It is particularly important for the userns mount case (when a sensible
nr_inodes maximum may not be enforced) that tmpfs user xattrs be subject
to memory cgroup limiting. Leave temporary buffer allocations as is,
but change the persistent simple xattr allocations from GFP_KERNEL to
GFP_KERNEL_ACCOUNT. This limits kernfs's cgroupfs too, but that's good.
(I had intended to send this change earlier, but had been confused by
shmem_alloc_inode() using GFP_KERNEL, and thought a discussion would be
needed to change that too: no, I was forgetting the SLAB_ACCOUNT on that
kmem_cache, which implicitly adds __GFP_ACCOUNT to all its allocations.)
Signed-off-by: Hugh Dickins <hughd@xxxxxxxxxx>
---
fs/xattr.c | 4 ++--
mm/shmem.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/fs/xattr.c b/fs/xattr.c
index 2d607542281b..efd4736bc94b 100644
--- a/fs/xattr.c
+++ b/fs/xattr.c
@@ -1093,7 +1093,7 @@ struct simple_xattr *simple_xattr_alloc(const void *value, size_t size)
if (len < sizeof(*new_xattr))
return NULL;
- new_xattr = kvmalloc(len, GFP_KERNEL);
+ new_xattr = kvmalloc(len, GFP_KERNEL_ACCOUNT);
if (!new_xattr)
return NULL;
@@ -1217,7 +1217,7 @@ struct simple_xattr *simple_xattr_set(struct simple_xattrs *xattrs,
if (!new_xattr)
return ERR_PTR(-ENOMEM);
- new_xattr->name = kstrdup(name, GFP_KERNEL);
+ new_xattr->name = kstrdup(name, GFP_KERNEL_ACCOUNT);
if (!new_xattr->name) {
simple_xattr_free(new_xattr);
return ERR_PTR(-ENOMEM);
diff --git a/mm/shmem.c b/mm/shmem.c
index b782edeb69aa..11298c797cdc 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -3616,7 +3616,7 @@ static int shmem_initxattrs(struct inode *inode,
len = strlen(xattr->name) + 1;
new_xattr->name = kmalloc(XATTR_SECURITY_PREFIX_LEN + len,
- GFP_KERNEL);
+ GFP_KERNEL_ACCOUNT);
if (!new_xattr->name) {
kvfree(new_xattr);
break;
--
2.35.3
