The program is made in a way that everytime an execve syscall
is executed it prints Hello, BPF World!
This is inspired and based on the code example for the book
Linux Observability with BPF [1], load_bpf_file() has been
removed after commit ceb5dea56543 ("samples: bpf: Remove
bpf_load loader completely"), so the old version can not
work in the latest mainline kernel.
Since it is very simple and useful for newbies, I think it is
necessary to be upstreamed.
[1] https://github.com/bpftools/linux-observability-with-bpf/tree/master/code/chapter-2/hello_world
Signed-off-by: Tiezhu Yang <[email protected]>
---
samples/bpf/Makefile | 3 +++
samples/bpf/hello_kern.c | 14 ++++++++++++++
samples/bpf/hello_user.c | 42 ++++++++++++++++++++++++++++++++++++++++++
3 files changed, 59 insertions(+)
create mode 100644 samples/bpf/hello_kern.c
create mode 100644 samples/bpf/hello_user.c
diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
index 45ceca4..fd17cbd 100644
--- a/samples/bpf/Makefile
+++ b/samples/bpf/Makefile
@@ -55,6 +55,7 @@ tprogs-y += task_fd_query
tprogs-y += xdp_sample_pkts
tprogs-y += ibumad
tprogs-y += hbm
+tprogs-y += hello
# Libbpf dependencies
LIBBPF = $(TOOLS_PATH)/lib/bpf/libbpf.a
@@ -113,6 +114,7 @@ task_fd_query-objs := task_fd_query_user.o $(TRACE_HELPERS)
xdp_sample_pkts-objs := xdp_sample_pkts_user.o
ibumad-objs := ibumad_user.o
hbm-objs := hbm.o $(CGROUP_HELPERS)
+hello-objs := hello_user.o $(TRACE_HELPERS)
# Tell kbuild to always build the programs
always-y := $(tprogs-y)
@@ -174,6 +176,7 @@ always-y += ibumad_kern.o
always-y += hbm_out_kern.o
always-y += hbm_edt_kern.o
always-y += xdpsock_kern.o
+always-y += hello_kern.o
ifeq ($(ARCH), arm)
# Strip all except -D__LINUX_ARM_ARCH__ option needed to handle linux
diff --git a/samples/bpf/hello_kern.c b/samples/bpf/hello_kern.c
new file mode 100644
index 0000000..b841029
--- /dev/null
+++ b/samples/bpf/hello_kern.c
@@ -0,0 +1,14 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <linux/bpf.h>
+#include <bpf/bpf_helpers.h>
+
+SEC("tracepoint/syscalls/sys_enter_execve")
+int trace_enter_execve(void *ctx)
+{
+ static const char msg[] = "Hello, BPF World!\n";
+
+ bpf_trace_printk(msg, sizeof(msg));
+ return 0;
+}
+
+char _license[] SEC("license") = "GPL";
diff --git a/samples/bpf/hello_user.c b/samples/bpf/hello_user.c
new file mode 100644
index 0000000..9423bbb
--- /dev/null
+++ b/samples/bpf/hello_user.c
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <stdio.h>
+#include <bpf/libbpf.h>
+#include "trace_helpers.h"
+
+int main(int argc, char **argv)
+{
+ struct bpf_link *link = NULL;
+ struct bpf_program *prog;
+ struct bpf_object *obj;
+
+ obj = bpf_object__open_file("hello_kern.o", NULL);
+ if (libbpf_get_error(obj)) {
+ fprintf(stderr, "ERROR: opening BPF object file failed\n");
+ return 0;
+ }
+
+ if (bpf_object__load(obj)) {
+ fprintf(stderr, "ERROR: loading BPF object file failed\n");
+ goto cleanup;
+ }
+
+ prog = bpf_object__find_program_by_name(obj, "trace_enter_execve");
+ if (!prog) {
+ fprintf(stderr, "ERROR: finding a prog in obj file failed\n");
+ goto cleanup;
+ }
+
+ link = bpf_program__attach(prog);
+ if (libbpf_get_error(link)) {
+ fprintf(stderr, "ERROR: bpf_program__attach failed\n");
+ link = NULL;
+ goto cleanup;
+ }
+
+ read_trace_pipe();
+
+cleanup:
+ bpf_link__destroy(link);
+ bpf_object__close(obj);
+ return 0;
+}
--
2.1.0
On Thu, Feb 4, 2021 at 3:42 AM Tiezhu Yang <[email protected]> wrote:
>
> The program is made in a way that everytime an execve syscall
> is executed it prints Hello, BPF World!
>
> This is inspired and based on the code example for the book
> Linux Observability with BPF [1], load_bpf_file() has been
> removed after commit ceb5dea56543 ("samples: bpf: Remove
> bpf_load loader completely"), so the old version can not
> work in the latest mainline kernel.
>
> Since it is very simple and useful for newbies, I think it is
> necessary to be upstreamed.
I wonder how much value we will get from this sample. If the user is
able to compile and try the hello world, they are sure able to compile
other code in samples/bpf. Also, this code doesn't use BPF skeleton,
which is the recommended way to write BPF programs. Maybe an
minimal example with BPF skeleton will add more value here?
Thanks,
Song
>
> [1] https://github.com/bpftools/linux-observability-with-bpf/tree/master/code/chapter-2/hello_world
>
> Signed-off-by: Tiezhu Yang <[email protected]>
> ---
> samples/bpf/Makefile | 3 +++
> samples/bpf/hello_kern.c | 14 ++++++++++++++
> samples/bpf/hello_user.c | 42 ++++++++++++++++++++++++++++++++++++++++++
> 3 files changed, 59 insertions(+)
> create mode 100644 samples/bpf/hello_kern.c
> create mode 100644 samples/bpf/hello_user.c
>
> diff --git a/samples/bpf/Makefile b/samples/bpf/Makefile
> index 45ceca4..fd17cbd 100644
> --- a/samples/bpf/Makefile
> +++ b/samples/bpf/Makefile
> @@ -55,6 +55,7 @@ tprogs-y += task_fd_query
> tprogs-y += xdp_sample_pkts
> tprogs-y += ibumad
> tprogs-y += hbm
> +tprogs-y += hello
>
> # Libbpf dependencies
> LIBBPF = $(TOOLS_PATH)/lib/bpf/libbpf.a
> @@ -113,6 +114,7 @@ task_fd_query-objs := task_fd_query_user.o $(TRACE_HELPERS)
> xdp_sample_pkts-objs := xdp_sample_pkts_user.o
> ibumad-objs := ibumad_user.o
> hbm-objs := hbm.o $(CGROUP_HELPERS)
> +hello-objs := hello_user.o $(TRACE_HELPERS)
>
> # Tell kbuild to always build the programs
> always-y := $(tprogs-y)
> @@ -174,6 +176,7 @@ always-y += ibumad_kern.o
> always-y += hbm_out_kern.o
> always-y += hbm_edt_kern.o
> always-y += xdpsock_kern.o
> +always-y += hello_kern.o
>
> ifeq ($(ARCH), arm)
> # Strip all except -D__LINUX_ARM_ARCH__ option needed to handle linux
> diff --git a/samples/bpf/hello_kern.c b/samples/bpf/hello_kern.c
> new file mode 100644
> index 0000000..b841029
> --- /dev/null
> +++ b/samples/bpf/hello_kern.c
> @@ -0,0 +1,14 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include <linux/bpf.h>
> +#include <bpf/bpf_helpers.h>
> +
> +SEC("tracepoint/syscalls/sys_enter_execve")
> +int trace_enter_execve(void *ctx)
> +{
> + static const char msg[] = "Hello, BPF World!\n";
> +
> + bpf_trace_printk(msg, sizeof(msg));
> + return 0;
> +}
> +
> +char _license[] SEC("license") = "GPL";
> diff --git a/samples/bpf/hello_user.c b/samples/bpf/hello_user.c
> new file mode 100644
> index 0000000..9423bbb
> --- /dev/null
> +++ b/samples/bpf/hello_user.c
> @@ -0,0 +1,42 @@
> +// SPDX-License-Identifier: GPL-2.0
> +#include <stdio.h>
> +#include <bpf/libbpf.h>
> +#include "trace_helpers.h"
> +
> +int main(int argc, char **argv)
> +{
> + struct bpf_link *link = NULL;
> + struct bpf_program *prog;
> + struct bpf_object *obj;
> +
> + obj = bpf_object__open_file("hello_kern.o", NULL);
> + if (libbpf_get_error(obj)) {
> + fprintf(stderr, "ERROR: opening BPF object file failed\n");
> + return 0;
> + }
> +
> + if (bpf_object__load(obj)) {
> + fprintf(stderr, "ERROR: loading BPF object file failed\n");
> + goto cleanup;
> + }
> +
> + prog = bpf_object__find_program_by_name(obj, "trace_enter_execve");
> + if (!prog) {
> + fprintf(stderr, "ERROR: finding a prog in obj file failed\n");
> + goto cleanup;
> + }
> +
> + link = bpf_program__attach(prog);
> + if (libbpf_get_error(link)) {
> + fprintf(stderr, "ERROR: bpf_program__attach failed\n");
> + link = NULL;
> + goto cleanup;
> + }
> +
> + read_trace_pipe();
> +
> +cleanup:
> + bpf_link__destroy(link);
> + bpf_object__close(obj);
> + return 0;
> +}
> --
> 2.1.0
>
On Thu, Feb 4, 2021 at 11:27 AM Song Liu <[email protected]> wrote:
>
> On Thu, Feb 4, 2021 at 3:42 AM Tiezhu Yang <[email protected]> wrote:
> >
> > The program is made in a way that everytime an execve syscall
> > is executed it prints Hello, BPF World!
> >
> > This is inspired and based on the code example for the book
> > Linux Observability with BPF [1], load_bpf_file() has been
> > removed after commit ceb5dea56543 ("samples: bpf: Remove
> > bpf_load loader completely"), so the old version can not
> > work in the latest mainline kernel.
> >
> > Since it is very simple and useful for newbies, I think it is
> > necessary to be upstreamed.
>
> I wonder how much value we will get from this sample. If the user is
> able to compile and try the hello world, they are sure able to compile
> other code in samples/bpf. Also, this code doesn't use BPF skeleton,
> which is the recommended way to write BPF programs. Maybe an
> minimal example with BPF skeleton will add more value here?
>
I agree with Song. Plus, we already have similar simple examples that
are set up outside of kernel build infrastructure, which is simpler
for newbies to pick up. Please check [0]:
[0] https://github.com/libbpf/libbpf-bootstrap/tree/master/src
> Thanks,
> Song
>
>
>
> >
> > [1] https://github.com/bpftools/linux-observability-with-bpf/tree/master/code/chapter-2/hello_world
> >
> > Signed-off-by: Tiezhu Yang <[email protected]>
> > ---
> > samples/bpf/Makefile | 3 +++
> > samples/bpf/hello_kern.c | 14 ++++++++++++++
> > samples/bpf/hello_user.c | 42 ++++++++++++++++++++++++++++++++++++++++++
> > 3 files changed, 59 insertions(+)
> > create mode 100644 samples/bpf/hello_kern.c
> > create mode 100644 samples/bpf/hello_user.c
> >
[...]