Re: [linux-pm] [PATCH 1/9] PM / Hibernate: swap, switch to hibernate_io_handle

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

On Monday, June 21, 2010, Nigel Cunningham wrote:
> Hi Jiri.
> 
> On 22/06/10 01:21, Jiri Slaby wrote:
> > On 06/11/2010 11:46 AM, Nigel Cunningham wrote:
> >> On 02/06/10 18:52, Jiri Slaby wrote:
> >>> I addressed the comments I got on the previous RFC. I left the handles
> >>> in place, the functions in hibernate_io_ops now works on them. Further
> >>> I got rid of the memory barriers and minimized global variables as much
> >>> as possible. Comments welcome.
> >>
> >> I would like to hear the arguments for using these handles. I understand
> >> there may have been some previous discussion, but am unable to find it.
> >>
> >> It seems far more sensible to me to not pass around a handle that
> >> virtually nothing actually uses, and instead store and utilise the state
> >> in the place where it is actually useful. If we had more than one struct
> >> hibernate_io_handle in use at a time, I could understand going this way.
> >> As it stands, however...
> >
> > Hi, it I added that based on this: http://lkml.org/lkml/2010/3/24/458
> 
> Okay; thanks.
> 
> Looking at Pavel's comment is confusing. The variable you were adding 
> isn't "global static" (that's a contradiction in terms anyway). Its 
> scope is the file level.
> 
> Since the data is only used in this file, your change makes perfect 
> sense to me.
> 
> Rafael, Pavel: care to discuss this further?

Well.  Generally speaking, I like things as they are, except for patches [3/9]
and [4/9].

So, I'd like to take [1-2/9] and [5-9/9].  Jiri, do [6-7/9] need to be changed
substantially in case [3-4/9] are dropped?

Rafael