bpf: avoid holding freeze_mutex during mmap operation
PATCH 1/1 bpf: avoid holding freeze_mutex during mmap operation
From: Andrii Nakryiko <[email protected]>
[ Upstream commit bc27c52eea189e8f7492d40739b7746d67b65beb ]
We use map->freeze_mutex to prevent races between map_freeze() and memory mapping BPF map contents with writable permissions. The way we naively do this means we'll hold freeze_mutex for entire duration of all the mm and VMA manipulations, which is completely unnecessary. This can potentially also lead to deadlocks, as reported by syzbot in [0].
So, instead, hold freeze_mutex only during writeability checks, bump (proactively) "write active" count for the map, unlock the mutex and proceed with mmap logic. And only if something went wrong during mmap logic, then undo that "write active" counter increment.
[0] https://lore.kernel.org/bpf/[email protected]/
--- kernel/bpf/syscall.c | 17 ++++++++++------- 1 file changed, 10 insertions(+), 7 deletions(-)
kernel/bpf/syscall.c +10 -7
@@ -654,7 +654,7 @@ static const struct vm_operations_struct bpf_map_default_vmops = {static int bpf_map_mmap (struct file *filp,struct vm_area_struct *vma) {struct bpf_map *map = filp->private_data; -int err; +int err =0 ;if (!map->ops->map_mmap ||map_value_has_spin_lock (map) ||map_value_has_timer (map))
@@ -679,7 +679,12 @@ static int bpf_map_mmap(struct file *filp, struct vm_area_struct *vma) err = -EACCES ; goto out; } +bpf_map_write_active_inc (map); } +out: +mutex_unlock (&map->freeze_mutex); +if (err) +return err;/* set default open/close callbacks */ vma->vm_ops = &bpf_map_default_vmops;
@@ -690,13 +695,11 @@ static int bpf_map_mmap(struct file *filp, struct vm_area_struct *vma) vma->vm_flags &= ~VM_MAYWRITE ; err = map->ops->map_mmap (map, vma); -if (err) - goto out; +if (err) { +if (vma->vm_flags &VM_WRITE ) +bpf_map_write_active_dec (map); + } -if (vma->vm_flags &VM_MAYWRITE ) -bpf_map_write_active_inc (map); -out: -mutex_unlock (&map->freeze_mutex);return err; }
Signature · 1 line
2.47.1
Sasha Levin
maintainer[ Sasha's backport helper bot ]
Hi,
✅ All tests passed successfully. No issues detected. No action required from the submitter.
The upstream commit SHA1 provided is correct: bc27c52eea189e8f7492d40739b7746d67b65beb
WARNING: Author mismatch between patch and upstream commit:
Backport author: David Sauerwein<[email protected]> Commit author: Andrii Nakryiko<[email protected]>
Status in newer kernel trees:
6.14.y | Present (exact SHA1) 6.13.y | Present (different SHA1: 271e49f8a58e) 6.12.y | Present (different SHA1: d95607a5f2f9)
6.6.y | Present (different SHA1: 29cfda62ab4d) 6.1.y | Not found
Note: The patch differs from the upstream commit:
---
1: bc27c52eea189 ! 1: 74c702cd74209 bpf: avoid holding freeze_mutex during mmap operation
@@ Metadata
## Commit message ##
bpf: avoid holding freeze_mutex during mmap operation
+ [ Upstream commit bc27c52eea189e8f7492d40739b7746d67b65beb ]
+
We use map->freeze_mutex to prevent races between map_freeze() and memory mapping BPF map contents with writable permissions. The way we naively do this means we'll hold freeze_mutex for entire duration of all
@@ Commit message
+ Signed-off-by: Sasha Levin <[email protected]>
+ Signed-off-by: David Sauerwein <[email protected]>
## kernel/bpf/syscall.c ##
@@ kernel/bpf/syscall.c: static const struct vm_operations_struct bpf_map_default_vmops = {
@@ kernel/bpf/syscall.c: static const struct vm_operations_struct bpf_map_default_v
- int err;
+ int err = 0;
- if (!map->ops->map_mmap || !IS_ERR_OR_NULL(map->record))
- return -ENOTSUPP;
+ if (!map->ops->map_mmap || map_value_has_spin_lock(map) ||
+ map_value_has_timer(map))
@@ kernel/bpf/syscall.c: static int bpf_map_mmap(struct file *filp, struct vm_area_struct *vma)
err = -EACCES;
goto out;
@@ kernel/bpf/syscall.c: static int bpf_map_mmap(struct file *filp, struct vm_area_
/* set default open/close callbacks */
vma->vm_ops = &bpf_map_default_vmops;
@@ kernel/bpf/syscall.c: static int bpf_map_mmap(struct file *filp, struct vm_area_struct *vma)
- vm_flags_clear(vma, VM_MAYWRITE);
+ vma->vm_flags &= ~VM_MAYWRITE;
err = map->ops->map_mmap(map, vma);
- if (err)
@@ kernel/bpf/syscall.c: static int bpf_map_mmap(struct file *filp, struct vm_area_
+ bpf_map_write_active_dec(map);
+ }
-- if (vma->vm_flags & VM_WRITE)
+- if (vma->vm_flags & VM_MAYWRITE)
- bpf_map_write_active_inc(map);
-out:
- mutex_unlock(&map->freeze_mutex);
---
Results of testing on various branches:
| Branch | Patch Apply | Build Test | |---------------------------|-------------|------------| | stable/linux-5.15.y | Success | Success |