From 73449daf8f0db2697fb9f3f60b3a6d5c76a3c5fc Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Fri, 13 Jul 2018 21:49:50 -0700 Subject: filesystem-dax: Set page->index In support of enabling memory_failure() handling for filesystem-dax mappings, set ->index to the pgoff of the page. The rmap implementation requires ->index to bound the search through the vma interval tree. The index is set and cleared at dax_associate_entry() and dax_disassociate_entry() time respectively. Cc: Christoph Hellwig Cc: Matthew Wilcox Cc: Ross Zwisler Reviewed-by: Jan Kara Signed-off-by: Dan Williams Signed-off-by: Dave Jiang --- fs/dax.c | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 641192808bb6..4de11ed463ce 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -319,18 +319,27 @@ static unsigned long dax_radix_end_pfn(void *entry) for (pfn = dax_radix_pfn(entry); \ pfn < dax_radix_end_pfn(entry); pfn++) -static void dax_associate_entry(void *entry, struct address_space *mapping) +/* + * TODO: for reflink+dax we need a way to associate a single page with + * multiple address_space instances at different linear_page_index() + * offsets. + */ +static void dax_associate_entry(void *entry, struct address_space *mapping, + struct vm_area_struct *vma, unsigned long address) { - unsigned long pfn; + unsigned long size = dax_entry_size(entry), pfn, index; + int i = 0; if (IS_ENABLED(CONFIG_FS_DAX_LIMITED)) return; + index = linear_page_index(vma, address & ~(size - 1)); for_each_mapped_pfn(entry, pfn) { struct page *page = pfn_to_page(pfn); WARN_ON_ONCE(page->mapping); page->mapping = mapping; + page->index = index + i++; } } @@ -348,6 +357,7 @@ static void dax_disassociate_entry(void *entry, struct address_space *mapping, WARN_ON_ONCE(trunc && page_ref_count(page) > 1); WARN_ON_ONCE(page->mapping && page->mapping != mapping); page->mapping = NULL; + page->index = 0; } } @@ -701,7 +711,7 @@ static void *dax_insert_mapping_entry(struct address_space *mapping, new_entry = dax_radix_locked_entry(pfn, flags); if (dax_entry_size(entry) != dax_entry_size(new_entry)) { dax_disassociate_entry(entry, mapping, false); - dax_associate_entry(new_entry, mapping); + dax_associate_entry(new_entry, mapping, vmf->vma, vmf->address); } if (dax_is_zero_entry(entry) || dax_is_empty_entry(entry)) { -- cgit v1.2.3 From c2a7d2a115525d3501d38e23d24875a79a07e15e Mon Sep 17 00:00:00 2001 From: Dan Williams Date: Fri, 13 Jul 2018 21:50:16 -0700 Subject: filesystem-dax: Introduce dax_lock_mapping_entry() In preparation for implementing support for memory poison (media error) handling via dax mappings, implement a lock_page() equivalent. Poison error handling requires rmap and needs guarantees that the page->mapping association is maintained / valid (inode not freed) for the duration of the lookup. In the device-dax case it is sufficient to simply hold a dev_pagemap reference. In the filesystem-dax case we need to use the entry lock. Export the entry lock via dax_lock_mapping_entry() that uses rcu_read_lock() to protect against the inode being freed, and revalidates the page->mapping association under xa_lock(). Cc: Christoph Hellwig Cc: Matthew Wilcox Cc: Ross Zwisler Cc: Jan Kara Signed-off-by: Dan Williams Signed-off-by: Dave Jiang --- fs/dax.c | 109 +++++++++++++++++++++++++++++++++++++++++++++++++--- include/linux/dax.h | 13 +++++++ 2 files changed, 116 insertions(+), 6 deletions(-) (limited to 'fs') diff --git a/fs/dax.c b/fs/dax.c index 4de11ed463ce..57ec272038da 100644 --- a/fs/dax.c +++ b/fs/dax.c @@ -226,8 +226,8 @@ static inline void *unlock_slot(struct address_space *mapping, void **slot) * * Must be called with the i_pages lock held. */ -static void *get_unlocked_mapping_entry(struct address_space *mapping, - pgoff_t index, void ***slotp) +static void *__get_unlocked_mapping_entry(struct address_space *mapping, + pgoff_t index, void ***slotp, bool (*wait_fn)(void)) { void *entry, **slot; struct wait_exceptional_entry_queue ewait; @@ -237,6 +237,8 @@ static void *get_unlocked_mapping_entry(struct address_space *mapping, ewait.wait.func = wake_exceptional_entry_func; for (;;) { + bool revalidate; + entry = __radix_tree_lookup(&mapping->i_pages, index, NULL, &slot); if (!entry || @@ -251,14 +253,31 @@ static void *get_unlocked_mapping_entry(struct address_space *mapping, prepare_to_wait_exclusive(wq, &ewait.wait, TASK_UNINTERRUPTIBLE); xa_unlock_irq(&mapping->i_pages); - schedule(); + revalidate = wait_fn(); finish_wait(wq, &ewait.wait); xa_lock_irq(&mapping->i_pages); + if (revalidate) + return ERR_PTR(-EAGAIN); } } -static void dax_unlock_mapping_entry(struct address_space *mapping, - pgoff_t index) +static bool entry_wait(void) +{ + schedule(); + /* + * Never return an ERR_PTR() from + * __get_unlocked_mapping_entry(), just keep looping. + */ + return false; +} + +static void *get_unlocked_mapping_entry(struct address_space *mapping, + pgoff_t index, void ***slotp) +{ + return __get_unlocked_mapping_entry(mapping, index, slotp, entry_wait); +} + +static void unlock_mapping_entry(struct address_space *mapping, pgoff_t index) { void *entry, **slot; @@ -277,7 +296,7 @@ static void dax_unlock_mapping_entry(struct address_space *mapping, static void put_locked_mapping_entry(struct address_space *mapping, pgoff_t index) { - dax_unlock_mapping_entry(mapping, index); + unlock_mapping_entry(mapping, index); } /* @@ -374,6 +393,84 @@ static struct page *dax_busy_page(void *entry) return NULL; } +static bool entry_wait_revalidate(void) +{ + rcu_read_unlock(); + schedule(); + rcu_read_lock(); + + /* + * Tell __get_unlocked_mapping_entry() to take a break, we need + * to revalidate page->mapping after dropping locks + */ + return true; +} + +bool dax_lock_mapping_entry(struct page *page) +{ + pgoff_t index; + struct inode *inode; + bool did_lock = false; + void *entry = NULL, **slot; + struct address_space *mapping; + + rcu_read_lock(); + for (;;) { + mapping = READ_ONCE(page->mapping); + + if (!dax_mapping(mapping)) + break; + + /* + * In the device-dax case there's no need to lock, a + * struct dev_pagemap pin is sufficient to keep the + * inode alive, and we assume we have dev_pagemap pin + * otherwise we would not have a valid pfn_to_page() + * translation. + */ + inode = mapping->host; + if (S_ISCHR(inode->i_mode)) { + did_lock = true; + break; + } + + xa_lock_irq(&mapping->i_pages); + if (mapping != page->mapping) { + xa_unlock_irq(&mapping->i_pages); + continue; + } + index = page->index; + + entry = __get_unlocked_mapping_entry(mapping, index, &slot, + entry_wait_revalidate); + if (!entry) { + xa_unlock_irq(&mapping->i_pages); + break; + } else if (IS_ERR(entry)) { + WARN_ON_ONCE(PTR_ERR(entry) != -EAGAIN); + continue; + } + lock_slot(mapping, slot); + did_lock = true; + xa_unlock_irq(&mapping->i_pages); + break; + } + rcu_read_unlock(); + + return did_lock; +} + +void dax_unlock_mapping_entry(struct page *page) +{ + struct address_space *mapping = page->mapping; + struct inode *inode = mapping->host; + + if (S_ISCHR(inode->i_mode)) + return; + + unlock_mapping_entry(mapping, page->index); +} + /* * Find radix tree entry at given index. If it points to an exceptional entry, * return it with the radix tree entry locked. If the radix tree doesn't diff --git a/include/linux/dax.h b/include/linux/dax.h index deb0f663252f..450b28db9533 100644 --- a/include/linux/dax.h +++ b/include/linux/dax.h @@ -88,6 +88,8 @@ int dax_writeback_mapping_range(struct address_space *mapping, struct block_device *bdev, struct writeback_control *wbc); struct page *dax_layout_busy_page(struct address_space *mapping); +bool dax_lock_mapping_entry(struct page *page); +void dax_unlock_mapping_entry(struct page *page); #else static inline bool bdev_dax_supported(struct block_device *bdev, int blocksize) @@ -119,6 +121,17 @@ static inline int dax_writeback_mapping_range(struct address_space *mapping, { return -EOPNOTSUPP; } + +static inline bool dax_lock_mapping_entry(struct page *page) +{ + if (IS_DAX(page->mapping->host)) + return true; + return false; +} + +static inline void dax_unlock_mapping_entry(struct page *page) +{ +} #endif int dax_read_lock(void); -- cgit v1.2.3