The orinoco_pci driver assigned the incorrect memory address to my NIC, a Netgear MA311 802.11b wireless card. Evidence is provided in the form of the output of ifconfig from both kernels included in this bug report. The result becomes that I am not able to connect to my network, but I can talk to the AP. I can't get a DHCP lease from my router, and I can't ping the AP. It's confusting, but this is what happens: The orionoco drivers load with no complaints. at this point, iwconfig returns correct information about the access point and its relationship to my network card. Now I need to ifconfig. I cannot contact the network at all, even if I manually specifiy IP. ping -b 255.255.255.255 returns only my address if the interface is turned on. yes, I've tried resetting new kernel features like "preemptible kernel" and funny new networking options; in fact, I sat both kernel configs side-by-side and made them parallel in configuration. I should also note that I have an ASUS board which features the nForce2 chipset, meaning it also includes the Soundstorm integrated sound (i810_audio [or snd-intel-8x0 for alsa users] in the kernel, nvaudio proprietary) and the built-in network interface (forcedeth in the kernel, nvnet proprietary), so I'm wondering if memory address issues could be due to forcedeth's presence? but I've tried blacklisting the module in hotplug, and it still doesn't work correctly, even though after startup, an "lsmod" returns that neither the proprietary nor the opensource module are present, and the wireless NIC is in the correct (eth0) device. Reproducible: Always Steps to Reproduce: 1. "emerge =sys-kernel/development-sources-2.6.7" 2. "genkernel --menuconfig all" and enable the orionoco wireless modules and their dependencies; 3. reset your bootloader to be able to switch between the multiple kernels 4. /sbin/restart 5. select your 2.6.7 kernel on next startup 6. expect errors with proprietary and/or third-party modules, they don't matter. 7. disable any other NICs (ifconfig wlan1 down, ifconfig eth1 down, et cetera) 8. be sure that only the orinoco card is capable of network access (lsmod, look for orinoco_pci et cetera, iwconfig, make sure returns correct information about AP.) 9. attempt to activate your orinoco interface if not already activated (ifconfig eth0 up/192.168.0.1, /etc/init.d/net.eth0 start, or such) 10. attempt to contact any computer on the network (ping -b 255.255.255.255) Actual Results: only my computer responded to the broadcast ping, at times, ping returned: connect: network is unreachable. Expected Results: it should have... worked? I should have gotten a response from, at the very least, my Access Point. Portage 2.0.50-r9 (default-x86-1.4, gcc-3.3.3, glibc-2.3.3.20040420-r0, 2.4.26-gentoo-r5) ================================================================= System uname: 2.4.26-gentoo-r5 i686 AMD Athlon(tm) XP 2700+ Gentoo Base System version 1.4.16 Autoconf: sys-devel/autoconf-2.59-r3 Automake: sys-devel/automake-1.8.3 ACCEPT_KEYWORDS="x86" AUTOCLEAN="yes" CFLAGS="-O2 -march=athlon-xp -pipe" CHOST="i686-pc-linux-gnu" COMPILER="gcc3" CONFIG_PROTECT="/etc /usr/X11R6/lib/X11/xkb /usr/kde/2/share/config /usr/kde/3.1/share/config /usr/kde/3.2/share/config /usr/kde/3/share/config /usr/share/config /var/qmail/control" CONFIG_PROTECT_MASK="/etc/gconf /etc/terminfo /etc/env.d" CXXFLAGS="-O2 -march=athlon-xp -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="autoaddcvs ccache sandbox" GENTOO_MIRRORS="http://mirror.tucdemonic.org/gentoo/ http://gentoo.ccccom.com http://mirrors.tds.net/gentoo http://open-systems.ufl.edu/mirrors/gentoo" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="" SYNC="rsync://earthman/gentoo-portage" USE="X aalib alsa apm arts avi berkdb bonobo cdr crypt cups directfb encode esd foomaticdb gdbm gif gnome gpm gtk gtk2 gtkhtml imlib java jpeg kde ldap libg++ libwww mad mikmod motif mozilla mpeg ncurses nls oggvorbis opengl oss pam pdflib perl png python qt quicktime readline sdl slang spell ssl svga tcltk tcpd truetype x86 xml2 xmms xv zlib"
Created attachment 37485 [details] /proc/interrupts on 2.4.26-r5, control reference This is the correct interrupt information
Created attachment 37486 [details] /proc/interrupts on 2.6.7, incorrect interrupt information this is the incorrect interrupt information, proof of bug.
Created attachment 37487 [details] more information, interrupts and memory address this is the result of ifconfig on the working kernel.
Created attachment 37488 [details] ifconfig from 2.6.7 this is the ifconfig from the broken kernel, you can see that the interrupt and memory addresses are different.
any developments so far?
A possible cause of the problem I see is that on 2.6.7, you don't have more than 16 irqs, causing the reuse of irq 5. This is caused becuase io-apic is somehow not enabled on 2.6.7. You are either missing a kernel option (smp?) or there is a problem with your motherboard and 2.6, the former seems more likely. See http://old.lwn.net/1998/0205/io-apic.html for more information. As this is not a gentoo problem, I request that a dev mark this as resolved.
I take that back, the problem is in the kernel: http://www.ussg.iu.edu/hypermail/linux/kernel/0404.1/0825.html. Update to the latest version of the gentoo kernel and see if it is fixed, or patch it yourself.
Well I'll try to find the patch and attach it to the bug. ;-)
Is this problem still present in latest version of the kernel?
This bug report is awaiting feedback from the reporter.
bug is fixed in 2.6.8.1
Ok - Closing this bug.
Reopening.
Closing as FIXED.
