summaryrefslogtreecommitdiff
path: root/mm/shmem.c
diff options
context:
space:
mode:
authorHugh Dickins <hughd@google.com>2023-09-30 06:32:40 +0300
committerAndrew Morton <akpm@linux-foundation.org>2023-10-19 00:34:13 +0300
commit3022fd7af9604d44ec43da8a4398872989599b18 (patch)
tree8ab0858bdd7ffc7b3a91479c28d2c4a1986df7c2 /mm/shmem.c
parent054a9f7ccd0a60607fb9bbe1e06ca671494971bf (diff)
downloadlinux-3022fd7af9604d44ec43da8a4398872989599b18.tar.xz
shmem: _add_to_page_cache() before shmem_inode_acct_blocks()
There has been a recurring problem, that when a tmpfs volume is being filled by racing threads, some fail with ENOSPC (or consequent SIGBUS or EFAULT) even though all allocations were within the permitted size. This was a problem since early days, but magnified and complicated by the addition of huge pages. We have often worked around it by adding some slop to the tmpfs size, but it's hard to say how much is needed, and some users prefer not to do that e.g. keeping sparse files in a tightly tailored tmpfs helps to prevent accidental writing to holes. This comes from the allocation sequence: 1. check page cache for existing folio 2. check and reserve from vm_enough_memory 3. check and account from size of tmpfs 4. if huge, check page cache for overlapping folio 5. allocate physical folio, huge or small 6. check and charge from mem cgroup limit 7. add to page cache (but maybe another folio already got in). Concurrent tasks allocating at the same position could deplete the size allowance and fail. Doing vm_enough_memory and size checks before the folio allocation was intentional (to limit the load on the page allocator from this source) and still has some virtue; but memory cgroup never did that, so I think it's better reordered to favour predictable behaviour. 1. check page cache for existing folio 2. if huge, check page cache for overlapping folio 3. allocate physical folio, huge or small 4. check and charge from mem cgroup limit 5. add to page cache (but maybe another folio already got in) 6. check and reserve from vm_enough_memory 7. check and account from size of tmpfs. The folio lock held from allocation onwards ensures that the !uptodate folio cannot be used by others, and can safely be deleted from the cache if checks 6 or 7 subsequently fail (and those waiting on folio lock already check that the folio was not truncated once they get the lock); and the early addition to page cache ensures that racers find it before they try to duplicate the accounting. Seize the opportunity to tidy up shmem_get_folio_gfp()'s ENOSPC retrying, which can be combined inside the new shmem_alloc_and_add_folio(): doing 2 splits twice (once huge, once nonhuge) is not exactly equivalent to trying 5 splits (and giving up early on huge), but let's keep it simple unless more complication proves necessary. Userfaultfd is a foreign country: they do things differently there, and for good reason - to avoid mmap_lock deadlock. Leave ordering in shmem_mfill_atomic_pte() untouched for now, but I would rather like to mesh it better with shmem_get_folio_gfp() in the future. Link: https://lkml.kernel.org/r/22ddd06-d919-33b-1219-56335c1bf28e@google.com Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: Carlos Maiolino <cem@kernel.org> Cc: Christian Brauner <brauner@kernel.org> Cc: Chuck Lever <chuck.lever@oracle.com> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Dave Chinner <dchinner@redhat.com> Cc: Jan Kara <jack@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Tim Chen <tim.c.chen@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Diffstat (limited to 'mm/shmem.c')
-rw-r--r--mm/shmem.c229
1 files changed, 118 insertions, 111 deletions
diff --git a/mm/shmem.c b/mm/shmem.c
index 670f77de3238..269cd3c1110f 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -789,13 +789,11 @@ static int shmem_add_to_page_cache(struct folio *folio,
xas_store(&xas, folio);
if (xas_error(&xas))
goto unlock;
- if (folio_test_pmd_mappable(folio)) {
- count_vm_event(THP_FILE_ALLOC);
+ if (folio_test_pmd_mappable(folio))
__lruvec_stat_mod_folio(folio, NR_SHMEM_THPS, nr);
- }
- mapping->nrpages += nr;
__lruvec_stat_mod_folio(folio, NR_FILE_PAGES, nr);
__lruvec_stat_mod_folio(folio, NR_SHMEM, nr);
+ mapping->nrpages += nr;
unlock:
xas_unlock_irq(&xas);
} while (xas_nomem(&xas, gfp));
@@ -1612,25 +1610,17 @@ static struct folio *shmem_alloc_hugefolio(gfp_t gfp,
struct shmem_inode_info *info, pgoff_t index)
{
struct vm_area_struct pvma;
- struct address_space *mapping = info->vfs_inode.i_mapping;
- pgoff_t hindex;
struct folio *folio;
- hindex = round_down(index, HPAGE_PMD_NR);
- if (xa_find(&mapping->i_pages, &hindex, hindex + HPAGE_PMD_NR - 1,
- XA_PRESENT))
- return NULL;
-
- shmem_pseudo_vma_init(&pvma, info, hindex);
+ shmem_pseudo_vma_init(&pvma, info, index);
folio = vma_alloc_folio(gfp, HPAGE_PMD_ORDER, &pvma, 0, true);
shmem_pseudo_vma_destroy(&pvma);
- if (!folio)
- count_vm_event(THP_FILE_FALLBACK);
+
return folio;
}
static struct folio *shmem_alloc_folio(gfp_t gfp,
- struct shmem_inode_info *info, pgoff_t index)
+ struct shmem_inode_info *info, pgoff_t index)
{
struct vm_area_struct pvma;
struct folio *folio;
@@ -1642,36 +1632,101 @@ static struct folio *shmem_alloc_folio(gfp_t gfp,
return folio;
}
-static struct folio *shmem_alloc_and_acct_folio(gfp_t gfp, struct inode *inode,
- pgoff_t index, bool huge)
+static struct folio *shmem_alloc_and_add_folio(gfp_t gfp,
+ struct inode *inode, pgoff_t index,
+ struct mm_struct *fault_mm, bool huge)
{
+ struct address_space *mapping = inode->i_mapping;
struct shmem_inode_info *info = SHMEM_I(inode);
struct folio *folio;
- int nr;
- int err;
+ long pages;
+ int error;
if (!IS_ENABLED(CONFIG_TRANSPARENT_HUGEPAGE))
huge = false;
- nr = huge ? HPAGE_PMD_NR : 1;
- err = shmem_inode_acct_blocks(inode, nr);
- if (err)
- goto failed;
+ if (huge) {
+ pages = HPAGE_PMD_NR;
+ index = round_down(index, HPAGE_PMD_NR);
+
+ /*
+ * Check for conflict before waiting on a huge allocation.
+ * Conflict might be that a huge page has just been allocated
+ * and added to page cache by a racing thread, or that there
+ * is already at least one small page in the huge extent.
+ * Be careful to retry when appropriate, but not forever!
+ * Elsewhere -EEXIST would be the right code, but not here.
+ */
+ if (xa_find(&mapping->i_pages, &index,
+ index + HPAGE_PMD_NR - 1, XA_PRESENT))
+ return ERR_PTR(-E2BIG);
- if (huge)
folio = shmem_alloc_hugefolio(gfp, info, index);
- else
+ if (!folio)
+ count_vm_event(THP_FILE_FALLBACK);
+ } else {
+ pages = 1;
folio = shmem_alloc_folio(gfp, info, index);
- if (folio) {
- __folio_set_locked(folio);
- __folio_set_swapbacked(folio);
- return folio;
+ }
+ if (!folio)
+ return ERR_PTR(-ENOMEM);
+
+ __folio_set_locked(folio);
+ __folio_set_swapbacked(folio);
+
+ gfp &= GFP_RECLAIM_MASK;
+ error = mem_cgroup_charge(folio, fault_mm, gfp);
+ if (error) {
+ if (xa_find(&mapping->i_pages, &index,
+ index + pages - 1, XA_PRESENT)) {
+ error = -EEXIST;
+ } else if (huge) {
+ count_vm_event(THP_FILE_FALLBACK);
+ count_vm_event(THP_FILE_FALLBACK_CHARGE);
+ }
+ goto unlock;
}
- err = -ENOMEM;
- shmem_inode_unacct_blocks(inode, nr);
-failed:
- return ERR_PTR(err);
+ error = shmem_add_to_page_cache(folio, mapping, index, NULL, gfp);
+ if (error)
+ goto unlock;
+
+ error = shmem_inode_acct_blocks(inode, pages);
+ if (error) {
+ struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
+ long freed;
+ /*
+ * Try to reclaim some space by splitting a few
+ * large folios beyond i_size on the filesystem.
+ */
+ shmem_unused_huge_shrink(sbinfo, NULL, 2);
+ /*
+ * And do a shmem_recalc_inode() to account for freed pages:
+ * except our folio is there in cache, so not quite balanced.
+ */
+ spin_lock(&info->lock);
+ freed = pages + info->alloced - info->swapped -
+ READ_ONCE(mapping->nrpages);
+ if (freed > 0)
+ info->alloced -= freed;
+ spin_unlock(&info->lock);
+ if (freed > 0)
+ shmem_inode_unacct_blocks(inode, freed);
+ error = shmem_inode_acct_blocks(inode, pages);
+ if (error) {
+ filemap_remove_folio(folio);
+ goto unlock;
+ }
+ }
+
+ shmem_recalc_inode(inode, pages, 0);
+ folio_add_lru(folio);
+ return folio;
+
+unlock:
+ folio_unlock(folio);
+ folio_put(folio);
+ return ERR_PTR(error);
}
/*
@@ -1907,29 +1962,22 @@ static int shmem_get_folio_gfp(struct inode *inode, pgoff_t index,
struct vm_fault *vmf, vm_fault_t *fault_type)
{
struct vm_area_struct *vma = vmf ? vmf->vma : NULL;
- struct address_space *mapping = inode->i_mapping;
- struct shmem_inode_info *info = SHMEM_I(inode);
- struct shmem_sb_info *sbinfo;
struct mm_struct *fault_mm;
struct folio *folio;
- pgoff_t hindex;
- gfp_t huge_gfp;
int error;
- int once = 0;
- int alloced = 0;
+ bool alloced;
if (index > (MAX_LFS_FILESIZE >> PAGE_SHIFT))
return -EFBIG;
repeat:
if (sgp <= SGP_CACHE &&
- ((loff_t)index << PAGE_SHIFT) >= i_size_read(inode)) {
+ ((loff_t)index << PAGE_SHIFT) >= i_size_read(inode))
return -EINVAL;
- }
- sbinfo = SHMEM_SB(inode->i_sb);
+ alloced = false;
fault_mm = vma ? vma->vm_mm : NULL;
- folio = filemap_get_entry(mapping, index);
+ folio = filemap_get_entry(inode->i_mapping, index);
if (folio && vma && userfaultfd_minor(vma)) {
if (!xa_is_value(folio))
folio_put(folio);
@@ -1951,7 +1999,7 @@ repeat:
folio_lock(folio);
/* Has the folio been truncated or swapped out? */
- if (unlikely(folio->mapping != mapping)) {
+ if (unlikely(folio->mapping != inode->i_mapping)) {
folio_unlock(folio);
folio_put(folio);
goto repeat;
@@ -1986,65 +2034,38 @@ repeat:
return 0;
}
- if (!shmem_is_huge(inode, index, false,
- vma ? vma->vm_mm : NULL, vma ? vma->vm_flags : 0))
- goto alloc_nohuge;
+ if (shmem_is_huge(inode, index, false, fault_mm,
+ vma ? vma->vm_flags : 0)) {
+ gfp_t huge_gfp;
- huge_gfp = vma_thp_gfp_mask(vma);
- huge_gfp = limit_gfp_mask(huge_gfp, gfp);
- folio = shmem_alloc_and_acct_folio(huge_gfp, inode, index, true);
- if (IS_ERR(folio)) {
-alloc_nohuge:
- folio = shmem_alloc_and_acct_folio(gfp, inode, index, false);
+ huge_gfp = vma_thp_gfp_mask(vma);
+ huge_gfp = limit_gfp_mask(huge_gfp, gfp);
+ folio = shmem_alloc_and_add_folio(huge_gfp,
+ inode, index, fault_mm, true);
+ if (!IS_ERR(folio)) {
+ count_vm_event(THP_FILE_ALLOC);
+ goto alloced;
+ }
+ if (PTR_ERR(folio) == -EEXIST)
+ goto repeat;
}
- if (IS_ERR(folio)) {
- int retry = 5;
+ folio = shmem_alloc_and_add_folio(gfp, inode, index, fault_mm, false);
+ if (IS_ERR(folio)) {
error = PTR_ERR(folio);
+ if (error == -EEXIST)
+ goto repeat;
folio = NULL;
- if (error != -ENOSPC)
- goto unlock;
- /*
- * Try to reclaim some space by splitting a large folio
- * beyond i_size on the filesystem.
- */
- while (retry--) {
- int ret;
-
- ret = shmem_unused_huge_shrink(sbinfo, NULL, 1);
- if (ret == SHRINK_STOP)
- break;
- if (ret)
- goto alloc_nohuge;
- }
goto unlock;
}
- hindex = round_down(index, folio_nr_pages(folio));
-
- if (sgp == SGP_WRITE)
- __folio_set_referenced(folio);
-
- error = mem_cgroup_charge(folio, fault_mm, gfp);
- if (error) {
- if (folio_test_pmd_mappable(folio)) {
- count_vm_event(THP_FILE_FALLBACK);
- count_vm_event(THP_FILE_FALLBACK_CHARGE);
- }
- goto unacct;
- }
-
- error = shmem_add_to_page_cache(folio, mapping, hindex, NULL, gfp);
- if (error)
- goto unacct;
-
- folio_add_lru(folio);
- shmem_recalc_inode(inode, folio_nr_pages(folio), 0);
+alloced:
alloced = true;
-
if (folio_test_pmd_mappable(folio) &&
DIV_ROUND_UP(i_size_read(inode), PAGE_SIZE) <
folio_next_index(folio) - 1) {
+ struct shmem_sb_info *sbinfo = SHMEM_SB(inode->i_sb);
+ struct shmem_inode_info *info = SHMEM_I(inode);
/*
* Part of the large folio is beyond i_size: subject
* to shrink under memory pressure.
@@ -2062,6 +2083,8 @@ alloc_nohuge:
spin_unlock(&sbinfo->shrinklist_lock);
}
+ if (sgp == SGP_WRITE)
+ folio_set_referenced(folio);
/*
* Let SGP_FALLOC use the SGP_WRITE optimization on a new folio.
*/
@@ -2085,11 +2108,6 @@ clear:
/* Perhaps the file has been truncated since we checked */
if (sgp <= SGP_CACHE &&
((loff_t)index << PAGE_SHIFT) >= i_size_read(inode)) {
- if (alloced) {
- folio_clear_dirty(folio);
- filemap_remove_folio(folio);
- shmem_recalc_inode(inode, 0, 0);
- }
error = -EINVAL;
goto unlock;
}
@@ -2100,25 +2118,14 @@ out:
/*
* Error recovery.
*/
-unacct:
- shmem_inode_unacct_blocks(inode, folio_nr_pages(folio));
-
- if (folio_test_large(folio)) {
- folio_unlock(folio);
- folio_put(folio);
- goto alloc_nohuge;
- }
unlock:
+ if (alloced)
+ filemap_remove_folio(folio);
+ shmem_recalc_inode(inode, 0, 0);
if (folio) {
folio_unlock(folio);
folio_put(folio);
}
- if (error == -ENOSPC && !once++) {
- shmem_recalc_inode(inode, 0, 0);
- goto repeat;
- }
- if (error == -EEXIST)
- goto repeat;
return error;
}