Skip to contentSkip to conversation

libbpf/bpftool: support merging split BTFs

PATCH Superseded Josef Bacik 5 participants 8 messages Feb 24 – Mar 3

Subsystems (5)
Kselftest BPF BPF selftests libbpf +1 more subsystems

v5 of 6 Superseded Latest is v6 →

lore
Share & tools
lore
#0 lore raw

v5 cover libbpf/bpftool: support merging split BTFs

v1: https://lore.kernel.org/bpf/[email protected]/
v2: https://lore.kernel.org/bpf/[email protected]/
v3: https://lore.kernel.org/bpf/[email protected]/
v4: https://lore.kernel.org/bpf/[email protected]/

v1->v2:
- Added a btf__dedup() call to btf__add_btf() to ensure that we don't have duplicate types in the merged BTF.
v2->v3:
- AI review got confused about the UAF comment, so the comment was expanded to clarify the UAF potential.
- Fixed potential clobbering of errno in the error path.
v3->v4:
- Fixed a potential silent corruption pointed out by the AI review bot. v4->v5:
- Addressed Andrii's comments for 1/3.
- Addressed Alan and Quentin's comments for 2/3.
- Addressed Alan's comments for 3/3.
- Added my Signed-off-by for the third patch.
- Made sure to validate everything still worked.

--- Original email ---

Hello,

I'm extending systing to do introspection on vfio devices, which requires having the structs I need from the kernel available in userspace. Normally these are loadable modules, but in the case of vfio there's multiple structs across multiple modules. Normally you'd do the following to generate your vmlinux.h with a module

bpftool btf dump file /sys/kernel/btf/<module> format c \
--base /sys/kernel/btf/vmlinux > vmlinux.h

but if you need multiple modules you have to hack together multiple dumps and merge them together. This patch series adds support for merging multiple BTF sources together, so you can do

bpftool btf dump file /sys/kernel/btf/<module1> \
	file /sys/kernel/btf/<module2> format c \
	--base /sys/kernel/btf/vmlinux > vmlinux.h

I tested this with my usecase and it works. Thanks,

Josef

Josef Bacik (3):
libbpf: support appending split BTF in btf__add_btf()
bpftool: support merging multiple module BTFs in btf dump selftests/bpf: add test for btf__add_btf() with split BTF sources

 .../bpf/bpftool/Documentation/bpftool-btf.rst |  11 +-
 tools/bpf/bpftool/bash-completion/bpftool     |   6 +-
 tools/bpf/bpftool/btf.c                       | 128 ++++++++++++++++--
 tools/lib/bpf/btf.c                           |  37 +++--
 .../selftests/bpf/prog_tests/btf_write.c      | 111 +++++++++++++++
 5 files changed, 267 insertions(+), 26 deletions(-)
Signature · 1 line
2.53.0
#1 lore raw

PATCH v5 1/3 libbpf: support appending split BTF in btf__add_btf()

btf__add_btf() currently rejects split BTF sources with -ENOTSUP. This prevents merging types from multiple kernel module BTFs that are all split against the same vmlinux base.

Extend btf__add_btf() to handle split BTF sources by:

- Replacing the blanket -ENOTSUP with a validation that src and dst share the same base BTF pointer when both are split, returning -EOPNOTSUPP on mismatch.

- Computing src_start_id from the source's base to distinguish base type ID references (which must remain unchanged) from split type IDs (which must be remapped to new positions in the destination).

- Using src_btf->nr_types instead of btf__type_cnt()-1 for the type count, which is correct for both split and non-split sources.

- Skipping base string offsets (< start_str_off) during the string rewrite loop, mirroring the type ID skip pattern. Since src and dst share the same base BTF, base string offsets are already valid and need no remapping.

For non-split sources the behavior is identical: src_start_id is 1, the type_id < 1 guard is never true (VOID is already skipped), and the remapping formula reduces to the original. start_str_off is 0 so no string offsets are skipped.

Assisted-by Claude:claude-opus-4-6
---
 tools/lib/bpf/btf.c | 37 +++++++++++++++++++++++++++++--------
 1 file changed, 29 insertions(+), 8 deletions(-)
tools/lib/bpf/btf.c +29 -8
index 83fe79ffcb8f..be8cc4ee83fe 100644
@@ -2004,12 +2004,18 @@ int btf__add_btf(struct btf *btf, const struct btf *src_btf)
 {
 	struct btf_pipe p = { .src = src_btf, .dst = btf };
 	int data_sz, sz, cnt, i, err, old_strs_len;
+	__u32 src_start_id;
 	__u32 *off;
 	void *t;
 
-	/* appending split BTF isn't supported yet */
-	if (src_btf->base_btf)
-		return libbpf_err(-ENOTSUP);
+	/* When appending split BTF, the destination must share the same base
+	 * BTF so that base type ID references remain valid.
+	 */
+	if (src_btf->base_btf &&
+	    src_btf->base_btf != btf->base_btf)
+		return libbpf_err(-EOPNOTSUPP);
+
+	src_start_id = src_btf->base_btf ? btf__type_cnt(src_btf->base_btf) : 1;
 
 	/* deconstruct BTF, if necessary, and invalidate raw_data */
 	if (btf_ensure_modifiable(btf))
@@ -2021,7 +2027,7 @@ int btf__add_btf(struct btf *btf, const struct btf *src_btf)
 	old_strs_len = btf->hdr->str_len;
 
 	data_sz = src_btf->hdr->type_len;
-	cnt = btf__type_cnt(src_btf) - 1;
+	cnt = src_btf->nr_types;
 
 	/* pre-allocate enough memory for new types */
 	t = btf_add_type_mem(btf, data_sz);
@@ -2060,6 +2066,16 @@ int btf__add_btf(struct btf *btf, const struct btf *src_btf)
 		if (err)
 			goto err_out;
 		while ((str_off = btf_field_iter_next(&it))) {
+			/* For split BTF sources, string offsets referencing
+			 * the base BTF (< start_str_off) remain unchanged
+			 * since dst shares the same base.  Only remap
+			 * strings in the split portion.  For non-split
+			 * sources start_str_off is 0 so all offsets except
+			 * the empty string (handled inside btf_rewrite_str)
+			 * are remapped as before.
+			 */
+			if (*str_off < src_btf->start_str_off)
+				continue;
 			err = btf_rewrite_str(&p, str_off);
 			if (err)
 				goto err_out;
@@ -2074,11 +2090,16 @@ int btf__add_btf(struct btf *btf, const struct btf *src_btf)
 			if (!*type_id) /* nothing to do for VOID references */
 				continue;
 
-			/* we haven't updated btf's type count yet, so
-			 * btf->start_id + btf->nr_types - 1 is the type ID offset we should
-			 * add to all newly added BTF types
+			/* For split BTF sources, type IDs referencing the
+			 * base BTF (< src_start_id) remain unchanged since
+			 * dst shares the same base.  Only remap IDs in the
+			 * split portion.  For non-split sources src_start_id
+			 * is 1 so all IDs are remapped as before.
 			 */
-			*type_id += btf->start_id + btf->nr_types - 1;
+			if (*type_id < src_start_id)
+				continue;
+
+			*type_id += btf->start_id + btf->nr_types - src_start_id;
 		}
 
 		/* go to next type data and type offset index entry */
Signature · 1 line
2.53.0
#2 lore raw

PATCH v5 2/3 bpftool: support merging multiple module BTFs in btf dump

In reply to Josef Bacik

Add support for specifying multiple file sources in 'bpftool btf dump' to generate a single C header containing types from vmlinux plus multiple kernel modules:

bpftool btf dump file /sys/kernel/btf/mod1 file /sys/kernel/btf/mod2 format c

This is useful for BPF programs that need to access types defined in kernel modules. Previously this required a separate bpftool invocation for each module, producing separate headers that could not be combined due to overlapping vmlinux type definitions.

The implementation collects all file paths, then for the multi-file case creates an empty split BTF on the vmlinux base and iteratively merges each module's types into it via btf__add_btf(). The single-file code path is preserved exactly to avoid any regression risk.

Auto-detection of vmlinux as the base BTF from sysfs paths works as before. If vmlinux itself appears in the file list it is skipped with a warning since its types are already provided by the base.

Assisted-by Claude:claude-opus-4-6
---
 .../bpf/bpftool/Documentation/bpftool-btf.rst |  11 +-
 tools/bpf/bpftool/bash-completion/bpftool     |   6 +-
 tools/bpf/bpftool/btf.c                       | 128 ++++++++++++++++--
 3 files changed, 127 insertions(+), 18 deletions(-)
tools/bpf/bpftool/Documentation/bpftool-btf.rst +7 -4
index d47dddc2b4ee..cf75a7fa2d6b 100644
@@ -27,7 +27,7 @@ BTF COMMANDS
 | **bpftool** **btf dump** *BTF_SRC* [**format** *FORMAT*] [**root_id** *ROOT_ID*]
 | **bpftool** **btf help**
 |
-| *BTF_SRC* := { **id** *BTF_ID* | **prog** *PROG* | **map** *MAP* [{**key** | **value** | **kv** | **all**}] | **file** *FILE* }
+| *BTF_SRC* := { **id** *BTF_ID* | **prog** *PROG* | **map** *MAP* [{**key** | **value** | **kv** | **all**}] | **file** *FILE* [**file** *FILE*]... }
 | *FORMAT* := { **raw** | **c** [**unsorted**] }
 | *MAP* := { **id** *MAP_ID* | **pinned** *FILE* }
 | *PROG* := { **id** *PROG_ID* | **pinned** *FILE* | **tag** *PROG_TAG* | **name** *PROG_NAME* }
@@ -58,9 +58,12 @@ bpftool btf dump *BTF_SRC* [format *FORMAT*] [root_id *ROOT_ID*]
     When **prog** is provided, it's expected that program has associated BTF
     object with BTF types.
 
-    When specifying *FILE*, an ELF file is expected, containing .BTF section
-    with well-defined BTF binary format data, typically produced by clang or
-    pahole.
+    When specifying *FILE*, an ELF file or a raw BTF file (e.g. from
+    ``/sys/kernel/btf/``) is expected.  Multiple **file** arguments may be
+    given to merge BTF from several kernel modules into a single output.
+    When sysfs paths are used, vmlinux BTF is loaded automatically as the
+    base; if vmlinux itself appears in the file list it is skipped.
+    A base BTF can also be specified explicitly with **-B**.
 
     **format** option can be used to override default (raw) output format. Raw
     (**raw**) or C-syntax (**c**) output formats are supported. With C-style
tools/bpf/bpftool/bash-completion/bpftool +5 -1
index a28f0cc522e4..babb0d4e9753 100644
@@ -961,10 +961,14 @@ _bpftool()
                         *)
                             # emit extra options
                             case ${words[3]} in
-                                id|file)
+                                id)
                                     COMPREPLY=( $( compgen -W "root_id" -- "$cur" ) )
                                     _bpftool_once_attr 'format'
                                     ;;
+                                file)
+                                    COMPREPLY=( $( compgen -W "root_id file" -- "$cur" ) )
+                                    _bpftool_once_attr 'format'
+                                    ;;
                                 map|prog)
                                     if [[ ${words[3]} == "map" ]] && [[ $cword == 6 ]]; then
                                         COMPREPLY+=( $( compgen -W "key value kv all" -- "$cur" ) )
tools/bpf/bpftool/btf.c +115 -13

154 lines not on this page · raw message

Signature · 1 line
2.53.0
#3 lore raw

Re: [PATCH v5 2/3] bpftool: support merging multiple module BTFs in btf dump

The commit message says vmlinux is "skipped with a warning" but the code filters vmlinux entries silently. In do_dump(), the first file is filtered with:
Quoted 3 lines

Auto-detection of vmlinux as the base BTF from sysfs paths works as before. If vmlinux itself appears in the file list it is skipped with a warning since its types are already provided by the base.

The commit message says vmlinux is "skipped with a warning" but the code filters vmlinux entries silently. In do_dump(), the first file is filtered with:

    if (strcmp(*argv, sysfs_vmlinux) != 0)
        files[nr_files++] = *argv;

and subsequent files with:

    if (strcmp(*argv, sysfs_vmlinux) == 0) {
        NEXT_ARG();
        continue;
    }

Neither path emits a p_info() or any other diagnostic. The v4 version of this patch had a p_info("skipping %s (already loaded as base)") call which was dropped during restructuring for v5. Should the warning be restored, or should the commit message be updated to match the actual behavior?

Quoted 3 lines
tools/bpf/bpftool/btf.c +0 -0

[ ... ]

tools/bpf/bpftool/btf.c static int do_dump(int argc, char **argv) · line 1019

Quoted 6 lines
tools/bpf/bpftool/btf.c +6 -0
+		while (argc && strcmp(*argv, "file") == 0) {
+			NEXT_ARG();
+			if (!REQ_ARGS(1)) {
+				err = -EINVAL;
+				goto done;
+			}

Quentin Monnet asked in v4 review why strcmp() is used here instead of is_prefix(), which is the standard pattern throughout bpftool for keyword matching:

https://lore.kernel.org/bpf/[email protected]/

The initial "file" keyword is matched via is_prefix(src, "file") at the top of do_dump(), so abbreviated forms work for the first file argument but not for subsequent ones. Was this intentional?

tools/bpf/bpftool/btf.c static int do_dump(int argc, char **argv) · line 1024

Quoted 5 lines
tools/bpf/bpftool/btf.c +5 -0
+			/* Filter out any sysfs vmlinux entries. */
+			if (strcmp(*argv, sysfs_vmlinux) == 0) {
+				NEXT_ARG();
+				continue;
+			}

[ ... ]

---
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/22361062191

AI-authorship-score: high
AI-authorship-explanation: The commit explicitly declares AI assistance via the Assisted-by: Claude:claude-opus-4-6 tag.
issues-found: 2
issue-severity-score: low
issue-severity-explanation: Commit message inaccuracy about a missing warning and an unaddressed maintainer review comment about keyword matching convention; neither causes functional problems.

#4 lore raw

Re: [PATCH v5 2/3] bpftool: support merging multiple module BTFs in btf dump

2026-02-24 17:18 UTC+0000 ~ [email protected]

Quoted 30 lines
Quoted 3 lines

Auto-detection of vmlinux as the base BTF from sysfs paths works as before. If vmlinux itself appears in the file list it is skipped with a warning since its types are already provided by the base.

The commit message says vmlinux is "skipped with a warning" but the code filters vmlinux entries silently. In do_dump(), the first file is filtered with:

    if (strcmp(*argv, sysfs_vmlinux) != 0)
        files[nr_files++] = *argv;

and subsequent files with:

    if (strcmp(*argv, sysfs_vmlinux) == 0) {
        NEXT_ARG();
        continue;
    }

Neither path emits a p_info() or any other diagnostic. The v4 version of this patch had a p_info("skipping %s (already loaded as base)") call which was dropped during restructuring for v5. Should the warning be restored, or should the commit message be updated to match the actual behavior?

Quoted 3 lines
tools/bpf/bpftool/btf.c +0 -0

[ ... ]

Quoted 10 lines

tools/bpf/bpftool/btf.c static int do_dump(int argc, char **argv) · line 1019

Quoted 6 lines
tools/bpf/bpftool/btf.c +6 -0
+		while (argc && strcmp(*argv, "file") == 0) {
+			NEXT_ARG();
+			if (!REQ_ARGS(1)) {
+				err = -EINVAL;
+				goto done;
+			}

Quentin Monnet asked in v4 review why strcmp() is used here instead of is_prefix(), which is the standard pattern throughout bpftool for keyword matching:

And the question still stands :)
The rest of the bpftool patch looks good to me, thank you for the changes.

Quoted 13 lines

https://lore.kernel.org/bpf/[email protected]/

The initial "file" keyword is matched via is_prefix(src, "file") at the top of do_dump(), so abbreviated forms work for the first file argument but not for subsequent ones. Was this intentional?

Quoted 5 lines
tools/bpf/bpftool/btf.c +5 -0
+			/* Filter out any sysfs vmlinux entries. */
+			if (strcmp(*argv, sysfs_vmlinux) == 0) {
+				NEXT_ARG();
+				continue;
+			}

[ ... ]

#5 lore raw

Re: [PATCH v5 2/3] bpftool: support merging multiple module BTFs in btf dump

Quoted 39 lines · Quentin Monnet

2026-02-24 17:18 UTC+0000 ~ [email protected]

Quoted 37 lines
Quoted 3 lines

Auto-detection of vmlinux as the base BTF from sysfs paths works as before. If vmlinux itself appears in the file list it is skipped with a warning since its types are already provided by the base.

The commit message says vmlinux is "skipped with a warning" but the code filters vmlinux entries silently. In do_dump(), the first file is filtered with:

    if (strcmp(*argv, sysfs_vmlinux) != 0)
        files[nr_files++] = *argv;

and subsequent files with:

    if (strcmp(*argv, sysfs_vmlinux) == 0) {
        NEXT_ARG();
        continue;
    }

Neither path emits a p_info() or any other diagnostic. The v4 version of this patch had a p_info("skipping %s (already loaded as base)") call which was dropped during restructuring for v5. Should the warning be restored, or should the commit message be updated to match the actual behavior?

Quoted 3 lines
tools/bpf/bpftool/btf.c +0 -0

[ ... ]

Quoted 6 lines
tools/bpf/bpftool/btf.c +6 -0
+            while (argc && strcmp(*argv, "file") == 0) {
+                    NEXT_ARG();
+                    if (!REQ_ARGS(1)) {
+                            err = -EINVAL;
+                            goto done;
+                    }
Quoted 7 lines
Quoted 3 lines

Quentin Monnet asked in v4 review why strcmp() is used here instead of is_prefix(), which is the standard pattern throughout bpftool for keyword matching:

And the question still stands :)
The rest of the bpftool patch looks good to me, thank you for the changes.

I was thinking to patch it up while applying, but I'm not sure whether just that strcmp needs to change or all of them below as well. Please fix and respin.

pw-bot: cr

#6 lore raw

PATCH v5 3/3 selftests/bpf: add test for btf__add_btf() with split BTF sources

In reply to Josef Bacik

Add a test that verifies btf__add_btf() correctly handles merging multiple split BTF objects that share the same base BTF. The test creates two sibling split BTFs on a common base, merges them into a combined split BTF, and validates that base type references are preserved while split type references are properly remapped.

Assisted-by Claude:claude-opus-4-6
---
 .../selftests/bpf/prog_tests/btf_write.c      | 111 ++++++++++++++++++
 1 file changed, 111 insertions(+)
tools/testing/selftests/bpf/prog_tests/btf_write.c +111 -0
index 6e36de1302fc..5c84723cf254 100644
@@ -497,10 +497,121 @@ static void test_btf_add_btf()
 	btf__free(btf2);
 }
 
+static void test_btf_add_btf_split()
+{
+	struct btf *base = NULL, *split1 = NULL, *split2 = NULL;
+	struct btf *combined = NULL;
+	int id, err;
+
+	/* Create a base BTF with an INT and a PTR to it */
+	base = btf__new_empty();
+	if (!ASSERT_OK_PTR(base, "base"))
+		return;
+
+	id = btf__add_int(base, "int", 4, BTF_INT_SIGNED);
+	ASSERT_EQ(id, 1, "base_int_id");
+	id = btf__add_ptr(base, 1);
+	ASSERT_EQ(id, 2, "base_ptr_id");
+
+	/* base has 2 types, type IDs 1..2 */
+	ASSERT_EQ(btf__type_cnt(base), 3, "base_type_cnt");
+
+	/* Create split1 on base: a STRUCT referencing base's int (ID 1) */
+	split1 = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(split1, "split1"))
+		goto cleanup;
+
+	id = btf__add_struct(split1, "s1", 4);
+	/* split types start at base_type_cnt = 3 */
+	ASSERT_EQ(id, 3, "split1_struct_id");
+	btf__add_field(split1, "x", 1, 0, 0); /* refers to base int */
+
+	id = btf__add_ptr(split1, 3);
+	ASSERT_EQ(id, 4, "split1_ptr_id"); /* ptr to the struct (split self-ref) */
+
+	/* Add a typedef "int_alias" -> base int in split1, which will be
+	 * duplicated in split2 to test that btf__dedup() merges them.
+	 */
+	id = btf__add_typedef(split1, "int_alias", 1);
+	ASSERT_EQ(id, 5, "split1_typedef_id");
+
+	/* Create split2 on base: a TYPEDEF referencing base's ptr (ID 2) */
+	split2 = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(split2, "split2"))
+		goto cleanup;
+
+	id = btf__add_typedef(split2, "int_ptr", 2); /* refers to base ptr */
+	ASSERT_EQ(id, 3, "split2_typedef_id");
+
+	id = btf__add_struct(split2, "s2", 8);
+	ASSERT_EQ(id, 4, "split2_struct_id");
+	btf__add_field(split2, "p", 3, 0, 0); /* refers to split2's own typedef */
+
+	/* Same "int_alias" typedef as split1 - should be deduped away */
+	id = btf__add_typedef(split2, "int_alias", 1);
+	ASSERT_EQ(id, 5, "split2_dup_typedef_id");
+
+	/* Create combined split BTF on same base and merge both */
+	combined = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(combined, "combined"))
+		goto cleanup;
+
+	/* Merge split1: its types (3,4,5) should land at IDs 3,4,5 */
+	id = btf__add_btf(combined, split1);
+	if (!ASSERT_GE(id, 0, "add_split1"))
+		goto cleanup;
+	ASSERT_EQ(id, 3, "split1_first_id");
+
+	/* Merge split2: its types (3,4,5) should be remapped to 6,7,8 */
+	id = btf__add_btf(combined, split2);
+	if (!ASSERT_GE(id, 0, "add_split2"))
+		goto cleanup;
+	ASSERT_EQ(id, 6, "split2_first_id");
+
+	/* Before dedup: base (2) + split1 (3) + split2 (3) = 8 types + void */
+	ASSERT_EQ(btf__type_cnt(combined), 9, "pre_dedup_type_cnt");
+
+	VALIDATE_RAW_BTF(
+		combined,
+		/* base types (IDs 1-2) */
+		"[1] INT 'int' size=4 bits_offset=0 nr_bits=32 encoding=SIGNED",
+		"[2] PTR '(anon)' type_id=1",
+
+		/* split1 types (IDs 3-5): base refs unchanged */
+		"[3] STRUCT 's1' size=4 vlen=1\n"
+		"\t'x' type_id=1 bits_offset=0",      /* refers to base int=1 */
+		"[4] PTR '(anon)' type_id=3",          /* refers to split1's struct=3 */
+		"[5] TYPEDEF 'int_alias' type_id=1",   /* refers to base int=1 */
+
+		/* split2 types (IDs 6-8): remapped from 3,4,5 to 6,7,8 */
+		"[6] TYPEDEF 'int_ptr' type_id=2",     /* base ptr=2, unchanged */
+		"[7] STRUCT 's2' size=8 vlen=1\n"
+		"\t'p' type_id=6 bits_offset=0",       /* split2 typedef: 3->6 */
+		"[8] TYPEDEF 'int_alias' type_id=1");   /* dup of [5] */
+
+	/* Dedup to mirror the bpftool merge flow; should remove the
+	 * duplicate "int_alias" typedef.
+	 */
+	err = btf__dedup(combined, NULL);
+	if (!ASSERT_OK(err, "dedup"))
+		goto cleanup;
+
+	/* After dedup: one int_alias removed, so 7 types + void */
+	ASSERT_EQ(btf__type_cnt(combined), 8, "dedup_type_cnt");
+
+cleanup:
+	btf__free(combined);
+	btf__free(split2);
+	btf__free(split1);
+	btf__free(base);
+}
+
 void test_btf_write()
 {
 	if (test__start_subtest("btf_add"))
 		test_btf_add();
 	if (test__start_subtest("btf_add_btf"))
 		test_btf_add_btf();
+	if (test__start_subtest("btf_add_btf_split"))
+		test_btf_add_btf_split();
 }
Signature · 1 line
2.53.0
Reviewed-by on v5 3/3
#7 lore raw

Re: [PATCH v5 3/3] selftests/bpf: add test for btf__add_btf() with split BTF sources

Quoted 8 lines · Josef Bacik

Add a test that verifies btf__add_btf() correctly handles merging multiple split BTF objects that share the same base BTF. The test creates two sibling split BTFs on a common base, merges them into a combined split BTF, and validates that base type references are preserved while split type references are properly remapped.

Assisted-by Claude:claude-opus-4-6
Quoted 130 lines
---
 .../selftests/bpf/prog_tests/btf_write.c      | 111 ++++++++++++++++++
 1 file changed, 111 insertions(+)
tools/testing/selftests/bpf/prog_tests/btf_write.c +111 -0
index 6e36de1302fc..5c84723cf254 100644
@@ -497,10 +497,121 @@ static void test_btf_add_btf()
 	btf__free(btf2);
 }
 
+static void test_btf_add_btf_split()
+{
+	struct btf *base = NULL, *split1 = NULL, *split2 = NULL;
+	struct btf *combined = NULL;
+	int id, err;
+
+	/* Create a base BTF with an INT and a PTR to it */
+	base = btf__new_empty();
+	if (!ASSERT_OK_PTR(base, "base"))
+		return;
+
+	id = btf__add_int(base, "int", 4, BTF_INT_SIGNED);
+	ASSERT_EQ(id, 1, "base_int_id");
+	id = btf__add_ptr(base, 1);
+	ASSERT_EQ(id, 2, "base_ptr_id");
+
+	/* base has 2 types, type IDs 1..2 */
+	ASSERT_EQ(btf__type_cnt(base), 3, "base_type_cnt");
+
+	/* Create split1 on base: a STRUCT referencing base's int (ID 1) */
+	split1 = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(split1, "split1"))
+		goto cleanup;
+
+	id = btf__add_struct(split1, "s1", 4);
+	/* split types start at base_type_cnt = 3 */
+	ASSERT_EQ(id, 3, "split1_struct_id");
+	btf__add_field(split1, "x", 1, 0, 0); /* refers to base int */
+
+	id = btf__add_ptr(split1, 3);
+	ASSERT_EQ(id, 4, "split1_ptr_id"); /* ptr to the struct (split self-ref) */
+
+	/* Add a typedef "int_alias" -> base int in split1, which will be
+	 * duplicated in split2 to test that btf__dedup() merges them.
+	 */
+	id = btf__add_typedef(split1, "int_alias", 1);
+	ASSERT_EQ(id, 5, "split1_typedef_id");
+
+	/* Create split2 on base: a TYPEDEF referencing base's ptr (ID 2) */
+	split2 = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(split2, "split2"))
+		goto cleanup;
+
+	id = btf__add_typedef(split2, "int_ptr", 2); /* refers to base ptr */
+	ASSERT_EQ(id, 3, "split2_typedef_id");
+
+	id = btf__add_struct(split2, "s2", 8);
+	ASSERT_EQ(id, 4, "split2_struct_id");
+	btf__add_field(split2, "p", 3, 0, 0); /* refers to split2's own typedef */
+
+	/* Same "int_alias" typedef as split1 - should be deduped away */
+	id = btf__add_typedef(split2, "int_alias", 1);
+	ASSERT_EQ(id, 5, "split2_dup_typedef_id");
+
+	/* Create combined split BTF on same base and merge both */
+	combined = btf__new_empty_split(base);
+	if (!ASSERT_OK_PTR(combined, "combined"))
+		goto cleanup;
+
+	/* Merge split1: its types (3,4,5) should land at IDs 3,4,5 */
+	id = btf__add_btf(combined, split1);
+	if (!ASSERT_GE(id, 0, "add_split1"))
+		goto cleanup;
+	ASSERT_EQ(id, 3, "split1_first_id");
+
+	/* Merge split2: its types (3,4,5) should be remapped to 6,7,8 */
+	id = btf__add_btf(combined, split2);
+	if (!ASSERT_GE(id, 0, "add_split2"))
+		goto cleanup;
+	ASSERT_EQ(id, 6, "split2_first_id");
+
+	/* Before dedup: base (2) + split1 (3) + split2 (3) = 8 types + void */
+	ASSERT_EQ(btf__type_cnt(combined), 9, "pre_dedup_type_cnt");
+
+	VALIDATE_RAW_BTF(
+		combined,
+		/* base types (IDs 1-2) */
+		"[1] INT 'int' size=4 bits_offset=0 nr_bits=32 encoding=SIGNED",
+		"[2] PTR '(anon)' type_id=1",
+
+		/* split1 types (IDs 3-5): base refs unchanged */
+		"[3] STRUCT 's1' size=4 vlen=1\n"
+		"\t'x' type_id=1 bits_offset=0",      /* refers to base int=1 */
+		"[4] PTR '(anon)' type_id=3",          /* refers to split1's struct=3 */
+		"[5] TYPEDEF 'int_alias' type_id=1",   /* refers to base int=1 */
+
+		/* split2 types (IDs 6-8): remapped from 3,4,5 to 6,7,8 */
+		"[6] TYPEDEF 'int_ptr' type_id=2",     /* base ptr=2, unchanged */
+		"[7] STRUCT 's2' size=8 vlen=1\n"
+		"\t'p' type_id=6 bits_offset=0",       /* split2 typedef: 3->6 */
+		"[8] TYPEDEF 'int_alias' type_id=1");   /* dup of [5] */
+
+	/* Dedup to mirror the bpftool merge flow; should remove the
+	 * duplicate "int_alias" typedef.
+	 */
+	err = btf__dedup(combined, NULL);
+	if (!ASSERT_OK(err, "dedup"))
+		goto cleanup;
+
+	/* After dedup: one int_alias removed, so 7 types + void */
+	ASSERT_EQ(btf__type_cnt(combined), 8, "dedup_type_cnt");
+
+cleanup:
+	btf__free(combined);
+	btf__free(split2);
+	btf__free(split1);
+	btf__free(base);
+}
+
 void test_btf_write()
 {
 	if (test__start_subtest("btf_add"))
 		test_btf_add();
 	if (test__start_subtest("btf_add_btf"))
 		test_btf_add_btf();
+	if (test__start_subtest("btf_add_btf_split"))
+		test_btf_add_btf_split();
 }