Skip to contentSkip to conversation

permit merging all dwarf cu's for clang lto built binary

PATCH Superseded Yonghong Song 1 participant 5 messages Mar 28, 2021

Subsystems (2)

v2 of 3 Superseded Latest is v3 →

lore
Share & tools
lore
#0 lore raw

v2 cover permit merging all dwarf cu's for clang lto built binary

For vmlinux built with clang thin-lto or lto for latest bpf-next, there exist cross cu debuginfo type references. For example,
compile unit 1:
tag 10: type A
compile unit 2:
...
refer to type A (tag 10 in compile unit 1)
I only checked a few but have seen type A may be a simple type like "unsigned char" or a complex type like an array of base types. I am using latest llvm trunk and bpf-next. I suspect llvm12 or linus tree >= 5.12 rc2 should be able to exhibit the issue as well. Both thin-lto and lto have the same issues.

Current pahole cannot handle this. It will report types cannot be found error. Bill Wendling has attempted to fix the issue with [1] by permitting all tags/types are hashed to the same hash table and then process cu's one by one. This does not really work. The reason is that each cu resolves types locally so for the above example we may have
compile unit 1:
type A : type_id = 10
compile unit 2:
refer to type A : type A will be resolved as type id = 10
But id 10 refers to compile unit 1, we will get either out of bound type id or incorrect one.

This patch set is a continuation of Bill's work. We still increase the hashtable size and traverse all cu's before recoding and finalization. But instead of creating one-to-one mapping between debuginfo cu and pahole cu, we just create one pahole cu, which should solve the above incorrect type id issue.

Patch #1 and #2 are refactoring the existing code and Patch #3 added logic to premit merging all debuginfo cu's into one pahole cu. The detection for whether merging is desirable is done by checking the existence of "clang" compiler and its "lto" option in dwarf producer tag.

The following linux patch is needed to put clang compiler flags into dwarf. It will be submitted separately

Makefile +6 -0
index d4784d181123..ab0119beb42d 100644
@@ -839,6 +839,12 @@ dwarf-version-$(CONFIG_DEBUG_INFO_DWARF5) := 5
 DEBUG_CFLAGS   += -gdwarf-$(dwarf-version-y)
 endif
 
+# gcc emits compilation flags in dwarf DW_AT_producer by default
+# while clang needs explicit flag. Add this flag explicitly.
+ifdef CONFIG_CC_IS_CLANG
+DEBUG_CFLAGS   += -grecord-gcc-switches
+endif
+
 ifdef CONFIG_DEBUG_INFO_REDUCED
 DEBUG_CFLAGS   += $(call cc-option, -femit-struct-debug-baseonly) \
                   $(call cc-option,-fno-var-tracking)

Changelogs:
v1 -> v2:
. removed "--merge_cus" option, relied on detections on
clang compiler and its lto flags.

Yonghong Song (3):
dwarf_loader: permits flexible HASHTAGS__BITS
dwarf_loader: factor out common code to initialize a cu
dwarf_loader: permit merging all dwarf cu's for clang lto built binary

 dwarf_loader.c | 209 +++++++++++++++++++++++++++++++++++++++++--------
 1 file changed, 175 insertions(+), 34 deletions(-)
Signature · 1 line
2.30.2
#1 lore raw

PATCH v2 1/3 dwarf_loader: permits flexible HASHTAGS__BITS

Currently, types/tags hash table has fixed HASHTAGS__BITS = 15. That means the number of buckets will be 1UL << 15 = 32768. In my experiments, a thin-LTO built vmlinux has roughly 9M entries in types table and 5.2M entries in tags table. So the number of buckets is too less for an efficient lookup. This patch refactored the code to allow the number of buckets to be changed.

In addition, currently hashtags__fn(key) return value is assigned to uint16_t. Change to uint32_t as in a later patch the number of hashtag bits can be increased to be more than 16.

---
 dwarf_loader.c | 48 +++++++++++++++++++++++++++++++++++++-----------
 1 file changed, 37 insertions(+), 11 deletions(-)
dwarf_loader.c +37 -11
index c106919..5a1e860 100644
@@ -50,7 +50,12 @@ struct strings *strings;
 #define DW_FORM_implicit_const 0x21
 #endif
 
-#define hashtags__fn(key) hash_64(key, HASHTAGS__BITS)
+static uint32_t hashtags__bits = 15;
+
+static uint32_t hashtags__fn(Dwarf_Off key)
+{
+	return hash_64(key, hashtags__bits);
+}
 
 bool no_bitfield_type_recode = true;
 
@@ -102,9 +107,6 @@ static void dwarf_tag__set_spec(struct dwarf_tag *dtag, dwarf_off_ref spec)
 	*(dwarf_off_ref *)(dtag + 1) = spec;
 }
 
-#define HASHTAGS__BITS 15
-#define HASHTAGS__SIZE (1UL << HASHTAGS__BITS)
-
 #define obstack_chunk_alloc malloc
 #define obstack_chunk_free free
 
@@ -118,22 +120,41 @@ static void *obstack_zalloc(struct obstack *obstack, size_t size)
 }
 
 struct dwarf_cu {
-	struct hlist_head hash_tags[HASHTAGS__SIZE];
-	struct hlist_head hash_types[HASHTAGS__SIZE];
+	struct hlist_head *hash_tags;
+	struct hlist_head *hash_types;
 	struct obstack obstack;
 	struct cu *cu;
 	struct dwarf_cu *type_unit;
 };
 
-static void dwarf_cu__init(struct dwarf_cu *dcu)
+static int dwarf_cu__init(struct dwarf_cu *dcu)
 {
+	uint64_t hashtags_size = 1UL << hashtags__bits;
+	dcu->hash_tags = malloc(sizeof(struct hlist_head) * hashtags_size);
+	if (!dcu->hash_tags)
+		return -ENOMEM;
+
+	dcu->hash_types = malloc(sizeof(struct hlist_head) * hashtags_size);
+	if (!dcu->hash_types) {
+		free(dcu->hash_tags);
+		return -ENOMEM;
+	}
+
 	unsigned int i;
-	for (i = 0; i < HASHTAGS__SIZE; ++i) {
+	for (i = 0; i < hashtags_size; ++i) {
 		INIT_HLIST_HEAD(&dcu->hash_tags[i]);
 		INIT_HLIST_HEAD(&dcu->hash_types[i]);
 	}
 	obstack_init(&dcu->obstack);
 	dcu->type_unit = NULL;
+	return 0;
+}
+
+static void dwarf_cu__delete(struct cu *cu)
+{
+	struct dwarf_cu *dcu = cu->priv;
+	free(dcu->hash_tags);
+	free(dcu->hash_types);
 }
 
 static void hashtags__hash(struct hlist_head *hashtable,
@@ -151,7 +172,7 @@ static struct dwarf_tag *hashtags__find(const struct hlist_head *hashtable,
 
 	struct dwarf_tag *tpos;
 	struct hlist_node *pos;
-	uint16_t bucket = hashtags__fn(id);
+	uint32_t bucket = hashtags__fn(id);
 	const struct hlist_head *head = hashtable + bucket;
 
 	hlist_for_each_entry(tpos, pos, head, hash_node) {
@@ -2429,7 +2450,9 @@ static int cus__load_debug_types(struct cus *cus, struct conf_load *conf,
 			}
 			cu->little_endian = ehdr.e_ident[EI_DATA] == ELFDATA2LSB;
 
-			dwarf_cu__init(dcup);
+			if (dwarf_cu__init(dcup) != 0)
+				return DWARF_CB_ABORT;
+
 			dcup->cu = cu;
 			/* Funny hack.  */
 			dcup->type_unit = dcup;
@@ -2521,7 +2544,9 @@ static int cus__load_module(struct cus *cus, struct conf_load *conf,
 
 		struct dwarf_cu dcu;
 
-		dwarf_cu__init(&dcu);
+		if (dwarf_cu__init(&dcu) != 0)
+			return DWARF_CB_ABORT;
+
 		dcu.cu = cu;
 		dcu.type_unit = type_cu ? &type_dcu : NULL;
 		cu->priv = &dcu;
@@ -2672,5 +2697,6 @@ struct debug_fmt_ops dwarf__ops = {
 	.tag__decl_file	     = dwarf_tag__decl_file,
 	.tag__decl_line	     = dwarf_tag__decl_line,
 	.tag__orig_id	     = dwarf_tag__orig_id,
+	.cu__delete	     = dwarf_cu__delete,
 	.has_alignment_info  = true,
 };
Signature · 1 line
2.30.2
#2 lore raw

PATCH v2 2/3 dwarf_loader: factor out common code to initialize a cu

In reply to Yonghong Song

Both cus__load_debug_types() and cus__load_module() created new cu's followed by initialization. The initialization codes are identical so let us refactor into a common function which can be used later as well when dealing with merging cu's.

---
 dwarf_loader.c | 45 ++++++++++++++++++++-------------------------
 1 file changed, 20 insertions(+), 25 deletions(-)
dwarf_loader.c +20 -25
index 5a1e860..aa6372a 100644
@@ -2411,6 +2411,23 @@ static int finalize_cu_immediately(struct cus *cus, struct cu *cu,
 	return lsk;
 }
 
+static int cu__set_common(struct cu *cu, struct conf_load *conf,
+			  Dwfl_Module *mod, Elf *elf)
+{
+	cu->uses_global_strings = true;
+	cu->elf = elf;
+	cu->dwfl = mod;
+	cu->extra_dbg_info = conf ? conf->extra_dbg_info : 0;
+	cu->has_addr_info = conf ? conf->get_addr_info : 0;
+
+	GElf_Ehdr ehdr;
+	if (gelf_getehdr(elf, &ehdr) == NULL)
+		return DWARF_CB_ABORT;
+
+	cu->little_endian = ehdr.e_ident[EI_DATA] == ELFDATA2LSB;
+	return 0;
+}
+
 static int cus__load_debug_types(struct cus *cus, struct conf_load *conf,
 				 Dwfl_Module *mod, Dwarf *dw, Elf *elf,
 				 const char *filename,
@@ -2434,22 +2451,11 @@ static int cus__load_debug_types(struct cus *cus, struct conf_load *conf,
 
 			cu = cu__new("", pointer_size, build_id,
 				     build_id_len, filename);
-			if (cu == NULL) {
+			if (cu == NULL ||
+			    cu__set_common(cu, conf, mod, elf) != 0) {
 				return DWARF_CB_ABORT;
 			}
 
-			cu->uses_global_strings = true;
-			cu->elf = elf;
-			cu->dwfl = mod;
-			cu->extra_dbg_info = conf ? conf->extra_dbg_info : 0;
-			cu->has_addr_info = conf ? conf->get_addr_info : 0;
-
-			GElf_Ehdr ehdr;
-			if (gelf_getehdr(elf, &ehdr) == NULL) {
-				return DWARF_CB_ABORT;
-			}
-			cu->little_endian = ehdr.e_ident[EI_DATA] == ELFDATA2LSB;
-
 			if (dwarf_cu__init(dcup) != 0)
 				return DWARF_CB_ABORT;
 
@@ -2528,19 +2534,8 @@ static int cus__load_module(struct cus *cus, struct conf_load *conf,
 		const char *name = attr_string(cu_die, DW_AT_name);
 		struct cu *cu = cu__new(name ?: "", pointer_size,
 					build_id, build_id_len, filename);
-		if (cu == NULL)
-			return DWARF_CB_ABORT;
-		cu->uses_global_strings = true;
-		cu->elf = elf;
-		cu->dwfl = mod;
-		cu->extra_dbg_info = conf ? conf->extra_dbg_info : 0;
-		cu->has_addr_info = conf ? conf->get_addr_info : 0;
-
-		GElf_Ehdr ehdr;
-		if (gelf_getehdr(elf, &ehdr) == NULL) {
+		if (cu == NULL || cu__set_common(cu, conf, mod, elf) != 0)
 			return DWARF_CB_ABORT;
-		}
-		cu->little_endian = ehdr.e_ident[EI_DATA] == ELFDATA2LSB;
 
 		struct dwarf_cu dcu;
 
Signature · 1 line
2.30.2
#3 lore raw

PATCH v2 3/3 dwarf_loader: permit merging all dwarf cu's for clang lto built binary

In reply to Yonghong Song

For vmlinux built with clang thin-lto or lto, there exist cross cu type references. For example, the below can happen:
compile unit 1:
tag 10: type A
compile unit 2:
...
refer to type A (tag 10 in compile unit 1)
I only checked a few but have seen type A may be a simple type like "unsigned char" or a complex type like an array of base types.

To resolve this issue, the tag DW_AT_producer of the first DW_TAG_compile_unit is checked. If the binary is built with clang lto, all debuginfo dwarf cu's will be merged into one pahole cu which will resolve the above cross-cu tag reference issue. To test whether a binary is built with clang lto or not, The "clang version" and "-flto" will be checked against DW_AT_producer string for the first 5 debuginfo cu's. The reason is that a few linux files disabled lto for various reasons.

Merging cu's will create a single cu with lots of types, tags and functions. For example with clang thin-lto built vmlinux, I saw 9M entries in types table, 5.2M in tags table. The below are pahole wallclock time for different hashbits:
command line: time pahole -J vmlinux

      # of hashbits            wallclock time in seconds
          15                       460
          16                       255
          17                       131
          18                       97
          19                       75
          20                       69
          21                       64
          22                       62
          23                       58
          24                       64

The problem is with hashtags__find(), esp. the loop

    uint32_t bucket = hashtags__fn(id);
    const struct hlist_head *head = hashtable + bucket;
    hlist_for_each_entry(tpos, pos, head, hash_node) {
            if (tpos->id == id)
                    return tpos;
    }

Say we have 9M types and (1 << 15) buckets, that means each bucket will have roughly 64 elements. So each lookup will traverse the loop 32 iterations on average.

If we have 1 << 21 buckets, then each buckets will have 4 elements, and the average number of loop iterations for hashtags__find() will be 2.

Note that the number of hashbits 24 makes performance worse than 23. The reason could be that 23 hashbits can cover 8M buckets (close to 9M for the number of entries in types table). Higher number of hash bits allocates more memory and becomes less cache efficient compared to 23 hashbits.

This patch picks # of hashbits 21 as the starting value and will try to allocate memory based on that, if memory allocation fails, we will go with less hashbits until we reach hashbits 15 which is the default for non merge-cu case.

---
 dwarf_loader.c | 120 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 120 insertions(+)
dwarf_loader.c +120 -0
index aa6372a..1e63ca9 100644
@@ -51,6 +51,7 @@ struct strings *strings;
 #endif
 
 static uint32_t hashtags__bits = 15;
+static uint32_t max_hashtags__bits = 21;
 
 static uint32_t hashtags__fn(Dwarf_Off key)
 {
@@ -2484,6 +2485,115 @@ static int cus__load_debug_types(struct cus *cus, struct conf_load *conf,
 	return 0;
 }
 
+static bool cus__merging_cu(Dwarf *dw)
+{
+	uint8_t pointer_size, offset_size;
+	Dwarf_Off off = 0, noff;
+	size_t cuhl;
+	int cnt = 0;
+
+	/*
+	 * Just checking the first cu is not enough.
+	 * In linux, some C files may have LTO is disabled, e.g.,
+	 *   e242db40be27  x86, vdso: disable LTO only for vDSO
+	 *   d2dcd3e37475  x86, cpu: disable LTO for cpu.c
+	 * Fortunately, disabling LTO for a particular file in a LTO build
+	 * is rather an exception. Iterating 5 cu's to check whether
+	 * LTO is used or not should be enough.
+	 */
+	while (dwarf_nextcu(dw, off, &noff, &cuhl, NULL, &pointer_size,
+			    &offset_size) == 0) {
+		Dwarf_Die die_mem;
+		Dwarf_Die *cu_die = dwarf_offdie(dw, off + cuhl, &die_mem);
+
+		if (cu_die == NULL)
+			break;
+
+		if (++cnt > 5)
+			break;
+
+		const char *producer = attr_string(cu_die, DW_AT_producer);
+		if (strstr(producer, "clang version") != NULL &&
+		    strstr(producer, "-flto") != NULL)
+			return true;
+
+		off = noff;
+	}
+
+	return false;
+}
+
+static int cus__merge_and_process_cu(struct cus *cus, struct conf_load *conf,
+				     Dwfl_Module *mod, Dwarf *dw, Elf *elf,
+				     const char *filename,
+				     const unsigned char *build_id,
+				     int build_id_len,
+				     struct dwarf_cu *type_dcu)
+{
+	uint8_t pointer_size, offset_size;
+	struct dwarf_cu *dcu = NULL;
+	Dwarf_Off off = 0, noff;
+	struct cu *cu = NULL;
+	size_t cuhl;
+
+	while (dwarf_nextcu(dw, off, &noff, &cuhl, NULL, &pointer_size,
+			    &offset_size) == 0) {
+		Dwarf_Die die_mem;
+		Dwarf_Die *cu_die = dwarf_offdie(dw, off + cuhl, &die_mem);
+
+		if (cu_die == NULL)
+			break;
+
+		if (cu == NULL) {
+			cu = cu__new("", pointer_size, build_id, build_id_len,
+				     filename);
+			if (cu == NULL || cu__set_common(cu, conf, mod, elf) != 0)
+				return 1;
+
+			dcu = malloc(sizeof(struct dwarf_cu));
+			if (dcu == NULL)
+				return 1;
+
+			/* Merged cu tends to need a lot more memory.
+			 * Let us start with max_hashtags__bits and
+			 * go down to find a proper hashtag bit value.
+			 */
+			uint32_t default_hbits = hashtags__bits;
+			for (hashtags__bits = max_hashtags__bits;
+			     hashtags__bits >= default_hbits;
+			     hashtags__bits--) {
+				if (dwarf_cu__init(dcu) == 0)
+					break;
+			}
+			if (hashtags__bits < default_hbits)
+				return 1;
+
+			dcu->cu = cu;
+			dcu->type_unit = type_dcu;
+			cu->priv = dcu;
+			cu->dfops = &dwarf__ops;
+			cu->language = attr_numeric(cu_die, DW_AT_language);
+		}
+
+		Dwarf_Die child;
+		if (dwarf_child(cu_die, &child) == 0) {
+			if (die__process_unit(&child, cu) != 0)
+				return 1;
+		}
+
+		off = noff;
+	}
+
+	/* process merged cu */
+	if (cu__recode_dwarf_types(cu) != LSK__KEEPIT)
+		return 1;
+	if (finalize_cu_immediately(cus, cu, dcu, conf)
+	    == LSK__STOP_LOADING)
+		return 1;
+
+	return 0;
+}
+
 static int cus__load_module(struct cus *cus, struct conf_load *conf,
 			    Dwfl_Module *mod, Dwarf *dw, Elf *elf,
 			    const char *filename)
@@ -2518,6 +2628,15 @@ static int cus__load_module(struct cus *cus, struct conf_load *conf,
 		}
 	}
 
+	if (cus__merging_cu(dw)) {
+		res = cus__merge_and_process_cu(cus, conf, mod, dw, elf, filename,
+						build_id, build_id_len,
+						type_cu ? &type_dcu : NULL);
+		if (res)
+			return res;
+		goto out;
+	}
+
 	while (dwarf_nextcu(dw, off, &noff, &cuhl, NULL, &pointer_size,
 			    &offset_size) == 0) {
 		Dwarf_Die die_mem;
@@ -2557,6 +2676,7 @@ static int cus__load_module(struct cus *cus, struct conf_load *conf,
 		off = noff;
 	}
 
+out:
 	if (type_lsk == LSK__DELETE)
 		cu__delete(type_cu);
 
Signature · 1 line
2.30.2
#4 lore raw

Re: [PATCH dwarves v2 3/3] dwarf_loader: permit merging all dwarf cu's for clang lto built binary

Quoted 140 lines · Yonghong Song

For vmlinux built with clang thin-lto or lto, there exist cross cu type references. For example, the below can happen:
compile unit 1:
tag 10: type A
compile unit 2:
...
refer to type A (tag 10 in compile unit 1)
I only checked a few but have seen type A may be a simple type like "unsigned char" or a complex type like an array of base types.

To resolve this issue, the tag DW_AT_producer of the first DW_TAG_compile_unit is checked. If the binary is built with clang lto, all debuginfo dwarf cu's will be merged into one pahole cu which will resolve the above cross-cu tag reference issue. To test whether a binary is built with clang lto or not, The "clang version" and "-flto" will be checked against DW_AT_producer string for the first 5 debuginfo cu's. The reason is that a few linux files disabled lto for various reasons.

Merging cu's will create a single cu with lots of types, tags and functions. For example with clang thin-lto built vmlinux, I saw 9M entries in types table, 5.2M in tags table. The below are pahole wallclock time for different hashbits:
command line: time pahole -J vmlinux

       # of hashbits            wallclock time in seconds
           15                       460
           16                       255
           17                       131
           18                       97
           19                       75
           20                       69
           21                       64
           22                       62
           23                       58
           24                       64

The problem is with hashtags__find(), esp. the loop

     uint32_t bucket = hashtags__fn(id);
     const struct hlist_head *head = hashtable + bucket;
     hlist_for_each_entry(tpos, pos, head, hash_node) {
             if (tpos->id == id)
                     return tpos;
     }

Say we have 9M types and (1 << 15) buckets, that means each bucket will have roughly 64 elements. So each lookup will traverse the loop 32 iterations on average.

If we have 1 << 21 buckets, then each buckets will have 4 elements, and the average number of loop iterations for hashtags__find() will be 2.

Note that the number of hashbits 24 makes performance worse than 23. The reason could be that 23 hashbits can cover 8M buckets (close to 9M for the number of entries in types table). Higher number of hash bits allocates more memory and becomes less cache efficient compared to 23 hashbits.

This patch picks # of hashbits 21 as the starting value and will try to allocate memory based on that, if memory allocation fails, we will go with less hashbits until we reach hashbits 15 which is the default for non merge-cu case.

---
  dwarf_loader.c | 120 +++++++++++++++++++++++++++++++++++++++++++++++++
  1 file changed, 120 insertions(+)
dwarf_loader.c +55 -0
index aa6372a..1e63ca9 100644
@@ -51,6 +51,7 @@ struct strings *strings;
  #endif
 
  static uint32_t hashtags__bits = 15;
+static uint32_t max_hashtags__bits = 21;
 
  static uint32_t hashtags__fn(Dwarf_Off key)
  {
@@ -2484,6 +2485,115 @@ static int cus__load_debug_types(struct cus *cus, struct conf_load *conf,
  	return 0;
  }
 
+static bool cus__merging_cu(Dwarf *dw)
+{
+	uint8_t pointer_size, offset_size;
+	Dwarf_Off off = 0, noff;
+	size_t cuhl;
+	int cnt = 0;
+
+	/*
+	 * Just checking the first cu is not enough.
+	 * In linux, some C files may have LTO is disabled, e.g.,
+	 *   e242db40be27  x86, vdso: disable LTO only for vDSO
+	 *   d2dcd3e37475  x86, cpu: disable LTO for cpu.c
+	 * Fortunately, disabling LTO for a particular file in a LTO build
+	 * is rather an exception. Iterating 5 cu's to check whether
+	 * LTO is used or not should be enough.
+	 */
+	while (dwarf_nextcu(dw, off, &noff, &cuhl, NULL, &pointer_size,
+			    &offset_size) == 0) {
+		Dwarf_Die die_mem;
+		Dwarf_Die *cu_die = dwarf_offdie(dw, off + cuhl, &die_mem);
+
+		if (cu_die == NULL)
+			break;
+
+		if (++cnt > 5)
+			break;
+
+		const char *producer = attr_string(cu_die, DW_AT_producer);
+		if (strstr(producer, "clang version") != NULL &&
+		    strstr(producer, "-flto") != NULL)
+			return true;
+
+		off = noff;
+	}
+
+	return false;
+}
+
+static int cus__merge_and_process_cu(struct cus *cus, struct conf_load *conf,
+				     Dwfl_Module *mod, Dwarf *dw, Elf *elf,
+				     const char *filename,
+				     const unsigned char *build_id,
+				     int build_id_len,
+				     struct dwarf_cu *type_dcu)
+{
+	uint8_t pointer_size, offset_size;
+	struct dwarf_cu *dcu = NULL;
+	Dwarf_Off off = 0, noff;
+	struct cu *cu = NULL;
+	size_t cuhl;
+
+	while (dwarf_nextcu(dw, off, &noff, &cuhl, NULL, &pointer_size,
+			    &offset_size) == 0) {
+		Dwarf_Die die_mem;

dwarf_loader.c static int cus__load_debug_types(struct cus *cus, struct conf_load *conf, · line 2551

Quoted 10 lines
dwarf_loader.c +10 -0
+		Dwarf_Die *cu_die = dwarf_offdie(dw, off + cuhl, &die_mem);
+
+		if (cu_die == NULL)
+			break;
+
+		if (cu == NULL) {
+			cu = cu__new("", pointer_size, build_id, build_id_len,
+				     filename);
+			if (cu == NULL || cu__set_common(cu, conf, mod, elf) != 0)
+				return 1;

This is an oversight. I forgot to change back to DWARF_CB_ABORT after some experiments. I just sent out v3 to correct this and added the kernel patch link which is needed to make this auto lto discovery work.

Quoted 5 lines
dwarf_loader.c +5 -0
+
+			dcu = malloc(sizeof(struct dwarf_cu));
+			if (dcu == NULL)
+				return 1;
+

[...]