From: Mike Kravetz mike.kravetz@oracle.com
commit fd4aed8d985a3236d0877ff6d0c80ad39d4ce81a upstream.
Ackerley Tng reported an issue with hugetlbfs fallocate as noted in the Closes tag. The issue showed up after the conversion of hugetlb page cache lookup code to use page_cache_next_miss. User visible effects are:
- hugetlbfs fallocate incorrectly returns -EEXIST if pages are presnet in the file. - hugetlb pages will not be included in core dumps if they need to be brought in via GUP. - userfaultfd UFFDIO_COPY will not notice pages already present in the cache. It may try to allocate a new page and potentially return ENOMEM as opposed to EEXIST.
Revert the use page_cache_next_miss() in hugetlb code.
IMPORTANT NOTE FOR STABLE BACKPORTS: This patch will apply cleanly to v6.3. However, due to the change of filemap_get_folio() return values, it will not function correctly. This patch must be modified for stable backports.
[dan.carpenter@linaro.org: fix hugetlbfs_pagecache_present()] Link: https://lkml.kernel.org/r/efa86091-6a2c-4064-8f55-9b44e1313015@moroto.mounta... Link: https://lkml.kernel.org/r/20230621212403.174710-2-mike.kravetz@oracle.com Fixes: d0ce0e47b323 ("mm/hugetlb: convert hugetlb fault paths to use alloc_hugetlb_folio()") Signed-off-by: Mike Kravetz mike.kravetz@oracle.com Signed-off-by: Dan Carpenter dan.carpenter@linaro.org Reported-by: Ackerley Tng ackerleytng@google.com Closes: https://lore.kernel.org/linux-mm/cover.1683069252.git.ackerleytng@google.com Reviewed-by: Sidhartha Kumar sidhartha.kumar@oracle.com Cc: Erdem Aktas erdemaktas@google.com Cc: Greg Kroah-Hartman gregkh@linuxfoundation.org Cc: Matthew Wilcox willy@infradead.org Cc: Muchun Song songmuchun@bytedance.com Cc: Vishal Annapurve vannapurve@google.com Signed-off-by: Andrew Morton akpm@linux-foundation.org Signed-off-by: Sidhartha Kumar sidhartha.kumar@oracle.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org --- fs/hugetlbfs/inode.c | 8 +++----- mm/hugetlb.c | 12 ++++++------ 2 files changed, 9 insertions(+), 11 deletions(-)
--- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -821,7 +821,6 @@ static long hugetlbfs_fallocate(struct f */ struct folio *folio; unsigned long addr; - bool present;
cond_resched();
@@ -845,10 +844,9 @@ static long hugetlbfs_fallocate(struct f mutex_lock(&hugetlb_fault_mutex_table[hash]);
/* See if already present in mapping to avoid alloc/free */ - rcu_read_lock(); - present = page_cache_next_miss(mapping, index, 1) != index; - rcu_read_unlock(); - if (present) { + folio = filemap_get_folio(mapping, index); + if (!IS_ERR(folio)) { + folio_put(folio); mutex_unlock(&hugetlb_fault_mutex_table[hash]); hugetlb_drop_vma_policy(&pseudo_vma); continue; --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -5731,13 +5731,13 @@ static bool hugetlbfs_pagecache_present( { struct address_space *mapping = vma->vm_file->f_mapping; pgoff_t idx = vma_hugecache_offset(h, vma, address); - bool present; + struct folio *folio;
- rcu_read_lock(); - present = page_cache_next_miss(mapping, idx, 1) != idx; - rcu_read_unlock(); - - return present; + folio = filemap_get_folio(mapping, idx); + if (IS_ERR(folio)) + return false; + folio_put(folio); + return true; }
int hugetlb_add_to_page_cache(struct folio *folio, struct address_space *mapping,