tools/bpf/bpftool: Reset vmlinux BTF after struct_ops commands
A message that is not in the archive (sent off-list or not yet received): [email protected]
PATCH v2 1/2 tools/bpf/bpftool: Reset vmlinux BTF after map commands
get_map_kv_btf() caches the vmlinux BTF object when a map uses btf_vmlinux_value_type_id. map dump released that object when the command completed, but left the global pointer stale.
The same cached object can also be returned to print_key_value(), which freed it directly. That leaves btf_vmlinux dangling before the command cleanup path runs.
Use free_map_kv_btf() for per-entry cleanup, and reset the cached btf_vmlinux pointer when the map command releases the object. This keeps batch mode from reusing a freed BTF object.
--- tools/bpf/bpftool/map.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)
tools/bpf/bpftool/map.c +9 -3
@@ -790,6 +790,12 @@ static int maps_have_btf(int *fds, int nb_fds)static struct btf *btf_vmlinux; +static void free_btf_vmlinux (void ) +{ +btf__free (btf_vmlinux); + btf_vmlinux =NULL ; +} +static int get_map_kv_btf (const struct bpf_map_info *info,struct btf **btf) {int err =0 ;
@@ -958,7 +964,7 @@ static int do_dump(int argc, char **argv)close (fds[i]); exit_free:free (fds); -btf__free (btf_vmlinux); +free_btf_vmlinux ();return err; }
@@ -1049,7 +1055,7 @@ static void print_key_value(struct bpf_map_info *info, void *key, btf_wtr =get_btf_writer ();if (!btf_wtr) {p_info ("failed to create json writer for btf. falling back to plain output" ); -btf__free (btf); +free_map_kv_btf (btf); btf =NULL ;print_entry_plain (info, key, value); }else {
@@ -1065,7 +1071,7 @@ static void print_key_value(struct bpf_map_info *info, void *key, }else {print_entry_plain (info, key, value); } -btf__free (btf); +free_map_kv_btf (btf); }static int do_lookup (int argc,char **argv)
Signature · 1 line
2.51.0
[email protected]
AI reviewRe: [PATCH v2 1/2] tools/bpf/bpftool: Reset vmlinux BTF after map commands
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Unbounded memory leak of vmlinux BTF object during bpftool cgroup tree execution. --
commit c2cdea6fdb621f7e2e4a217204d6fd383a327434
Author: Yichong Chen <[email protected]>
tools/bpf/bpftool: Reset vmlinux BTF after map commands
This commit fixes a use-after-free in bpftool map commands by correctly resetting the global btf_vmlinux pointer when the map command releases the object. This prevents batch mode from reusing a dangling pointer.
tools/bpf/bpftool/map.c static int maps_have_btf(int *fds, int nb_fds) · line 797
Quoted 13 lines
tools/bpf/bpftool/map.c +5
-0@@ -790,6 +790,12 @@ static int maps_have_btf(int *fds, int nb_fds)static struct btf *btf_vmlinux; +static void free_btf_vmlinux (void ) +{ +btf__free (btf_vmlinux); + btf_vmlinux =NULL ; +}
[Severity: High]
This isn't a bug introduced by this patch, but while you are fixing the btf_vmlinux cleanup here, does tools/bpf/bpftool/cgroup.c need similar attention?
Looking at do_show_tree_fn() in cgroup.c, it unconditionally allocates a new BTF object over the global pointer:
btf_vmlinux = libbpf_find_kernel_btf();
for (i = 0; i < ARRAY_SIZE(cgroup_attach_types); i++) {Since do_show_tree_fn() is called by nftw() for every directory in the cgroup tree, it overwrites the global pointer without freeing the previous object. Could this lead to an unbounded memory leak and potential OOM killer invocation when walking extensive cgroup trees?
The same pattern also appears in do_show() for batch mode:
btf_vmlinux = libbpf_find_kernel_btf();
for (i = 0; i < ARRAY_SIZE(cgroup_attach_types); i++) {Would it make sense to reuse the new free_btf_vmlinux() logic to plug this leak as well?
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
[email protected]
AI reviewRe: [PATCH v2 1/2] tools/bpf/bpftool: Reset vmlinux BTF after map commands
In reply to Yichong ChenThis looks like a bug fix for the cached vmlinux BTF handling that was introduced when map dumping started caching btf_vmlinux via get_map_kv_btf()/free_map_kv_btf(). Should this carry a Fixes: tag?
Quoted 8 lines
get_map_kv_btf()caches the vmlinux BTF object when a map usesbtf_vmlinux_value_type_id. map dump released that object when the command completed, but left the global pointer stale.The same cached object can also be returned to
print_key_value(), which freed it directly. That leavesbtf_vmlinuxdangling before the command cleanup path runs.
Quoted 5 lines
Use
free_map_kv_btf()for per-entry cleanup, and reset the cachedbtf_vmlinuxpointer when the map command releases the object. This keeps batch mode from reusing a freed BTF object.Signed-off-by Yichong Chen <[email protected]>
This looks like a bug fix for the cached vmlinux BTF handling that was introduced when map dumping started caching btf_vmlinux via get_map_kv_btf()/free_map_kv_btf(). Should this carry a Fixes: tag?
---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/27927483612
Re: [PATCH v2 1/2] tools/bpf/bpftool: Reset vmlinux BTF after map commands
In reply to Yichong ChenQuoted 8 lines · Yichong Chen
get_map_kv_btf()caches the vmlinux BTF object when a map usesbtf_vmlinux_value_type_id. map dump released that object when the command completed, but left the global pointer stale.The same cached object can also be returned to
print_key_value(), which freed it directly. That leavesbtf_vmlinuxdangling before the command cleanup path runs.
Quoted 5 lines
Use
free_map_kv_btf()for per-entry cleanup, and reset the cachedbtf_vmlinuxpointer when the map command releases the object. This keeps batch mode from reusing a freed BTF object.Signed-off-by Yichong Chen <[email protected]>
You added a Fixes tag to another patch. Should we add one here as well?
PATCH v2 2/2 tools/bpf/bpftool: Reset vmlinux BTF after struct_ops commands
In reply to a message not in the archivestruct_ops frees the global btf_vmlinux object.
In batch mode, a later struct_ops command can reuse stale state.
Reset the BTF pointer and cached map info state.
--- tools/bpf/bpftool/struct_ops.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-)
tools/bpf/bpftool/struct_ops.c +10 -1
@@ -21,6 +21,15 @@ static __u32 map_info_alloc_len;static struct btf *btf_vmlinux;static __s32 map_info_type_id; +static void free_btf_vmlinux (void ) +{ +btf__free (btf_vmlinux); + btf_vmlinux =NULL ; + map_info_type =NULL ; + map_info_alloc_len =0 ; + map_info_type_id =0 ; +} +struct res {unsigned int nr_maps;unsigned int nr_errs;
@@ -642,7 +651,7 @@ int do_struct_ops(int argc, char **argv) err =cmd_select (cmds, argc, argv, do_help); -btf__free (btf_vmlinux); +free_btf_vmlinux ();return err; }
Signature · 1 line
2.51.0
[email protected]
AI reviewThank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [Medium] Unconditional memory leak of the global btf_vmlinux object during cgroup tree traversal and batch mode. --
commit bd1d63c4728bc77ece052cc69d67e204a7e59411
Author: Yichong Chen <[email protected]>
tools/bpf/bpftool: Reset vmlinux BTF after struct_ops commands
struct_ops frees the global btf_vmlinux object. In batch mode, a later struct_ops command can reuse stale state. This patch resets the BTF pointer and cached map info state.
Quoted 7 lines
tools/bpf/bpftool/struct_ops.c +0
-0@@ -21,6 +21,15 @@ static __u32 map_info_alloc_len;static struct btf *btf_vmlinux;static __s32 map_info_type_id;
tools/bpf/bpftool/struct_ops.c static __u32 map_info_alloc_len; · line 32
Quoted 10 lines
tools/bpf/bpftool/struct_ops.c +9
-0+static void free_btf_vmlinux (void ) +{ +btf__free (btf_vmlinux); + btf_vmlinux =NULL ; + map_info_type =NULL ; + map_info_alloc_len =0 ; + map_info_type_id =0 ; +} +
[Severity: Medium]
This isn't a bug introduced by this patch, but since we are addressing BTF memory management and stale state in batch mode, could we also look at similar leaks in cgroup.c?
In tools/bpf/bpftool/cgroup.c, libbpf_find_kernel_btf() is called unconditionally to initialize the btf_vmlinux pointer in both do_show() and do_show_tree_fn(), without any subsequent cleanup.
For example, in the tree traversal:
tools/bpf/bpftool/cgroup.c:do_show_tree_fn() {
...
btf_vmlinux = libbpf_find_kernel_btf();
for (i = 0; i < ARRAY_SIZE(cgroup_attach_types); i++)
show_bpf_progs(cgroup_fd, cgroup_attach_types[i], ftw->level);
...
}Because do_show_tree_fn() is called by nftw() for every directory in the cgroup tree, it repeatedly allocates a new BTF object and overwrites the pointer without freeing the previous instance.
Could this cause an O(N) memory leak scaling with the number of cgroups, potentially leading to OOM crashes of the bpftool process during traversal or when running cgroup commands in batch mode?
[ ... ]
Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=2
[email protected]
AI reviewThis looks like a fix for a stale-state reuse problem in do_struct_ops(), where btf__free() left btf_vmlinux and the cached map_info_type/ map_info_type_id dangling for a later struct_ops command in…
Quoted 9 lines
tools/bpf/bpftool: Reset vmlinux BTF afterstruct_opscommands
struct_opsfrees the globalbtf_vmlinuxobject.In batch mode, a later
struct_opscommand can reuse stale state.Reset the BTF pointer and cached map info state.
Signed-off-by Yichong Chen <[email protected]>
This looks like a fix for a stale-state reuse problem in do_struct_ops(), where btf__free() left btf_vmlinux and the cached map_info_type/ map_info_type_id dangling for a later struct_ops command in batch mode.
The cached state and the do_struct_ops() free path date back to when struct_ops.c was first added. Should this carry a Fixes: tag?
---
AI reviewed your patch. Please fix the bug or email reply why it's not a bug.
See: https://github.com/kernel-patches/vmtest/blob/master/ci/claude/README.md
CI run summary: https://github.com/kernel-patches/bpf/actions/runs/27927483612