Am Mittwoch, 19. September 2001 01:15 schrieb Matthew Dharm:
> At this point, I think the "best" answer is to say:
>
>       Don't swap over USB for a 2.4.x kernel.
>
> I think this issue is better addressed in 2.5, for stability reasons.

Unfortunately this is not limited to swapping.
Functionally writing out dirty pages while memory is low is equivalent to 
swapping. It is possible for memory to be low and there to be no pages to be 
written to swap space.
Depending on the exact behaviour of the vm subsystem it might even be 
possible to have a lockup during ordinary operations.
The sequence is like this:

first page written out -> usb storage allocates memory -> vm selects second 
page to be written to usb storage -> scsi layer finds the request queue to be 
full

I am not sure what happens at this point. We might be lucky and the vm 
subsystem writes out pages to other devices or frees clean pages. In fact 
this is very likely, but it is probably not guaranteed.
Unless I am mistaken there is a global limit on the number of pending io 
request. If all these are destined for usb storage, we can only hope that 
there are clean freeable pages.

        Regards
                Oliver

_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to