xorg-server with the Intel driver for 4 years and the recent xorg-server/driver updates totally broke it. I couldn't wait for a fix (wife's computer - we know who really runs this place...) so I bought a new cad and moved on. Is this the same old ReiserFS that you've been running? It's not the new fangled versions is it? It sounded like the time your boot goes crazy might be when everything is getting mounted? Is that about the right time? What do you think the last message is that's good? Does the machine have a serial port and do you have access to a second machine that you could capture and save the early boot messages? That might be helpful. I know many machine now don't support that. There may be ways to do something similar over the network or USB if the boot gets that far. No serial port Yeah, I understand. I wonder what the kernel devs use these days for this issue? I know with my machine, which has a serial port, I was able to capture some stuff for Ingo Molnar that helped him out, but I couldn't do it today on 6 of the 7 machines in my house. Cheers, Mark
Re: [gentoo-user] Re: Fresh install on reboot endless scrolling text
On Tue, Jun 9, 2009 at 6:52 PM, Harry Putnamrea...@newsguy.com wrote: Roy Wright r...@wright.org writes: [...] I really have no idea what to change when I boot off the install media. So you have grub installed, the grub menu appears, you select a boot option, boot starts, the usual kernel messages flash by, you get to the second part of the boot where the lines look like: blah... [OK] Then something goes crazy. Yes, just the first few `blah [ok]' lines... maybe 3. Assuming that's the scenario, boot with your install media, enter the chroot, then take a peek at the logs in /var/log. Hopefully you can find a hint. In progress on you suggestion now. But no syslog is running in the chroot.(is it?) But still looking around. And just incase its something related to the newest kernel I've downgraded to 2.6.29-r2 and building from an old .config of 2.6.28. (not genkernel on this one) I think he was hoping that you would find something in your chroot/var/log/messages file or something like that. boot from the CD, mount the hard drive like you were doing an install, and then poke around and see if anything got written and is still there. Just hunting for clues. Do you have a digital camera? If so maybe it's fast enough to capture the screen while it's scrolling and make the message semi-readable even though they aren't to your eyes?
Re: [gentoo-user] KVM and no keyboard at start of boot
On Tue, 09 Jun 2009 18:03:52 -0500 Harry Putnam rea...@newsguy.com wrote: Summary: What can I do to get my keyboard recognized (through a kvm switch) right at bootup. I mean like when the grub prompt comes up. Details: I've had this curious problem for some time now. Over at least several kernels but I think beginning with changing from one KVM to another a few mnths ago. ... So how can I figure out what gets enabled by the time bootup reaches login prompt that is not enabled when grub screen comes up? Something finally allows the kvm pass through to be recognized but not until I've reached the login prompt. But linux kernel isn't loaded or used in any way when grub screen comes up - grub is loading it as a last step of it's execution, so any kernel configuration settings should be completely irrelevant here. Either that, or I completely missed the point. -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
[gentoo-user] Re: KVM and no keyboard at start of boot
Mike Kazantsev mk.frag...@gmail.com writes: Something finally allows the kvm pass through to be recognized but not until I've reached the login prompt. But linux kernel isn't loaded or used in any way when grub screen comes up - grub is loading it as a last step of it's execution, so any kernel configuration settings should be completely irrelevant here. Ha... now I remember someone saying that when I posted about this quite some time ago. Here is the thing though. There was a time when my KVM switch was recognized at the grub boot prompt. So something has changed. Are you suggesting something inside grub can be tweaked? I've mentioned that I changed KVM switches, but with the new one, it does work by the time the login prompt comes up. So if I could identify what it is in the kernel that allowes it to work at the point where the kernel takes over (login prompt), then maybe I could enable that aspect somehow inside an initramfs, and be able to have the KVM recognized at the grub prompt by booting with an initrd.
[gentoo-user] Re: Fresh install on reboot endless scrolling text
Mark Knecht markkne...@gmail.com writes: And just incase its something related to the newest kernel I've downgraded to 2.6.29-r2 and building from an old .config of 2.6.28. (not genkernel on this one) I think he was hoping that you would find something in your chroot/var/log/messages file or something like that. boot from the CD, mount the hard drive like you were doing an install, and then poke around and see if anything got written and is still there. Just hunting for clues. Do you have a digital camera? If so maybe it's fast enough to capture the screen while it's scrolling and make the message semi-readable even though they aren't to your eyes? Just reporting some kind of fix here. I downgraded to 2.6.29-r2 kernel from `r5' and built it from my old .config from a different system. (with a few changes in accord with the difference in the kernels [ 2.6.28 to 2.6.29-r2 ]. On reboot it just worked no endlessly scrolling text. So maybe its someting related to 2.6.29-r5 kernel or related to building that kernel with genkernel. At any rate, it seems a bit too opaque of a problem to continue trying to sort out, now that I got something working. Thanks for the tips posters.
Re: [gentoo-user] Re: Fresh install on reboot endless scrolling text
On Tue, 09 Jun 2009 20:52:26 -0500 Harry Putnam rea...@newsguy.com wrote: Roy Wright r...@wright.org writes: So you have grub installed, the grub menu appears, you select a boot option, boot starts, the usual kernel messages flash by, you get to the second part of the boot where the lines look like: blah... [OK] Then something goes crazy. Yes, just the first few `blah [ok]' lines... maybe 3. Okay, the syslog is probably not one of them, so /var/log shouldn't be too populated, except for rc.log, which should be enabled specifically in /etc/rc.conf (w/ baselayout-2) or /etc/conf.d/rc - look for rc_logger line or something similar with older baselayout. Then there's also interactive boot feature. Just press I as soon as you see first gentoo [ok] messages and the boot will proceed step-by-step asking about every init script encountered. In baselayout-2' rc.conf there's option rc_interactiveto enable it. I believe it was in baselayout-1 as well, but you can try upgading to 2 if it's not there. If that won't work try passing init=/bin/bash as a kernel parameter (from grub) to drop right into shell on start, then do ls /etc/runlevels/boot and try starting each script one-by-one until one of them fails like you described. I believe there's also 1 or single parameter, that can instruct init not to proceed past the first runlevel, which might be a cleaner way than init=/bin/bash, but might as well try to start the same faulty script if it's encountered early enough. And then there's a low-level hack w/ sysrq key (SysRq-K, if i recall correctly) to drop into shell, which can be used at the same time as I mode I've described above. You can try disabling one of the scripts mentioned at the beginning of linked log file (my rc.log) by removing +x from them (just removing symlink from runlevel probably won't do, since other scripts might start it as a dependency) to see if boot proceeds a bit further w/o mentioned crash. rc.log: http://dpaste.com/53552/ -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
Re: [gentoo-user] Re: KVM and no keyboard at start of boot
On Tue, 09 Jun 2009 21:55:49 -0500 Harry Putnam rea...@newsguy.com wrote: Mike Kazantsev mk.frag...@gmail.com writes: Something finally allows the kvm pass through to be recognized but not until I've reached the login prompt. But linux kernel isn't loaded or used in any way when grub screen comes up - grub is loading it as a last step of it's execution, so any kernel configuration settings should be completely irrelevant here. So if I could identify what it is in the kernel that allowes it to work at the point where the kernel takes over (login prompt), then maybe I could enable that aspect somehow inside an initramfs, and be able to have the KVM recognized at the grub prompt by booting with an initrd. But it's the grub that loads initrd and linux kernel is actually the one using it, not the grub, so you won't get it until you drop out from grub already. You can try updating grub itself though, 0.X tree ebuilds are accessible for both stable and unstable gentoo arch, but there's 1.X and even live ebuild masked in the portage tree, prehaps they would be able to work with newer hardware? I also wonder, does BIOS recognize this KVM, can you access it? -- Mike Kazantsev // fraggod.net signature.asc Description: PGP signature
Re: [gentoo-user] Atheros kernel driver and my wireless access point setup
Jason Carson schrieb: Hey Norman, I tried changing my /etc/conf.d/net to what you suggested but it resulted in the following errors... *Bringing up interface wlan0 * Configuring wireless network for wlan0 * Scanning for access points *no access points found * Failed to configure wireless for wlan0 Then because wlan0 wouldn't start I got a bunch of errors like... * Error: cannot start named as net.wlan0 would not start * Error: cannot start sshd as net.wlan0 would not start * Error: cannot start apache2 as net.wlan0 would not start etc... This is still rc-script output... do not use the rc-script. Do not start it! Your errors would be, because you have to detach these services from wlan0 and let them use the bridge. Here is some output from my box: mimir ~ # rc-update -s -v | grep net local | default nonetwork net.br0 | default net.eth0 | default net.lo | boot netmount | default mimir ~ # ifconfig wlan0 wlan0 Protokoll:Ethernet Hardware Adresse 00:80:48:5e:57:3d inet6 Adresse: fe80::280:48ff:fe5e:573d/64 Gültigkeitsbereich:Verbindung UP BROADCAST RUNNING PROMISC MULTICAST MTU:1500 Metric:1 RX packets:48 errors:0 dropped:0 overruns:0 frame:0 TX packets:80 errors:0 dropped:0 overruns:0 carrier:0 Kollisionen:0 Sendewarteschlangenlänge:1000 RX bytes:7344 (7.1 KiB) TX bytes:14190 (13.8 KiB) mimir ~ # ls /etc/init.d/net.* /etc/init.d/net.br0 /etc/init.d/net.eth0 /etc/init.d/net.lo As you can see, there is no rc-script on wlan, but it is initialized an running in AP mode. Regards Norman Thanks for the clarification Norman :-) ok, I removed net.wlan0 so it doesn't start up anymore. My computer booted up and all the services are working the way they should, however I am having a problem getting hostapd to start. Here is the error I get when I tried to start up hostapd... penguin ~ # /etc/init.d/hostapd start * Bringing up interface wlan0 * Configuring wireless network for wlan0 * Scanning for access points * no access points found * Failed to configure wireless for wlan0 * ERROR: net.wlan0 failed to start * ERROR: cannot start hostapd as net.wlan0 would not start It's suppose to be an access point, not scanning for one so do you have any idea what I should do now?
[gentoo-user] Re: 64-bit kernel on a 32-bit installation
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Mark Knecht wrote: Hi, I lurk on the LKML, say hi once in awhile, ask a question once in awhile, and try to read at least the interesting to a non-programmer posts. I was curious about this one that came up today. Seems like this is a natural for Gentoo. I have a Gentoo 64-bit setup but have had lots of troubles over the years (far less now though) with web media and other things that need to be more Windows compatible. (I do audio work with my Gentoo boxes - interface to studios and a few bands, etc) I've found that my 32-bit Gentoo installations have been more compatible than 64-bit. Outside stuff like Java is better. In general when I have a problem I wonder if it's because I'm running 64-bit. How would one go about building a 64-bit kernel on a 32-bit machine with Gentoo? I presume that's mostly just how I configure the kernel, along with maybe some cross-compile options? Are there any projects going on in this area where I might become a test case? Wiki? Docs? Do others see value - getting 64-bit memory management, new CPU flags, etc., but keeping the apps 32-bit for compatibility? Take care, Mark Personally, I am using a 64-bit kernel with a 32-bit userland. My setup is a bit more complicated than the usual, because I have a x86_64-pc-linux-gnu-gcc that will build 32-bit as well as 64-bit binaries. The simpler version of what I use is: # emerge crossdev # crossdev -t x86_64-pc-linux-gnu Then, you can use something like the following to actually build a 64-bit kernel (personally, I always use out-of-tree builds, and create a GNUmakefile that calls the Makefile in the current directory with all the options I want): (in the kernel build directory) # make -C /path/to/sources O=`pwd` ARCH=x86 \ CROSS_COMPILE=x86_64-pc-linux-gnu- \ menuconfig I have found that just about everything works perfectly in my 64-bit kernel with 32-bit userland, *except* VirtualBox, which I have to run the 64-bit version of from a chroot. I also personally handle all external kernel modules, and add them to package.provided when necessary, so portage doesn't have to think about them. PS: I was going to outline all the patches, etc. that I needed for a multilib gcc/glibc, but then realized that you probably didn't need that much detail. - -- ABCD -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.11 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkovRjcACgkQOypDUo0oQOrMdgCfXRiLDyg1IH8d9fA+WodUjWO8 PRMAnihXrPy3VZBYhRF7LzWVivKl2eIb =dD3A -END PGP SIGNATURE-
[gentoo-user] Re: 64-bit kernel on a 32-bit installation
On 06/10/2009 02:44 AM, Mark Knecht wrote: Do others see value - getting 64-bit memory management, new CPU flags, etc., but keeping the apps 32-bit for compatibility? Personally, no. I see more value in a true multilib Gentoo. Unfortunately, Gentoo's multilib is fake. You can't build 32-bit libs and packages but have to download emul binary packages instead. It is by far better (IMO) being able to compile problematic packages as 32-bit instead of running a 32-bit userland in a 64-bit kernel.