Re: [PATCH] task_struct colouring ...

[Davide Libenzi]

On Fri, 30 Nov 2001, Alan Cox wrote:

> > So You like the idea of stocking structure pointers inside CPU registers
> > or I missed Your point ?
> > The proposed implementation is "uniform" between architectures, that's my
> > point.
>
> An uniform implementation for a totally non uniform set of processors. Not
> actually useful. The x86 is one of the few cpus so short of registers that
> current in a global register is not a win performancewise.

The point is why store kernel pointers in global registers when You can
achieve the same functionality, with a smaller patch, that does not need
to be recoded for each CPU, without using global registers.

> The cache behaviour also heavily depends on the processor. In paticular the
> problem with having to align stacks to do current tricks is absent on non
> x86 processors so they can use properly coloured stacks.
>
> current is far too critical to generalise

This is the diff for current :

 static inline struct task_struct * get_current(void)
 {
-       struct task_struct *current;
-       __asm__("andl %%esp,%0; ":"=r" (current) : "0" (~8191UL));
-       return current;
+       unsigned long *tskptr;
+       __asm__("andl %%esp,%0; ":"=r" (tskptr) : "0" (~8191UL));
+       return (struct task_struct *) *tskptr;
 }

- Davide