Re: [RFC][PATCH] PM / Hibernate: Modify signature used to mark swap

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Saturday, October 02, 2010, Nigel Cunningham wrote:
> Hi.
> 
> On 30/09/10 08:02, Rafael J. Wysocki wrote:
> > On Wednesday, September 29, 2010, Nigel Cunningham wrote:
> >> Hi.
> >>
> >> On 30/09/10 07:13, Rafael J. Wysocki wrote:
> >>> From: Rafael J. Wysocki<rjw@sisk.pl>
> >>> Subject: PM / Hibernate: Modify signature used to mark swap
> >>>
> >>> Since we are adding compression to the kernel's hibernate code,
> >>> change signature used by it to mark swap spaces, so that earlier
> >>> kernels don't attempt to restore compressed images they cannot
> >>> handle.
> >>>
> >>> Signed-off-by: Rafael J. Wysocki<rjw@sisk.pl>
> >>> ---
> >>>    kernel/power/swap.c |    6 +++---
> >>>    1 file changed, 3 insertions(+), 3 deletions(-)
> >>>
> >>> Index: linux-2.6/kernel/power/swap.c
> >>> ===================================================================
> >>> --- linux-2.6.orig/kernel/power/swap.c
> >>> +++ linux-2.6/kernel/power/swap.c
> >>> @@ -29,7 +29,7 @@
> >>>
> >>>    #include "power.h"
> >>>
> >>> -#define SWSUSP_SIG	"S1SUSPEND"
> >>> +#define HIBERNATE_SIG	"LINHIB0001"
> >>>
> >>>    /*
> >>>     *	The swap map is a data structure used for keeping track of each page
> >>> @@ -195,7 +195,7 @@ static int mark_swapfiles(struct swap_ma
> >>>    	if (!memcmp("SWAP-SPACE",swsusp_header->sig, 10) ||
> >>>    	    !memcmp("SWAPSPACE2",swsusp_header->sig, 10)) {
> >>>    		memcpy(swsusp_header->orig_sig,swsusp_header->sig, 10);
> >>
> >> if no compression
> >>
> >>> -		memcpy(swsusp_header->sig,SWSUSP_SIG, 10);
> >>
> >> else
> >>
> >>> +		memcpy(swsusp_header->sig, HIBERNATE_SIG, 10);
> >>
> >> ??
> >
> > I thought about that, but we'll need to drop the old signature when the image
> > format changes (I think it will after your patch series) anyway.  And the
> > benefit is not really worth it IMO (it only affects people using the in-kernel
> > hibernation on x86-64).
> 
> Okay. I'm just a little weary because if everything I want to get in 
> gets in, we're going to then be changing this signature multiple times. 

That most likely is correct.

> I want to eventually reach a situation where we have a 'proper' header 
> section to the image that stores (among other things) whether 
> compression is enabled and which cryptoapi algorithm is used, allowing a 
> person to en/disable compression at runtime without recompiling or 
> rebooting. I suppose if I seek to put that stuff in earlier rather than 
> later, that will simplify things.

I don't have a problem with changing the signature.  If the image format
changes, it's probably the best we can do.

> Oh, while I'm writing to you: let's not worry about getting my patches 
> in for 2.6.37. I won't be working on them much today, and want to put 
> the time into making sure they're right before they get merged. As you 
> said - let's seek to get things right first time.

OK

Thanks,
Rafael