oops in serial_out in 2.6.5
I'm (trying to) run Debian's 2.6.5-mckinley-smp on an hp rx2600. It oopses at startup in setserial. The console is ttyS0, which points to the management processor.
Loading the saved-state of the serial devices...
Cannot set serial info: Device or resource busy/dev/ttyS0 at 0x0000 (irq = 49) is a 16550A/dev/ttyS1 at 0x0000 (irq = 60) is a 16550A
Unable to handle kernel NULL pointer dereferencesetserial[203]: Oops 8804682956800 [1]
Pid: 203, CPU 0, comm: setserial
psr : 0000101008026018 ifs : 800000000000040a ip : [<a0000001002b8690>] Tainted: GF
ip is at serial_out+0xb0/0x1a0
unat: 0000000000000000 pfs : 000000000000038a rsc : 0000000000000003
rnat: 0000000000000040 bsps: e0000040fe1e4680 pr : a400000000169965
ldrs: 0000000000000000 ccv : 0000000000000002 fpsr: 0009804c0270033f csd : 0000000000000000 ssd : 0000000000000000
b0 : a0000001002bc890 b6 : a000000100002d70 b7 : a0000001002bc340 f6 : 1003ecccccccccccccccd f7 : 000000000000000000000 f8 : 1003e000000000101426b f9 : 1003e0000000005064c17 f10 : 1003e0000000051eb851f f11 : 1003e0000000000080000 r1 : a000000100790000 r2 : 0000000000000002 r3 : 0000000000000000 r8 : a0000001006b9373 r9 : a0000001006b9372 r10 : a000000100642820 r11 : a000000100642820 r12 : e00000405142fd70 r13 : e000004051428000 r14 : e00000405142fd80 r15 : a0000001006b9360 r16 : 0000000000000002 r17 : 0010000000000000 r18 : 0000000000000001 r19 : 0000000000000002 r20 : 0000000000000001 r21 : 0000000000000000 r22 : 0000000000000000 r23 : a0000001006428f8 r24 : a0000001006b935c r25 : a000000100642820 r26 : a000000100642820 r27 : e0000040509b3ab0 r28 : e00000405142fd88 r29 : e0000040509b3aa8 r30 : 0000000000000001 r31 : e00000405142fd84
Call Trace:
[<a000000100019800>] show_stack+0x80/0xa0
spà0000405142f940 bspà000040514292f8
[<a00000010003f910>] die+0x190/0x260
spà0000405142fb10 bspà000040514292c0
[<a000000100059e80>] ia64_do_page_fault+0x1e0/0x980
spà0000405142fb10 bspà00004051429258
[<a000000100011d20>] ia64_leave_kernel+0x0/0x280
spà0000405142fba0 bspà00004051429258
[<a0000001002b8690>] serial_out+0xb0/0x1a0
spà0000405142fd70 bspà00004051429208
[<a0000001002bc890>] serial8250_startup+0x550/0x6c0
spà0000405142fd70 bspà000040514291c8
[<a0000001002af980>] uart_startup+0x240/0x4e0
spà0000405142fd70 bspà00004051429178
[<a0000001002b1cf0>] uart_set_info+0x3f0/0xb20
spà0000405142fd90 bspà00004051429090
[<a0000001002b39f0>] uart_ioctl+0x2f0/0x3a0
spà0000405142fe20 bspà00004051429048
[<a0000001002841d0>] tty_ioctl+0x870/0xa60
spà0000405142fe20 bspà00004051428ff0
[<a000000100159850>] sys_ioctl+0x350/0x820
spà0000405142fe20 bspà00004051428f60
[<a000000100011b80>] ia64_ret_from_syscall+0x0/0x20
spà0000405142fe30 bspà00004051428f60
/dev/ttyS2 at 0x0000 (irq = 57) is a 16550A/etc/rcS.d/S46setserial: line 152: 203 Segmentation fault ${SETSERIAL} -z $device $args
In case it helps here are some of the TOC dump fields:
<6>Entered OS INIT handler. PSPÿe301a0
Delaying for 5 seconds...
NaT bits 0000000000000000 pr 80000001ff606ca9 b0 a000000100018b90 ia64_pal_call_static+0x70/0xa0 ar.rsc 0000000000000003 cr.iip e00000003fae0da2 0xe00000003fae0da2 cr.ipsr 0000141008020018 cr.ifs 8000000000003060 xip a000000100018ba0 ia64_pal_call_static+0x80/0xa0 xpsr 0000101008026018 xfs 8000000000003060 b1 80000000ffe0d3d0 0x80000000ffe0d3d0
Backtrace of pid 204 (setserial)
Call Trace:
[<a000000100087520>] schedule+0x1080/0x1720
spà0000405142fda0 bspà00004051429208
[<a0000001000398e0>] __down_interruptible+0x280/0x460
spà0000405142fdb0 bspà00004051429190
[<a0000001002b57a0>] uart_get+0x340/0x380
spà0000405142fde0 bspà00004051429148
[<a0000001002b5860>] uart_open+0x80/0x320
spà0000405142fde0 bspà00004051429100
[<a000000100281ad0>] tty_open+0x550/0x8a0
spà0000405142fde0 bspà00004051429098
[<a000000100140b30>] chrdev_open+0x390/0x700
spà0000405142fdf0 bspà00004051429038
[<a000000100125560>] dentry_open+0x3a0/0x5a0
spà0000405142fe00 bspà00004051428fb8
[<a0000001001251a0>] filp_open+0xc0/0xe0
spà0000405142fe00 bspà00004051428f88
[<a000000100126030>] sys_open+0xb0/0x140
spà0000405142fe30 bspà00004051428f08
[<a000000100011b80>] ia64_ret_from_syscall+0x0/0x20
spà0000405142fe30 bspà00004051428f08INIT dump complete. Please reboot now.
Signature · 1 line
Martin
What's in your /var/lib/setserial/autoserial.conf? What does the
kernel print out for serial ports on bootup? I'm guessing they're out of sync. It would be nice if we could stop the oops, but I'm guessing the quicker fix is to reconfigure setserial to "kernel" configuration. Your ttyS0 on irq 49 is likely the builtin UART, not the MP UART. If that fixes it, please report back the differences in config versus kernel serial port listings. Thanks,
Alex
Quoted 102 lines · Martin Pool
I'm (trying to) run Debian's 2.6.5-mckinley-smp on an hp rx2600. It oopses at startup in setserial. The console is ttyS0, which points to the management processor.
Loading the saved-state of the serial devices...
Cannot set serial info: Device or resource busy/dev/ttyS0at0x0000(irq = 49) is a 16550A/dev/ttyS1at0x0000(irq = 60) is a 16550AUnable to handle kernel NULL pointer dereferencesetserial[203]: Oops 8804682956800 [1]Pid: 203, CPU 0, comm: setserial
psr : 0000101008026018 ifs :800000000000040aip : [<a0000001002b8690>] Tainted: GF
ip is atserial_out+0xb0/0x1a0
unat: 0000000000000000 pfs :000000000000038arsc : 0000000000000003
rnat: 0000000000000040 bsps:e0000040fe1e4680pr :a400000000169965
ldrs: 0000000000000000 ccv : 0000000000000002 fpsr:0009804c0270033fcsd : 0000000000000000 ssd : 0000000000000000
b0 :a0000001002bc890b6 :a000000100002d70b7 :a0000001002bc340f6 :1003ecccccccccccccccdf7 : 000000000000000000000 f8 :1003e000000000101426bf9 :1003e0000000005064c17f10 :1003e0000000051eb851ff11 :1003e0000000000080000r1 :a000000100790000r2 : 0000000000000002 r3 : 0000000000000000 r8 :a0000001006b9373r9 :a0000001006b9372r10 :a000000100642820r11 :a000000100642820r12 :e00000405142fd70r13 :e000004051428000r14 :e00000405142fd80r15 :a0000001006b9360r16 : 0000000000000002 r17 : 0010000000000000 r18 : 0000000000000001 r19 : 0000000000000002 r20 : 0000000000000001 r21 : 0000000000000000 r22 : 0000000000000000 r23 :a0000001006428f8r24 :a0000001006b935cr25 :a000000100642820r26 :a000000100642820r27 :e0000040509b3ab0r28 :e00000405142fd88r29 :e0000040509b3aa8r30 : 0000000000000001 r31 :e00000405142fd84Call Trace: [<a000000100019800>] show_stack+0x80/0xa0 spà0000405142f940 bspà000040514292f8 [<a00000010003f910>] die+0x190/0x260 spà0000405142fb10 bspà000040514292c0 [<a000000100059e80>] ia64_do_page_fault+0x1e0/0x980 spà0000405142fb10 bspà00004051429258 [<a000000100011d20>] ia64_leave_kernel+0x0/0x280 spà0000405142fba0 bspà00004051429258 [<a0000001002b8690>] serial_out+0xb0/0x1a0 spà0000405142fd70 bspà00004051429208 [<a0000001002bc890>] serial8250_startup+0x550/0x6c0 spà0000405142fd70 bspà000040514291c8 [<a0000001002af980>] uart_startup+0x240/0x4e0 spà0000405142fd70 bspà00004051429178 [<a0000001002b1cf0>] uart_set_info+0x3f0/0xb20 spà0000405142fd90 bspà00004051429090 [<a0000001002b39f0>] uart_ioctl+0x2f0/0x3a0 spà0000405142fe20 bspà00004051429048 [<a0000001002841d0>] tty_ioctl+0x870/0xa60 spà0000405142fe20 bspà00004051428ff0 [<a000000100159850>] sys_ioctl+0x350/0x820 spà0000405142fe20 bspà00004051428f60 [<a000000100011b80>] ia64_ret_from_syscall+0x0/0x20 spà0000405142fe30 bspà00004051428f60 /dev/ttyS2 at 0x0000 (irq = 57) is a 16550A
/etc/rcS.d/S46setserial: line 152: 203 Segmentation fault ${SETSERIAL} -z $device $argsIn case it helps here are some of the TOC dump fields:
<6>Entered OS INIT handler. PSPÿe301a0
Delaying for 5 seconds...NaT bits 0000000000000000 pr 80000001ff606ca9 b0 a000000100018b90 ia64_pal_call_static+0x70/0xa0 ar.rsc 0000000000000003 cr.iip e00000003fae0da2 0xe00000003fae0da2 cr.ipsr 0000141008020018 cr.ifs 8000000000003060 xip a000000100018ba0 ia64_pal_call_static+0x80/0xa0 xpsr 0000101008026018 xfs 8000000000003060 b1 80000000ffe0d3d0 0x80000000ffe0d3d0Backtrace of pid 204 (setserial)
Call Trace: [<a000000100087520>] schedule+0x1080/0x1720 spà0000405142fda0 bspà00004051429208 [<a0000001000398e0>] __down_interruptible+0x280/0x460 spà0000405142fdb0 bspà00004051429190 [<a0000001002b57a0>] uart_get+0x340/0x380 spà0000405142fde0 bspà00004051429148 [<a0000001002b5860>] uart_open+0x80/0x320 spà0000405142fde0 bspà00004051429100 [<a000000100281ad0>] tty_open+0x550/0x8a0 spà0000405142fde0 bspà00004051429098 [<a000000100140b30>] chrdev_open+0x390/0x700 spà0000405142fdf0 bspà00004051429038 [<a000000100125560>] dentry_open+0x3a0/0x5a0 spà0000405142fe00 bspà00004051428fb8 [<a0000001001251a0>] filp_open+0xc0/0xe0 spà0000405142fe00 bspà00004051428f88 [<a000000100126030>] sys_open+0xb0/0x140 spà0000405142fe30 bspà00004051428f08 [<a000000100011b80>] ia64_ret_from_syscall+0x0/0x20 spà0000405142fe30 bspà00004051428f08INIT dump complete. Please reboot now.
Signature · 1 line
Alex Williamson HP Linux & Open Source Lab
Quoted 3 lines · Martin Pool
I'm (trying to) run Debian's 2.6.5-mckinley-smp on an hp rx2600. It oopses at startup in setserial. The console is ttyS0, which points to the management processor.
I tend to purge setserial, it's just worthless.
Signature · 6 lines
"Next the statesmen will invent cheap lies, putting the blame upon the nation that is attacked, and every man will be glad of those conscience-soothing falsities, and will diligently study them, and refuse to examine any refutations of them; and thus he will by and by convince himself that the war is just, and will thank God for the better sleep he enjoys after this process of grotesque self-deception." -- Mark Twain
Quoted 8 lines · Alex Williamson
What's in your
/var/lib/setserial/autoserial.conf? What does the
kernel print out for serial ports on bootup? I'm guessing they're out of sync. It would be nice if we could stop the oops, but I'm guessing the quicker fix is to reconfigure setserial to "kernel" configuration. Your ttyS0 on irq 49 is likely the builtin UART, not the MP UART. If that fixes it, please report back the differences in config versus kernel serial port listings. Thanks,
Autoserial.conf has
###PORT STATE GENERATED USING AUTOSAVE-ONCE###
###AUTOSAVE-ONCE###
###AUTOSAVE-ONCE###
###AUTOSAVE###
#
# If you want to configure this file by hand, use
# dpkg --configure setserial # and change the configuration mode of the file to MANUAL. If you do not do this# this file may be overwritten automatically the next time you upgrade the # package. # /dev/ttyS0 uart 16550A port 0x0000 irq 59 baud_base 115200 spd_normal skip_test auto_irq /dev/ttyS1 uart 16550A port 0x0000 irq 60 baud_base 115200 spd_normal skip_test auto_irq /dev/ttyS2 uart 16550A port 0x0000 irq 57 baud_base 115200 spd_normal skip_test /dev/ttyS3 uart 16550A port 0x0000 irq 57 baud_base 115200 spd_normal skip_test /dev/ttyS4 uart 16550A port 0x0000 irq 57 baud_base 115200 spd_normal skip_test /dev/ttyS5 uart 16550A port 0x0000 irq 57 baud_base 115200 spd_normal skip_test
The kernel reported
Apr 13 16:18:25 kate kernel: Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled Apr 13 16:18:25 kate kernel: ttyS0 at MMIO 0xf4050000 (irq = 49) is a 16550A Apr 13 16:18:25 kate kernel: GSI 0x22(high,edge) -> CPU 0x0000 vector 75 Apr 13 16:18:25 kate kernel: ttyS1 at MMIO 0xff5e0000 (irq = 75) is a 16550A Apr 13 16:18:25 kate kernel: GSI 0x23(high,edge) -> CPU 0x0000 vector 76 Apr 13 16:18:25 kate kernel: ttyS2 at MMIO 0xff5e2000 (irq = 76) is a 16550A Apr 13 16:18:25 kate kernel: PCI: Found IRQ 49 for device 0000:e0:01.0 Apr 13 16:18:25 kate kernel: IOSAPIC: vector 49 -> CPU 0x0100, enabled Apr 13 16:18:25 kate kernel: ttyS3 at MMIO 0xf4051000 (irq = 49) is a 16450 Apr 13 16:18:25 kate kernel: PCI: Found IRQ 49 for device 0000:e0:01.1 Apr 13 16:18:25 kate kernel: IOSAPIC: vector 49 -> CPU 0x0000, enabled Apr 13 16:18:25 kate kernel: Couldn't register serial port 0000:e0:01.1: -28
I got my machine working again by just removing setserial. Does it ever do anything useful on legacy-free ia64? I wouldn't think people ever set IRQs or IO addresses by hand after bootup.
Signature · 1 line
Martin
Quoted 8 lines · Martin Pool
/dev/ttyS0uart 16550A port0x0000irq 59baud_base115200spd_normalskip_testauto_irq/dev/ttyS1uart 16550A port0x0000irq 60baud_base115200spd_normalskip_testauto_irq/dev/ttyS2uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS3uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS4uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS5uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test
Notice that setserial is missing some smarts about MMIO UARTS (see the "port 0x0000" above). I poked at this a while back, but it seemed sort of ugly to fix, IIRC. Removing setserial seemed to be the easiest workaround.
Quoted 12 lines
Apr 13 16:18:25 kate kernel: Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled Apr 13 16:18:25 kate kernel: ttyS0 at MMIO 0xf4050000 (irq = 49) is a 16550A Apr 13 16:18:25 kate kernel: GSI 0x22(high,edge) -> CPU 0x0000 vector 75 Apr 13 16:18:25 kate kernel: ttyS1 at MMIO 0xff5e0000 (irq = 75) is a 16550A Apr 13 16:18:25 kate kernel: GSI 0x23(high,edge) -> CPU 0x0000 vector 76 Apr 13 16:18:25 kate kernel: ttyS2 at MMIO 0xff5e2000 (irq = 76) is a 16550A Apr 13 16:18:25 kate kernel: PCI: Found IRQ 49 for device 0000:e0:01.0 Apr 13 16:18:25 kate kernel: IOSAPIC: vector 49 -> CPU 0x0100, enabled Apr 13 16:18:25 kate kernel: ttyS3 at MMIO 0xf4051000 (irq = 49) is a 16450 Apr 13 16:18:25 kate kernel: PCI: Found IRQ 49 for device 0000:e0:01.1 Apr 13 16:18:25 kate kernel: IOSAPIC: vector 49 -> CPU 0x0000, enabled Apr 13 16:18:25 kate kernel: Couldn't register serial port 0000:e0:01.1: -28
Quoted 10 lines · Bjorn Helgaas
Quoted 8 lines · Martin Pool
/dev/ttyS0uart 16550A port0x0000irq 59baud_base115200spd_normalskip_testauto_irq/dev/ttyS1uart 16550A port0x0000irq 60baud_base115200spd_normalskip_testauto_irq/dev/ttyS2uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS3uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS4uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test/dev/ttyS5uart 16550A port0x0000irq 57baud_base115200spd_normalskip_test
Quoted 4 lines
Notice that setserial is missing some smarts about MMIO UARTS (see the "port
0x0000" above). I poked at this a while back, but it seemed sort of ugly to fix, IIRC. Removing setserial seemed to be the easiest workaround.
I have done that, and filed a plea for Debian to stop making it autorun, #243841.
I see the same request has been open for hppa for >6 months though. :-(
Signature · 1 line
Martin