yes.
> Setting NR_TASKS to much greater than 4080 causes an overflow in
> arch/i386/kernel/head.S it tries it store a value of ((12+2*NR_TASKS)*8-1)
> (with NR_TASKS=4080 this becomes 65375, just within the 2^16 of a word)
> in a 16bit work. This value appears to by loaded by a
> lgdt gdt_descr
> instruction. I don't know x86 assembly and at this point it goes straight
> over my head.
>
> I am writing a large parallel IO system a would like to use kernel
> threads (glibc2) to handle the IO connections and make use of multiple
> processors.
>
> Any thoughts?
I think mingo had an experimental patch the worked around this, you can try
looking back of linux-kernel archive for this.
(Do you _really_ need >4k processes?)
-Chris
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu