| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20080722130727.GA13917@elte.hu> Date: Tue, 22 Jul 2008 15:07:27 +0200 From: Ingo Molnar <mingo@...e.hu> To: Greg KH <gregkh@...e.de>, Jeff Garzik <jgarzik@...ox.com> Cc: linux-kernel@...r.kernel.org, netdev@...r.kernel.org Subject: [bug] hard hang with CONFIG_USB_USBNET=y hi Greg, Jeff, -tip randconfig based boot testing ran into a weird bootup hard hang with one specific config: http://redhat.com/~mingo/misc/config-Tue_Jul_22_10_06_14_CEST_2008.bad the hard hang just happens some time into the bootup - usually somewhere around starting up crond. I first suspected a recent regression, so i tried to bisect it - just to discover that it happens with v2.6.26 too and probably goes way back into the past. So i did a ".config bisection" trying to figure out which specific .config option (out of thousands) triggers the hang. After a few hours of .config bisection i have arrived to CONFIG_USB_USBNET=y - enabling _any_ driver there (i tried CONFIG_USB_NET_ZAURUS=y or CONFIG_USB_NET_PLUSB=y) causes the hard hang. See the delta patch below between a 'good' and 'bad' config. Also see further below a diff between the good and bad serial log output. The same bzImage boots fine on another testsystem. The full 'bad' bootlog can be found at: http://redhat.com/~mingo/misc/log-Tue_Jul_22_10_06_14_CEST_2008.bad What do you suggest me to debug this problem? CONFIG_USB_USBNET=y is supposed to work just fine even if no such hardware is present, correct? Old ISA drivers sometimes fail to boot because they have certain assumptions and poke ports randomly - but CONFIG_USB_USBNET is supposed to work just fine, correct? The hang is fully reproducible so i can try any debug patch or other test-patch. If you think it's worthwile i can try a bisection back into the ancient past. There's no APIC on this box so no NMI watchdog can be used to punch through the hard lockup. It has the feeling of an IRQ flood. (but no kernel printks happen after the lockup, even after hours of waiting) Ingo --- .config.good 2008-07-22 14:25:08.000000000 +0200 +++ .config.bad 2008-07-22 14:35:00.000000000 +0200 CONFIG_USB_RTL8150=y CONFIG_USB_USBNET=y # CONFIG_USB_NET_AX8817X is not set -# CONFIG_USB_NET_CDCETHER is not set +CONFIG_USB_NET_CDCETHER=y # CONFIG_USB_NET_DM9601 is not set # CONFIG_USB_NET_GL620A is not set # CONFIG_USB_NET_NET1080 is not set @@ -904,7 +904,7 @@ CONFIG_USB_USBNET=y # CONFIG_USB_NET_MCS7830 is not set # CONFIG_USB_NET_RNDIS_HOST is not set # CONFIG_USB_NET_CDC_SUBSET is not set -# CONFIG_USB_NET_ZAURUS is not set +CONFIG_USB_NET_ZAURUS=y CONFIG_WAN=y CONFIG_LANMEDIA=y CONFIG_HDLC=y [ ... uninteresting delta like timestamp differences omitted ... ] --- dmesg.good 2008-07-22 14:40:54.000000000 +0200 +++ dmesg.bad 2008-07-22 14:48:56.000000000 +0200 @@ -1,4 +1,4 @@ -Linux version 2.6.26-dirty (mingo@...ne) (gcc version 4.2.2) #21529 SMP Tue Jul 22 14:21:04 CEST 2008 +Linux version 2.6.26-dirty (mingo@...ne) (gcc version 4.2.2) #21530 SMP Tue Jul 22 14:25:39 CEST 2008 BIOS-provided physical RAM map: BIOS-e820: 0000000000000000 - 000000000009f800 (usable) BIOS-e820: 000000000009f800 - 00000000000a0000 (reserved) @@ -1076,6 +1076,12 @@ calling usb_rtl8150_init+0x0/0x30 rtl8150: rtl8150 based usb-ethernet driver v0.6.2 (2004/08/27) usbcore: registered new interface driver rtl8150 initcall usb_rtl8150_init+0x0/0x30 returned 0 after 7 msecs +calling cdc_init+0x0/0x20 +usbcore: registered new interface driver cdc_ether +initcall cdc_init+0x0/0x20 returned 0 after 5 msecs +calling zaurus_init+0x0/0x20 +usbcore: registered new interface driver zaurus +initcall zaurus_init+0x0/0x20 returned 0 after 4 msecs calling usbnet_init+0x0/0x30 initcall usbnet_init+0x0/0x30 returned 0 after 0 msecs calling init_netconsole+0x0/0x1d0 @@ -1281,8 +1287,8 @@ ether gadget: using random self ethernet ether gadget: using random host ethernet address usb0: Ethernet Gadget, version: May Day 2005 usb0: using dummy_udc, OUT ep-b IN ep-a STATUS ep-c -usb0: MAC ee:cf:d5:ad:92:32 -usb0: HOST MAC b6:e4:17:99:38:77 +usb0: MAC fa:96:e7:05:b1:96 +usb0: HOST MAC fa:7d:7d:f2:06:1a initcall init+0x0/0x10 returned 0 after 19 msecs calling i8042_init+0x0/0x400 pnp: the driver 'i8042 kbd' has been registered @@ -1332,10 +1338,10 @@ initcall amd756_init+0x0/0x20 returned 0 calling i2c_amd8111_init+0x0/0x20 initcall i2c_amd8111_init+0x0/0x20 returned 0 after 0 msecs calling nforce2_init+0x0/0x20 -usb 2-1: new high speed USB device using dummy_hcd and address 2 -initcall nforce2_init+0x0/0x20 returned 0 after 3 msecs +initcall nforce2_init+0x0/0x20 returned 0 after 0 msecs calling i2c_piix4_init+0x0/0x20 piix4_smbus 0000:00:07.3: Found 0000:00:07.3 device +usb 2-1: new high speed USB device using dummy_hcd and address 2 initcall i2c_piix4_init+0x0/0x20 returned 0 after 30 msecs calling i2c_sis630_init+0x0/0x20 initcall i2c_sis630_init+0x0/0x20 returned 0 after 0 msecs @@ -1357,6 +1363,7 @@ usb 2-1: configuration #1 chosen from 1 initcall sensors_adm1025_init+0x0/0x10<6>usb0: high speed config #1: 100 mA, Ethernet Gadget, using CDC Ethernet returned 0 after 23 msecs calling sensors_adm1029_init+0x0/0x10 +usb1: register 'cdc_ether' at usb-dummy_hcd-1, CDC Ethernet Device, fa:7d:7d:f2:06:1a usb 2-1: New USB device found, idVendor=0525, idProduct=a4a1 usb 2-1: New USB device strings: Mfr=1, Product=2, SerialNumber=0 usb 2-1: Product: Ethernet Gadget -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@...r.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists