* [PATCH bpf-next v3 0/2] bpf: enable some functions in cgroup programs @ 2024-07-25 0:14 technoboy85 2024-07-25 0:14 ` [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs technoboy85 2024-07-25 0:14 ` [PATCH bpf-next v3 2/2] bpf: allow bpf_current_task_under_cgroup() with BPF_CGROUP_* technoboy85 0 siblings, 2 replies; 5+ messages in thread From: technoboy85 @ 2024-07-25 0:14 UTC (permalink / raw) To: Andrii Nakryiko, Alexei Starovoitov, Daniel Borkmann, Steven Rostedt, Masami Hiramatsu, bpf, linux-trace-kernel Cc: linux-kernel, Matteo Croce From: Matteo Croce <teknoraver@meta.com> Enable some BPF kfuncs and the helper bpf_current_task_under_cgroup() for program types BPF_CGROUP_*. These will be used by systemd-networkd: https://github.com/systemd/systemd/pull/32212 Matteo Croce (2): bpf: enable generic kfuncs for BPF_CGROUP_* programs bpf: allow bpf_current_task_under_cgroup() with BPF_CGROUP_* include/linux/bpf.h | 1 + kernel/bpf/cgroup.c | 2 ++ kernel/bpf/helpers.c | 29 +++++++++++++++++++++++++++++ kernel/trace/bpf_trace.c | 23 ----------------------- 4 files changed, 32 insertions(+), 23 deletions(-) -- 2.45.2 ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs 2024-07-25 0:14 [PATCH bpf-next v3 0/2] bpf: enable some functions in cgroup programs technoboy85 @ 2024-07-25 0:14 ` technoboy85 2024-08-12 16:48 ` Matteo Croce 2024-07-25 0:14 ` [PATCH bpf-next v3 2/2] bpf: allow bpf_current_task_under_cgroup() with BPF_CGROUP_* technoboy85 1 sibling, 1 reply; 5+ messages in thread From: technoboy85 @ 2024-07-25 0:14 UTC (permalink / raw) To: Andrii Nakryiko, Alexei Starovoitov, Daniel Borkmann, Steven Rostedt, Masami Hiramatsu, bpf, linux-trace-kernel Cc: linux-kernel, Matteo Croce, Andrii Nakryiko From: Matteo Croce <teknoraver@meta.com> These kfuncs are enabled even in BPF_PROG_TYPE_TRACING, so they should be safe also in BPF_CGROUP_* programs. Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Matteo Croce <teknoraver@meta.com> --- kernel/bpf/helpers.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c index d02ae323996b..0d1d97d968b0 100644 --- a/kernel/bpf/helpers.c +++ b/kernel/bpf/helpers.c @@ -3052,6 +3052,12 @@ static int __init kfunc_init(void) ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &generic_kfunc_set); ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_STRUCT_OPS, &generic_kfunc_set); ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_SYSCALL, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SKB, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_DEVICE, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK_ADDR, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SYSCTL, &generic_kfunc_set); + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCKOPT, &generic_kfunc_set); ret = ret ?: register_btf_id_dtor_kfuncs(generic_dtors, ARRAY_SIZE(generic_dtors), THIS_MODULE); -- 2.45.2 ^ permalink raw reply related [flat|nested] 5+ messages in thread
* Re: [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs 2024-07-25 0:14 ` [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs technoboy85 @ 2024-08-12 16:48 ` Matteo Croce 2024-08-12 21:48 ` Andrii Nakryiko 0 siblings, 1 reply; 5+ messages in thread From: Matteo Croce @ 2024-08-12 16:48 UTC (permalink / raw) To: Andrii Nakryiko, Alexei Starovoitov, Daniel Borkmann, Steven Rostedt, Masami Hiramatsu, bpf, linux-trace-kernel Cc: linux-kernel, Matteo Croce, Andrii Nakryiko Il giorno gio 25 lug 2024 alle ore 02:14 <technoboy85@gmail.com> ha scritto: > --- a/kernel/bpf/helpers.c > +++ b/kernel/bpf/helpers.c > @@ -3052,6 +3052,12 @@ static int __init kfunc_init(void) > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &generic_kfunc_set); > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_STRUCT_OPS, &generic_kfunc_set); > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_SYSCALL, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SKB, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_DEVICE, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK_ADDR, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SYSCTL, &generic_kfunc_set); > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCKOPT, &generic_kfunc_set); > ret = ret ?: register_btf_id_dtor_kfuncs(generic_dtors, > ARRAY_SIZE(generic_dtors), > THIS_MODULE); This seems not enough, some kfuncs like bpf_cgroup_from_id are still rejected. To fix this we need also this chunk: --- a/kernel/bpf/btf.c +++ b/kernel/bpf/btf.c @@ -8309,7 +8319,11 @@ static int bpf_prog_type_to_kfunc_hook(enum bpf_prog_type prog_type) case BPF_PROG_TYPE_SYSCALL: return BTF_KFUNC_HOOK_SYSCALL; case BPF_PROG_TYPE_CGROUP_SKB: + case BPF_PROG_TYPE_CGROUP_SOCK: + case BPF_PROG_TYPE_CGROUP_DEVICE: case BPF_PROG_TYPE_CGROUP_SOCK_ADDR: + case BPF_PROG_TYPE_CGROUP_SYSCTL: + case BPF_PROG_TYPE_CGROUP_SOCKOPT: return BTF_KFUNC_HOOK_CGROUP_SKB; case BPF_PROG_TYPE_SCHED_ACT: return BTF_KFUNC_HOOK_SCHED_ACT; but even with this it won't work, because bpf_prog_type_to_kfunc_hook() aliases many program types with a single hook, and btf_kfunc_id_set_contains() will fail to check if the kfunc is in the set. One solution could be to extend the btf_kfunc_hook enum with an entry for every CGROUP program type, but a thing I wanted to avoid is to let this enum proliferate in this way. I wish to group all the CGROUP_ program types into a single hook, and perhaps drop the "SKB" in "BTF_KFUNC_HOOK_CGROUP_SKB" which will have no meaning anymore. Ideas? -- Matteo Croce perl -e 'for($t=0;;$t++){print chr($t*($t>>8|$t>>13)&255)}' |aplay ^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs 2024-08-12 16:48 ` Matteo Croce @ 2024-08-12 21:48 ` Andrii Nakryiko 0 siblings, 0 replies; 5+ messages in thread From: Andrii Nakryiko @ 2024-08-12 21:48 UTC (permalink / raw) To: Matteo Croce Cc: Alexei Starovoitov, Daniel Borkmann, Steven Rostedt, Masami Hiramatsu, bpf, linux-trace-kernel, linux-kernel, Matteo Croce, Andrii Nakryiko On Mon, Aug 12, 2024 at 9:49 AM Matteo Croce <technoboy85@gmail.com> wrote: > > Il giorno gio 25 lug 2024 alle ore 02:14 <technoboy85@gmail.com> ha scritto: > > --- a/kernel/bpf/helpers.c > > +++ b/kernel/bpf/helpers.c > > @@ -3052,6 +3052,12 @@ static int __init kfunc_init(void) > > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_XDP, &generic_kfunc_set); > > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_STRUCT_OPS, &generic_kfunc_set); > > ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_SYSCALL, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SKB, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_DEVICE, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCK_ADDR, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SYSCTL, &generic_kfunc_set); > > + ret = ret ?: register_btf_kfunc_id_set(BPF_PROG_TYPE_CGROUP_SOCKOPT, &generic_kfunc_set); > > ret = ret ?: register_btf_id_dtor_kfuncs(generic_dtors, > > ARRAY_SIZE(generic_dtors), > > THIS_MODULE); > > This seems not enough, some kfuncs like bpf_cgroup_from_id are still rejected. > To fix this we need also this chunk: > > --- a/kernel/bpf/btf.c > +++ b/kernel/bpf/btf.c > @@ -8309,7 +8319,11 @@ static int bpf_prog_type_to_kfunc_hook(enum > bpf_prog_type prog_type) > case BPF_PROG_TYPE_SYSCALL: > return BTF_KFUNC_HOOK_SYSCALL; > case BPF_PROG_TYPE_CGROUP_SKB: > + case BPF_PROG_TYPE_CGROUP_SOCK: > + case BPF_PROG_TYPE_CGROUP_DEVICE: > case BPF_PROG_TYPE_CGROUP_SOCK_ADDR: > + case BPF_PROG_TYPE_CGROUP_SYSCTL: > + case BPF_PROG_TYPE_CGROUP_SOCKOPT: > return BTF_KFUNC_HOOK_CGROUP_SKB; > case BPF_PROG_TYPE_SCHED_ACT: > return BTF_KFUNC_HOOK_SCHED_ACT; > > but even with this it won't work, because > bpf_prog_type_to_kfunc_hook() aliases many program types with a single > hook, and btf_kfunc_id_set_contains() will fail to check if the kfunc > is in the set. > > One solution could be to extend the btf_kfunc_hook enum with an entry > for every CGROUP program type, but a thing I wanted to avoid is to let > this enum proliferate in this way. > I wish to group all the CGROUP_ program types into a single hook, and > perhaps drop the "SKB" in "BTF_KFUNC_HOOK_CGROUP_SKB" which will have > no meaning anymore. > > Ideas? Given BPF_PROG_TYPE_CGROUP_SKB and BPF_PROG_TYPE_CGROUP_SOCK_ADDR are already grouped into the same hook type, and bpf_sock_addr_set_sun_path() is meant to be enabled for BPF_PROG_TYPE_CGROUP_SOCK_ADDR (but really it's enabled for BPF_PROG_TYPE_CGROUP_SKB as well), I guess it's fine to just have one CGROUP hook type for all those cgroup-related program types and just rely on BTF type matching logic. I'd say that register_btf_kfunc_id_set() should just accept enum bpf_kfunc_hook instead of enum bpf_prog_type (I don't see where we remember or use prog_type beyond just translating that to bpf_kfunc_hook), but that can be left for follow up clean ups. > > -- > Matteo Croce > > perl -e 'for($t=0;;$t++){print chr($t*($t>>8|$t>>13)&255)}' |aplay ^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH bpf-next v3 2/2] bpf: allow bpf_current_task_under_cgroup() with BPF_CGROUP_* 2024-07-25 0:14 [PATCH bpf-next v3 0/2] bpf: enable some functions in cgroup programs technoboy85 2024-07-25 0:14 ` [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs technoboy85 @ 2024-07-25 0:14 ` technoboy85 1 sibling, 0 replies; 5+ messages in thread From: technoboy85 @ 2024-07-25 0:14 UTC (permalink / raw) To: Andrii Nakryiko, Alexei Starovoitov, Daniel Borkmann, Steven Rostedt, Masami Hiramatsu, bpf, linux-trace-kernel Cc: linux-kernel, Matteo Croce, Andrii Nakryiko From: Matteo Croce <teknoraver@meta.com> The helper bpf_current_task_under_cgroup() currently is only allowed for tracing programs. Allow its usage also in the BPF_CGROUP_* program types. Move the code from kernel/trace/bpf_trace.c to kernel/bpf/helpers.c, so it compiles also without CONFIG_BPF_EVENTS. This will be used in systemd-networkd to monitor the sysctl writes, and filter it's own writes from others: https://github.com/systemd/systemd/pull/32212 Acked-by: Andrii Nakryiko <andrii@kernel.org> Signed-off-by: Matteo Croce <teknoraver@meta.com> --- include/linux/bpf.h | 1 + kernel/bpf/cgroup.c | 2 ++ kernel/bpf/helpers.c | 23 +++++++++++++++++++++++ kernel/trace/bpf_trace.c | 23 ----------------------- 4 files changed, 26 insertions(+), 23 deletions(-) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 7ad37cbdc815..3820fcf360b6 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -3202,6 +3202,7 @@ extern const struct bpf_func_proto bpf_sock_hash_update_proto; extern const struct bpf_func_proto bpf_get_current_cgroup_id_proto; extern const struct bpf_func_proto bpf_get_current_ancestor_cgroup_id_proto; extern const struct bpf_func_proto bpf_get_cgroup_classid_curr_proto; +extern const struct bpf_func_proto bpf_current_task_under_cgroup_proto; extern const struct bpf_func_proto bpf_msg_redirect_hash_proto; extern const struct bpf_func_proto bpf_msg_redirect_map_proto; extern const struct bpf_func_proto bpf_sk_redirect_hash_proto; diff --git a/kernel/bpf/cgroup.c b/kernel/bpf/cgroup.c index 8ba73042a239..e7113d700b87 100644 --- a/kernel/bpf/cgroup.c +++ b/kernel/bpf/cgroup.c @@ -2581,6 +2581,8 @@ cgroup_current_func_proto(enum bpf_func_id func_id, const struct bpf_prog *prog) case BPF_FUNC_get_cgroup_classid: return &bpf_get_cgroup_classid_curr_proto; #endif + case BPF_FUNC_current_task_under_cgroup: + return &bpf_current_task_under_cgroup_proto; default: return NULL; } diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c index 0d1d97d968b0..8502cfed2926 100644 --- a/kernel/bpf/helpers.c +++ b/kernel/bpf/helpers.c @@ -2458,6 +2458,29 @@ __bpf_kfunc long bpf_task_under_cgroup(struct task_struct *task, return ret; } +BPF_CALL_2(bpf_current_task_under_cgroup, struct bpf_map *, map, u32, idx) +{ + struct bpf_array *array = container_of(map, struct bpf_array, map); + struct cgroup *cgrp; + + if (unlikely(idx >= array->map.max_entries)) + return -E2BIG; + + cgrp = READ_ONCE(array->ptrs[idx]); + if (unlikely(!cgrp)) + return -EAGAIN; + + return task_under_cgroup_hierarchy(current, cgrp); +} + +const struct bpf_func_proto bpf_current_task_under_cgroup_proto = { + .func = bpf_current_task_under_cgroup, + .gpl_only = false, + .ret_type = RET_INTEGER, + .arg1_type = ARG_CONST_MAP_PTR, + .arg2_type = ARG_ANYTHING, +}; + /** * bpf_task_get_cgroup1 - Acquires the associated cgroup of a task within a * specific cgroup1 hierarchy. The cgroup1 hierarchy is identified by its diff --git a/kernel/trace/bpf_trace.c b/kernel/trace/bpf_trace.c index cd098846e251..cc9301054f84 100644 --- a/kernel/trace/bpf_trace.c +++ b/kernel/trace/bpf_trace.c @@ -798,29 +798,6 @@ const struct bpf_func_proto bpf_task_pt_regs_proto = { .ret_btf_id = &bpf_task_pt_regs_ids[0], }; -BPF_CALL_2(bpf_current_task_under_cgroup, struct bpf_map *, map, u32, idx) -{ - struct bpf_array *array = container_of(map, struct bpf_array, map); - struct cgroup *cgrp; - - if (unlikely(idx >= array->map.max_entries)) - return -E2BIG; - - cgrp = READ_ONCE(array->ptrs[idx]); - if (unlikely(!cgrp)) - return -EAGAIN; - - return task_under_cgroup_hierarchy(current, cgrp); -} - -static const struct bpf_func_proto bpf_current_task_under_cgroup_proto = { - .func = bpf_current_task_under_cgroup, - .gpl_only = false, - .ret_type = RET_INTEGER, - .arg1_type = ARG_CONST_MAP_PTR, - .arg2_type = ARG_ANYTHING, -}; - struct send_signal_irq_work { struct irq_work irq_work; struct task_struct *task; -- 2.45.2 ^ permalink raw reply related [flat|nested] 5+ messages in thread
end of thread, other threads:[~2024-08-12 21:49 UTC | newest] Thread overview: 5+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2024-07-25 0:14 [PATCH bpf-next v3 0/2] bpf: enable some functions in cgroup programs technoboy85 2024-07-25 0:14 ` [PATCH bpf-next v3 1/2] bpf: enable generic kfuncs for BPF_CGROUP_* programs technoboy85 2024-08-12 16:48 ` Matteo Croce 2024-08-12 21:48 ` Andrii Nakryiko 2024-07-25 0:14 ` [PATCH bpf-next v3 2/2] bpf: allow bpf_current_task_under_cgroup() with BPF_CGROUP_* technoboy85
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).