Re: [2.6.16-gitX] initcall at 0xffffffff804615d1: returned with error code -1

[Andrew Morton <akpm@osdl.org>]

Voluspa <lista1@telia.com> wrote:
>
> 
> CC-ing author since I'm unsure about domain responsibility. The commit
> "[PATCH] initcall failure reporting" AKA
> c1cda48af8b330a23206eceef3bd030b53c979cd immediately triggered:
> 
> --- dmesg-cd02b966bfcad12d1b2e265dc8dbc331d4c184c4      2006-03-30 08:41:30.000000000 +0200
> +++ dmesg-c1cda48af8b330a23206eceef3bd030b53c979cd      2006-03-30 09:12:31.000000000 +0200
> [...]
> @@ -83,10 +83,12 @@
>  SCSI subsystem initialized
>  PCI: Using ACPI for IRQ routing
>  PCI: If a device doesn't work, try "pci=routeirq".  If it helps, post a report
> +initcall at 0xffffffff804615d1: returned with error code -1
>  pnp: 00:06: ioport range 0x600-0x60f has been reserved
>  pnp: 00:06: ioport range 0x1c0-0x1cf has been reserved
>  pnp: 00:06: ioport range 0x4d0-0x4d1 has been reserved
>  pnp: 00:06: ioport range 0xfe10-0xfe11 could not be reserved
> +initcall at 0xffffffff804704fc: returned with error code 2
>  PCI: Failed to allocate mem resource #6:20000@f0000000 for 0000:01:00.0
>  PCI: Bridge: 0000:00:01.0
>    IO window: disabled.
> 
> Since then (now at 2.6.16-git18) the last/lower one has vanished, but the
> first one remains. The patch says that the initcall function should be
> printed, but it seems to need some debugging option set. Please advice if
> this is of interest (the addresses do not stay constant). 
> 

If you have CONFIG_KALLSYMS enabled then the kernel should print the name
of the function at that address.  It's pretty hard to turn off
CONFIG_KALLSYMS, actually.

If you _do_ have CONFIG_KALLSYMS enabled then something bad has happened -
you can look up those addresses with `nm -n vmlinux > foo' then searching
through foo.  (Or use addr2line, or gdb, or other things...)