The patch titled
Subject: hugetlb: fix hugetlb cgroup refcounting during vma split
has been added to the -mm tree. Its filename is
hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split.patch
This patch should soon appear at
https://ozlabs.org/~akpm/mmots/broken-out/hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split.patch
and later at
https://ozlabs.org/~akpm/mmotm/broken-out/hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split.patch
Before you just go and hit "reply", please:
a) Consider who else should be cc'ed
b) Prefer to cc a suitable mailing list as well
c) Ideally: find the original patch on the mailing list and do a
reply-to-all to that, adding suitable additional cc's
*** Remember to use Documentation/process/submit-checklist.rst when testing your code ***
The -mm tree is included into linux-next and is updated
there every 3-4 working days
------------------------------------------------------
From: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Subject: hugetlb: fix hugetlb cgroup refcounting during vma split
Guillaume Morin reported hitting the following WARNING followed by GPF or
NULL pointer deference either in cgroups_destroy or in the kill_css path.:
percpu ref (css_release) <= 0 (-1) after switching to atomic
WARNING: CPU: 23 PID: 130 at lib/percpu-refcount.c:196 percpu_ref_switch_to_atomic_rcu+0x127/0x130
CPU: 23 PID: 130 Comm: ksoftirqd/23 Kdump: loaded Tainted: G O 5.10.60 #1
RIP: 0010:percpu_ref_switch_to_atomic_rcu+0x127/0x130
Call Trace:
rcu_core+0x30f/0x530
rcu_core_si+0xe/0x10
__do_softirq+0x103/0x2a2
? sort_range+0x30/0x30
run_ksoftirqd+0x2b/0x40
smpboot_thread_fn+0x11a/0x170
kthread+0x10a/0x140
? kthread_create_worker_on_cpu+0x70/0x70
ret_from_fork+0x22/0x30
Upon further examination, it was discovered that the css structure was
associated with hugetlb reservations.
For private hugetlb mappings the vma points to a reserve map that contains
a pointer to the css. At mmap time, reservations are set up and a
reference to the css is taken. This reference is dropped in the vma close
operation; hugetlb_vm_op_close. However, if a vma is split no additional
reference to the css is taken yet hugetlb_vm_op_close will be called twice
for the split vma resulting in an underflow.
Fix by taking another reference in hugetlb_vm_op_open. Note that the
reference is only taken for the owner of the reserve map. In the more
common fork case, the pointer to the reserve map is cleared for non-owning
vmas.
Link: https://lkml.kernel.org/r/20210830215015.155224-1-mike.kravetz@xxxxxxxxxx
Fixes: e9fe92ae0cd2 ("hugetlb_cgroup: add reservation accounting for
private mappings")
Signed-off-by: Mike Kravetz <mike.kravetz@xxxxxxxxxx>
Reported-by: Guillaume Morin <guillaume@xxxxxxxxxxx>
Suggested-by: Guillaume Morin <guillaume@xxxxxxxxxxx>
Tested-by: Guillaume Morin <guillaume@xxxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
include/linux/hugetlb_cgroup.h | 12 ++++++++++++
mm/hugetlb.c | 4 +++-
2 files changed, 15 insertions(+), 1 deletion(-)
--- a/include/linux/hugetlb_cgroup.h~hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split
+++ a/include/linux/hugetlb_cgroup.h
@@ -121,6 +121,13 @@ static inline void hugetlb_cgroup_put_rs
css_put(&h_cg->css);
}
+static inline void resv_map_dup_hugetlb_cgroup_uncharge_info(
+ struct resv_map *resv_map)
+{
+ if (resv_map->css)
+ css_get(resv_map->css);
+}
+
extern int hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
struct hugetlb_cgroup **ptr);
extern int hugetlb_cgroup_charge_cgroup_rsvd(int idx, unsigned long nr_pages,
@@ -199,6 +206,11 @@ static inline void hugetlb_cgroup_put_rs
{
}
+static inline void resv_map_dup_hugetlb_cgroup_uncharge_info(
+ struct resv_map *resv_map)
+{
+}
+
static inline int hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
struct hugetlb_cgroup **ptr)
{
--- a/mm/hugetlb.c~hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split
+++ a/mm/hugetlb.c
@@ -4106,8 +4106,10 @@ static void hugetlb_vm_op_open(struct vm
* after this open call completes. It is therefore safe to take a
* new reference here without additional locking.
*/
- if (resv && is_vma_resv_set(vma, HPAGE_RESV_OWNER))
+ if (resv && is_vma_resv_set(vma, HPAGE_RESV_OWNER)) {
+ resv_map_dup_hugetlb_cgroup_uncharge_info(resv);
kref_get(&resv->refs);
+ }
}
static void hugetlb_vm_op_close(struct vm_area_struct *vma)
_
Patches currently in -mm which might be from mike.kravetz@xxxxxxxxxx are
hugetlb-simplify-prep_compound_gigantic_page-ref-count-racing-code.patch
hugetlb-drop-ref-count-earlier-after-page-allocation.patch
hugetlb-before-freeing-hugetlb-page-set-dtor-to-appropriate-value.patch
hugetlb-fix-hugetlb-cgroup-refcounting-during-vma-split.patch
