Re: skb_over_panic from b43 after a few hours
On Friday 19 February 2010 16:40:40 Larry Finger wrote: [8018c2d4] skb_put+0x74/0x90 [80c9a8e4] b43_dma_rx+0x338/0x444 [b43] [80c87420] b43_controller_restart+0x7a0/0x974 [b43] The traceback indicates a controller restart. Does your log show any reason for that event? That may help me understand why skb-tail is past skb-end. Note that on bcm47xx backtraces are _extremely_ unreliable. As b43_controller_restart does not call b43_dma_rx, yet it shows it this way in the backtrace, I think that line is just invalid. -- Greetings, Michael. ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: skb_over_panic from b43 after a few hours
On Friday 19 February 2010 16:40:40 Larry Finger wrote: [8018c2d4] skb_put+0x74/0x90 [80c9a8e4] b43_dma_rx+0x338/0x444 [b43] [80c87420] b43_controller_restart+0x7a0/0x974 [b43] The traceback indicates a controller restart. Does your log show any reason for that event? That may help me understand why skb-tail is past skb-end. @Larry: I activated CONFIG_KERNEL_KALLSYMS in the build settings of OpenWRT as told in an old ticker to get this traceback. Poorly, there is no more output from the kernel than this. Note that on bcm47xx backtraces are _extremely_ unreliable. As b43_controller_restart does not call b43_dma_rx, yet it shows it this way in the backtrace, I think that line is just invalid. @Michael: Okay, is there any method to get more reliable information about the kernel panic (strace, ...) or a idea to faster reproduce the kernel panic than waiting? Stefan ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: skb_over_panic from b43 after a few hours
On Friday 19 February 2010 21:28:01 Stefan Wahren wrote: @Michael: Okay, is there any method to get more reliable information about the kernel panic (strace, ...) or a idea to faster reproduce the kernel panic than waiting? I think the backtrace is pretty good as-is, if you ignore the controller_restart line. It's pretty obvious that this is an skb overflow panic caused by a received packet. There's nothing in the trace that falsifies this, as far as I can see. So what if you set up another device in monitor mode and simply capture all packets until the error happens again? If the panic happened by receiving a malformed packet, it should be obvious to see then. -- Greetings, Michael. ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: skb_over_panic from b43 after a few hours
I think the backtrace is pretty good as-is, if you ignore the controller_restart line. It's pretty obvious that this is an skb overflow panic caused by a received packet. There's nothing in the trace that falsifies this, as far as I can see. So what if you set up another device in monitor mode and simply capture all packets until the error happens again? If the panic happened by receiving a malformed packet, it should be obvious to see then. Okay, i will try it and post the results. Stefan ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
Re: skb_over_panic from b43 after a few hours
Hi, it's me again. The problem still exists. I had no idea to get it better reproduceable. Please tell me what information do you need to fix it. Thanks in advice. Stefan Stefan Wahren schrieb: Hi, i'm using OpenWRT on an ASUS WL-500gP V2 with the built-in Broadcom BCM3302. After a few hours the b43 driver crashes with a skb_over_panic and i need to reboot the device. The crash happend also, if there is no traffic over the wireless interface. The wireless interface is configured in AP mode and the encryption psk2. The used kernel (Linux OpenWrt 2.6.30.10 #1 Tue Feb 2 01:15:42 CET 2010 mips GNU/Linux) throws the following output: skb_over_panic: text:80c9a8e4 len:2374 put:2374 head:80e79000 data:80e79040 tail:0x80e79986 end:0x80e79980 dev:NULL Kernel bug detected[#1]: Cpu 0 $ 0 : 1000f800 0079 0001 $ 4 : 802880c0 271f 271f $ 8 : 4000 802959b0 0001 $12 : 000f 8022a1d0 7ffb537b $16 : 00e79040 80e79040 0928 80e9a9a0 $20 : 80d84c80 0019 a0df7190 80ca1518 $24 : 0002 80151108 $28 : 80ea 80ea1e60 001b 8018c2d4 Hi: Lo: 0077 epc : 8018c2d4 skb_put+0x74/0x90 Not tainted ra: 8018c2d4 skb_put+0x74/0x90 Status: 1000f803KERNEL EXL IE Cause : 00800024 PrId : 00029029 (Broadcom BCM3302) Modules linked in: pl2303 option ftdi_sio usb_storage usbserial ohci_hcd nf_nat_ tftp nf_conntrack_tftp nf_nat_irc nf_conntrack_irc nf_nat_ftp nf_conntrack_ftp i pt_MASQUERADE iptable_nat nf_nat xt_NOTRACK iptable_raw xt_state nf_conntrack_ip v4 nf_defrag_ipv4 nf_conntrack ehci_hcd sd_mod pppoe pppox ipt_REJECT xt_TCPMSS ipt_LOG xt_multiport xt_mac xt_limit iptable_mangle iptable_filter ip_tables xt_ tcpudp x_tables ext2 ext3 jbd tun ppp_async ppp_generic slhc vfat fat b43 nls_ut f8 nls_iso8859_15 nls_iso8859_1 nls_cp850 nls_cp437 usbcore scsi_mod nls_base ma c80211 cfg80211 compat_firmware_class compat crc_ccitt arc4 aes_generic deflate ecb cbc switch_robo switch_core diag Process compirq/5-b43 (pid: 812, threadinfo=80ea, task=81f19188, tls=000 0) Stack : 80c9a8e4 0946 0946 80e79000 80e79040 80e79986 80e79980 80266f84 80d84c80 0019 80c9a8e4 80287b88 80286170 1000f800 00fe 80ea1fe0 f800 80d92d2c 80284000 80ca1518 7f816db4 0001 80c7f400 8000 0001 fffc 802c7860 0001 fffe efff 80c87420 81f19188 80c8 80e9ee08 8001c63c 80ea 80ea1f18 8021db88 ... Call Trace: [8018c2d4] skb_put+0x74/0x90 [80c9a8e4] b43_dma_rx+0x338/0x444 [b43] [80c87420] b43_controller_restart+0x7a0/0x974 [b43] Code: afab001c 0c00268d afa20020 020d 080630b6 8fbf002c 0120 1021 03e8 Disabling lock debugging due to kernel taint Stefan ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev
skb_over_panic from b43 after a few hours
Hi, i'm using OpenWRT on an ASUS WL-500gP V2 with the built-in Broadcom BCM3302. After a few hours the b43 driver crashes with a skb_over_panic and i need to reboot the device. The crash happend also, if there is no traffic over the wireless interface. The wireless interface is configured in AP mode and the encryption psk2. The used kernel (Linux OpenWrt 2.6.30.10 #1 Tue Feb 2 01:15:42 CET 2010 mips GNU/Linux) throws the following output: skb_over_panic: text:80c9a8e4 len:2374 put:2374 head:80e79000 data:80e79040 tail:0x80e79986 end:0x80e79980 dev:NULL Kernel bug detected[#1]: Cpu 0 $ 0 : 1000f800 0079 0001 $ 4 : 802880c0 271f 271f $ 8 : 4000 802959b0 0001 $12 : 000f 8022a1d0 7ffb537b $16 : 00e79040 80e79040 0928 80e9a9a0 $20 : 80d84c80 0019 a0df7190 80ca1518 $24 : 0002 80151108 $28 : 80ea 80ea1e60 001b 8018c2d4 Hi: Lo: 0077 epc : 8018c2d4 skb_put+0x74/0x90 Not tainted ra: 8018c2d4 skb_put+0x74/0x90 Status: 1000f803KERNEL EXL IE Cause : 00800024 PrId : 00029029 (Broadcom BCM3302) Modules linked in: pl2303 option ftdi_sio usb_storage usbserial ohci_hcd nf_nat_ tftp nf_conntrack_tftp nf_nat_irc nf_conntrack_irc nf_nat_ftp nf_conntrack_ftp i pt_MASQUERADE iptable_nat nf_nat xt_NOTRACK iptable_raw xt_state nf_conntrack_ip v4 nf_defrag_ipv4 nf_conntrack ehci_hcd sd_mod pppoe pppox ipt_REJECT xt_TCPMSS ipt_LOG xt_multiport xt_mac xt_limit iptable_mangle iptable_filter ip_tables xt_ tcpudp x_tables ext2 ext3 jbd tun ppp_async ppp_generic slhc vfat fat b43 nls_ut f8 nls_iso8859_15 nls_iso8859_1 nls_cp850 nls_cp437 usbcore scsi_mod nls_base ma c80211 cfg80211 compat_firmware_class compat crc_ccitt arc4 aes_generic deflate ecb cbc switch_robo switch_core diag Process compirq/5-b43 (pid: 812, threadinfo=80ea, task=81f19188, tls=000 0) Stack : 80c9a8e4 0946 0946 80e79000 80e79040 80e79986 80e79980 80266f84 80d84c80 0019 80c9a8e4 80287b88 80286170 1000f800 00fe 80ea1fe0 f800 80d92d2c 80284000 80ca1518 7f816db4 0001 80c7f400 8000 0001 fffc 802c7860 0001 fffe efff 80c87420 81f19188 80c8 80e9ee08 8001c63c 80ea 80ea1f18 8021db88 ... Call Trace: [8018c2d4] skb_put+0x74/0x90 [80c9a8e4] b43_dma_rx+0x338/0x444 [b43] [80c87420] b43_controller_restart+0x7a0/0x974 [b43] Code: afab001c 0c00268d afa20020 020d 080630b6 8fbf002c 0120 1021 03e8 Disabling lock debugging due to kernel taint Stefan ___ Bcm43xx-dev mailing list Bcm43xx-dev@lists.berlios.de https://lists.berlios.de/mailman/listinfo/bcm43xx-dev