Re: swap question

Benjamin C R LaHaise (blah@dot.superaje.com)
Mon, 25 Aug 1997 23:01:26 +0000 ( )

Messages sorted by: [ date ][ thread ][ subject ][ author ]
Next message: Benjamin C R LaHaise: "dentry_operations vs inode_operations"
Previous message: Jeff Wiegley: "Re: AMD K6 and unresolved symbols.."

On Mon, 25 Aug 1997, Bill Hawes wrote:
...
> If it's not being cleared for performance reasons, then all accesses to
> the value should be made only after testing the PG_swap_cache bit to see
> whether it's actually valid. (I think it would be simpler to just clear
> it though.)

> Have I missed something here?

The only places I see that read pg_swap_entry are in swap.h and both are
protected by testing the PG_swap_cache bit. Please don't start clearing
this field... In my personal tree I've moved it into a union in struct
page that overlaps with some other fields (our struct page is already huge
compared to other unices). There are a couple of other cases where we can
make these savings.

Cheers,
-ben

Next message: Benjamin C R LaHaise: "dentry_operations vs inode_operations"
Previous message: Jeff Wiegley: "Re: AMD K6 and unresolved symbols.."