bpf: Force kprobe multi expected_attach_type for kprobe_multi link
Jiri Olsa
reviewerPATCH 1/1 bpf: Force kprobe multi expected_attach_type for kprobe_multi link
We currently allow to create perf link for program with expected_attach_type == BPF_TRACE_KPROBE_MULTI.
This will cause crash when we call helpers like get_attach_cookie or get_func_ip in such program, because it will call the kprobe_multi's version (current->bpf_ctx context setup) of those helpers while it expects perf_link's current->bpf_ctx context setup.
Making sure that we use BPF_TRACE_KPROBE_MULTI expected_attach_type only for programs attaching through kprobe_multi link.
--- kernel/bpf/syscall.c | 5 +++++ 1 file changed, 5 insertions(+)
kernel/bpf/syscall.c +5 -0
@@ -4675,6 +4675,11 @@ static int link_create(union bpf_attr *attr, bpfptr_t uattr) ret =bpf_perf_link_attach (attr, prog);break ;case BPF_PROG_TYPE_KPROBE : +if (prog->expected_attach_type ==BPF_TRACE_KPROBE_MULTI && + attr->link_create.attach_type !=BPF_TRACE_KPROBE_MULTI ) { + ret = -EINVAL ; + goto out; + }if (attr->link_create.attach_type ==BPF_PERF_EVENT ) ret =bpf_perf_link_attach (attr, prog);else
Signature · 1 line
2.40.1
Quoted 8 lines · Jiri Olsa
We currently allow to create perf link for program with
expected_attach_type==BPF_TRACE_KPROBE_MULTI.This will cause crash when we call helpers like
get_attach_cookieorget_func_ipin such program, because it will call the kprobe_multi's version (current->bpf_ctxcontext setup) of those helpers while it expects perf_link'scurrent->bpf_ctxcontext setup.
Quoted 5 lines
Making sure that we use
BPF_TRACE_KPROBE_MULTIexpected_attach_typeonly for programs attaching throughkprobe_multilink.Fixes ca74823c6e16 ("bpf: Add cookie support to programs attached with kprobe multi link")Signed-off-by Jiri Olsa <[email protected]>
Quoted 23 lines · Jiri Olsa
We currently allow to create perf link for program with
expected_attach_type==BPF_TRACE_KPROBE_MULTI.This will cause crash when we call helpers like
get_attach_cookieorget_func_ipin such program, because it will call the kprobe_multi's version (current->bpf_ctxcontext setup) of those helpers while it expects perf_link'scurrent->bpf_ctxcontext setup.Making sure that we use
BPF_TRACE_KPROBE_MULTIexpected_attach_typeonly for programs attaching throughkprobe_multilink.Fixes ca74823c6e16 ("bpf: Add cookie support to programs attached with kprobe multi link")Signed-off-by Jiri Olsa <[email protected]>--- kernel/bpf/syscall.c | 5 +++++ 1 file changed, 5 insertions(+)kernel/bpf/syscall.c +0
-0
kernel/bpf/syscall.c static int link_create(union bpf_attr *attr, bpfptr_t uattr) · line 4682
Quoted 8 lines
kernel/bpf/syscall.c +5
-0@@ -4675,6 +4675,11 @@ static int link_create(union bpf_attr *attr, bpfptr_t uattr) ret =bpf_perf_link_attach (attr, prog);break ;case BPF_PROG_TYPE_KPROBE : +if (prog->expected_attach_type ==BPF_TRACE_KPROBE_MULTI && + attr->link_create.attach_type !=BPF_TRACE_KPROBE_MULTI ) { + ret = -EINVAL ; + goto out; + }
there is a separate expected attach type validation switch above this, shouldn't this go there? We also have bpf_prog_attach_check_attach_type() call above as well, and tbh by now I'm not sure why we have like three places to check conditions like this... But I'd put this check in either bpf_prog_attach_check_attach_type() or in the dedicated switch for attach_type checks.
Quoted 6 lines
kernel/bpf/syscall.c +0
-1if (attr->link_create.attach_type ==BPF_PERF_EVENT ) ret =bpf_perf_link_attach (attr, prog);else --2.40.1
Quoted 33 lines · Andrii Nakryiko
Quoted 23 lines · Jiri Olsa
We currently allow to create perf link for program with
expected_attach_type==BPF_TRACE_KPROBE_MULTI.This will cause crash when we call helpers like
get_attach_cookieorget_func_ipin such program, because it will call the kprobe_multi's version (current->bpf_ctxcontext setup) of those helpers while it expects perf_link'scurrent->bpf_ctxcontext setup.Making sure that we use
BPF_TRACE_KPROBE_MULTIexpected_attach_typeonly for programs attaching throughkprobe_multilink.Fixes ca74823c6e16 ("bpf: Add cookie support to programs attached with kprobe multi link")Signed-off-by Jiri Olsa <[email protected]>--- kernel/bpf/syscall.c | 5 +++++ 1 file changed, 5 insertions(+)kernel/bpf/syscall.c +0
-0kernel/bpf/syscall.c static int link_create(union bpf_attr *attr, bpfptr_t uattr) · line 4682
Quoted 8 lines
kernel/bpf/syscall.c +5
-0@@ -4675,6 +4675,11 @@ static int link_create(union bpf_attr *attr, bpfptr_t uattr) ret =bpf_perf_link_attach (attr, prog);break ;case BPF_PROG_TYPE_KPROBE : +if (prog->expected_attach_type ==BPF_TRACE_KPROBE_MULTI && + attr->link_create.attach_type !=BPF_TRACE_KPROBE_MULTI ) { + ret = -EINVAL ; + goto out; + }
Quoted 8 lines
there is a separate expected attach type validation switch above this, shouldn't this go there? We also have
bpf_prog_attach_check_attach_type()call above as well, and tbh by now I'm not sure why we have like three places to check conditions like this... But I'd put this check in eitherbpf_prog_attach_check_attach_type()or in the dedicated switch forattach_typechecks.
bpf_prog_attach_check_attach_type looks good, will move it there
thanks,
jirka