Re: RFC: Kill -ERESTART_RESTARTBLOCK.

[Roman Zippel]

Hi,

On Thu, 24 Nov 2005, David Woodhouse wrote:

> I'm looking at implementing ppoll() and pselect(), and the existing
> restartblock stuff isn't sufficient for that -- we don't get to store
> enough args, and although it could possibly be expanded, I just don't
> much like restarting syscalls that way.

What arguments do you need to store?
Instead of messing with the signal delivery it may be better to slightly 
change the restart logic. Instead of calling a separate function, we could 
call the original function with all the arguments, which would reduce the 
state required to be saved.
In the case of nanosleep this could look like this:

sys_nanosleep()
{
	...
	if (restart_block) {
		expire = restart_block->arg0;
	} else {
		if (copy_from_user(&t, rqtp, sizeof(t)))
		...
	}
	expire = schedule_timeout_interruptible(expire);
	if (expire) {
		...
		restart_block->arg0 = expire;
		ret = -ERESTART_RESTARTBLOCK;
	}
}

AFAICT only the timeout argument needs to saved over a restart, the rest 
can be reinitialized from the original arguments.
The main problem would be to prevent an incorrect restart, e.g. if the 
debugger pokes around in the registers.

bye, Roman
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/