Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri 25-05-18 12:43:00, Andrew Morton wrote: > On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > > > we might consider this for 4.17 although I don't know if there's anything > > currently broken. Stable backports should be more important, but will have > > to > > be

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri 25-05-18 12:43:00, Andrew Morton wrote: > On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > > > we might consider this for 4.17 although I don't know if there's anything > > currently broken. Stable backports should be more important, but will have > > to > > be reviewed

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/28/2018 09:21 AM, Michal Hocko wrote: > On Fri 25-05-18 12:43:00, Andrew Morton wrote: >> On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: >> >>> we might consider this for 4.17 although I don't know if there's anything >>> currently broken. Stable backports should

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/28/2018 09:21 AM, Michal Hocko wrote: > On Fri 25-05-18 12:43:00, Andrew Morton wrote: >> On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: >> >>> we might consider this for 4.17 although I don't know if there's anything >>> currently broken. Stable backports should be more

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/25/2018 10:48 PM, Vlastimil Babka wrote: > On 05/25/2018 09:43 PM, Andrew Morton wrote: >> On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: >> >>> we might consider this for 4.17 although I don't know if there's anything >>> currently broken. Stable backports should

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/25/2018 10:48 PM, Vlastimil Babka wrote: > On 05/25/2018 09:43 PM, Andrew Morton wrote: >> On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: >> >>> we might consider this for 4.17 although I don't know if there's anything >>> currently broken. Stable backports should be more

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri, May 25, 2018 at 03:08:53PM +0200, Vlastimil Babka wrote: > In __alloc_pages_slowpath() we reset zonelist and preferred_zoneref for > allocations that can ignore memory policies. The zonelist is obtained from > current CPU's node. This is a problem for __GFP_THISNODE allocations that want >

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri, May 25, 2018 at 03:08:53PM +0200, Vlastimil Babka wrote: > In __alloc_pages_slowpath() we reset zonelist and preferred_zoneref for > allocations that can ignore memory policies. The zonelist is obtained from > current CPU's node. This is a problem for __GFP_THISNODE allocations that want >

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/25/2018 09:43 PM, Andrew Morton wrote: > On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > >> we might consider this for 4.17 although I don't know if there's anything >> currently broken. Stable backports should be more important, but will have to >> be reviewed

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On 05/25/2018 09:43 PM, Andrew Morton wrote: > On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > >> we might consider this for 4.17 although I don't know if there's anything >> currently broken. Stable backports should be more important, but will have to >> be reviewed carefully, as

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > we might consider this for 4.17 although I don't know if there's anything > currently broken. Stable backports should be more important, but will have to > be reviewed carefully, as the code went through many changes. >

Re: [PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

On Fri, 25 May 2018 15:08:53 +0200 Vlastimil Babka wrote: > we might consider this for 4.17 although I don't know if there's anything > currently broken. Stable backports should be more important, but will have to > be reviewed carefully, as the code went through many changes. > BTW I think that

[PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

In __alloc_pages_slowpath() we reset zonelist and preferred_zoneref for allocations that can ignore memory policies. The zonelist is obtained from current CPU's node. This is a problem for __GFP_THISNODE allocations that want to allocate on a different node, e.g. because the allocating thread has

[PATCH] mm, page_alloc: do not break __GFP_THISNODE by zonelist reset

In __alloc_pages_slowpath() we reset zonelist and preferred_zoneref for allocations that can ignore memory policies. The zonelist is obtained from current CPU's node. This is a problem for __GFP_THISNODE allocations that want to allocate on a different node, e.g. because the allocating thread has