2.6.13-rc1-git2 Oops on rmmod serial_core

2.6.13-rc1-git2 Oops on rmmod serial_core

[Kenneth Parrish]

The serial support code newly allows selecting the max number of serial
ports. This computer has one s-port so i typed '1'.

 config SERIAL_8250_NR_UARTS
 int "Maximum number of 8250/16550 serial ports"
 depends on SERIAL_8250
 default "4"
 help
 Set this to the number of serial ports you want the driver
 to support.  This includes any ports discovered via ACPI or
 PCI enumeration and any ports that may be added at run-time
 via hot-plug, or any ISA multi-port serial cards.

I found the value needs to be at least '4' or we oops at rmmod
serial_core.  2.6.13-rc1-git2 kernels with values from one through four
were tested.
--
All progress is based upon a universal innate desire on the part of
every organism to live beyond its income.
                -- Samuel Butler
--- MultiMail/Linux v0.46

Re: 2.6.13-rc1-git2 Oops on rmmod serial_core

[Russell King]

On Sat, Jul 02, 2005 at 06:57:54AM +0000, Kenneth Parrish wrote:
> The serial support code newly allows selecting the max number of serial
> ports. This computer has one s-port so i typed '1'.
> 
>  config SERIAL_8250_NR_UARTS
>  int "Maximum number of 8250/16550 serial ports"
>  depends on SERIAL_8250
>  default "4"
>  help
>  Set this to the number of serial ports you want the driver
>  to support.  This includes any ports discovered via ACPI or
>  PCI enumeration and any ports that may be added at run-time
>  via hot-plug, or any ISA multi-port serial cards.
> 
> I found the value needs to be at least '4' or we oops at rmmod
> serial_core.  2.6.13-rc1-git2 kernels with values from one through four
> were tested.

It's actually worse than that.  If you set it to less than 4 on an x86
PC, it'll stamp over memory.  All is not lost though - it's fixed in:

http://www.kernel.org/git/gitweb.cgi?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=44454bcdb90532b372c74e3546043d8a3a468939

I'd advise people _not_ to run -git2.  If they do and they see any
oddities, please don't bother reporting them unless you can reproduce
them with a later version or with this set to at least '4'.  This
problem can cause oopses in other parts of the kernel.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 Serial core