[0/3] mm: Some rework on zap_details

Message

Peter Xu Dec. 8, 2020, 2:50 a.m. UTC

Posted this small series out to rework zap_details a bit, before adding
something new in.  Hopefully it makes things slighly clearer.

Smoke test only.  Please have a look, thanks.

Peter Xu (3):
  mm: Drop first_index/last_index in zap_details
  mm: Introduce zap_details.zap_flags
  mm: Introduce ZAP_FLAG_SKIP_SWAP

 include/linux/mm.h | 33 ++++++++++++++++++++++++---
 mm/memory.c        | 57 ++++++++++++++++++----------------------------
 2 files changed, 52 insertions(+), 38 deletions(-)

Comments

Peter Xu Dec. 15, 2020, 8:39 p.m. UTC | #1

On Mon, Dec 07, 2020 at 09:50:19PM -0500, Peter Xu wrote:
> Posted this small series out to rework zap_details a bit, before adding
> something new in.  Hopefully it makes things slighly clearer.
> 
> Smoke test only.  Please have a look, thanks.
> 
> Peter Xu (3):
>   mm: Drop first_index/last_index in zap_details
>   mm: Introduce zap_details.zap_flags
>   mm: Introduce ZAP_FLAG_SKIP_SWAP

Some more information on "before adding something new in" - the new flag as a
reference but not yet posted anywhere...

https://github.com/xzpeter/linux/commit/102790511a1a25b7be841fb2059a8c8d8f1a87b2

I still think this small series worths to be reviewed/merged even before that
new flag, because patch 1 should definitely be something good to have, patch 2
prepares for patch 3 and the new bit (which can be seen as optional), but patch
3 should help make things clearer rather than using non-null "details" pointer
to show "whether we should skip swap entries", which can be easily misused IMHO
when someone accidentally replaced one "details==NULL" with some valid pointer.

Of course I can repost these series with the larger one when time comes too,
but I'd still like to at least get a NO early if there is...

Thanks,