search for: zone_idx

...assume the rest - * is movable. + * is movable. As we can have some pages in the movable zone + * that are only considered movable for memory offlining (esp., + * PageHWPoison and PageOffline that will be skipped), we + * perform this optimization only for memory offlining. */ - if (zone_idx(zone) == ZONE_MOVABLE) + if ((flags & MEMORY_OFFLINE) && zone_idx(zone) == ZONE_MOVABLE) continue; /* -- 2.26.2

Currently, virtio-mem does not really support ZONE_MOVABLE. While it allows to online fully plugged memory blocks to ZONE_MOVABLE, it does not allow to online partially-plugged memory blocks to ZONE_MOVABLE and will never consider such memory blocks when unplugging memory. This might be surprising for users (especially, if onlining suddenly fails). Let's support partially plugged memory

Currently, virtio-mem does not really support ZONE_MOVABLE. While it allows to online fully plugged memory blocks to ZONE_MOVABLE, it does not allow to online partially-plugged memory blocks to ZONE_MOVABLE and will never consider such memory blocks when unplugging memory. This might be surprising for users (especially, if onlining suddenly fails). Let's support partially plugged memory

...6 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -5853,6 +5853,7 @@ void __ref memmap_init_zone_device(struct zone *zone, { unsigned long pfn, end_pfn = start_pfn + size; struct pglist_data *pgdat = zone->zone_pgdat; + struct vmem_altmap *altmap = pgmap_altmap(pgmap); unsigned long zone_idx = zone_idx(zone); unsigned long start = jiffies; int nid = pgdat->node_id; @@ -5865,9 +5866,7 @@ void __ref memmap_init_zone_device(struct zone *zone, * of the pages reserved for the memmap, so we can just jump to * the end of that region and start processing the device pages. */ -...

...++ b/mm/page_isolation.c >> @@ -58,16 +58,12 @@ static int set_migratetype_isolate(struct page *page, int migratetype, int isol_ >> spin_unlock_irqrestore(&zone->lock, flags); >> if (!ret) { >> drain_all_pages(zone); >> - } else { >> - WARN_ON_ONCE(zone_idx(zone) == ZONE_MOVABLE); >> - >> - if ((isol_flags & REPORT_FAILURE) && unmovable) >> - /* >> - * printk() with zone->lock held will likely trigger a >> - * lockdep splat, so defer it here. >> - */ >> - dump_page(unmovable, &quot...

...d_count += thp_nr_pages(page) + page_has_private(page); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index fab5e97dc9ca..397b0765de97 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -6070,6 +6070,9 @@ void __ref memmap_init_zone_device(struct zone *zone, __init_single_page(page, pfn, zone_idx, nid); + /* ZONE_DEVICE pages start with a zero reference count. */ + set_page_count(page, 0); + /* * Mark page reserved as it will need to wait for onlining * phase for it to be fully associated with a zone. diff --git a/mm/swap.c b/mm/swap.c index e7bdf094f76a..769644fe70ef 100644...

...d_count += thp_nr_pages(page) + page_has_private(page); diff --git a/mm/page_alloc.c b/mm/page_alloc.c index eb0962976a0e..039b88bb3978 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -6154,6 +6154,9 @@ void __ref memmap_init_zone_device(struct zone *zone, __init_single_page(page, pfn, zone_idx, nid); + /* ZONE_DEVICE pages start with a zero reference count. */ + set_page_count(page, 0); + /* * Mark page reserved as it will need to wait for onlining * phase for it to be fully associated with a zone. diff --git a/mm/swap.c b/mm/swap.c index 0eb057141a04..82cbe069dcf3 100644...

@Andrew, @Mst, I suggest the whole series (including the virtio-mem change) goes via the -mm tree. Currently, virtio-mem does not really support ZONE_MOVABLE. While it allows to online fully plugged memory blocks to ZONE_MOVABLE, it does not allow to online partially-plugged memory blocks to ZONE_MOVABLE and will never consider such memory blocks when unplugging memory. This might be surprising

For 5.10. Patch #1-#4,#6 have RBs or ACKs, patch #5 is virtio-mem stuff maintained by me. This should go via the -mm tree. --- When introducing virtio-mem, the semantics of ZONE_MOVABLE were rather unclear, which is why we special-cased ZONE_MOVABLE such that partially plugged blocks would never end up in ZONE_MOVABLE. Now that the semantics are much clearer (and are documented in patch #6),

For 5.10. Patch #1-#4,#6 have RBs or ACKs, patch #5 is virtio-mem stuff maintained by me. This should go via the -mm tree. --- When introducing virtio-mem, the semantics of ZONE_MOVABLE were rather unclear, which is why we special-cased ZONE_MOVABLE such that partially plugged blocks would never end up in ZONE_MOVABLE. Now that the semantics are much clearer (and are documented in patch #6),

@Andrew can we give this a churn and consider it for v5.9 in case there are no more comments? Patch #1-#4,#6 have RBss or ACKs, patch #5 is virtio-mem stuff maintained by me (and MST is aware). --- When introducing virtio-mem, the semantics of ZONE_MOVABLE were rather unclear, which is why we special-cased ZONE_MOVABLE such that partially plugged blocks would never end up in ZONE_MOVABLE. Now

@Andrew can we give this a churn and consider it for v5.9 in case there are no more comments? Patch #1-#4,#6 have RBss or ACKs, patch #5 is virtio-mem stuff maintained by me (and MST is aware). --- When introducing virtio-mem, the semantics of ZONE_MOVABLE were rather unclear, which is why we special-cased ZONE_MOVABLE such that partially plugged blocks would never end up in ZONE_MOVABLE. Now

@Andrew can we give this a churn and consider it for v5.9 in case there are no more comments? Patch #1-#4 have RB's, patch #5 is virtio-mem stuff maintained by me, patch #6 is just a doc update. --- Currently, virtio-mem does not really support ZONE_MOVABLE. While it allows to online fully plugged memory blocks to ZONE_MOVABLE, it does not allow to online partially-plugged memory blocks to

@Andrew can we give this a churn and consider it for v5.9 in case there are no more comments? Patch #1-#4 have RB's, patch #5 is virtio-mem stuff maintained by me, patch #6 is just a doc update. --- Currently, virtio-mem does not really support ZONE_MOVABLE. While it allows to online fully plugged memory blocks to ZONE_MOVABLE, it does not allow to online partially-plugged memory blocks to

Matthew Wilcox, Ira Weiny, and others have complained that ZONE_DEVICE struct page reference counting is ugly because they are "free" when the reference count is one instead of zero. This leads to explicit checks for ZONE_DEVICE pages in places like put_page(), GUP, THP splitting, and page migration which have to adjust the expected reference count when determining if the page is

This is still an RFC because after looking at the pmem/dax code some more, I realized that the ZONE_DEVICE struct pages are being inserted into the process' page tables with vmf_insert_mixed() and a zero refcount on the ZONE_DEVICE struct page. This is sort of OK because insert_pfn() increments the reference count on the pgmap which is what prevents memunmap_pages() from freeing the struct

Hi Dan, Jérôme and Jason, below is a series that cleans up the dev_pagemap interface so that it is more easily usable, which removes the need to wrap it in hmm and thus allowing to kill a lot of code Note: this series is on top of Linux 5.2-rc5 and has some minor conflicts with the hmm tree that are easy to resolve. Diffstat summary: 32 files changed, 361 insertions(+), 1012 deletions(-) Git

Hi Dan, Jérôme and Jason, below is a series that cleans up the dev_pagemap interface so that it is more easily usable, which removes the need to wrap it in hmm and thus allowing to kill a lot of code Note: this series is on top of the rdma/hmm branch + the dev_pagemap releas fix series from Dan that went into 5.2-rc5. Git tree: git://git.infradead.org/users/hch/misc.git