- Nov 2024
-
elixir.bootlin.com elixir.bootlin.com
-
#ifdef CONFIG_ARCH_HAS_HUGEPD static unsigned long hugepte_addr_end(unsigned long addr, unsigned long end, unsigned long sz) { unsigned long __boundary = (addr + sz) & ~(sz-1); return (__boundary - 1 < end - 1) ? __boundary : end; } static int gup_hugepte(pte_t *ptep, unsigned long sz, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { unsigned long pte_end; struct page *page; struct folio *folio; pte_t pte; int refs; pte_end = (addr + sz) & ~(sz-1); if (pte_end < end) end = pte_end; pte = huge_ptep_get(ptep); if (!pte_access_permitted(pte, flags & FOLL_WRITE)) return 0; /* hugepages are never "special" */ VM_BUG_ON(!pfn_valid(pte_pfn(pte))); page = nth_page(pte_page(pte), (addr & (sz - 1)) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr); folio = try_grab_folio(page, refs, flags); if (!folio) return 0; if (unlikely(pte_val(pte) != pte_val(ptep_get(ptep)))) { gup_put_folio(folio, refs, flags); return 0; } if (!folio_fast_pin_allowed(folio, flags)) { gup_put_folio(folio, refs, flags); return 0; } if (!pte_write(pte) && gup_must_unshare(NULL, flags, &folio->page)) { gup_put_folio(folio, refs, flags); return 0; } *nr += refs; folio_set_referenced(folio); return 1; } static int gup_huge_pd(hugepd_t hugepd, unsigned long addr, unsigned int pdshift, unsigned long end, unsigned int flags, struct page **pages, int *nr) { pte_t *ptep; unsigned long sz = 1UL << hugepd_shift(hugepd); unsigned long next; ptep = hugepte_offset(hugepd, addr, pdshift); do { next = hugepte_addr_end(addr, end, sz); if (!gup_hugepte(ptep, sz, addr, end, flags, pages, nr)) return 0; } while (ptep++, addr = next, addr != end); return 1; } #else static inline int gup_huge_pd(hugepd_t hugepd, unsigned long addr, unsigned int pdshift, unsigned long end, unsigned int flags, struct page **pages, int *nr) { return 0; } #endif /* CONFIG_ARCH_HAS_HUGEPD */ static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { struct page *page; struct folio *folio; int refs; if (!pmd_access_permitted(orig, flags & FOLL_WRITE)) return 0; if (pmd_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; return __gup_device_huge_pmd(orig, pmdp, addr, end, flags, pages, nr); } page = nth_page(pmd_page(orig), (addr & ~PMD_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr); folio = try_grab_folio(page, refs, flags); if (!folio) return 0; if (unlikely(pmd_val(orig) != pmd_val(*pmdp))) { gup_put_folio(folio, refs, flags); return 0; } if (!folio_fast_pin_allowed(folio, flags)) { gup_put_folio(folio, refs, flags); return 0; } if (!pmd_write(orig) && gup_must_unshare(NULL, flags, &folio->page)) { gup_put_folio(folio, refs, flags); return 0; } *nr += refs; folio_set_referenced(folio); return 1; } static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { struct page *page; struct folio *folio; int refs; if (!pud_access_permitted(orig, flags & FOLL_WRITE)) return 0; if (pud_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; return __gup_device_huge_pud(orig, pudp, addr, end, flags, pages, nr); } page = nth_page(pud_page(orig), (addr & ~PUD_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr); folio = try_grab_folio(page, refs, flags); if (!folio) return 0; if (unlikely(pud_val(orig) != pud_val(*pudp))) { gup_put_folio(folio, refs, flags); return 0; } if (!folio_fast_pin_allowed(folio, flags)) { gup_put_folio(folio, refs, flags); return 0; } if (!pud_write(orig) && gup_must_unshare(NULL, flags, &folio->page)) { gup_put_folio(folio, refs, flags); return 0; } *nr += refs; folio_set_referenced(folio); return 1; } static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { int refs; struct page *page; struct folio *folio; if (!pgd_access_permitted(orig, flags & FOLL_WRITE)) return 0; BUILD_BUG_ON(pgd_devmap(orig)); page = nth_page(pgd_page(orig), (addr & ~PGDIR_MASK) >> PAGE_SHIFT); refs = record_subpages(page, addr, end, pages + *nr); folio = try_grab_folio(page, refs, flags); if (!folio) return 0; if (unlikely(pgd_val(orig) != pgd_val(*pgdp))) { gup_put_folio(folio, refs, flags); return 0; } if (!pgd_write(orig) && gup_must_unshare(NULL, flags, &folio->page)) { gup_put_folio(folio, refs, flags); return 0; } if (!folio_fast_pin_allowed(folio, flags)) { gup_put_folio(folio, refs, flags); return 0; } *nr += refs; folio_set_referenced(folio); return 1; } static int gup_pmd_range(pud_t *pudp, pud_t pud, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { unsigned long next; pmd_t *pmdp; pmdp = pmd_offset_lockless(pudp, pud, addr); do { pmd_t pmd = pmdp_get_lockless(pmdp); next = pmd_addr_end(addr, end); if (!pmd_present(pmd)) return 0; if (unlikely(pmd_trans_huge(pmd) || pmd_huge(pmd) || pmd_devmap(pmd))) { /* See gup_pte_range() */ if (pmd_protnone(pmd)) return 0; if (!gup_huge_pmd(pmd, pmdp, addr, next, flags, pages, nr)) return 0; } else if (unlikely(is_hugepd(__hugepd(pmd_val(pmd))))) { /* * architecture have different format for hugetlbfs * pmd format and THP pmd format */ if (!gup_huge_pd(__hugepd(pmd_val(pmd)), addr, PMD_SHIFT, next, flags, pages, nr)) return 0; } else if (!gup_pte_range(pmd, pmdp, addr, next, flags, pages, nr)) return 0; } while (pmdp++, addr = next, addr != end); return 1; } static int gup_pud_range(p4d_t *p4dp, p4d_t p4d, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { unsigned long next; pud_t *pudp; pudp = pud_offset_lockless(p4dp, p4d, addr); do { pud_t pud = READ_ONCE(*pudp); next = pud_addr_end(addr, end); if (unlikely(!pud_present(pud))) return 0; if (unlikely(pud_huge(pud) || pud_devmap(pud))) { if (!gup_huge_pud(pud, pudp, addr, next, flags, pages, nr)) return 0; } else if (unlikely(is_hugepd(__hugepd(pud_val(pud))))) { if (!gup_huge_pd(__hugepd(pud_val(pud)), addr, PUD_SHIFT, next, flags, pages, nr)) return 0; } else if (!gup_pmd_range(pudp, pud, addr, next, flags, pages, nr)) return 0; } while (pudp++, addr = next, addr != end); return 1; } static int gup_p4d_range(pgd_t *pgdp, pgd_t pgd, unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { unsigned long next; p4d_t *p4dp; p4dp = p4d_offset_lockless(pgdp, pgd, addr); do { p4d_t p4d = READ_ONCE(*p4dp); next = p4d_addr_end(addr, end); if (p4d_none(p4d)) return 0; BUILD_BUG_ON(p4d_huge(p4d)); if (unlikely(is_hugepd(__hugepd(p4d_val(p4d))))) { if (!gup_huge_pd(__hugepd(p4d_val(p4d)), addr, P4D_SHIFT, next, flags, pages, nr)) return 0; } else if (!gup_pud_range(p4dp, p4d, addr, next, flags, pages, nr)) return 0; } while (p4dp++, addr = next, addr != end); return 1; } static void gup_pgd_range(unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { unsigned long next; pgd_t *pgdp; pgdp = pgd_offset(current->mm, addr); do { pgd_t pgd = READ_ONCE(*pgdp); next = pgd_addr_end(addr, end); if (pgd_none(pgd)) return; if (unlikely(pgd_huge(pgd))) { if (!gup_huge_pgd(pgd, pgdp, addr, next, flags, pages, nr)) return; } else if (unlikely(is_hugepd(__hugepd(pgd_val(pgd))))) { if (!gup_huge_pd(__hugepd(pgd_val(pgd)), addr, PGDIR_SHIFT, next, flags, pages, nr)) return; } else if (!gup_p4d_range(pgdp, pgd, addr, next, flags, pages, nr)) return; } while (pgdp++, addr = next, addr != end); } #else static inline void gup_pgd_range(unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { }
policy use functions for gup_huge pte policy code function above (not right above, gotta scroll probably to find it)
-
static int internal_get_user_pages_fast(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages) { unsigned long len, end; unsigned long nr_pinned; int locked = 0; int ret; if (WARN_ON_ONCE(gup_flags & ~(FOLL_WRITE | FOLL_LONGTERM | FOLL_FORCE | FOLL_PIN | FOLL_GET | FOLL_FAST_ONLY | FOLL_NOFAULT | FOLL_PCI_P2PDMA | FOLL_HONOR_NUMA_FAULT))) return -EINVAL; if (gup_flags & FOLL_PIN) mm_set_has_pinned_flag(¤t->mm->flags); if (!(gup_flags & FOLL_FAST_ONLY)) might_lock_read(¤t->mm->mmap_lock); start = untagged_addr(start) & PAGE_MASK; len = nr_pages << PAGE_SHIFT; if (check_add_overflow(start, len, &end)) return -EOVERFLOW; if (end > TASK_SIZE_MAX) return -EFAULT; if (unlikely(!access_ok((void __user *)start, len))) return -EFAULT; nr_pinned = lockless_pages_from_mm(start, end, gup_flags, pages); if (nr_pinned == nr_pages || gup_flags & FOLL_FAST_ONLY) return nr_pinned; /* Slow path: try to get the remaining pages with get_user_pages */ start += nr_pinned << PAGE_SHIFT; pages += nr_pinned; ret = __gup_longterm_locked(current->mm, start, nr_pages - nr_pinned, pages, &locked, gup_flags | FOLL_TOUCH | FOLL_UNLOCKABLE); if (ret < 0) { /* * The caller has to unpin the pages we already pinned so * returning -errno is not an option */ if (nr_pinned) return nr_pinned; return ret; } return ret + nr_pinned; } /** * get_user_pages_fast_only() - pin user pages in memory * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * * Like get_user_pages_fast() except it's IRQ-safe in that it won't fall back to * the regular GUP. * * If the architecture does not support this function, simply return with no * pages pinned. * * Careful, careful! COW breaking can go either way, so a non-write * access can get ambiguous page results. If you call this function without * 'write' set, you'd better be sure that you're ok with that ambiguity. */ int get_user_pages_fast_only(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { /* * Internally (within mm/gup.c), gup fast variants must set FOLL_GET, * because gup fast is always a "pin with a +1 page refcount" request. * * FOLL_FAST_ONLY is required in order to match the API description of * this routine: no fall back to regular ("slow") GUP. */ if (!is_valid_gup_args(pages, NULL, &gup_flags, FOLL_GET | FOLL_FAST_ONLY)) return -EINVAL; return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); } EXPORT_SYMBOL_GPL(get_user_pages_fast_only); /** * get_user_pages_fast() - pin user pages in memory * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * * Attempt to pin user pages in memory without taking mm->mmap_lock. * If not successful, it will fall back to taking the lock and * calling get_user_pages(). * * Returns number of pages pinned. This may be fewer than the number requested. * If nr_pages is 0 or negative, returns 0. If no pages were pinned, returns * -errno. */ int get_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { /* * The caller may or may not have explicitly set FOLL_GET; either way is * OK. However, internally (within mm/gup.c), gup fast variants must set * FOLL_GET, because gup fast is always a "pin with a +1 page refcount" * request. */ if (!is_valid_gup_args(pages, NULL, &gup_flags, FOLL_GET)) return -EINVAL; return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); } EXPORT_SYMBOL_GPL(get_user_pages_fast); /** * pin_user_pages_fast() - pin user pages in memory without taking locks * * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * * Nearly the same as get_user_pages_fast(), except that FOLL_PIN is set. See * get_user_pages_fast() for documentation on the function arguments, because * the arguments here are identical. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please * see Documentation/core-api/pin_user_pages.rst for further details. * * Note that if a zero_page is amongst the returned pages, it will not have * pins in it and unpin_user_page() will not remove pins from it. */ int pin_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { if (!is_valid_gup_args(pages, NULL, &gup_flags, FOLL_PIN)) return -EINVAL; return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); } EXPORT_SYMBOL_GPL(pin_user_pages_fast); /** * pin_user_pages_remote() - pin pages of a remote process * * @mm: mm_struct of target mm * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * @locked: pointer to lock flag indicating whether lock is held and * subsequently whether VM_FAULT_RETRY functionality can be * utilised. Lock must initially be held. * * Nearly the same as get_user_pages_remote(), except that FOLL_PIN is set. See * get_user_pages_remote() for documentation on the function arguments, because * the arguments here are identical. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please * see Documentation/core-api/pin_user_pages.rst for details. * * Note that if a zero_page is amongst the returned pages, it will not have * pins in it and unpin_user_page*() will not remove pins from it. */ long pin_user_pages_remote(struct mm_struct *mm, unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, int *locked) { int local_locked = 1; if (!is_valid_gup_args(pages, locked, &gup_flags, FOLL_PIN | FOLL_TOUCH | FOLL_REMOTE)) return 0; return __gup_longterm_locked(mm, start, nr_pages, pages, locked ? locked : &local_locked, gup_flags); } EXPORT_SYMBOL(pin_user_pages_remote); /** * pin_user_pages() - pin user pages in memory for use by other devices * * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * * Nearly the same as get_user_pages(), except that FOLL_TOUCH is not set, and * FOLL_PIN is set. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please * see Documentation/core-api/pin_user_pages.rst for details. * * Note that if a zero_page is amongst the returned pages, it will not have * pins in it and unpin_user_page*() will not remove pins from it. */ long pin_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages) { int locked = 1; if (!is_valid_gup_args(pages, NULL, &gup_flags, FOLL_PIN)) return 0; return __gup_longterm_locked(current->mm, start, nr_pages, pages, &locked, gup_flags); } EXPORT_SYMBOL(pin_user_pages); /* * pin_user_pages_unlocked() is the FOLL_PIN variant of * get_user_pages_unlocked(). Behavior is the same, except that this one sets * FOLL_PIN and rejects FOLL_GET. * * Note that if a zero_page is amongst the returned pages, it will not have * pins in it and unpin_user_page*() will not remove pins from it. */ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages, struct page **pages, unsigned int gup_flags) { int locked = 0; if (!is_valid_gup_args(pages, NULL, &gup_flags, FOLL_PIN | FOLL_TOUCH | FOLL_UNLOCKABLE)) return 0; return __gup_longterm_locked(current->mm, start, nr_pages, pages, &locked, gup_flags); }
fast gup functions
-
/** * unpin_user_pages() - release an array of gup-pinned pages. * @pages: array of pages to be marked dirty and released. * @npages: number of pages in the @pages array. * * For each page in the @pages array, release the page using unpin_user_page(). * * Please see the unpin_user_page() documentation for details. */ void unpin_user_pages(struct page **pages, unsigned long npages) { unsigned long i; struct folio *folio; unsigned int nr; /* * If this WARN_ON() fires, then the system *might* be leaking pages (by * leaving them pinned), but probably not. More likely, gup/pup returned * a hard -ERRNO error to the caller, who erroneously passed it here. */ if (WARN_ON(IS_ERR_VALUE(npages))) return; sanity_check_pinned_pages(pages, npages); for (i = 0; i < npages; i += nr) { folio = gup_folio_next(pages, npages, i, &nr); gup_put_folio(folio, nr, FOLL_PIN); } }
gup unpin function, not actual logic
-
void unpin_user_page_range_dirty_lock(struct page *page, unsigned long npages, bool make_dirty) { unsigned long i; struct folio *folio; unsigned int nr; for (i = 0; i < npages; i += nr) { folio = gup_folio_range_next(page, npages, i, &nr); if (make_dirty && !folio_test_dirty(folio)) { folio_lock(folio); folio_mark_dirty(folio); folio_unlock(folio); } gup_put_folio(folio, nr, FOLL_PIN); } }
unpin logic but for dirty pages
-
static void __maybe_unused undo_dev_pagemap(int *nr, int nr_start, unsigned int flags, struct page **pages) { while ((*nr) - nr_start) { struct page *page = pages[--(*nr)]; ClearPageReferenced(page); if (flags & FOLL_PIN) unpin_user_page(page); else put_page(page); } }
policy use function that undoes mapping
-
#ifdef CONFIG_STACK_GROWSUP return vma_lookup(mm, addr); #else static volatile unsigned long next_warn; struct vm_area_struct *vma; unsigned long now, next; vma = find_vma(mm, addr); if (!vma || (addr >= vma->vm_start)) return vma; /* Only warn for half-way relevant accesses */ if (!(vma->vm_flags & VM_GROWSDOWN)) return NULL; if (vma->vm_start - addr > 65536) return NULL; /* Let's not warn more than once an hour.. */ now = jiffies; next = next_warn; if (next && time_before(now, next)) return NULL; next_warn = now + 60*60*HZ; /* Let people know things may have changed. */ pr_warn("GUP no longer grows the stack in %s (%d): %lx-%lx (%lx)\n", current->comm, task_pid_nr(current), vma->vm_start, vma->vm_end, addr); dump_stack(); return NULL;
helper func to lookup vma(virtual mem area) that warns per hour about half way relevant acc and changes in stack
-
void unpin_user_pages_dirty_lock(struct page **pages, unsigned long npages, bool make_dirty) { unsigned long i; struct folio *folio; unsigned int nr; if (!make_dirty) { unpin_user_pages(pages, npages); return; } sanity_check_pinned_pages(pages, npages); for (i = 0; i < npages; i += nr) { folio = gup_folio_next(pages, npages, i, &nr); /* * Checking PageDirty at this point may race with * clear_page_dirty_for_io(), but that's OK. Two key * cases: * * 1) This code sees the page as already dirty, so it * skips the call to set_page_dirty(). That could happen * because clear_page_dirty_for_io() called * page_mkclean(), followed by set_page_dirty(). * However, now the page is going to get written back, * which meets the original intention of setting it * dirty, so all is well: clear_page_dirty_for_io() goes * on to call TestClearPageDirty(), and write the page * back. * * 2) This code sees the page as clean, so it calls * set_page_dirty(). The page stays dirty, despite being * written back, so it gets written back again in the * next writeback cycle. This is harmless. */ if (!folio_test_dirty(folio)) { folio_lock(folio); folio_mark_dirty(folio); folio_unlock(folio); } gup_put_folio(folio, nr, FOLL_PIN); } }
unpins and dirties page
-
static inline struct folio *gup_folio_next(struct page **list, unsigned long npages, unsigned long i, unsigned int *ntails) { struct folio *folio = page_folio(list[i]); unsigned int nr; for (nr = i + 1; nr < npages; nr++) { if (page_folio(list[nr]) != folio) break; } *ntails = nr - i; return folio; }
gets folio of next page along with reference to end of folio
-
static inline struct folio *gup_folio_range_next(struct page *start, unsigned long npages, unsigned long i, unsigned int *ntails) { struct page *next = nth_page(start, i); struct folio *folio = page_folio(next); unsigned int nr = 1; if (folio_test_large(folio)) nr = min_t(unsigned int, npages - i, folio_nr_pages(folio) - folio_page_idx(folio, next)); *ntails = nr; return folio; }
gets the folio of the next page from start to 'i' range. also gets the tail folio/reference
-
void unpin_user_page(struct page *page) { sanity_check_pinned_pages(&page, 1); gup_put_folio(page_folio(page), 1, FOLL_PIN); } EXPORT_SYMBOL(unpin_user_page);
actual policy use logic
-
if (!put_devmap_managed_page_refs(&folio->page, refs)) folio_put_refs(folio, refs);
Definitely a vital and straightforward policy use section of gup that simples places a reference on the folio
-
if (flags & FOLL_PIN) { if (is_zero_folio(folio)) return; node_stat_mod_folio(folio, NR_FOLL_PIN_RELEASED, refs); if (folio_test_large(folio)) atomic_sub(refs, &folio->_pincount); else refs *= GUP_PIN_COUNTING_BIAS; }
Checks if the folio is zero/large
-
if (folio_test_large(folio)) atomic_add(refs, &folio->_pincount); else folio_ref_add(folio, refs * (GUP_PIN_COUNTING_BIAS - 1))
maintaining reference counts. Part of policy logic most likely
-
- Oct 2024
-
elixir.bootlin.com elixir.bootlin.com
-
if (dtc->wb_thresh < 2 * wb_stat_error()) { wb_reclaimable = wb_stat_sum(wb, WB_RECLAIMABLE); dtc->wb_dirty = wb_reclaimable + wb_stat_sum(wb, WB_WRITEBACK); } else { wb_reclaimable = wb_stat(wb, WB_RECLAIMABLE); dtc->wb_dirty = wb_reclaimable + wb_stat(wb, WB_WRITEBACK); }
This is a configuration policy that does a more accurate calculation on the number of reclaimable pages and dirty pages when the threshold for the dirty pages in the writeback context is lower than 2 times the maximal error of a stat counter.
-
static long wb_min_pause(struct bdi_writeback *wb, long max_pause, unsigned long task_ratelimit, unsigned long dirty_ratelimit, int *nr_dirtied_pause
This function is an algorithmic policy that determines the minimum throttle time for a process between consecutive writeback operations for dirty pages based on heuristics. It is used for balancing the load of the I/O subsystems so that there will not be excessive I/O operations that impact the performance of the system.
-
if (!laptop_mode && nr_reclaimable > gdtc->bg_thresh && !writeback_in_progress(wb)) wb_start_background_writeback(wb);
This is a configuration policy that determines whether to start background writeout. The code here indicates that if laptop_mode, which will reduce disk activity for power saving, is not set, then when the number of dirty pages reaches the bg_thresh threshold, the system starts writing back pages.
-
if (thresh > dirty) return 1UL << (ilog2(thresh - dirty) >> 1);
This implements a configuration policy that determines the interval for the kernel to wake up and check for dirty pages that need to be written back to disk.
-
limit -= (limit - thresh) >> 5;
This is a configuration policy that determines how much should the limit be updated. The limit controls the amount of dirty memory allowed in the system.
-
if (dirty <= dirty_freerun_ceiling(thresh, bg_thresh) && (!mdtc || m_dirty <= dirty_freerun_ceiling(m_thresh, m_bg_thresh))) { unsigned long intv; unsigned long m_intv; free_running: intv = dirty_poll_interval(dirty, thresh); m_intv = ULONG_MAX; current->dirty_paused_when = now; current->nr_dirtied = 0; if (mdtc) m_intv = dirty_poll_interval(m_dirty, m_thresh); current->nr_dirtied_pause = min(intv, m_intv); break; } /* Start writeback even when in laptop mode */ if (unlikely(!writeback_in_progress(wb))) wb_start_background_writeback(wb); mem_cgroup_flush_foreign(wb); /* * Calculate global domain's pos_ratio and select the * global dtc by default. */ if (!strictlimit) { wb_dirty_limits(gdtc); if ((current->flags & PF_LOCAL_THROTTLE) && gdtc->wb_dirty < dirty_freerun_ceiling(gdtc->wb_thresh, gdtc->wb_bg_thresh)) /* * LOCAL_THROTTLE tasks must not be throttled * when below the per-wb freerun ceiling. */ goto free_running; } dirty_exceeded = (gdtc->wb_dirty > gdtc->wb_thresh) && ((gdtc->dirty > gdtc->thresh) || strictlimit); wb_position_ratio(gdtc); sdtc = gdtc; if (mdtc) { /* * If memcg domain is in effect, calculate its * pos_ratio. @wb should satisfy constraints from * both global and memcg domains. Choose the one * w/ lower pos_ratio. */ if (!strictlimit) { wb_dirty_limits(mdtc); if ((current->flags & PF_LOCAL_THROTTLE) && mdtc->wb_dirty < dirty_freerun_ceiling(mdtc->wb_thresh, mdtc->wb_bg_thresh)) /* * LOCAL_THROTTLE tasks must not be * throttled when below the per-wb * freerun ceiling. */ goto free_running; } dirty_exceeded |= (mdtc->wb_dirty > mdtc->wb_thresh) && ((mdtc->dirty > mdtc->thresh) || strictlimit); wb_position_ratio(mdtc); if (mdtc->pos_ratio < gdtc->pos_ratio) sdtc = mdtc; }
This is an algorithmic policy that determines whether the process can run freely or a throttle is needed to control the rate of the writeback by checking if the number of dirty pages exceed the average of the global threshold and background threshold.
-
shift = dirty_ratelimit / (2 * step + 1); if (shift < BITS_PER_LONG) step = DIV_ROUND_UP(step >> shift, 8); else step = 0; if (dirty_ratelimit < balanced_dirty_ratelimit) dirty_ratelimit += step; else dirty_ratelimit -= step;
This is a configuration policy that determines how much we should increase/decrease the dirty_ratelimit, which controls the rate that processors write dirty pages back to storage.
-
ratelimit_pages = dirty_thresh / (num_online_cpus() * 32); if (ratelimit_pages < 16) ratelimit_pages = 16;
This is a configuration policy that dynamically determines the rate that kernel can write dirty pages back to storage in a single writeback cycle.
-
t = wb_dirty / (1 + bw / roundup_pow_of_two(1 + HZ / 8));
This implements a configuration policy that determines the maximum time that the kernel should wait between writeback operations for dirty pages. This ensures that dirty pages are flushed to disk within a reasonable time frame and control the risk of data loss in case of a system crash.
-
if (IS_ENABLED(CONFIG_CGROUP_WRITEBACK) && mdtc) {
This is a configuration policy that controls whether to update the limit in the control group. The config enables support for controlling the writeback of dirty pages on a per-cgroup basis in the Linux kernel. This allows for better resource management and improved performance.
-
-
elixir.bootlin.com elixir.bootlin.com
-
if (si->cluster_info) { if (!scan_swap_map_try_ssd_cluster(si, &offset, &scan_base)) goto scan; } else if (unlikely(!si->cluster_nr--)) {
Algorithmic policy decision: If SSD, use SSD wear-leveling friendly algorithm. Otherwise, use HDD algo which minimizes seek times. Potentially, certain access patterns may make one of these algorithms less effective (i.e. in the case of wear leveling, the swap is constantly full)
-
scan_base = offset = si->lowest_bit; last_in_cluster = offset + SWAPFILE_CLUSTER - 1; /* Locate the first empty (unaligned) cluster */ for (; last_in_cluster <= si->highest_bit; offset++) { if (si->swap_map[offset]) last_in_cluster = offset + SWAPFILE_CLUSTER; else if (offset == last_in_cluster) { spin_lock(&si->lock); offset -= SWAPFILE_CLUSTER - 1; si->cluster_next = offset; si->cluster_nr = SWAPFILE_CLUSTER - 1; goto checks; } if (unlikely(--latency_ration < 0)) { cond_resched(); latency_ration = LATENCY_LIMIT; } }
Here, (when using HDDs), a policy is implemented that places the swapped page in the first available slot. This is supposed to reduce seek time in spinning drives, as it encourages having swap entries near each other.
-
/* * Even if there's no free clusters available (fragmented), * try to scan a little more quickly with lock held unless we * have scanned too many slots already. */ if (!scanned_many) { unsigned long scan_limit; if (offset < scan_base) scan_limit = scan_base; else scan_limit = si->highest_bit; for (; offset <= scan_limit && --latency_ration > 0; offset++) { if (!si->swap_map[offset]) goto checks; } }
Here we have a configuration policy where we do another smaller scan as long as we haven't exhausted our latency_ration. Another alternative could be yielding early in anticipation that we aren't going to find a free slot.
-
/* * select a random position to start with to help wear leveling * SSD */ for_each_possible_cpu(cpu) { per_cpu(*p->cluster_next_cpu, cpu) = get_random_u32_inclusive(1, p->highest_bit); }
An algorithmic (random) policy is used here to spread swap pages around an SSD to help with wear leveling instead of writing to the same area of an SSD often.
-
cluster_list_add_tail(&si->discard_clusters, si->cluster_info, idx);
A policy decision is made to add a cluster to the discard list on a first-come, first-served basis. However, this approach could be enhanced by prioritizing certain clusters higher on the list based on their 'importance.' This 'importance' could be defined by how closely a cluster is related to other pages in the swap. By doing so, the system can reduce seek time as mentioned on line 817.
-
if (unlikely(--latency_ration < 0)) { cond_resched(); latency_ration = LATENCY_LIMIT; scanned_many = true; } if (swap_offset_available_and_locked(si, offset)) goto checks; } offset = si->lowest_bit; while (offset < scan_base) { if (unlikely(--latency_ration < 0)) { cond_resched(); latency_ration = LATENCY_LIMIT; scanned_many = true; }
Here, a policy decision is made to fully replenish the latency_ration with the LATENCY_LIMIT and then yield back to the scheduler if we've exhausted it. This makes it so that when scheduled again, we have the full LATENCY_LIMIT to do a scan. Alternative policies could grow/shrink this to find a better heuristic instead of fully replenishing each time.
Marked config/value as awe're replacing latency_ration with a compiletime-defined limit.
-
while (scan_swap_map_ssd_cluster_conflict(si, offset)) { /* take a break if we already got some slots */ if (n_ret) goto done; if (!scan_swap_map_try_ssd_cluster(si, &offset, &scan_base)) goto scan;
Here, a policy decision is made to stop scanning if some slots were already found. Other policy decisions could be made to keep scanning or take into account how long the scan took or how many pages were found.
-
else if (!cluster_list_empty(&si->discard_clusters)) { /* * we don't have free cluster but have some clusters in * discarding, do discard now and reclaim them, then * reread cluster_next_cpu since we dropped si->lock */ swap_do_scheduled_discard(si); *scan_base = this_cpu_read(*si->cluster_next_cpu); *offset = *scan_base; goto new_cluster;
This algorithmic policy discards + reclaims pages as-needed whenever there is no free cluster. Other policies could be explored that do this preemptively in order to avoid the cost of doing it here.
-
#ifdef CONFIG_THP_SWAP
This is a build-time flag that configures how hugepages are handled when swapped. When defined, it swaps them in one piece, while without it splits them into smaller units and swaps those units.
-
if (swap_flags & SWAP_FLAG_DISCARD_ONCE) p->flags &= ~SWP_PAGE_DISCARD; else if (swap_flags & SWAP_FLAG_DISCARD_PAGES) p->flags &= ~SWP_AREA_DISCARD
This is a configuration policy decision where a sysadmin can pass flags to sys_swapon() to control the behavior discards are handled. If DISCARD_ONCE is set, a flag which "discard[s] swap area at swapon-time" is unset, and if DISCARD_PAGES is set, a flag which "discard[s] page-clusters after use" is unset.
-
-
elixir.bootlin.com elixir.bootlin.com
-
randomize_stack_top
This function uses a configuration policy to enable Address Space Layout Randomization (ASLR) for a specific process if PF_RANDOMIZE flag is set. It randomly arranges the positions of stack of a process to help defend certain attacks by making memory addresses unpredictable.
-
-
elixir.bootlin.com elixir.bootlin.com
-
if (prev_class) { if (can_merge(prev_class, pages_per_zspage, objs_per_zspage)) { pool->size_class[i] = prev_class; continue; } }
This is an algorithmic policy. A
zs_poolmaintainszs_pages of differentsize_class. However, somesize_classes share exactly same characteristics, namelypages_per_zspageandobjs_per_zspage. Recall the other annotation of mine, it searched freezspages bysize_class:zspage = find_get_zspage(class);. Thus, grouping different classes improves memory utilization. -
zspage = find_get_zspage(class); if (likely(zspage)) { obj = obj_malloc(pool, zspage, handle); /* Now move the zspage to another fullness group, if required */ fix_fullness_group(class, zspage); record_obj(handle, obj); class_stat_inc(class, ZS_OBJS_INUSE, 1); goto out; }
This is an algorithmic policy. Instead of immediately allocating new zspages for each memory request, the algorithm first attempts to find and reuse existing partially filled zspages from a given size class by invoking the find_get_zspage(class) function. It also updates the corresponding fullness groups.
-
-
elixir.bootlin.com elixir.bootlin.com
-
1
This is a configuration policy that sets the timeout between retries if vmap_pages_range() fails. This could be tunable variable.
-
if (!(flags & VM_ALLOC)) area->addr = kasan_unpoison_vmalloc(area->addr, requested_size, KASAN_VMALLOC_PROT_NORMAL);
This is an algorithmic policy. This is an optimization that prevents duplicate marks of accessibility. Only pages allocated without
VM_ALLOC(e.g. ioremap) was not set accessible (unpoison), thus requiring explicit setting here. -
100U
This is an configuration policy that determines 100 pages are the upper limit for the bulk-allocator. However, the implementation of
alloc_pages_bulk_array_mempolicydoes not explicitly limit in the implementation. So I believe it is an algorithmic policy related to some sort of optimization. -
if (!order) {
This is an algorithmic policy determines that only use the bulk allocator for order-0 pages (non-super pages). Maybe the bulk allocator could be applied to super pages to speed up allocation. Currently I haven't seen the reason why it cannot be applied.
-
if (likely(count <= VMAP_MAX_ALLOC)) { mem = vb_alloc(size, GFP_KERNEL); if (IS_ERR(mem)) return NULL; addr = (unsigned long)mem; } else { struct vmap_area *va; va = alloc_vmap_area(size, PAGE_SIZE, VMALLOC_START, VMALLOC_END, node, GFP_KERNEL, VMAP_RAM); if (IS_ERR(va)) return NULL; addr = va->va_start; mem = (void *)addr; }
This is an algorithmic policy that determines whether to use a more efficient
vl_allocwhich does not involve complex virtual-to-physical mappings. Unlike the latteralloc_vmap_area,vb_allocdoes not need to traverse the rb-tree of free vmap areas. It simply find a larger enough block fromvmap_block_queue. -
VMAP_PURGE_THRESHOLD
The threshold VMAP_PURGE_THRESHOLD is a configuration policy that could be tuned by machine learning. Setting this threshold lower reduces purging activities while setting it higher reduces framentation.
-
if (!(force_purge
This is an algorithmic policy that prevents purging blocks with considerable amount of "usable memory" unless requested with force_purge.
-
resched_threshold = lazy_max_pages() << 1;
The assignment of resched_threshold and lines 1776-1777 are configuration policies to determine fewer than which number of lazily-freed pages it should yield CPU temporarily to higher-priority tasks.
-
log = fls(num_online_cpus());
This heuristic scales lazy_max_pages logarithmically, which is a configuration policy. Alternatively, machine learning could determine the optimal scaling function—whether linear, logarithmic, square-root, or another approach.
-
32UL * 1024 * 1024
This is a configuration policy that decides to always returns multiples of 32 MB worth of pages. This could be a configurable variable rather than a fixed magic number.
-
- Sep 2024
-
elixir.bootlin.com elixir.bootlin.com
-
if (lruvec->file_cost + lruvec->anon_cost > lrusize / 4) { lruvec->file_cost /= 2; lruvec->anon_cost /= 2; }
It is a configuration policy. The policy here is to adjust the cost of current page. The cost means the overhead for kernel to operate the page if the page is swapped out. Kernel adopts a decay policy. Specifically, if current cost is greater than lrusize/4, its cost will be divided by 2. If kernel has no this policy, after a long-term running, the cost of each page is very high. Kernel might mistakenly reserve pages which are frequently visited long time ago but are inactive currently. It might cause performance degradation. The value (lrusize/4) here has a trade-off between performance and sensitivity. For example, if the value is too small, kernel will frequently adjust the priority of each page, resulting in process's performance degradation. If the value is too large, kernel might be misleaded by historical data, causing wrongly swapping currently popular pages and further performance degradation.
-
if (megs < 16) page_cluster = 2; else page_cluster = 3;
It is a configuration policy. The policy here is to determine the size of page cluster. "2" and "3" is determined externally to the kernel. Page cluster is the actual execution unit when swapping. If the machine is small-memory, kernel executes swap operation on 4 (2^2) pages for each time. Otherwise, kernel operats 8 (2^3) for each time. The rational here is to avoid much memory pressure for small-memory system and to improve performance for large-memory system.
-
-
elixir.bootlin.com elixir.bootlin.com
-
mod_memcg_page_state(area->pages[i], MEMCG_VMALLOC, 1);
It is a configuration policy. It is used to count the physical page for memory control group. The policy here is one physical page corresponds to one count to the memory control group. If using huge page, the value here might not be 1,
-
schedule_timeout_uninterruptible(1);
It is a configuration policy. If kernel cannot allocate an enough virtual space with alignment, while nofail is specified to disallow failure, the kernel will let the process to sleep for 1 time slot. In this period, the process cannot be interrupted and quit the schedule queue of CPU. The "1" here is a configuration parameter, made externally to the kernel. It is not large enough for latency-sensitive process and is not small enough to retry frequently.
-
if (array_size > PAGE_SIZE) { area->pages = __vmalloc_node(array_size, 1, nested_gfp, node, area->caller); } else { area->pages = kmalloc_node(array_size, nested_gfp, node); }
It is an algorithmic policy and also a configuration policy. The algorithmic policy here is to maxmize the data locality of virtual memory address, by letting the space be in continuous virtual pages. If the demanded size of virtual memory is larger than one page, the kernel will allocate multiple continuous pages for it by using __vmalloc_node(). Otherwise, the kernel will call kmalloc_node() to place it in a single page. On the other hand, to allocate continuous pages, we should invoke __vmalloc_node() by declaring align = 1, which is a configuration policy.
-
if (!counters) return; if (v->flags & VM_UNINITIALIZED) return;
It is an algorithmic policy. The show_numa_info() is used for printing how the virtual memory of v (vm_struct) distributes its pages across numa nodes. The two if conditions are used for filtering the invalid v. The first if means the virtual memory has not been associated with physical page. The second if means the virtual memory has not been initialized yet.
-
if (page) copied = copy_page_to_iter_nofault(page, offset, length, iter); else copied = zero_iter(iter, length);
It is an algorithmic policy. The policy here is to avoid lock/unlock overhead by using copy_page_to_iter_nofault() function plus if-else branch. Because copy_page_to_iter_nofault() is based on no page fault, we need to check the physical page is still alive. So the kernel uses an if condition to guarantee copy_page_to_iter_nofault() without page fault. On the other hand, if the page is not valid, kernel chooses to fill the iter using zero value instead of returning an error. I think it is because it can make invocation more convenient. The caller does not need to do a verification of the return value and does not need further actions to handle it.
-
if (base + last_end < vmalloc_start + last_end) goto overflow; /* * Fitting base has not been found. */ if (va == NULL) goto overflow;
It is an algorithmic policy. The two if conditions are used for checking the feasibility of the allocation. The first if means the adress of the allocation is beyond the allowable address, causing overflow. In this case, since the base address is decided from higher one to smaller one, reducing base address cannot work. Note that the base address is dependent on the va. The second if means we don't find a valid va which has appropriate base address. So the policy here is going to overflow branch to re-allocate va list.
-
if (base + start < va->va_start) { va = node_to_va(rb_prev(&va->rb_node)); base = pvm_determine_end_from_reverse(&va, align) - end; term_area = area; continue; }
It is an algorithmic policy. It is used for checking whether the start address of current va is valid. The policy here is to change a new va, different from the solution of "end address check" (explained in my other annotation). The reason is that we find the base address from high address to low address. If we continue to reduce the base address, the if condition will be always wrong. Changing a new va with a lower start address is the only solution.
-
if (base + end > va->va_end) { base = pvm_determine_end_from_reverse(&va, align) - end; term_area = area; continue; }
It is an algorithmic policy. The outer loop is an endless loop until finding a vaild virtual memory space satisfying the requirements. The if condition here is used to guarantee the space in va is large enough. If not, we continue to scan the memory space from high address to low address while satisfying the alignment requirement. Specifically, tuning the base address to a lower address. And then, it will retry for verification.
-
va = kmem_cache_zalloc(vmap_area_cachep, GFP_NOWAIT); if (WARN_ON_ONCE(!va)) continue;
It is an algorithmic policy. The for loop is used for transferring the virtual memory managment by using from vm_struct (link list) to vm_area (tree?). Inside the for loop, if we cannot allocate a new vmap_area from vmap_area_cachep (va = NULL), the program will simply print a warning and skip current tmp, instread of retry or anything to guarantee the integrity.
-
if (!spin_trylock(&vmap_area_lock)) return false;
It is an algorithmic policy. The vmalloc_dump_obj() function is used for printing the information of specified virtual memory. Due to concurrency, before manipulating the critical state, we must obtain the lock. The policy here is spin_trylock. It only tries once. If fails, the function will simply returns false. I think the policy here is to avoid long-term waiting to improve the performance.
-
-
elixir.bootlin.com elixir.bootlin.com
-
static inline unsigned int calc_slab_order(unsigned int size, unsigned int min_order, unsigned int max_order, unsigned int fract_leftover) { unsigned int order; for (order = min_order; order <= max_order; order++) { unsigned int slab_size = (unsigned int)PAGE_SIZE << order; unsigned int rem; rem = slab_size % size; if (rem <= slab_size / fract_leftover) break; } return order; }
Code to choose how many pages to allocate for a new slab to minimize wasted space from the remainder
-
-
elixir.bootlin.com elixir.bootlin.com
-
if (folio) { pgoff_t start; rcu_read_lock(); start = page_cache_next_miss(ractl->mapping, index + 1, max_pages); rcu_read_unlock(); if (!start || start - index > max_pages) return; ra->start = start; ra->size = start - index; /* old async_size */ ra->size += req_size; ra->size = get_next_ra_size(ra, max_pages); ra->async_size = ra->size; goto readit;
Read ahead policy for marked folios
-
- Aug 2024
-
elixir.bootlin.com elixir.bootlin.com
-
if (hugetlb_cgroup_disabled())
This probably not an interesting policy decision for ldos. It is a feature flag for the running OS. But if cgroups were decided by policy then this flag would be controlled by the cgroup decision.
-
-
git.doit.wisc.edu git.doit.wisc.edu
-
static long get_nr_to_scan(struct lruvec *lruvec, struct scan_control *sc, bool can_swap)
figure out how many pages to scan.
-
static void prepare_scan_control(pg_data_t *pgdat, struct scan_control *sc) {
sets up the struct scan_control. Most of the value come from elsewhere but this function seems to bring it all together.
-
-
git.doit.wisc.edu git.doit.wisc.edu
-
ksm_thread_pages_to_scan
how many pages to scan. This is what many of the other config values are trying to get at.
-
function to compute a EWA for the number of pages to scan. Uses many of the config parameters from sysfs
-
- Jul 2024
-
git.doit.wisc.edu git.doit.wisc.edu
-
static void scan_time_advisor(void)
This function is calculating the pages to scan based on a other metrics such as cpu consumed. Seems like a good place for a learned algorithm.
-
- Jun 2024
-
git.doit.wisc.edu git.doit.wisc.edu
-
/* * The larger the object size is, the more slabs we want on the partial * list to avoid pounding the page allocator excessively. */ s->min_partial = min_t(unsigned long, MAX_PARTIAL, ilog2(s->size) / 2); s->min_partial = max_t(unsigned long, MIN_PARTIAL, s->min_partial);
A policy decision about how often we may have to go to the page allocator.
-
/* * calculate_sizes() determines the order and the distribution of data within * a slab object. */ 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 static int calculate_sizes(struct kmem_cache *s) {
computes a several values for the allocator based on the size and flags of the allocator being created.
-
#ifndef CONFIG_SLUB_TINY static inline int alloc_kmem_cache_cpus(struct kmem_cache *s)
Depending on the CONFIG_SLUB_TINY should ther be an active slab for each CPU?
-
static inline int calculate_order(unsigned int size) { unsigned int order; unsigned int min_objects; unsigned int max_objects; unsigned int min_order; min_objects = slub_min_objects; if (!min_objects) {
calculate the order (power of two number of pages) that each slab in this allocator should have.
-
max_objects = max(order_objects(slub_max_order, size), 1U); min_objects = min(min_objects, max_objects); min_order = max_t(unsigned int, slub_min_order, get_order(min_objects * size)); if (order_objects(min_order, size) > MAX_OBJS_PER_PAGE) return get_order(size * MAX_OBJS_PER_PAGE) - 1;
Policy calculation
-
s->flags & __OBJECT_POISON
apply policy
-
s->flags & SLAB_RED_ZONE)
debug plicy
-
if (s->flags & __CMPXCHG_DOUBLE) {
Config fastpath?
-
if (s->flags & __CMPXCHG_DOUBLE) {
Fast path config
-
set the number of slabs per cpu
-
-
git.doit.wisc.edu git.doit.wisc.edu
-
int calculate_normal_threshold(struct zone *zone) { int threshold; int mem; /* memory in 128 MB units */ /* * The threshold scales with the number of processors and the amount * of memory per zone. More memory means that we can defer updates for * longer, more processors could lead to more contention. * fls() is used to have a cheap way of logarithmic scaling. * * Some sample thresholds: * * Threshold Processors (fls) Zonesize fls(mem)+1 * ------------------------------------------------------------------ * 8 1 1 0.9-1 GB 4 * 16 2 2 0.9-1 GB 4 * 20 2 2 1-2 GB 5 * 24 2 2 2-4 GB 6 * 28 2 2 4-8 GB 7 * 32 2 2 8-16 GB 8 * 4 2 2 <128M 1 * 30 4 3 2-4 GB 5 * 48 4 3 8-16 GB 8 * 32 8 4 1-2 GB 4 * 32 8 4 0.9-1GB 4 * 10 16 5 <128M 1 * 40 16 5 900M 4 * 70 64 7 2-4 GB 5 * 84 64 7 4-8 GB 6 * 108 512 9 4-8 GB 6 * 125 1024 10 8-16 GB 8 * 125 1024 10 16-32 GB 9 */ mem = zone_managed_pages(zone) >> (27 - PAGE_SHIFT); threshold = 2 * fls(num_online_cpus()) * (1 + fls(mem)); /* * Maximum threshold is 125 */ threshold = min(125, threshold); return threshold; }
a "magic" formula for computing the amount of memory per zone.
-
- May 2024
-
git.doit.wisc.edu git.doit.wisc.edu
-
extern struct cma cma_areas[MAX_CMA_AREAS];
policy use?
-
-
git.doit.wisc.edu git.doit.wisc.edu
-
ARRAY_SIZE(cma_areas)
Policy use - cma_areas size
-
ARRAY_SIZE(cma_areas)
Policy use - cma_areas size
-
cma_bitmap_maxno(cma)
policy application -= maxno of bitmaps
-
-
git.doit.wisc.edu git.doit.wisc.edu
-
mapsize = PAGE_ALIGN(mem_section_usage_size()) >> PAGE_SHIFT;
Policy application: choosing map size for page usage?
-