Re: cvs commit: src/sys/bus/iicbus iicbb_if.m iicbus_if.m src/sys/bus/isa isa_if.m src/sys/bus/isa/i386 isa_compat.c src/sys/bus/pci pci.c pci_if.m src/sys/bus/smbus smbus_if.m src/sys/conf Makefile.i386 kmod.mk src/sys/dev/misc/dec ...

[YONETANI Tomokazu <qhwt@xxxxxxxxxxxxx>]

Hi,

On Sun, Nov 16, 2003 at 04:54:40PM -0800, Jeroen Ruigrok/asmodai wrote:
> asmodai     2003/11/16 16:54:40 PST
> 
> DragonFly src repository
> 
>   Modified files:
>     sys/bus/iicbus       iicbb_if.m iicbus_if.m 
>     sys/bus/isa          isa_if.m 
>     sys/bus/isa/i386     isa_compat.c 
>     sys/bus/pci          pci.c pci_if.m 
>     sys/bus/smbus        smbus_if.m 
>     sys/conf             Makefile.i386 kmod.mk 
>     sys/dev/misc/dec     mcclock_if.m 
>     sys/dev/netif/mii_layer miibus_if.m 
>     sys/kern             bus_if.m device_if.m subr_bus.c 
>     sys/sys              bus.h bus_private.h 
>   Log:
>   Factor out the object system from new-bus so that it can be used by
>   non-device code.
>   
>   Re-implement the method dispatch to improve efficiency. The new system
>   takes about 40ns for a method dispatch on a 300Mhz PII which is only 10ns
>   slower than a direct function call on the same hardware.
>   
>   This is all based on Doug Rabson's work in April and May of 2000.
>   
>   Revision  Changes    Path
>   1.3       +3 -1      src/sys/bus/iicbus/iicbb_if.m
>   1.3       +3 -1      src/sys/bus/iicbus/iicbus_if.m
>   1.4       +1 -2      src/sys/bus/isa/isa_if.m
>   1.4       +1 -1      src/sys/bus/isa/i386/isa_compat.c
>   1.6       +1 -1      src/sys/bus/pci/pci.c
>   1.3       +3 -1      src/sys/bus/pci/pci_if.m
>   1.3       +3 -1      src/sys/bus/smbus/smbus_if.m
>   1.11      +2 -2      src/sys/conf/Makefile.i386
>   1.8       +2 -2      src/sys/conf/kmod.mk
>   1.3       +3 -1      src/sys/dev/misc/dec/mcclock_if.m
>   1.3       +3 -1      src/sys/dev/netif/mii_layer/miibus_if.m
>   1.3       +3 -1      src/sys/kern/bus_if.m
>   1.3       +3 -1      src/sys/kern/device_if.m
>   1.5       +31 -213   src/sys/kern/subr_bus.c
>   1.3       +5 -15     src/sys/sys/bus.h
>   1.3       +7 -27     src/sys/sys/bus_private.h

After this commit, my DragonFly laptop panics in early stage of
boot process. I inserted a small printf() at the beginning of
kobj_delete() to see what's the problem, and I got:

 Preloaded elf kernel "/kernel.new" at 0xc03af000.
 Preloaded userconfig_script "/boot/kernel.conf" at 0xc03af28c.
 Preloaded elf module "/modules.new/splash_bmp.ko" at 0xc03af2dc.
 Preloaded elf module "/modules.new/vesa.ko" at 0xc03af38c.
 Preloaded elf module "/modules.new/linux.ko" at 0xc03af438.
 Preloaded elf module "/modules.new/miibus.ko" at 0xc03af4e4.
 Preloaded elf module "/modules.new/if_fxp.ko" at 0xc03af590.
 Preloaded elf module "/modules.new/if_wi.ko" at 0xc03af63c.
 Preloaded elf module "/modules.new/snd_pcm.ko" at 0xc03af6e8.
 Preloaded elf module "/modules.new/snd_t4dwave.ko" at 0xc03af794.
 Preloaded elf module "/modules.new/usb.ko" at 0xc03af844.
 Preloaded elf module "/modules.new/agp.ko" at 0xc03af8ec.
 VESA: v2.0, 16384k memory, flags:0x0, mode table:0xc030f480 (1000040)
 VESA: Trident CYBER 8820
 Pentium Pro MTRR support enabled
 obj: 0xc776b300, obj->ops: 0xc774bdb0, obj->ops->cls: 0xc027b560
 Using $PIR table, 7 entries at 0xc00f0200
 obj: 0xc776b348, obj->ops: 0xc774bdb0, obj->ops->cls: 0xc027b560
 npx0: <math processor> on motherboard
 npx0: INT 16 interface
 obj: 0xc776b3d8, obj->ops: 0xc774bdb0, obj->ops->cls: 0xc027b560
 acpi0: <TOSHIB 750     > on motherboard
 installed MI handler for int 9
 acpi0: power button is handled as a fixed feature programming model.
 Timecounter "ACPI-safe"  frequency 3579545 Hz
 obj: 0xc776b4b0, obj->ops: 0xc774bdb0, obj->ops->cls: 0xc027b560
 acpi_timer0: <24-bit timer at 3.579545MHz> port 0xd808-0xd80b on acpi0
 obj: 0xc776b4f8, obj->ops: 0xc774bdb0, obj->ops->cls: 0xc027b560
 obj: 0xc776b4f8, obj->ops: 0xc774e1b0, obj->ops->cls: 0xc0273954
 obj: 0xc776b4f8, obj->ops: 0, obj->ops->cls: 0
 
 
 Fatal trap 12: page fault while in kernel mode
 fault virtual address   = 0x800
 fault code              = supervisor read, page not present
 instruction pointer     = 0x8:0xc01831c9
 stack pointer           = 0x10:0xc03d1c34
 frame pointer           = 0x10:0xc03d1c4c
 code segment            = base 0x0, limit 0xfffff, type 0x1b
                         = DPL 0, pres 1, def32 1, gran 1
 processor eflags        = interrupt enabled, resume, IOPL = 0
 current process         = 0 (swapper)
 current thread          = pri 12
 interrupt mask          = net tty bio cam
 trap number             = 12
 panic: page fault
 Uptime: 0s
 Terminate ACPI
 Automatic reboot in 15 seconds - press a key on the console to abort
 --> Press a key on the console to reboot,
 --> or switch off the system now.

Seems like obj->ops was NULL, so obj->ops->cls dereferenced a NULL pointer.
DDB didn't show a nice traceback(again). The lines immediately following
the "acpi_timer0" when I ran the working kernel are:
 acpi_cpu0: <CPU> on acpi0
 acpi_tz0: <thermal zone> on acpi0
 acpi_button0: <Power Button> on acpi0
 acpi_lid0: <Control Method Lid Switch> on acpi0
 acpi_cmbat0: <Control method Battery> on acpi0
 acpi_acad0: <AC adapter> on acpi0
 acpi_toshiba0: <Toshiba Notebook Extra HCI driver> on acpi0

Regards.