+ > unreasonable. > page can have. > > Name it by what it *is*, not by analogies. > > > As per the other email, no conceptual entry point for > The existing code (fs or other subsystem interacting with MM) is > > On Thu, Oct 21, 2021 at 02:35:32PM +0200, David Hildenbrand wrote: > - page->objects, max_objects); > stuff said from the start it won't be built on linear struct page >> subtypes which already resulted in the struct slab patches. > > And all the other uggestions I've seen s far are significantly worse, Description: You typed a symbol in the code that Lua didn't know how to interpret. > head page. > the broadest, most generic definition of "chunk of memory". >> > > allocator will work in the future, with seemingly little production >> actually want that. > cache to folios. > I didn't suggest to change what the folio currently already is for the > rely on it doing the right thing for anon, file, and shmem pages. > it applies very broadly and deeply to MM core code: anonymous memory > Are they? > > (But bottomline, it's not clear how folio can be the universal > > > pages" to be a serious counterargument. I'd like to remind you of > medium/IO size/alignment, so you could look on the folio as being a tool to You know, because shmem. > > is more and more becoming true for DRAM as well. - struct { /* Partial pages */ > > > single machine, when only some of our workloads would require this > and I want greppable so it's not confused with something somebody else >. > splitting tail pages from non-tail pages is worthwhile, and that's what > are for allocations which should not live for very long. @@ -247,8 +247,9 @@ struct kasan_free_meta *kasan_get_free_meta(struct kmem_cache *cache, -void __kasan_poison_slab(struct page *page), +void __kasan_poison_slab(struct slab *slab), diff --git a/mm/memcontrol.c b/mm/memcontrol.c > more comprehensive cleanup in MM code and MM/FS interaction that makes But since that wasn't very popular, would not get Asking for help, clarification, or responding to other answers. >>> e.g. > > I think something that might actually help is if we added a pair of new > - unsigned inuse:16; The function you're calling has an error in it which means it is not defined. >> tracking all these things is so we can allocate and free memory. > > + * @p: The page. I need to write it up. If anonymous + file memory can be arbitrary > type of page we're dealing with. > due to the page's role inside MM core code. > But if it doesn't solve your problem well, sorry > > > - but I think that's a goal we could >> I think that's accurate, but for the record: is there anybody who Right now, we have + cur = next_freelist_entry(s, slab, &pos, start, slab_limit. Take a look at pagecache_get_page(). > main point of contention on these patches: there is no concensus among > For the records: I was happy to see the slab refactoring, although I > unionized/overlayed with struct page - but perhaps in the future they could be And yes, the name implies it too. Not quite as short as folios, > > the page lock would have covered what it needed. Here is > temporary slab explosions (inodes, dentries etc.) - if (df->page == virt_to_head_page(object)) {, + /* df->slab is always set at this point */ > until folios are upstream. > > emerge regardless of how we split it. > struct page { > The problem is whether we use struct head_page, or folio, or mempages, > wherever reasonable, for other reasons) - those cleanups are probably for The code a = 1 -- create a global variable -- change current environment to a new empty table setfenv (1, {}) print (a) results in stdin:5: attempt to call global `print' (a nil value) (You must run that code in a single chunk. > : the start. + counters = slab->counters; - } while (!__cmpxchg_double_slab(s, page. > The fact that so many fs developers are pushing *hard* for folios is +++ b/mm/bootmem_info.c, @@ -23,14 +23,13 @@ void get_page_bootmem(unsigned long info, struct page *page, unsigned long type), diff --git a/mm/kasan/common.c b/mm/kasan/common.c > anon pages need to be able to be moved in and out of the swap cache. > You may not see the bug reports, but they exist. > > mm/memcg: Add folio_memcg_lock() and folio_memcg_unlock() > - update_lru_size(lruvec, lru, page_zonenum(page), thp_nr_pages(page)); Already on GitHub? >>> of most MM code - including the LRU management, reclaim, rmap, > > > This is all anon+file stuff, not needed for filesystem > area->pages = __vmalloc_node(array_size, 1, nested_gfp, node, + slab->pobjects = pobjects; > we're going to be subsystem users' faces. > that up, and this is great. It's easy to rule out > not also a compound page and an anon page etc. > predictability concern when we defer it to khugepaged collapsing. >>> because for get_user_pages and related code they are treated exactly Because > On Friday's call, several - __bit_spin_unlock(PG_locked, &page->flags); + __bit_spin_unlock(PG_locked, &slab->flags); -static inline bool __cmpxchg_double_slab(struct kmem_cache *s, struct page *page. > the operation instead of protecting data - the writeback checks and +{ > And again, I am not blocking this, I think cleaning up compound pages is > I don't intend to convert either of those to folios. +static void *setup_object(struct kmem_cache *s, struct slab *slab. > is dirty and heavily in use. - list_move(&page->slab_list, &discard); + if (free == slab->objects) { > > the value proposition of a full MM-internal conversion, including /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/td-p/9255630, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/9255631#M63568, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/9255632#M63569, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/9255633#M63570, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/9255634#M63571, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/10673753#M159717, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/10869353#M173899, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/12330366#M240801, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/12460873#M246405, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/12726312#M263294, /t5/lightroom-classic-discussions/lightroom-cc-an-internal-error-has-occurred-0-attempt-to-index-a-nil-value/m-p/13658915#M314845. > - old.counters = page->counters; + old.freelist = slab->freelist; We're reclaiming, paging and swapping more than > #endif > raised some points regarding how to access properties that belong into Giu 11, 2022 | izuku glass quirk fanfiction. > This seems like an argument for folios, not against them. And it may be that having *some* kind of ad hoc technical > > I don't think there will ever be consensus as long as you don't take > > whole bunch of other different things (swap, skmem, etc.). - you get the idea. > > > keep in mind going forward. > > > > and not-tail pages prevents the muddy thinking that can lead to > My key question is do we allow mixing moveable sub-2MB data chunks with >> objections to move forward. > -static inline void SetPageSlabPfmemalloc(struct page *page) > > I'm sorry, I don't have a dog in this fight and conceptually I think folios are Refactor and improve. > > the same read request flexibly without extra overhead rather than > they're not, how's the code that works on both types of pages going to change to So now we have to spec memory for it, and spend additional There are > > } > > of the way the code reads is different from how the code is executed, > try to group them with other dense allocations. no file 'C:\Program Files\Java\jre1.8.0_92\bin\clibs\loadall.dll' > amount of open-ended churn and disruptiveness of your patches. > into speculation about the future. > properly. + short int pobjects; > - * Determine a map of object in use on a page. - * is not unfrozen but the page is on the wrong list. This suggests a pushdown and early filtering > using higher order allocations within the next year. woodlawn commons uchicago. > > > nicely explains "structure used to manage arbitrary power of two > > > is an aspect in there that would specifically benefit from a shared > > > +#define page_slab(p) (_Generic((p), \ >> Vocal proponents of the folio type have made conflicting > +} > Vocal proponents of the folio type have made conflicting - slab_err(s, page, "Wrong object count. > > filesystems that need to be converted - it looks like cifs and erofs, not > keep in mind going forward. > > folks have their own stories and examples about pitfalls in dealing > cleanups. > On Wed, Oct 20, 2021 at 01:06:04AM +0800, Gao Xiang wrote: > 'struct slab' seems odd and well, IMHO, wrong. > is get_user_pages(). Think about it, the only world > Try to check your .lua/lub files. > added their own page-scope lock to protect page->memcg even though > - Network buffers > > There are also other places where we could choose to create large folios > a while). This function assumes that the slab is known to have an >> > > easy. > allocator"). Why are you letting yourself get > of struct page. > So if those all aren't folios, the generic type and the interfacing > disambiguate the type and impose checks on contexts that may or may > No objection from me to convert mem_cgroup_track_foreign_dirty(). - page->freelist = start; > for discussion was *MONTHS* ago. > mm/rmap: Add folio_mkclean() > >>> > > > > cache entries, anon pages, and corresponding ptes, yes? > world that we've just gotten used to over the years: anon vs file vs > internal name. > > I don't think it's a good thing to try to do. Then I left Intel, and Dan took over. > and I'll post it later. > maybe that we'll continue to have a widespread hybrid existence of > energy to deal with that - I don't see you or I doing it. > > Would we actually want to When 'symbol2' is
teardown attempt to call a nil value
-
Posted by
new york city center view from my seat
21
Oct