Most recent kernel where this bug did not occur: 2.6.16.2 Distribution: Source Mage GNU/Linux Hardware Environment: dmlb2000@dmlb2000:~$ lspci 00:00.0 Host bridge: VIA Technologies, Inc. VT82C693A/694x [Apollo PRO133x] (rev c4) 00:01.0 PCI bridge: VIA Technologies, Inc. VT82C598/694x [Apollo MVP3/Pro133x AGP] 00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 40) 00:07.1 IDE interface: VIA Technologies, Inc. VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06) 00:07.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 16) 00:07.3 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller (rev 16) 00:07.4 Bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40) 00:09.0 Ethernet controller: Linksys NC100 Network Everywhere Fast Ethernet 10/100 (rev 11) 00:0a.0 Ethernet controller: Accton Technology Corporation SMC2-1211TX (rev 10) 00:0b.0 Mass storage controller: Silicon Image, Inc. PCI0680 Ultra ATA-133 Host Controller (rev 02) 01:00.0 VGA compatible controller: nVidia Corporation NV11 [GeForce2 MX/MX 400] (rev a1) dmlb2000@dmlb2000:~$ lsmod Module Size Used by ipt_MASQUERADE 3072 1 ipt_REJECT 4096 4 ipt_LOG 5376 7 xt_tcpudp 3072 13 xt_state 1920 11 xt_pkttype 1792 4 iptable_raw 1792 0 xt_CLASSIFY 2048 0 xt_CONNMARK 2176 0 xt_connmark 1792 0 ipt_owner 1920 0 ipt_recent 8716 0 ipt_iprange 1792 0 xt_physdev 2192 0 ipt_policy 2816 7 ipt_multiport 2304 4 xt_conntrack 2176 0 iptable_mangle 2304 1 ip_nat_irc 2176 0 ip_nat_tftp 1664 0 ip_nat_ftp 2816 0 iptable_nat 6916 1 ip_nat 13588 5 ipt_MASQUERADE,ip_nat_irc,ip_nat_tftp,ip_nat_ftp ,iptable_nat ip_conntrack_irc 5144 1 ip_nat_irc ip_conntrack_tftp 3348 1 ip_nat_tftp ip_conntrack_ftp 5660 1 ip_nat_ftp ip_conntrack 40280 13 ipt_MASQUERADE,xt_state,xt_CONNMARK,xt_connmark ,xt_conntrack,ip_nat_irc,ip_nat_tftp,ip_nat_ftp,iptable_nat,ip_nat,ip_conntrack_ irc,ip_conntrack_tftp,ip_conntrack_ftp nfnetlink 4888 2 ip_nat,ip_conntrack iptable_filter 2432 1 ip_tables 10204 4 iptable_raw,iptable_mangle,iptable_nat,iptable_f ilter x_tables 9348 18 ipt_MASQUERADE,ipt_REJECT,ipt_LOG,xt_tcpudp,xt_ state,xt_pkttype,xt_CLASSIFY,xt_CONNMARK,xt_connmark,ipt_owner,ipt_recent,ipt_ip range,xt_physdev,ipt_policy,ipt_multiport,xt_conntrack,iptable_nat,ip_tables nfsd 182788 1 exportfs 4352 1 nfsd lockd 49640 2 nfsd sunrpc 116516 2 nfsd,lockd ipv6 203584 14 ext2 49416 1 mbcache 6404 1 ext2 i2c_viapro 6932 0 via_agp 7680 1 uhci_hcd 24848 0 8139too 21248 0 8250_pnp 8448 0 evbug 2560 0 mii 4352 1 8139too shpchp 35552 0 pci_hotplug 21180 1 shpchp usbcore 95904 2 uhci_hcd evdev 7040 0 tulip 43424 0 agpgart 23240 1 via_agp psmouse 31496 0 serio_raw 5380 0 via686a 13448 0 hwmon 2308 1 via686a i2c_isa 3584 1 via686a i2c_core 15232 3 i2c_viapro,via686a,i2c_isa pcspkr 2692 0 rtc 9524 0 parport_pc 31728 0 parport 27848 1 parport_pc siimage 9216 0 [permanent] floppy 51920 0 8250 19180 1 8250_pnp serial_core 15104 1 8250 I'll create more attachments with better output of hardware Software Environment: pretty standard basesystem gnu/linux distro... I'll attach a list of all packages and their versions Problem Description: network doesn't send or recieve with 2.6.16.11 the other kernel I checked it with is 2.6.16.9 and the problems exists there too dmlb2000@dmlb2000:~$ cat /proc/cmdline root=/dev/hdb3 ro vga=791 noapic I've tried removing the noapic but then the network devices don't even show up in ifconfig -a on any of the kernels including the one I'm running now. The problem with 2.6.16.9 and above seems (when passing noapic) that there isn't any traffic sent over my network sending or recieving on any interface (there's two). I can assign ip addresses and create routes but no traffic is sent. all these problems disapear when I remove smp, apic and local apic. Steps to reproduce: 1) compile smp kernel 2.6.16.9 or above on my hardware 2) boot the new kernel I'm kind of disapointed that a 'maintenance/security fix' crippled my system, how did this happen?
Created attachment 7949 [details] colon separated package listing
Created attachment 7950 [details] lshw output
Created attachment 7951 [details] lspci -vvv output
Created attachment 7952 [details] kernel config with bug
(please reply via reply-to-all, rather than via bugzilla web forms) bugme-daemon@bugzilla.kernel.org wrote: > > http://bugzilla.kernel.org/show_bug.cgi?id=6439 > > Summary: apic network smp problems > Kernel Version: 2.6.16.11 > Status: NEW > Severity: normal > Owner: other_other@kernel-bugs.osdl.org > Submitter: dmlb2000@gmail.com > > > Most recent kernel where this bug did not occur: 2.6.16.2 > Distribution: Source Mage GNU/Linux > Hardware Environment: > dmlb2000@dmlb2000:~$ lspci > 00:00.0 Host bridge: VIA Technologies, Inc. VT82C693A/694x [Apollo PRO133x] (rev c4) > 00:01.0 PCI bridge: VIA Technologies, Inc. VT82C598/694x [Apollo MVP3/Pro133x AGP] > 00:07.0 ISA bridge: VIA Technologies, Inc. VT82C686 [Apollo Super South] (rev 40) > 00:07.1 IDE interface: VIA Technologies, Inc. > VT82C586A/B/VT82C686/A/B/VT823x/A/C PIPC Bus Master IDE (rev 06) > 00:07.2 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller > (rev 16) > 00:07.3 USB Controller: VIA Technologies, Inc. VT82xxxxx UHCI USB 1.1 Controller > (rev 16) > 00:07.4 Bridge: VIA Technologies, Inc. VT82C686 [Apollo Super ACPI] (rev 40) > 00:09.0 Ethernet controller: Linksys NC100 Network Everywhere Fast Ethernet > 10/100 (rev 11) > 00:0a.0 Ethernet controller: Accton Technology Corporation SMC2-1211TX (rev 10) > 00:0b.0 Mass storage controller: Silicon Image, Inc. PCI0680 Ultra ATA-133 Host > Controller (rev 02) > 01:00.0 VGA compatible controller: nVidia Corporation NV11 [GeForce2 MX/MX 400] > (rev a1) > dmlb2000@dmlb2000:~$ lsmod > Module Size Used by > ipt_MASQUERADE 3072 1 > ipt_REJECT 4096 4 > ipt_LOG 5376 7 > xt_tcpudp 3072 13 > xt_state 1920 11 > xt_pkttype 1792 4 > iptable_raw 1792 0 > xt_CLASSIFY 2048 0 > xt_CONNMARK 2176 0 > xt_connmark 1792 0 > ipt_owner 1920 0 > ipt_recent 8716 0 > ipt_iprange 1792 0 > xt_physdev 2192 0 > ipt_policy 2816 7 > ipt_multiport 2304 4 > xt_conntrack 2176 0 > iptable_mangle 2304 1 > ip_nat_irc 2176 0 > ip_nat_tftp 1664 0 > ip_nat_ftp 2816 0 > iptable_nat 6916 1 > ip_nat 13588 5 ipt_MASQUERADE,ip_nat_irc,ip_nat_tftp,ip_nat_ftp > ,iptable_nat > ip_conntrack_irc 5144 1 ip_nat_irc > ip_conntrack_tftp 3348 1 ip_nat_tftp > ip_conntrack_ftp 5660 1 ip_nat_ftp > ip_conntrack 40280 13 ipt_MASQUERADE,xt_state,xt_CONNMARK,xt_connmark > ,xt_conntrack,ip_nat_irc,ip_nat_tftp,ip_nat_ftp,iptable_nat,ip_nat,ip_conntrack_ > irc,ip_conntrack_tftp,ip_conntrack_ftp > nfnetlink 4888 2 ip_nat,ip_conntrack > iptable_filter 2432 1 > ip_tables 10204 4 iptable_raw,iptable_mangle,iptable_nat,iptable_f > ilter > x_tables 9348 18 ipt_MASQUERADE,ipt_REJECT,ipt_LOG,xt_tcpudp,xt_ > state,xt_pkttype,xt_CLASSIFY,xt_CONNMARK,xt_connmark,ipt_owner,ipt_recent,ipt_ip > range,xt_physdev,ipt_policy,ipt_multiport,xt_conntrack,iptable_nat,ip_tables > nfsd 182788 1 > exportfs 4352 1 nfsd > lockd 49640 2 nfsd > sunrpc 116516 2 nfsd,lockd > ipv6 203584 14 > ext2 49416 1 > mbcache 6404 1 ext2 > i2c_viapro 6932 0 > via_agp 7680 1 > uhci_hcd 24848 0 > 8139too 21248 0 > 8250_pnp 8448 0 > evbug 2560 0 > mii 4352 1 8139too > shpchp 35552 0 > pci_hotplug 21180 1 shpchp > usbcore 95904 2 uhci_hcd > evdev 7040 0 > tulip 43424 0 > agpgart 23240 1 via_agp > psmouse 31496 0 > serio_raw 5380 0 > via686a 13448 0 > hwmon 2308 1 via686a > i2c_isa 3584 1 via686a > i2c_core 15232 3 i2c_viapro,via686a,i2c_isa > pcspkr 2692 0 > rtc 9524 0 > parport_pc 31728 0 > parport 27848 1 parport_pc > siimage 9216 0 [permanent] > floppy 51920 0 > 8250 19180 1 8250_pnp > serial_core 15104 1 8250 > I'll create more attachments with better output of hardware > Software Environment: > pretty standard basesystem gnu/linux distro... > I'll attach a list of all packages and their versions > > Problem Description: > network doesn't send or recieve with 2.6.16.11 the other kernel I checked it > with is 2.6.16.9 and the problems exists there too > > dmlb2000@dmlb2000:~$ cat /proc/cmdline > root=/dev/hdb3 ro vga=791 noapic > > I've tried removing the noapic but then the network devices don't even show up > in ifconfig -a on any of the kernels including the one I'm running now. > > The problem with 2.6.16.9 and above seems (when passing noapic) that there isn't > any traffic sent over my network sending or recieving on any interface (there's > two). I can assign ip addresses and create routes but no traffic is sent. > > all these problems disapear when I remove smp, apic and local apic. > > Steps to reproduce: > 1) compile smp kernel 2.6.16.9 or above on my hardware > 2) boot the new kernel > > I'm kind of disapointed that a 'maintenance/security fix' crippled my system, > how did this happen? > I've looked back through the 2.6.16.3->2.6.16.9 changelogs and cannot immediately see anything which could have caused this. Are you really sure that 2.6.16.3 worked OK? If so, are you able to pinpoint a particular release which caused the regression? Thanks.
On Tue, Apr 25, 2006 at 03:25:24PM -0700, Andrew Morton wrote: > I've looked back through the 2.6.16.3->2.6.16.9 changelogs and cannot > immediately see anything which could have caused this. I don't either. Any way you can use git to bisect the different patches between 2.6.16.3 and 2.6.16.11? There are not many at all so it shouldn't take very long to narrow it down.
Andrew, so I'm supposed to reply via email? and not use the fine bugzilla web interface? ;) Also I narrowed it down to between 16.8-9... so 2.6.16.8 works and 2.6.16.9 doesn't work. There's a lot in that patch where should I start ripping it apart?
> I've looked back through the 2.6.16.3->2.6.16.9 changelogs and cannot > immediately see anything which could have caused this. > > Are you really sure that 2.6.16.3 worked OK? If so, are you able to > pinpoint a particular release which caused the regression? Okay, after about 20+ kernel compiles of differing versions and differing reverse patches 2.6.16.11 seems to work okay, just as it is, meaning this bug is invalid. I compiled manually instead of using the package manager to make the tests faster and it seems that's what was making everything work. I'm testing this very thing now. In the package there are several environmental vars set when going through the compile (LDFLAGS, CFLAGS, CPPFLAGS, CXXFLAGS, etc). These are the only things I can figure breaking the kernel. These flags also change and can various user defined options. I know compiling for a ppc the kernel will not link if LDFLAGS has -s in it. Since, I'm also the maintainer of this package, should the package just unset all those variables for every arch just to be safe? - David Brown
Okay now I can't reproduce the problem... and since everything works no sense in keeping the bug around... /me is thoroughly confused...