Skip to contentSkip to conversation

How about adding a name for bpftool self created maps?

DISCUSSION Hangbin Liu 2 participants 6 messages Jun 21, 2022 – Aug 5, 2022

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

Hi Quentin,

When I using bpftool map list to show what maps user using. bpftool will show maps including self created maps. Apart from the "pid_iter.rodata", there are also array maps without name, which makes it hard to filter out what maps user using. e.g.

# bpftool map list
63: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B
65: array  name pid_iter.rodata  flags 0x480
         key 4B  value 4B  max_entries 1  memlock 4096B
         btf_id 98  frozen
         pids bpftool(10572)
66: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B

So do you have plan to add a special name for the bpftool self created maps? Or do you know if there is a way to filter out these maps?

Thanks
Hangbin

#1 lore raw
Quoted 18 lines · Hangbin Liu

Hi Quentin,

When I using bpftool map list to show what maps user using. bpftool will show maps including self created maps. Apart from the "pid_iter.rodata", there are also array maps without name, which makes it hard to filter out what maps user using. e.g.

# bpftool map list
63: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B
65: array  name pid_iter.rodata  flags 0x480
         key 4B  value 4B  max_entries 1  memlock 4096B
         btf_id 98  frozen
         pids bpftool(10572)
66: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B
Quoted 2 lines

So do you have plan to add a special name for the bpftool self created maps? Or do you know if there is a way to filter out these maps?

Hi Hangbin,

No plan currently. Adding names has been suggested before, but it's not compatible with some older kernels that don't support map names [0]. Maybe one solution would be to probe the kernel for map name support, and to add a name if supported.

Other than this I'm not aware of a reliable way to filter out these maps at the moment. This could probably be done in bpftool since we should have the ids of the self-generated maps. But I think I'd rather find a way to add map names, if possible. It would make it easier to recognise/filter these maps on regular listing, whereas a new option would be harder for users to discover.

Quentin

[0] https://lore.kernel.org/bpf/CAEf4BzY66WPKQbDe74AKZ6nFtZjq5e+G3Ji2egcVytB9R6_sGQ@mail.gmail.com/

#2 lore raw
Quoted 18 lines · Quentin Monnet
Quoted 15 lines

When I using bpftool map list to show what maps user using. bpftool will show maps including self created maps. Apart from the "pid_iter.rodata", there are also array maps without name, which makes it hard to filter out what maps user using. e.g.

# bpftool map list
63: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B
65: array  name pid_iter.rodata  flags 0x480
         key 4B  value 4B  max_entries 1  memlock 4096B
         btf_id 98  frozen
         pids bpftool(10572)
66: array  flags 0x0
         key 4B  value 32B  max_entries 1  memlock 4096B
Quoted 2 lines

So do you have plan to add a special name for the bpftool self created maps? Or do you know if there is a way to filter out these maps?

Quoted 6 lines

Hi Hangbin,

No plan currently. Adding names has been suggested before, but it's not compatible with some older kernels that don't support map names [0]. Maybe one solution would be to probe the kernel for map name support, and to add a name if supported.

Hi Quentin,

Thanks for the reply. Probe the kernel and add name if supported makes sense to me.

Thanks
Hangbin

Quoted 11 lines

Other than this I'm not aware of a reliable way to filter out these maps at the moment. This could probably be done in bpftool since we should have the ids of the self-generated maps. But I think I'd rather find a way to add map names, if possible. It would make it easier to recognise/filter these maps on regular listing, whereas a new option would be harder for users to discover.

Quentin

[0] https://lore.kernel.org/bpf/CAEf4BzY66WPKQbDe74AKZ6nFtZjq5e+G3Ji2egcVytB9R6_sGQ@mail.gmail.com/

#3 lore raw
In reply to Quentin Monnet
Quoted 6 lines · Quentin Monnet

Hi Hangbin,

No plan currently. Adding names has been suggested before, but it's not compatible with some older kernels that don't support map names [0]. Maybe one solution would be to probe the kernel for map name support, and to add a name if supported.

Hi Quentin,

I looked into this issue this week. And I have some questions. Can we just use the probe_kern_prog_name() function directly? e.g.

tools/lib/bpf/libbpf.c +4 -1
index e89cc9c885b3..f7d1580cd54e 100644
@@ -4476,7 +4476,10 @@ static int probe_kern_global_data(void)
        };
        int ret, map, insn_cnt = ARRAY_SIZE(insns);
 
-       map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
+       if (probe_kern_prog_name() > 0)
+               map = bpf_map_create(BPF_MAP_TYPE_ARRAY, "global_data", sizeof(int), 32, 1, NULL);
+       else
+               map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
        if (map < 0) {
                ret = -errno;
                cp = libbpf_strerror_r(ret, errmsg, sizeof(errmsg));

I know the map name and prog name supports are not in the same patch. But they are added to kernel in one patch series. I doubt any one will backport them separately.

And I also have a question about function probe_kern_prog_name(). I only saw it created a prog with name "test". But I didn't find the function check if the prog are really has name "test". If a old kernel doesn't support prog name, I think it will just ignore the name field. No?

Another way I think we can use to probe if kernel supports map name is try to attach a kprobe/bpf_obj_name_cpy. If attach success, the kernel should support the prog/map name. WDYT?

Thanks
Hangbin

#4 lore raw
Quoted 30 lines · Hangbin Liu
Quoted 6 lines · Quentin Monnet

Hi Hangbin,

No plan currently. Adding names has been suggested before, but it's not compatible with some older kernels that don't support map names [0]. Maybe one solution would be to probe the kernel for map name support, and to add a name if supported.

Hi Quentin,

I looked into this issue this week. And I have some questions. Can we just use the probe_kern_prog_name() function directly? e.g.

tools/lib/bpf/libbpf.c +4 -1
index e89cc9c885b3..f7d1580cd54e 100644
@@ -4476,7 +4476,10 @@ static int probe_kern_global_data(void)
        };
        int ret, map, insn_cnt = ARRAY_SIZE(insns);
 
-       map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
+       if (probe_kern_prog_name() > 0)
+               map = bpf_map_create(BPF_MAP_TYPE_ARRAY, "global_data", sizeof(int), 32, 1, NULL);
+       else
+               map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
        if (map < 0) {
                ret = -errno;
                cp = libbpf_strerror_r(ret, errmsg, sizeof(errmsg));
Quoted 2 lines

I know the map name and prog name supports are not in the same patch. But they are added to kernel in one patch series. I doubt any one will backport them separately.

Hi! It would look much cleaner to have something specific to map names. It does not have to be a dedicated probe in my opinion, maybe we can just try loading with a name and retry if this fails with -EINVAL (a bit like we retry with another prog type in bpf_object__probe_loading(), if the first one fails). Something like this (not tested):

tools/lib/bpf/libbpf.c +4 -1
index 50d41815f431..abcafdf8ae7e 100644
@@ -4430,7 +4430,10 @@ static int probe_kern_global_data(void)
 	};
 	int ret, map, insn_cnt = ARRAY_SIZE(insns);
 
-	map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32,  1, NULL);
+	map = bpf_map_create(BPF_MAP_TYPE_ARRAY, "global_data", sizeof(int), 32, 1, NULL);
+	if (map < 0 && errno == EINVAL)
+		/* Retry without name */
+		map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
 	if (map < 0) {
 		ret = -errno;
 		cp = libbpf_strerror_r(ret, errmsg, sizeof(errmsg));

(Maybe with a small wrapper, given that we'd also need this in probe_prog_bind_map() and probe_kern_array_mmap() as well.)

Quoted 4 lines

And I also have a question about function probe_kern_prog_name(). I only saw it created a prog with name "test". But I didn't find the function check if the prog are really has name "test". If a old kernel doesn't support prog name, I think it will just ignore the name field. No?

No, "if (CHECK_ATTR(BPF_PROG_LOAD))" should fail in bpf_prog_load() in kernel/bpf/syscall.c, and the syscall should fail with -EINVAL.

If older kernels simply ignored the "name" field for programs and maps, we wouldn't have to probe or retry for the current case in the first place :).

Quoted 3 lines

Another way I think we can use to probe if kernel supports map name is try to attach a kprobe/bpf_obj_name_cpy. If attach success, the kernel should support the prog/map name. WDYT?

It's probably easier to try to load a map with a name. Also kprobes can be disabled, if I remember correctly.

Thanks,
Quentin

#5 lore raw
Quoted 23 lines · Quentin Monnet

Hi! It would look much cleaner to have something specific to map names. It does not have to be a dedicated probe in my opinion, maybe we can just try loading with a name and retry if this fails with -EINVAL (a bit like we retry with another prog type in bpf_object__probe_loading(), if the first one fails). Something like this (not tested):

tools/lib/bpf/libbpf.c +4 -1
index 50d41815f431..abcafdf8ae7e 100644
@@ -4430,7 +4430,10 @@ static int probe_kern_global_data(void)
 	};
 	int ret, map, insn_cnt = ARRAY_SIZE(insns);
 
-	map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32,  1, NULL);
+	map = bpf_map_create(BPF_MAP_TYPE_ARRAY, "global_data", sizeof(int), 32, 1, NULL);
+	if (map < 0 && errno == EINVAL)
+		/* Retry without name */
+		map = bpf_map_create(BPF_MAP_TYPE_ARRAY, NULL, sizeof(int), 32, 1, NULL);
 	if (map < 0) {
 		ret = -errno;
 		cp = libbpf_strerror_r(ret, errmsg, sizeof(errmsg));
Quoted 2 lines

(Maybe with a small wrapper, given that we'd also need this in probe_prog_bind_map() and probe_kern_array_mmap() as well.)

Ah, this looks more clean and easier.

Quoted 6 lines
Quoted 5 lines

And I also have a question about function probe_kern_prog_name(). I only saw it created a prog with name "test". But I didn't find the function check if the prog are really has name "test". If a old kernel doesn't support prog name, I think it will just ignore the name field. No?

Quoted 6 lines

No, "if (CHECK_ATTR(BPF_PROG_LOAD))" should fail in bpf_prog_load() in kernel/bpf/syscall.c, and the syscall should fail with -EINVAL.

If older kernels simply ignored the "name" field for programs and maps, we wouldn't have to probe or retry for the current case in the first place :).

Thanks for the explanation. I will try add a wrapper first.

Regards
Hangbin