lkml.org 
[lkml]   [2014]   [Dec]   [22]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH] proc: task_mmu: show page size in /proc/<pid>/numa_maps
On Mon, Dec 22, 2014 at 09:10:34AM -0800, Dave Hansen wrote:
> On 12/21/2014 02:28 PM, Rafael Aquini wrote:
> >>> > > I'm fine either way, it makes the change even simpler. Also, if we
> >>> > > decide to get rid of page_size != PAGE_SIZE condition I believe we can
> >>> > > also get rid of that "huge" hint being conditionally printed out too.
> >> >
> >> > That would break existing users of the "huge" flag. That makes it out
> >> > of the question, right?
> >> >
> > Yeah, but it sort of follows the same complaint Johannes did for the
> > conditional page size printouts. If we start to print out page size
> > deliberately for each map regardless their backing pages being PAGE_SIZE
> > long or bigger, I don't see much point on keep conditionally printing out
> > the 'huge' hint out.
>
> Because existing userspace might be relying on it. If we take the
> 'huge' hint out, userspace will break.
>
> > As I said before, I'm fine either way though I think
> > we can keep the current behaviour, and just disambiguate page sizes !=
> > PAGE_SIZE as in the current proposal.
>
> Unless we somehow have a (really good) handle on how many apps out there
> are reading and using 'huge', I think we have to keep the existing behavior.
>
Right. I definitely don't have anything better than what I already
presented which seems beaten by your argument, already.
Remaining question here is: should we print out 'pagesize' deliberately
or conditionally, only to disambiguate cases where page_size != PAGE_SIZE?

Have a very nice holidays folks!
-- Rafael


\
 
 \ /
  Last update: 2014-12-22 18:41    [W:0.066 / U:0.288 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site