[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] [PATCH] ACPI: move tables.c fully into .init.*
>>> On 19.09.12 at 09:56, Keir Fraser <keir.xen@xxxxxxxxx> wrote: > On 19/09/2012 08:29, "Jan Beulich" <JBeulich@xxxxxxxx> wrote: > >>>>> On 18.09.12 at 17:42, Keir Fraser <keir.xen@xxxxxxxxx> wrote: >>> On 18/09/2012 16:26, "Jan Beulich" <JBeulich@xxxxxxxx> wrote: >>> >>>> The only non-init item was the space reserved for the initial tables, >>>> but we can as well dynamically allocate that array. >>>> >>>> Signed-off-by: Jan Beulich <jbeulich@xxxxxxxx> >>> >>> Really worthwhile, versus having to round up the initial_tables[] allocation >>> to a whole number of pages? >> >> It was exactly one page in size already. > > Well, still I'm not sure about the style of converting static data to heap > data, just to be able to mark a whole unit init-only. If everything else is > already marked init inside that file already, is there a win? Yes - there a little of 1k of string literals, and there's no way (other than making them variables to be able to apply section attributes) to move them into .init.* - that's the purpose of the other instances where the same post processing is being applied. Of course, if you're only against the dynamic allocation, moving the array elsewhere would be another option (but would require making the symbol global, whereas here the symbol goes away altogether from the symbol tables). Yet another option would be to do the dynamic allocation where it was actually intended to be done, passing NULL here. The resizing there isn't being made use of anyway (and wouldn't work afaict), so we could as well do away with it and replace it by the simple allocation needed here (or simply fix it, considering that we might need it at some point). Jan _______________________________________________ Xen-devel mailing list Xen-devel@xxxxxxxxxxxxx http://lists.xen.org/xen-devel
|
Lists.xenproject.org is hosted with RackSpace, monitoring our |