wpa_supplicant 0.6.9 does not start on boot with the init-skript /etc/init.d/net.wlan0 After boot I need to run wpa_supplicant -B -iwlan0 -Dwext -c/etc/wpa_supplicant/wpa_supplicant.conf in a shell manually to get the connection running. A downgrade to version 0.6.4 solved the problem. Reproducible: Always Steps to Reproduce: -Boot the computer or killall wpa_supplicant and then run /etc/init.d/net.wlan0 restart Actual Results: no wireless connection Expected Results: wpa_supplicant should be running. kernel: gentoo-sources-2.6.29-r5, rt2500pci compiled as kernel module /ect/conf.d/net: modules_wlan0=( "!iwconfig" "wpa_supplicant" ) wpa_supplicant_wlan0=( "-Dwext" ) config_wlan0=( "192.168.0.3" ) routes_wlan0=( "default gw 192.168.0.1" ) dns_servers_wlan0=( "192.168.0.1" ) wpa_timeout_wlan0=60 associate_timeout_wlan0=60 config_eth0=( "192.168.1.12 netmask 255.255.255.0 brd 192.168.1.255" ) wpa_supplicant.conf: ctrl_interface=/var/run/wpa_supplicant ctrl_interface_group=0 ap_scan=1 network={ ssid="myssid" scan_ssid=1 key_mgmt=WPA-PSK pairwise=CCMP TKIP group=CCMP TKIP psk="mypsk" priority=-9999999 }
confirmed on my thinkpad X31.
Confirmed on my Thinkpad T22 with both a Netgear WG111v2 (rtl8187 chipset) and a PCMCIA card with a RaLink rt2400 chipset. Downgrade resolved issue. Additional info - following error msg is printed on a /etc/init.d/net.wlan0 start: * You either need to set a preferred_aps list in /etc/conf.d/wireless * preferred_aps="SSID1 SSID2" * and set associate_order_wlan0="forcepreferred" * or set associate_order_wlan0="forcepreferredonly" * or hardcode the SSID to "any" and let the driver find an Access Point * ssid_wlan0="any" * or configure defaulting to Ad-Hoc when Managed fails * adhoc_ssid_wlan0="WLAN" * or hardcode the SSID against the interface (not recommended) * ssid_wlan0="SSID" * Failed to configure wireless for wlan0
How about if you use baselayout 2.0 / openrc. Does that fix the problem?
Already using it here, emerge --info as follows: Portage 2.1.6.13 (default/linux/x86/2008.0, gcc-4.3.2, glibc-2.9_p20081201-r2, 2.6.30.1 i686) ================================================================= System uname: Linux-2.6.30.1-i686-Pentium_III_-Coppermine-with-glibc2.0 Timestamp of tree: Sat, 01 Aug 2009 03:00:20 +0000 app-shells/bash: 3.2_p39 dev-lang/python: 2.5.4-r3 dev-util/cmake: 2.6.4 sys-apps/baselayout: 2.0.1 sys-apps/openrc: 0.4.3-r3 sys-apps/sandbox: 1.6-r2 sys-devel/autoconf: 2.13, 2.63-r1 sys-devel/automake: 1.8.5-r3, 1.9.6-r2, 1.10.2 sys-devel/binutils: 2.18-r3 sys-devel/gcc-config: 1.4.1 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.27-r2 ACCEPT_KEYWORDS="x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-march=pentium3 -O2 -fomit-frame-pointer -pipe" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc /usr/share/config" CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/fonts/fonts.conf /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo /etc/udev/rules.d" CXXFLAGS="-march=pentium3 -O2 -fomit-frame-pointer -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="distlocks fixpackages parallel-fetch protect-owned sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://ftp.swin.edu.au/gentoo http://mirror.pacific.net.au/linux/Gentoo ftp://ftp.swin.edu.au/gentoo ftp://mirror.pacific.net.au/linux/Gentoo" LDFLAGS="-Wl,-O1" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_CONFIGROOT="/" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://server/gentoo-portage" USE="X a52 aac acpi alsa amazon amrnb amrwb aotuv archive ass auth bash-completion berkdb bittorrent branding bzip2 cairo cardbus cdaudio cddb cdparanoia cleartype clucene consolekit css cups dbus disk-partition divx djbfft dri dvd dvdnav embedded faad fam fbcondecor ffmpeg fftw firefox firefox3 gdbm gif glib glibc-omitfp glitz glut gmp gpm gstreamer gtk hal handbook iconv ipod jpeg kde kipi laptop libnotify lzma mad mikmod mmx mng mozilla mp2 mp3 mp4 mpeg mplayer musicbrainz mysql ncurses njb nonfsv4 nptl nptlonly nsplugin ogg opengl osdmenu oss pam pcmcia pcre pdf perl phonon plasma png ps python qt3 qt3support qt4 quicktime readline redland rtc samba scanner semantic-desktop shm sqlite srt sse ssl startup-notification svg sysfs syslog tcpd threads thumbnail tiff tremor truetype trusted usb vim-syntax visualization vnc vorbis webkit wifi win32codecs wma wmp x86 xanim xcb xcomposite xine xml xorg xpm xulrunner xv xvid zeroconf zip zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1 emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mmap_emul mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="evdev" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LIRC_DEVICES="serial sir" USERLAND="GNU" VIDEO_CARDS="savage vesa" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, FFLAGS, INSTALL_MASK, LANG, LC_ALL, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
(In reply to comment #3) > How about if you use baselayout 2.0 / openrc. > Does that fix the problem? > Hm, interesting. I already had baselayout-2.0.1. However, I upgraded openrc from sys-apps/openrc-0.3.0-r1 to sys-apps/openrc-0.4.3-r3 and sysvinit from sys-apps/sysvinit-2.86-r10 to sys-apps/sysvinit-2.86-r12 and now wpa_supplicant works as desired (i.e. autostarts). Here is my emerge --info: Portage 2.1.6.13 (default/linux/x86/2008.0, gcc-4.3.2, glibc-2.9_p20081201-r2, 2.6.27-gentoo-r10 i686) ================================================================= System uname: Linux-2.6.27-gentoo-r10-i686-Intel-R-_Pentium-R-_M_processor_1400MHz-with-glibc2.0 Timestamp of tree: Sun, 02 Aug 2009 19:00:01 +0000 app-shells/bash: 3.2_p39 dev-java/java-config: 2.1.8-r1 dev-lang/python: 2.5.4-r3 sys-apps/baselayout: 2.0.1 sys-apps/openrc: 0.4.3-r3 sys-apps/sandbox: 1.6-r2 sys-devel/autoconf: 2.63-r1 sys-devel/automake: 1.9.6-r2, 1.10.2 sys-devel/binutils: 2.18-r3 sys-devel/gcc-config: 1.4.1 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.27-r2 ACCEPT_KEYWORDS="x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=pentium-m -pipe -fomit-frame-pointer" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/ca-certificates.conf /etc/env.d /etc/env.d/java/ /etc/fonts/fonts.conf /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/sandbox.d /etc/terminfo /etc/texmf/language.dat.d /etc/texmf/language.def.d /etc/texmf/updmap.d /etc/texmf/web2c /etc/udev/rules.d" CXXFLAGS="-O2 -march=pentium-m -pipe -fomit-frame-pointer" DISTDIR="/usr/portage/distfiles" FEATURES="distlocks fixpackages parallel-fetch protect-owned sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://ftp.ucsb.edu/pub/mirrors/linux/gentoo/ " LDFLAGS="-Wl,-O1" LINGUAS="en" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_CONFIGROOT="/" PORTAGE_RSYNC_EXTRA_OPTS="--exclude-from=/etc/portage/rsync_excludes" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage/layman/sunrise /usr/local/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="X acl acpi alsa berkdb bluetooth bzip2 cdr cli cracklib crypt dbus dri dvd fortran fpu gdbm gpm gtk gtk2 hal iconv ipv6 isdnlog java jpeg kpathsea mmx mudflap ncurses nptl nptlonly openmp pam pcre perl png pppd python readline reflection session spl sse sse2 ssl sysfs tcpd truetype unicode x86 xorg zlib" ALSA_CARDS="intel8x0" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mmap_emul mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="evdev" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LINGUAS="en" USERLAND="GNU" VIDEO_CARDS="ati radeon vesa" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, FFLAGS, INSTALL_MASK, LANG, LC_ALL, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS Thanks!
(In reply to comment #3) > How about if you use baselayout 2.0 / openrc. > Does that fix the problem? > I encountered a similar bug two days ago on a fully updated ~amd64 system. Starting net.wlan0 would not start wpa_supplicant, so dhcp would fail. I eventually traced the problem to an incorrect path in /lib64/rc/cache/nettree which pointed to /sbin/wpa_supplicant rather than /usr/sbin/wpa_supplicant . Editing nettree fixed the problem.
> I encountered a similar bug two days ago on a fully updated ~amd64 system. > Starting net.wlan0 would not start wpa_supplicant, so dhcp would fail. I > eventually traced the problem to an incorrect path in /lib64/rc/cache/nettree > which pointed to /sbin/wpa_supplicant rather than /usr/sbin/wpa_supplicant . > Editing nettree fixed the problem. > Thank you, this post was a big help. I have two amd64 system, both running ~amd64. One is old, and has been upgraded many times, the other is a fresh install. The fresh install experienced this problem. The issue of wpa_supplicant not starting through net.wlan0 was resolved by fixing the file that Willian mentioned above. Thanks William, you saved me a lot of time.
Glad I could help, Anthony. I did a bit more hunting around, and this seems to be what happened: 1. During the initial setup of my machine, I installed openrc-0.4.3-r3, which created a bunch of scripts in /lib64/rc/net/ . 2. Shortly afterwards (likely at shutdown) the scripts in /lib64/rc/net/ were used to create /lib64/rc/cache/nettree . The wpa_supplicant.sh script checks if /usr/sbin/wpa_supplicant is executable (which it wasn't, because I hadn't installed it yet), and then defaults to /sbin/wpa_supplicant as the executable. 3. Later, I installed wpa_supplicant, but the cache file was never regenerated, and so still contained the default path. It seems that there should be some way to invalidate the nettree cache file, either when the net.* scripts are run (which might, unfortunately, remove all gain from having the cache in the first place) or in the wpa_supplicant ebuild. Alternatively, modifying /lib64/rc/net/wpa_supplicant.sh to default to an executable in /usr would solve this particular problem, but leave us open to similar ones in the future.
After a complete reinstall of gentoo (because of some mistakes I did while changing a file system to ext4 I don't want to talk about) I had this problem again and even modifying the path to wpa_supplicant in the nettree file didn't help. I reinstalled openrc and now everything works. Maybe this bug should get some higher priority, since I can imagine there are quite a few frustrated people out there. Plus, much of the solution is clear, right? Thanks
Apologies for the long time awol - the downgrade to wpa_supplicant-0.6.4 worked well enough until bothered updating it again. A reinstall of openrc-0.4.3-r3 fixed the problem for me - not sure what changed. If it helps, the original install date for openrc-0.4.3-r3 was 04-Jul-2009. By the changelog it looks like a bunch of updates went in just after that... (my fault for using ~arch?). Thanks to Atanas for the openrc upgrade infos. BTW: /lib/rc/cache/nettree still points to an incorrect wpa_supplicant file (never did try William's fix), so it is not *just* nettree related.
Does it work with the newest version af openrc in the tree?
(In reply to comment #11) > Does it work with the newest version af openrc in the tree? OK, so here is my full story. I have installed a fresh gentoo last week, which installed openrc-0.5.1 and wpa_supplicant-0.6.9. Wpa_supplicant-0.6.9 didn't work, even after I set the correct wpa_supplicant path in nettree. The older version of wpa_supplicant (0.6.4) did work. Then I installed again the newest wpa_supplicant, and AFTER that reinstalled openrc-0.5.1. After that wpa_supplicant worked (although wpa_gui couldn't communicate with wpa_cli). Now I just upgraded openrc to openrc-0.5.1-r1 and the result is that wpa_supplicant works (again wpa_gui says "Could not get status from wpa_supplicant"). However, this was maybe expected since I installed openrc AFTER wpa_wupplicant. Should I uninstall wpa_supplicant, reinstall openrc, then install wpa_supplicant again to see if it will work then?
(In reply to comment #11) > Does it work with the newest version af openrc in the tree? No. It does not work with openrc-0.6.1-r1.
*** Bug 317907 has been marked as a duplicate of this bug. ***
Does the symlink trick in bug 317907 fix the problem? If it does, we need to get a patch applied to openrc.
(In reply to comment #15) > Does the symlink trick in bug 317907 fix the problem? > > If it does, we need to get a patch applied to openrc. I didn't specifically test the symlink, but I did edit /lib64/rc/cache/nettree as suggested in commment 6. It works after fixing this file.
For me, editing nettree did not work. In fact, the wrong location (/sbin/wpa_supplicant) got reset once (I am not sure when), so I had to do it twice, until it persisted. Did not help anyway.. But then, remerging openrc did the trick, and everything works fine now. (using ~amd64, system up to date)
After a kernel upgrade to 2.6.34 I started to have the same problems with wpa_supplicant not starting. With 2.6.32 everything was fine. My wireless driver is b43. "/etc/init.d/net.wlan0 start --debug" produced a lot of interesting information and I managed to diff two kernels and track down the problem to "/lib/rcscripts/net/wpa_supplicant.sh" The script tries to identify if wlan0 did actually start by grepping through /proc/net/wireless (isn't it outdated and not recommended for future use?) Whatever I did with my kernel .config I couldn't get wlan0 appear in this file. It is always empty, even when WLAN is up. So a quick fix was to edit "wpa_supplicant_exists" function and replace "/proc/net/wireless" with "/proc/net/dev" There are more places, where /proc/net/wireless is used. Denis
Hi! For some reason "sysfs" files did not appear during previous boot. So now b43 creates /sys/class/net/$1/wireless directory and everything works fine. But /proc/net/wireless is still empty and I think it will be stay so.
(In reply to comment #18) > After a kernel upgrade to 2.6.34 I started to have the same problems with > wpa_supplicant not starting. With 2.6.32 everything was fine. My wireless > driver is b43. Hey Denis, what is your baselayout version? This bug has the same description like http://bugs.gentoo.org/show_bug.cgi?id=321663 But it seems baselayout-2 specific. If you use baselayout 1, is bug 321663 (my bug...) more similar to you?
Hi! My baselayout version is: $ emerge --info | grep base sys-apps/baselayout: 1.12.13 I would like to note once more that now my system works - /sys/class/net/$1/wireless directory is created, when b43 is loaded. But /proc/net/wireless stays empty all the time. History: I started to have problems after upgrade to 2.6.34 from 2.6.32 (gentoo-sources). In 2.6.32 everything was working and in 2.6.34 - wpa_supplicant wasn't starting. Since I always keep a previous kernel for backup I could start /etc/init.d/net.wlan0 with --debug option on both kernels and check an output for the differences - that is how I managed to find a place that stopped working (/proc/net/wireless is always empty on 2.6.34). Then I saw that /lib/rcscripts/net/wpa_supplicant.sh contains code that also checks "sysfs". So I recompiled my kernel with CONFIG_WIRELESS_EXT_SYSFS and everything started to work. So on my machine (2.6.34) if I want to use wireless (b43) and wpa_supplicant (-Dwext) I MUST turn CONFIG_WIRELESS_EXT_SYSFS on, because /proc/net/wireless is no longer filled with any data.
(In reply to comment #21) Sorry for the noise, but yeah! Exactly my problem! Turning on CONFIG_WIRELESS_EXT_SYSFS (deprecated?) works for me and /proc/net/wireless is empty with >=2.6.33 I'll post your solution in bug 321663. THANKS!
Is there still a bug here, or can we close it?
I just checked. This is still a bug. I am using net-wireless/wpa_supplicant-0.7.2-r3 and sys-kernel/gentoo-sources-2.6.34-r12 (the most current stable versions). When (the depreciated) CONFIG_WIRELESS_EXT_SYSFS is unset in the kernel, /etc/init.d/net.wlan0 does not start wpa_supplicant. The workaround is easy though: just set CONFIG_WIRELESS_EXT_SYSFS=y in the kernel. Manually starting wpa_supplicant after boot also works. Note that this bug and bug 321663 are closely related (one is probably a dupe of the other). I'll leave my copy of the kernel with "CONFIG_WIRELESS_EXT_SYSFS is not set" around for a little while longer in case there's some testing I can help do.
(In reply to comment #24) > Note that this bug and bug 321663 are closely related (one is probably a dupe > of the other). Sorry, but you are wrong: You decribe bug 321663, not this. This bug and 321663 are different: This bug was opened in 2009, one year before 2.6.33 was ready. This bug is an openrc / baseleyout-2 problem, 321663 is independent form the baselayout and was introduced by the kernel update (/proc/net/wireless is empty now). I think with the release of 2.6.33 this bugs mixed up (this bug has a general summary). And the workaround may work for both bugs (I never had this bug and I'm on baselayout-2 now). But the reasons for the bugs must be different. One may have to test this bug with 2.6.32 or lower.
(In reply to comment #25) > (In reply to comment #24) > > > Note that this bug and bug 321663 are closely related (one is probably a dupe > > of the other). > > Sorry, but you are wrong: You decribe bug 321663, not this. > This bug and 321663 are different: This bug was opened in 2009, one year > before 2.6.33 was ready. This bug is an openrc / baseleyout-2 problem, > 321663 is independent form the baselayout and was introduced by the kernel > update (/proc/net/wireless is empty now). > I think with the release of 2.6.33 this bugs mixed up (this bug has a > general summary). And the workaround may work for both bugs (I never had > this bug and I'm on baselayout-2 now). But the reasons for the bugs must be > different. > > One may have to test this bug with 2.6.32 or lower. What is the status of this second bug with updated systems? Feel free to reopen if still valid
