[PATCH v2 00/16] scripts/kallsyms: various cleanups and optimizations

[PATCH v2 00/16] scripts/kallsyms: various cleanups and optimizations

[Masahiro Yamada]

Changes in v2:
 - Add a new patch that shrinks the table before the sort.

Masahiro Yamada (16):
  scripts/kallsyms: remove unneeded #ifndef ARRAY_SIZE
  scripts/kallsyms: fix definitely-lost memory leak
  scripts/kallsyms: shrink table before sorting it
  scripts/kallsyms: set relative_base more effectively
  scripts/kallsyms: remove redundant is_arm_mapping_symbol()
  scripts/kallsyms: remove unneeded length check for prefix matching
  scripts/kallsyms: add sym_name() to mitigate cast ugliness
  scripts/kallsyms: replace prefix_underscores_count() with strspn()
  scripts/kallsyms: make find_token() return (unsigned char *)
  scripts/kallsyms: add const qualifiers where possible
  scripts/kallsyms: skip ignored symbols very early
  scripts/kallsyms: move more patterns to the ignored_prefixes array
  scripts/kallsyms: move ignored symbol types to is_ignored_symbol()
  scripts/kallsyms: make check_symbol_range() void function
  scripts/kallsyms: put check_symbol_range() calls close together
  scripts/kallsyms: remove redundant initializers

 scripts/kallsyms.c | 287 ++++++++++++++++++++++-----------------------
 1 file changed, 142 insertions(+), 145 deletions(-)

-- 
2.17.1

[PATCH v2 01/16] scripts/kallsyms: remove unneeded #ifndef ARRAY_SIZE

[Masahiro Yamada]

This is not defined in the standard headers. #ifndef is unneeded.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index ae6504d07fd6..918c2ba071b5 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -24,9 +24,7 @@
 #include <ctype.h>
 #include <limits.h>
 
-#ifndef ARRAY_SIZE
 #define ARRAY_SIZE(arr) (sizeof(arr) / sizeof(arr[0]))
-#endif
 
 #define KSYM_NAME_LEN		128
 
-- 
2.17.1

[PATCH v2 02/16] scripts/kallsyms: fix definitely-lost memory leak

[Masahiro Yamada]

build_initial_tok_table() overwrites unused sym_entry to shrink the
table size. Before the entry is overwritten, table[i].sym must be freed
since it is malloc'ed data.

This fixes the 'definitely lost' report from valgrind. I ran valgrind
against x86_64_defconfig of v5.4-rc8 kernel, and here is the summary:

[Before the fix]

  LEAK SUMMARY:
     definitely lost: 53,184 bytes in 2,874 blocks

[After the fix]

  LEAK SUMMARY:
     definitely lost: 0 bytes in 0 blocks

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 918c2ba071b5..79641874d860 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -487,6 +487,8 @@ static void build_initial_tok_table(void)
 				table[pos] = table[i];
 			learn_symbol(table[pos].sym, table[pos].len);
 			pos++;
+		} else {
+			free(table[i].sym);
 		}
 	}
 	table_cnt = pos;
-- 
2.17.1

[PATCH v2 03/16] scripts/kallsyms: shrink table before sorting it

[Masahiro Yamada]

Currently, build_initial_tok_table() trims unused symbols, which is
called after sort_symbol().

It is not efficient to sort the huge table that contains unused entries.
Shrink the table before sorting it.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2:
 - New patch

 scripts/kallsyms.c | 49 +++++++++++++++++++++++++++-------------------
 1 file changed, 29 insertions(+), 20 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 79641874d860..de986eda41a6 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -268,6 +268,30 @@ static int symbol_valid(struct sym_entry *s)
 	return 1;
 }
 
+/* remove all the invalid symbols from the table */
+static void shrink_table(void)
+{
+	unsigned int i, pos;
+
+	pos = 0;
+	for (i = 0; i < table_cnt; i++) {
+		if (symbol_valid(&table[i])) {
+			if (pos != i)
+				table[pos] = table[i];
+			pos++;
+		} else {
+			free(table[i].sym);
+		}
+	}
+	table_cnt = pos;
+
+	/* When valid symbol is not registered, exit to error */
+	if (!table_cnt) {
+		fprintf(stderr, "No valid symbol.\n");
+		exit(1);
+	}
+}
+
 static void read_map(FILE *in)
 {
 	while (!feof(in)) {
@@ -475,23 +499,13 @@ static void forget_symbol(unsigned char *symbol, int len)
 		token_profit[ symbol[i] + (symbol[i + 1] << 8) ]--;
 }
 
-/* remove all the invalid symbols from the table and do the initial token count */
+/* do the initial token count */
 static void build_initial_tok_table(void)
 {
-	unsigned int i, pos;
+	unsigned int i;
 
-	pos = 0;
-	for (i = 0; i < table_cnt; i++) {
-		if ( symbol_valid(&table[i]) ) {
-			if (pos != i)
-				table[pos] = table[i];
-			learn_symbol(table[pos].sym, table[pos].len);
-			pos++;
-		} else {
-			free(table[i].sym);
-		}
-	}
-	table_cnt = pos;
+	for (i = 0; i < table_cnt; i++)
+		learn_symbol(table[i].sym, table[i].len);
 }
 
 static void *find_token(unsigned char *str, int len, unsigned char *token)
@@ -614,12 +628,6 @@ static void optimize_token_table(void)
 
 	insert_real_symbols_in_table();
 
-	/* When valid symbol is not registered, exit to error */
-	if (!table_cnt) {
-		fprintf(stderr, "No valid symbol.\n");
-		exit(1);
-	}
-
 	optimize_result();
 }
 
@@ -756,6 +764,7 @@ int main(int argc, char **argv)
 		usage();
 
 	read_map(stdin);
+	shrink_table();
 	if (absolute_percpu)
 		make_percpus_absolute();
 	if (base_relative)
-- 
2.17.1

Re: [PATCH v2 03/16] scripts/kallsyms: shrink table before sorting it

[Olof Johansson]

Hi,

On Sat, Nov 23, 2019 at 8:05 AM Masahiro Yamada
<yamada.masahiro@socionext.com> wrote:
>
> Currently, build_initial_tok_table() trims unused symbols, which is
> called after sort_symbol().
>
> It is not efficient to sort the huge table that contains unused entries.
> Shrink the table before sorting it.
>
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

This started showing warnings on some 32-bit ARM platforms, due to
kallsyms_relative_base changing:

 kallsyms_relative_base:
-       PTR     _text - 0
+       PTR     _text - 0xfffffffffffffe20

The assembler started complaining:

.tmp_kallsyms1.S: Assembler messages:
.tmp_kallsyms1.S:15315: Warning: right operand is a bignum; integer 0 assumed

Also, I clearly see different output with this patch reverted and
applied; I would expect no actual difference if it was correct.

Can we please revert this for 5.5 while this is being sorted out?

To reproduce, build for example am200epdkit_defconfig for ARCH=arm. I
see it with GCC 8.2.0, binutils 2.30.0.


Thanks,

-Olof

[PATCH v2 04/16] scripts/kallsyms: set relative_base more effectively

[Masahiro Yamada]

Currently, record_relative_base() iterates over the entire table to
find the minimum address, but it is not efficient because we sort
the table anyway.

After sort_symbol(), the table is sorted by address. (kallsyms parses
the 'nm -n' output, so the data is already sorted by address, but this
commit does not rely on it.)

Move record_relative_base() after sort_symbols(), and take the first
non-absolute symbol value.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index de986eda41a6..c9efb67c6ecb 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -739,11 +739,15 @@ static void record_relative_base(void)
 {
 	unsigned int i;
 
-	relative_base = -1ULL;
 	for (i = 0; i < table_cnt; i++)
-		if (!symbol_absolute(&table[i]) &&
-		    table[i].addr < relative_base)
+		if (!symbol_absolute(&table[i])) {
+			/*
+			 * The table is sorted by address.
+			 * Take the first non-absolute symbol value.
+			 */
 			relative_base = table[i].addr;
+			return;
+		}
 }
 
 int main(int argc, char **argv)
@@ -767,9 +771,9 @@ int main(int argc, char **argv)
 	shrink_table();
 	if (absolute_percpu)
 		make_percpus_absolute();
+	sort_symbols();
 	if (base_relative)
 		record_relative_base();
-	sort_symbols();
 	optimize_token_table();
 	write_src();
 
-- 
2.17.1

[PATCH v2 05/16] scripts/kallsyms: remove redundant is_arm_mapping_symbol()

[Masahiro Yamada]

Since commit 6f00df24ee39 ("[PATCH] Strip local symbols from kallsyms"),
all symbols starting '$' are ignored.

is_arm_mapping_symbol() particularly ignores $a, $t, etc. but it is
redundant.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 19 ++++++-------------
 1 file changed, 6 insertions(+), 13 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index c9efb67c6ecb..14a50c8d3f34 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -74,16 +74,6 @@ static void usage(void)
 	exit(1);
 }
 
-/*
- * This ignores the intensely annoying "mapping symbols" found
- * in ARM ELF files: $a, $t and $d.
- */
-static int is_arm_mapping_symbol(const char *str)
-{
-	return str[0] == '$' && strchr("axtd", str[1])
-	       && (str[2] == '\0' || str[2] == '.');
-}
-
 static int check_symbol_range(const char *sym, unsigned long long addr,
 			      struct addr_range *ranges, int entries)
 {
@@ -139,10 +129,13 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 			return -1;
 
 	}
-	else if (toupper(stype) == 'U' ||
-		 is_arm_mapping_symbol(sym))
+	else if (toupper(stype) == 'U')
 		return -1;
-	/* exclude also MIPS ELF local symbols ($L123 instead of .L123) */
+	/*
+	 * Ignore generated symbols such as:
+	 *  - mapping symbols in ARM ELF files ($a, $t, and $d)
+	 *  - MIPS ELF local symbols ($L123 instead of .L123)
+	 */
 	else if (sym[0] == '$')
 		return -1;
 	/* exclude debugging symbols */
-- 
2.17.1

[PATCH v2 06/16] scripts/kallsyms: remove unneeded length check for prefix matching

[Masahiro Yamada]

l <= strlen(sym_name) is unnecessary for prefix matching.
strncmp() will do.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 14a50c8d3f34..a57636c6f84f 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -246,8 +246,7 @@ static int symbol_valid(struct sym_entry *s)
 	for (i = 0; special_prefixes[i]; i++) {
 		int l = strlen(special_prefixes[i]);
 
-		if (l <= strlen(sym_name) &&
-		    strncmp(sym_name, special_prefixes[i], l) == 0)
+		if (strncmp(sym_name, special_prefixes[i], l) == 0)
 			return 0;
 	}
 
-- 
2.17.1

[PATCH v2 07/16] scripts/kallsyms: add sym_name() to mitigate cast ugliness

[Masahiro Yamada]

sym_entry::sym is (unsigned char *) instead of (char *) because
kallsyms exploits the MSB for compression, and the characters are
used as the index of token_profit array.

However, it requires casting (unsigned char *) to (char *) in some
places since standard library functions such as strcmp(), strlen()
expect (char *).

Introduce a new helper, sym_name(), which advances the given pointer
by 1 and casts it to (char *).

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 29 ++++++++++++++++-------------
 1 file changed, 16 insertions(+), 13 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index a57636c6f84f..baa2fa5692b0 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -74,6 +74,11 @@ static void usage(void)
 	exit(1);
 }
 
+static char *sym_name(const struct sym_entry *s)
+{
+	return (char *)s->sym + 1;
+}
+
 static int check_symbol_range(const char *sym, unsigned long long addr,
 			      struct addr_range *ranges, int entries)
 {
@@ -154,7 +159,7 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 			"unable to allocate required amount of memory\n");
 		exit(EXIT_FAILURE);
 	}
-	strcpy((char *)s->sym + 1, sym);
+	strcpy(sym_name(s), sym);
 	s->sym[0] = stype;
 
 	s->percpu_absolute = 0;
@@ -215,7 +220,7 @@ static int symbol_valid(struct sym_entry *s)
 		NULL };
 
 	int i;
-	char *sym_name = (char *)s->sym + 1;
+	const char *name = sym_name(s);
 
 	/* if --all-symbols is not specified, then symbols outside the text
 	 * and inittext sections are discarded */
@@ -230,30 +235,28 @@ static int symbol_valid(struct sym_entry *s)
 		 * rules.
 		 */
 		if ((s->addr == text_range_text->end &&
-				strcmp(sym_name,
-				       text_range_text->end_sym)) ||
+		     strcmp(name, text_range_text->end_sym)) ||
 		    (s->addr == text_range_inittext->end &&
-				strcmp(sym_name,
-				       text_range_inittext->end_sym)))
+		     strcmp(name, text_range_inittext->end_sym)))
 			return 0;
 	}
 
 	/* Exclude symbols which vary between passes. */
 	for (i = 0; special_symbols[i]; i++)
-		if (strcmp(sym_name, special_symbols[i]) == 0)
+		if (strcmp(name, special_symbols[i]) == 0)
 			return 0;
 
 	for (i = 0; special_prefixes[i]; i++) {
 		int l = strlen(special_prefixes[i]);
 
-		if (strncmp(sym_name, special_prefixes[i], l) == 0)
+		if (strncmp(name, special_prefixes[i], l) == 0)
 			return 0;
 	}
 
 	for (i = 0; special_suffixes[i]; i++) {
-		int l = strlen(sym_name) - strlen(special_suffixes[i]);
+		int l = strlen(name) - strlen(special_suffixes[i]);
 
-		if (l >= 0 && strcmp(sym_name + l, special_suffixes[i]) == 0)
+		if (l >= 0 && strcmp(name + l, special_suffixes[i]) == 0)
 			return 0;
 	}
 
@@ -626,7 +629,7 @@ static void optimize_token_table(void)
 /* guess for "linker script provide" symbol */
 static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 {
-	const char *symbol = (char *)se->sym + 1;
+	const char *symbol = sym_name(se);
 	int len = se->len - 1;
 
 	if (len < 8)
@@ -696,8 +699,8 @@ static int compare_symbols(const void *a, const void *b)
 		return wa - wb;
 
 	/* sort by the number of prefix underscores */
-	wa = prefix_underscores_count((const char *)sa->sym + 1);
-	wb = prefix_underscores_count((const char *)sb->sym + 1);
+	wa = prefix_underscores_count(sym_name(sa));
+	wb = prefix_underscores_count(sym_name(sb));
 	if (wa != wb)
 		return wa - wb;
 
-- 
2.17.1

[PATCH v2 08/16] scripts/kallsyms: replace prefix_underscores_count() with strspn()

[Masahiro Yamada]

You can do equivalent things with strspn(). I do not see noticeable
performance difference.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 14 ++------------
 1 file changed, 2 insertions(+), 12 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index baa2fa5692b0..89cc7c098c51 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -661,16 +661,6 @@ static int may_be_linker_script_provide_symbol(const struct sym_entry *se)
 	return 0;
 }
 
-static int prefix_underscores_count(const char *str)
-{
-	const char *tail = str;
-
-	while (*tail == '_')
-		tail++;
-
-	return tail - str;
-}
-
 static int compare_symbols(const void *a, const void *b)
 {
 	const struct sym_entry *sa;
@@ -699,8 +689,8 @@ static int compare_symbols(const void *a, const void *b)
 		return wa - wb;
 
 	/* sort by the number of prefix underscores */
-	wa = prefix_underscores_count(sym_name(sa));
-	wb = prefix_underscores_count(sym_name(sb));
+	wa = strspn(sym_name(sa), "_");
+	wb = strspn(sym_name(sb), "_");
 	if (wa != wb)
 		return wa - wb;
 
-- 
2.17.1

[PATCH v2 09/16] scripts/kallsyms: make find_token() return (unsigned char *)

[Masahiro Yamada]

The callers of this function expect (unsigned char *). I do not see
a good reason to make this function return (void *).

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 89cc7c098c51..274a77bfbd63 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -503,7 +503,8 @@ static void build_initial_tok_table(void)
 		learn_symbol(table[i].sym, table[i].len);
 }
 
-static void *find_token(unsigned char *str, int len, unsigned char *token)
+static unsigned char *find_token(unsigned char *str, int len,
+				 unsigned char *token)
 {
 	int i;
 
-- 
2.17.1

[PATCH v2 10/16] scripts/kallsyms: add const qualifiers where possible

[Masahiro Yamada]

Add 'const' where a function does not write to the pointer dereferenes.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 28 ++++++++++++++--------------
 1 file changed, 14 insertions(+), 14 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 274a77bfbd63..056bde436540 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -170,11 +170,11 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 	return 0;
 }
 
-static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
-			   int entries)
+static int symbol_in_range(const struct sym_entry *s,
+			   const struct addr_range *ranges, int entries)
 {
 	size_t i;
-	struct addr_range *ar;
+	const struct addr_range *ar;
 
 	for (i = 0; i < entries; ++i) {
 		ar = &ranges[i];
@@ -186,14 +186,14 @@ static int symbol_in_range(struct sym_entry *s, struct addr_range *ranges,
 	return 0;
 }
 
-static int symbol_valid(struct sym_entry *s)
+static int symbol_valid(const struct sym_entry *s)
 {
 	/* Symbols which vary between passes.  Passes 1 and 2 must have
 	 * identical symbol lists.  The kallsyms_* symbols below are only added
 	 * after pass 1, they would be included in pass 2 when --all-symbols is
 	 * specified so exclude them to get a stable symbol list.
 	 */
-	static char *special_symbols[] = {
+	static const char * const special_symbols[] = {
 		"kallsyms_addresses",
 		"kallsyms_offsets",
 		"kallsyms_relative_base",
@@ -208,12 +208,12 @@ static int symbol_valid(struct sym_entry *s)
 		"_SDA2_BASE_",		/* ppc */
 		NULL };
 
-	static char *special_prefixes[] = {
+	static const char * const special_prefixes[] = {
 		"__crc_",		/* modversions */
 		"__efistub_",		/* arm64 EFI stub namespace */
 		NULL };
 
-	static char *special_suffixes[] = {
+	static const char * const special_suffixes[] = {
 		"_veneer",		/* arm */
 		"_from_arm",		/* arm */
 		"_from_thumb",		/* arm */
@@ -305,7 +305,7 @@ static void read_map(FILE *in)
 	}
 }
 
-static void output_label(char *label)
+static void output_label(const char *label)
 {
 	printf(".globl %s\n", label);
 	printf("\tALGN\n");
@@ -314,7 +314,7 @@ static void output_label(char *label)
 
 /* uncompress a compressed symbol. When this function is called, the best table
  * might still be compressed itself, so the function needs to be recursive */
-static int expand_symbol(unsigned char *data, int len, char *result)
+static int expand_symbol(const unsigned char *data, int len, char *result)
 {
 	int c, rlen, total=0;
 
@@ -339,7 +339,7 @@ static int expand_symbol(unsigned char *data, int len, char *result)
 	return total;
 }
 
-static int symbol_absolute(struct sym_entry *s)
+static int symbol_absolute(const struct sym_entry *s)
 {
 	return s->percpu_absolute;
 }
@@ -477,7 +477,7 @@ static void write_src(void)
 /* table lookup compression functions */
 
 /* count all the possible tokens in a symbol */
-static void learn_symbol(unsigned char *symbol, int len)
+static void learn_symbol(const unsigned char *symbol, int len)
 {
 	int i;
 
@@ -486,7 +486,7 @@ static void learn_symbol(unsigned char *symbol, int len)
 }
 
 /* decrease the count for all the possible tokens in a symbol */
-static void forget_symbol(unsigned char *symbol, int len)
+static void forget_symbol(const unsigned char *symbol, int len)
 {
 	int i;
 
@@ -504,7 +504,7 @@ static void build_initial_tok_table(void)
 }
 
 static unsigned char *find_token(unsigned char *str, int len,
-				 unsigned char *token)
+				 const unsigned char *token)
 {
 	int i;
 
@@ -517,7 +517,7 @@ static unsigned char *find_token(unsigned char *str, int len,
 
 /* replace a given token in all the valid symbols. Use the sampled symbols
  * to update the counts */
-static void compress_symbols(unsigned char *str, int idx)
+static void compress_symbols(const unsigned char *str, int idx)
 {
 	unsigned int i, len, size;
 	unsigned char *p1, *p2;
-- 
2.17.1

[PATCH v2 11/16] scripts/kallsyms: skip ignored symbols very early

[Masahiro Yamada]

Unless the address range matters, symbols can be ignored earlier,
which avoids unneeded memory allocation.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 113 +++++++++++++++++++++++++--------------------
 1 file changed, 62 insertions(+), 51 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 056bde436540..843615c1d384 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -18,6 +18,7 @@
  *
  */
 
+#include <stdbool.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -79,6 +80,64 @@ static char *sym_name(const struct sym_entry *s)
 	return (char *)s->sym + 1;
 }
 
+static bool is_ignored_symbol(const char *name, char type)
+{
+	static const char * const ignored_symbols[] = {
+		/*
+		 * Symbols which vary between passes. Passes 1 and 2 must have
+		 * identical symbol lists. The kallsyms_* symbols below are
+		 * only added after pass 1, they would be included in pass 2
+		 * when --all-symbols is specified so exclude them to get a
+		 * stable symbol list.
+		 */
+		"kallsyms_addresses",
+		"kallsyms_offsets",
+		"kallsyms_relative_base",
+		"kallsyms_num_syms",
+		"kallsyms_names",
+		"kallsyms_markers",
+		"kallsyms_token_table",
+		"kallsyms_token_index",
+		/* Exclude linker generated symbols which vary between passes */
+		"_SDA_BASE_",		/* ppc */
+		"_SDA2_BASE_",		/* ppc */
+		NULL
+	};
+
+	static const char * const ignored_prefixes[] = {
+		"__crc_",		/* modversions */
+		"__efistub_",		/* arm64 EFI stub namespace */
+		NULL
+	};
+
+	static const char * const ignored_suffixes[] = {
+		"_from_arm",		/* arm */
+		"_from_thumb",		/* arm */
+		"_veneer",		/* arm */
+		NULL
+	};
+
+	const char * const *p;
+
+	/* Exclude symbols which vary between passes. */
+	for (p = ignored_symbols; *p; p++)
+		if (!strcmp(name, *p))
+			return true;
+
+	for (p = ignored_prefixes; *p; p++)
+		if (!strncmp(name, *p, strlen(*p)))
+			return true;
+
+	for (p = ignored_suffixes; *p; p++) {
+		int l = strlen(name) - strlen(*p);
+
+		if (l >= 0 && !strcmp(name + l, *p))
+			return true;
+	}
+
+	return false;
+}
+
 static int check_symbol_range(const char *sym, unsigned long long addr,
 			      struct addr_range *ranges, int entries)
 {
@@ -118,6 +177,9 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 		return -1;
 	}
 
+	if (is_ignored_symbol(sym, stype))
+		return -1;
+
 	/* Ignore most absolute/undefined (?) symbols. */
 	if (strcmp(sym, "_text") == 0)
 		_text = s->addr;
@@ -188,38 +250,6 @@ static int symbol_in_range(const struct sym_entry *s,
 
 static int symbol_valid(const struct sym_entry *s)
 {
-	/* Symbols which vary between passes.  Passes 1 and 2 must have
-	 * identical symbol lists.  The kallsyms_* symbols below are only added
-	 * after pass 1, they would be included in pass 2 when --all-symbols is
-	 * specified so exclude them to get a stable symbol list.
-	 */
-	static const char * const special_symbols[] = {
-		"kallsyms_addresses",
-		"kallsyms_offsets",
-		"kallsyms_relative_base",
-		"kallsyms_num_syms",
-		"kallsyms_names",
-		"kallsyms_markers",
-		"kallsyms_token_table",
-		"kallsyms_token_index",
-
-	/* Exclude linker generated symbols which vary between passes */
-		"_SDA_BASE_",		/* ppc */
-		"_SDA2_BASE_",		/* ppc */
-		NULL };
-
-	static const char * const special_prefixes[] = {
-		"__crc_",		/* modversions */
-		"__efistub_",		/* arm64 EFI stub namespace */
-		NULL };
-
-	static const char * const special_suffixes[] = {
-		"_veneer",		/* arm */
-		"_from_arm",		/* arm */
-		"_from_thumb",		/* arm */
-		NULL };
-
-	int i;
 	const char *name = sym_name(s);
 
 	/* if --all-symbols is not specified, then symbols outside the text
@@ -241,25 +271,6 @@ static int symbol_valid(const struct sym_entry *s)
 			return 0;
 	}
 
-	/* Exclude symbols which vary between passes. */
-	for (i = 0; special_symbols[i]; i++)
-		if (strcmp(name, special_symbols[i]) == 0)
-			return 0;
-
-	for (i = 0; special_prefixes[i]; i++) {
-		int l = strlen(special_prefixes[i]);
-
-		if (strncmp(name, special_prefixes[i], l) == 0)
-			return 0;
-	}
-
-	for (i = 0; special_suffixes[i]; i++) {
-		int l = strlen(name) - strlen(special_suffixes[i]);
-
-		if (l >= 0 && strcmp(name + l, special_suffixes[i]) == 0)
-			return 0;
-	}
-
 	return 1;
 }
 
-- 
2.17.1

[PATCH v2 12/16] scripts/kallsyms: move more patterns to the ignored_prefixes array

[Masahiro Yamada]

Refactoring for shortening the code.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 12 ++----------
 1 file changed, 2 insertions(+), 10 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 843615c1d384..04a1dd16edcf 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -105,6 +105,8 @@ static bool is_ignored_symbol(const char *name, char type)
 	};
 
 	static const char * const ignored_prefixes[] = {
+		"$",			/* local symbols for ARM, MIPS, etc. */
+		".LASANPC",		/* s390 kasan local symbols */
 		"__crc_",		/* modversions */
 		"__efistub_",		/* arm64 EFI stub namespace */
 		NULL
@@ -198,19 +200,9 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 	}
 	else if (toupper(stype) == 'U')
 		return -1;
-	/*
-	 * Ignore generated symbols such as:
-	 *  - mapping symbols in ARM ELF files ($a, $t, and $d)
-	 *  - MIPS ELF local symbols ($L123 instead of .L123)
-	 */
-	else if (sym[0] == '$')
-		return -1;
 	/* exclude debugging symbols */
 	else if (stype == 'N' || stype == 'n')
 		return -1;
-	/* exclude s390 kasan local symbols */
-	else if (!strncmp(sym, ".LASANPC", 8))
-		return -1;
 
 	/* include the type field in the symbol name, so that it gets
 	 * compressed together */
-- 
2.17.1

[PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Masahiro Yamada]

Collect the ignored patterns to is_ignored_symbol().

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 30 +++++++++++++++---------------
 1 file changed, 15 insertions(+), 15 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 04a1dd16edcf..d90a6133d7b8 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -137,6 +137,21 @@ static bool is_ignored_symbol(const char *name, char type)
 			return true;
 	}
 
+	if (type == 'U' || type == 'u')
+		return true;
+	/* exclude debugging symbols */
+	if (type == 'N' || type == 'n')
+		return true;
+
+	if (toupper(type) == 'A') {
+		/* Keep these useful absolute symbols */
+		if (strcmp(name, "__kernel_syscall_via_break") &&
+		    strcmp(name, "__kernel_syscall_via_epc") &&
+		    strcmp(name, "__kernel_sigtramp") &&
+		    strcmp(name, "__gp"))
+			return true;
+	}
+
 	return false;
 }
 
@@ -188,21 +203,6 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 	else if (check_symbol_range(sym, s->addr, text_ranges,
 				    ARRAY_SIZE(text_ranges)) == 0)
 		/* nothing to do */;
-	else if (toupper(stype) == 'A')
-	{
-		/* Keep these useful absolute symbols */
-		if (strcmp(sym, "__kernel_syscall_via_break") &&
-		    strcmp(sym, "__kernel_syscall_via_epc") &&
-		    strcmp(sym, "__kernel_sigtramp") &&
-		    strcmp(sym, "__gp"))
-			return -1;
-
-	}
-	else if (toupper(stype) == 'U')
-		return -1;
-	/* exclude debugging symbols */
-	else if (stype == 'N' || stype == 'n')
-		return -1;
 
 	/* include the type field in the symbol name, so that it gets
 	 * compressed together */
-- 
2.17.1

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Finn Thain]

On Sun, 24 Nov 2019, Masahiro Yamada wrote:

> Collect the ignored patterns to is_ignored_symbol().
> 
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

This commit (887df76de67f5) caused a regression in my powerpc builds as it 
causes symbol names to disappear from backtraces:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 _einittext+0x3f9e5120/0x3feb71b8
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00055-g887df76de67f5 #18
NIP:  c00aef68 LR: c00af114 CTR: c001272c
REGS: c0705c40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00055-g887df76de67f5)
MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000

GPR00: 001f0100 c0705cf8 c06dc300 c070af1c c001258c 00000000 00000000 ef7fb5bc 
GPR08: 08800000 00000100 00000001 00000100 42000044 00000000 c0709040 00000004 
GPR16: 00000001 c06022b4 c058297c 00200002 ffff8cb9 00000000 c06d84a0 c0710000 
GPR24: c0710000 00000000 00000000 c070af1c c070af1c 00000000 c001258c 00000000 
NIP [c00aef68] _einittext+0x3f9e5120/0x3feb71b8
LR [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
Call Trace:
[c0705cf8] [ef006320] 0xef006320 (unreliable)
[c0705d38] [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
[c0705d48] [c00af158] _einittext+0x3f9e5310/0x3feb71b8
[c0705d68] [c0012768] _einittext+0x3f948920/0x3feb71b8
[c0705d78] [c0092c04] _einittext+0x3f9c8dbc/0x3feb71b8
[c0705d88] [c0092d18] _einittext+0x3f9c8ed0/0x3feb71b8
[c0705da8] [c0093a2c] _einittext+0x3f9c9be4/0x3feb71b8
[c0705de8] [c0580224] _einittext+0x3feb63dc/0x3feb71b8
[c0705e48] [c00382ec] _einittext+0x3f96e4a4/0x3feb71b8
[c0705e58] [c000d2a0] _einittext+0x3f943458/0x3feb71b8
[c0705e88] [c001353c] _einittext+0x3f9496f4/0x3feb71b8
--- interrupt: 901 at _einittext+0x3f941058/0x3feb71b8
    LR = _einittext+0x3f941058/0x3feb71b8
[c0705f50] [c06cc214] 0xc06cc214 (unreliable)
[c0705f60] [c057fa20] _einittext+0x3feb5bd8/0x3feb71b8
[c0705f70] [c005de48] _einittext+0x3f994000/0x3feb71b8
[c0705f90] [c005e050] _einittext+0x3f994208/0x3feb71b8
[c0705fa0] [c0004cc8] _einittext+0x3f93ae80/0x3feb71b8
[c0705fb0] [c069a36c] _einittext+0x3ffd0524/0x40000000
[c0705ff0] [00003500] 0x3500
Instruction dump:
7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78 
7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6 
---[ end trace a06fef4788747c72 ]---


Prior to that (e.g. 97261e1e2240f), I get backtraces like this:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 smp_call_function_many+0x318/0x320
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00054-g97261e1e2240f #20
NIP:  c00aef68 LR: c00af114 CTR: c001272c
REGS: c075dc40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00054-g97261e1e2240f)
MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000

GPR00: 001f0100 c075dcf8 c0733300 c0762f1c c001258c 00000000 00000000 ef7fb5bc 
GPR08: 04800000 00000100 00000001 00000100 42000044 00000000 c0761040 00000004 
GPR16: 00000001 c0658e58 c058297c 00200002 ffff8cb9 00000000 c072f4a0 c0760000 
GPR24: c0760000 00000000 00000000 c0762f1c c0762f1c 00000000 c001258c 00000000 
NIP [c00aef68] smp_call_function_many+0x318/0x320
LR [c00af114] smp_call_function+0x34/0x44
Call Trace:
[c075dcf8] [ef006320] 0xef006320 (unreliable)
[c075dd38] [c00af114] smp_call_function+0x34/0x44
[c075dd48] [c00af158] on_each_cpu+0x1c/0x4c
[c075dd68] [c0012768] tau_timeout_smp+0x3c/0x4c
[c075dd78] [c0092c04] call_timer_fn.isra.26+0x20/0x84
[c075dd88] [c0092d18] expire_timers+0xb0/0xc0
[c075dda8] [c0093a2c] run_timer_softirq+0xa4/0x1a4
[c075dde8] [c0580224] __do_softirq+0x11c/0x280
[c075de48] [c00382ec] irq_exit+0xc0/0xd4
[c075de58] [c000d2a0] timer_interrupt+0x154/0x260
[c075de88] [c001353c] ret_from_except+0x0/0x14
--- interrupt: 901 at arch_cpu_idle+0x24/0x78
    LR = arch_cpu_idle+0x24/0x78
[c075df50] [c0723214] 0xc0723214 (unreliable)
[c075df60] [c057fa20] default_idle_call+0x38/0x58
[c075df70] [c005de48] do_idle+0xd4/0x17c
[c075df90] [c005e054] cpu_startup_entry+0x24/0x28
[c075dfa0] [c0004cc8] rest_init+0xa8/0xbc
[c075dfb0] [c06f136c] start_kernel+0x40c/0x420
[c075dff0] [00003500] 0x3500
Instruction dump:
7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78 
7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6 
---[ end trace 784c7f15ecd23941 ]---

Has anyone else observed these problems (either the WARNING from 
smp_call_function_many() or the missing symbol names)?

What is the best way to fix this? Should I upgrade binutils?

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Finn Thain]

On Sun, 24 Nov 2019, Masahiro Yamada wrote:

> Collect the ignored patterns to is_ignored_symbol().
> 
> Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>

This commit (887df76de67f5) caused a regression in my powerpc builds as it 
causes symbol names to disappear from backtraces:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 _einittext+0x3f9e5120/0x3feb71b8
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00055-g887df76de67f5 #18
NIP:  c00aef68 LR: c00af114 CTR: c001272c
REGS: c0705c40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00055-g887df76de67f5)
MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000

GPR00: 001f0100 c0705cf8 c06dc300 c070af1c c001258c 00000000 00000000 ef7fb5bc 
GPR08: 08800000 00000100 00000001 00000100 42000044 00000000 c0709040 00000004 
GPR16: 00000001 c06022b4 c058297c 00200002 ffff8cb9 00000000 c06d84a0 c0710000 
GPR24: c0710000 00000000 00000000 c070af1c c070af1c 00000000 c001258c 00000000 
NIP [c00aef68] _einittext+0x3f9e5120/0x3feb71b8
LR [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
Call Trace:
[c0705cf8] [ef006320] 0xef006320 (unreliable)
[c0705d38] [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
[c0705d48] [c00af158] _einittext+0x3f9e5310/0x3feb71b8
[c0705d68] [c0012768] _einittext+0x3f948920/0x3feb71b8
[c0705d78] [c0092c04] _einittext+0x3f9c8dbc/0x3feb71b8
[c0705d88] [c0092d18] _einittext+0x3f9c8ed0/0x3feb71b8
[c0705da8] [c0093a2c] _einittext+0x3f9c9be4/0x3feb71b8
[c0705de8] [c0580224] _einittext+0x3feb63dc/0x3feb71b8
[c0705e48] [c00382ec] _einittext+0x3f96e4a4/0x3feb71b8
[c0705e58] [c000d2a0] _einittext+0x3f943458/0x3feb71b8
[c0705e88] [c001353c] _einittext+0x3f9496f4/0x3feb71b8
--- interrupt: 901 at _einittext+0x3f941058/0x3feb71b8
    LR = _einittext+0x3f941058/0x3feb71b8
[c0705f50] [c06cc214] 0xc06cc214 (unreliable)
[c0705f60] [c057fa20] _einittext+0x3feb5bd8/0x3feb71b8
[c0705f70] [c005de48] _einittext+0x3f994000/0x3feb71b8
[c0705f90] [c005e050] _einittext+0x3f994208/0x3feb71b8
[c0705fa0] [c0004cc8] _einittext+0x3f93ae80/0x3feb71b8
[c0705fb0] [c069a36c] _einittext+0x3ffd0524/0x40000000
[c0705ff0] [00003500] 0x3500
Instruction dump:
7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78 
7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6 
---[ end trace a06fef4788747c72 ]---


Prior to that (e.g. 97261e1e2240f), I get backtraces like this:

------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 smp_call_function_many+0x318/0x320
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00054-g97261e1e2240f #20
NIP:  c00aef68 LR: c00af114 CTR: c001272c
REGS: c075dc40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00054-g97261e1e2240f)
MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000

GPR00: 001f0100 c075dcf8 c0733300 c0762f1c c001258c 00000000 00000000 ef7fb5bc 
GPR08: 04800000 00000100 00000001 00000100 42000044 00000000 c0761040 00000004 
GPR16: 00000001 c0658e58 c058297c 00200002 ffff8cb9 00000000 c072f4a0 c0760000 
GPR24: c0760000 00000000 00000000 c0762f1c c0762f1c 00000000 c001258c 00000000 
NIP [c00aef68] smp_call_function_many+0x318/0x320
LR [c00af114] smp_call_function+0x34/0x44
Call Trace:
[c075dcf8] [ef006320] 0xef006320 (unreliable)
[c075dd38] [c00af114] smp_call_function+0x34/0x44
[c075dd48] [c00af158] on_each_cpu+0x1c/0x4c
[c075dd68] [c0012768] tau_timeout_smp+0x3c/0x4c
[c075dd78] [c0092c04] call_timer_fn.isra.26+0x20/0x84
[c075dd88] [c0092d18] expire_timers+0xb0/0xc0
[c075dda8] [c0093a2c] run_timer_softirq+0xa4/0x1a4
[c075dde8] [c0580224] __do_softirq+0x11c/0x280
[c075de48] [c00382ec] irq_exit+0xc0/0xd4
[c075de58] [c000d2a0] timer_interrupt+0x154/0x260
[c075de88] [c001353c] ret_from_except+0x0/0x14
--- interrupt: 901 at arch_cpu_idle+0x24/0x78
    LR = arch_cpu_idle+0x24/0x78
[c075df50] [c0723214] 0xc0723214 (unreliable)
[c075df60] [c057fa20] default_idle_call+0x38/0x58
[c075df70] [c005de48] do_idle+0xd4/0x17c
[c075df90] [c005e054] cpu_startup_entry+0x24/0x28
[c075dfa0] [c0004cc8] rest_init+0xa8/0xbc
[c075dfb0] [c06f136c] start_kernel+0x40c/0x420
[c075dff0] [00003500] 0x3500
Instruction dump:
7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78 
7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6 
---[ end trace 784c7f15ecd23941 ]---

Has anyone else observed these problems (either the WARNING from 
smp_call_function_many() or the missing symbol names)?

What is the best way to fix this? Should I upgrade binutils?

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Masahiro Yamada]

On Mon, Jul 20, 2020 at 10:46 AM Finn Thain <fthain@telegraphics.com.au> wrote:
>
> On Sun, 24 Nov 2019, Masahiro Yamada wrote:
>
> > Collect the ignored patterns to is_ignored_symbol().
> >
> > Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
>
> This commit (887df76de67f5) caused a regression in my powerpc builds as it
> causes symbol names to disappear from backtraces:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 _einittext+0x3f9e5120/0x3feb71b8
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00055-g887df76de67f5 #18
> NIP:  c00aef68 LR: c00af114 CTR: c001272c
> REGS: c0705c40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00055-g887df76de67f5)
> MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000
>
> GPR00: 001f0100 c0705cf8 c06dc300 c070af1c c001258c 00000000 00000000 ef7fb5bc
> GPR08: 08800000 00000100 00000001 00000100 42000044 00000000 c0709040 00000004
> GPR16: 00000001 c06022b4 c058297c 00200002 ffff8cb9 00000000 c06d84a0 c0710000
> GPR24: c0710000 00000000 00000000 c070af1c c070af1c 00000000 c001258c 00000000
> NIP [c00aef68] _einittext+0x3f9e5120/0x3feb71b8
> LR [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
> Call Trace:
> [c0705cf8] [ef006320] 0xef006320 (unreliable)
> [c0705d38] [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
> [c0705d48] [c00af158] _einittext+0x3f9e5310/0x3feb71b8
> [c0705d68] [c0012768] _einittext+0x3f948920/0x3feb71b8
> [c0705d78] [c0092c04] _einittext+0x3f9c8dbc/0x3feb71b8
> [c0705d88] [c0092d18] _einittext+0x3f9c8ed0/0x3feb71b8
> [c0705da8] [c0093a2c] _einittext+0x3f9c9be4/0x3feb71b8
> [c0705de8] [c0580224] _einittext+0x3feb63dc/0x3feb71b8
> [c0705e48] [c00382ec] _einittext+0x3f96e4a4/0x3feb71b8
> [c0705e58] [c000d2a0] _einittext+0x3f943458/0x3feb71b8
> [c0705e88] [c001353c] _einittext+0x3f9496f4/0x3feb71b8
> --- interrupt: 901 at _einittext+0x3f941058/0x3feb71b8
>     LR = _einittext+0x3f941058/0x3feb71b8
> [c0705f50] [c06cc214] 0xc06cc214 (unreliable)
> [c0705f60] [c057fa20] _einittext+0x3feb5bd8/0x3feb71b8
> [c0705f70] [c005de48] _einittext+0x3f994000/0x3feb71b8
> [c0705f90] [c005e050] _einittext+0x3f994208/0x3feb71b8
> [c0705fa0] [c0004cc8] _einittext+0x3f93ae80/0x3feb71b8
> [c0705fb0] [c069a36c] _einittext+0x3ffd0524/0x40000000
> [c0705ff0] [00003500] 0x3500
> Instruction dump:
> 7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78
> 7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6
> ---[ end trace a06fef4788747c72 ]---
>
>
> Prior to that (e.g. 97261e1e2240f), I get backtraces like this:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 smp_call_function_many+0x318/0x320
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00054-g97261e1e2240f #20
> NIP:  c00aef68 LR: c00af114 CTR: c001272c
> REGS: c075dc40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00054-g97261e1e2240f)
> MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000
>
> GPR00: 001f0100 c075dcf8 c0733300 c0762f1c c001258c 00000000 00000000 ef7fb5bc
> GPR08: 04800000 00000100 00000001 00000100 42000044 00000000 c0761040 00000004
> GPR16: 00000001 c0658e58 c058297c 00200002 ffff8cb9 00000000 c072f4a0 c0760000
> GPR24: c0760000 00000000 00000000 c0762f1c c0762f1c 00000000 c001258c 00000000
> NIP [c00aef68] smp_call_function_many+0x318/0x320
> LR [c00af114] smp_call_function+0x34/0x44
> Call Trace:
> [c075dcf8] [ef006320] 0xef006320 (unreliable)
> [c075dd38] [c00af114] smp_call_function+0x34/0x44
> [c075dd48] [c00af158] on_each_cpu+0x1c/0x4c
> [c075dd68] [c0012768] tau_timeout_smp+0x3c/0x4c
> [c075dd78] [c0092c04] call_timer_fn.isra.26+0x20/0x84
> [c075dd88] [c0092d18] expire_timers+0xb0/0xc0
> [c075dda8] [c0093a2c] run_timer_softirq+0xa4/0x1a4
> [c075dde8] [c0580224] __do_softirq+0x11c/0x280
> [c075de48] [c00382ec] irq_exit+0xc0/0xd4
> [c075de58] [c000d2a0] timer_interrupt+0x154/0x260
> [c075de88] [c001353c] ret_from_except+0x0/0x14
> --- interrupt: 901 at arch_cpu_idle+0x24/0x78
>     LR = arch_cpu_idle+0x24/0x78
> [c075df50] [c0723214] 0xc0723214 (unreliable)
> [c075df60] [c057fa20] default_idle_call+0x38/0x58
> [c075df70] [c005de48] do_idle+0xd4/0x17c
> [c075df90] [c005e054] cpu_startup_entry+0x24/0x28
> [c075dfa0] [c0004cc8] rest_init+0xa8/0xbc
> [c075dfb0] [c06f136c] start_kernel+0x40c/0x420
> [c075dff0] [00003500] 0x3500
> Instruction dump:
> 7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78
> 7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6
> ---[ end trace 784c7f15ecd23941 ]---
>
> Has anyone else observed these problems (either the WARNING from
> smp_call_function_many() or the missing symbol names)?
>
> What is the best way to fix this? Should I upgrade binutils?



I got a similar report before.

I'd like to know whether or not
this is the same issue as fixed by
7883a14339299773b2ce08dcfd97c63c199a9289


Does your problem happen on the latest kernel?
Which version of binutils are you using?


-- 
Best Regards
Masahiro Yamada

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Masahiro Yamada]

On Mon, Jul 20, 2020 at 10:46 AM Finn Thain <fthain@telegraphics.com.au> wrote:
>
> On Sun, 24 Nov 2019, Masahiro Yamada wrote:
>
> > Collect the ignored patterns to is_ignored_symbol().
> >
> > Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
>
> This commit (887df76de67f5) caused a regression in my powerpc builds as it
> causes symbol names to disappear from backtraces:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 _einittext+0x3f9e5120/0x3feb71b8
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00055-g887df76de67f5 #18
> NIP:  c00aef68 LR: c00af114 CTR: c001272c
> REGS: c0705c40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00055-g887df76de67f5)
> MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000
>
> GPR00: 001f0100 c0705cf8 c06dc300 c070af1c c001258c 00000000 00000000 ef7fb5bc
> GPR08: 08800000 00000100 00000001 00000100 42000044 00000000 c0709040 00000004
> GPR16: 00000001 c06022b4 c058297c 00200002 ffff8cb9 00000000 c06d84a0 c0710000
> GPR24: c0710000 00000000 00000000 c070af1c c070af1c 00000000 c001258c 00000000
> NIP [c00aef68] _einittext+0x3f9e5120/0x3feb71b8
> LR [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
> Call Trace:
> [c0705cf8] [ef006320] 0xef006320 (unreliable)
> [c0705d38] [c00af114] _einittext+0x3f9e52cc/0x3feb71b8
> [c0705d48] [c00af158] _einittext+0x3f9e5310/0x3feb71b8
> [c0705d68] [c0012768] _einittext+0x3f948920/0x3feb71b8
> [c0705d78] [c0092c04] _einittext+0x3f9c8dbc/0x3feb71b8
> [c0705d88] [c0092d18] _einittext+0x3f9c8ed0/0x3feb71b8
> [c0705da8] [c0093a2c] _einittext+0x3f9c9be4/0x3feb71b8
> [c0705de8] [c0580224] _einittext+0x3feb63dc/0x3feb71b8
> [c0705e48] [c00382ec] _einittext+0x3f96e4a4/0x3feb71b8
> [c0705e58] [c000d2a0] _einittext+0x3f943458/0x3feb71b8
> [c0705e88] [c001353c] _einittext+0x3f9496f4/0x3feb71b8
> --- interrupt: 901 at _einittext+0x3f941058/0x3feb71b8
>     LR = _einittext+0x3f941058/0x3feb71b8
> [c0705f50] [c06cc214] 0xc06cc214 (unreliable)
> [c0705f60] [c057fa20] _einittext+0x3feb5bd8/0x3feb71b8
> [c0705f70] [c005de48] _einittext+0x3f994000/0x3feb71b8
> [c0705f90] [c005e050] _einittext+0x3f994208/0x3feb71b8
> [c0705fa0] [c0004cc8] _einittext+0x3f93ae80/0x3feb71b8
> [c0705fb0] [c069a36c] _einittext+0x3ffd0524/0x40000000
> [c0705ff0] [00003500] 0x3500
> Instruction dump:
> 7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78
> 7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6
> ---[ end trace a06fef4788747c72 ]---
>
>
> Prior to that (e.g. 97261e1e2240f), I get backtraces like this:
>
> ------------[ cut here ]------------
> WARNING: CPU: 0 PID: 0 at kernel/smp.c:433 smp_call_function_many+0x318/0x320
> Modules linked in:
> CPU: 0 PID: 0 Comm: swapper/0 Not tainted 5.4.0-rc7-pmac-00054-g97261e1e2240f #20
> NIP:  c00aef68 LR: c00af114 CTR: c001272c
> REGS: c075dc40 TRAP: 0700   Not tainted  (5.4.0-rc7-pmac-00054-g97261e1e2240f)
> MSR:  00029032 <EE,ME,IR,DR,RI>  CR: 42000044  XER: 00000000
>
> GPR00: 001f0100 c075dcf8 c0733300 c0762f1c c001258c 00000000 00000000 ef7fb5bc
> GPR08: 04800000 00000100 00000001 00000100 42000044 00000000 c0761040 00000004
> GPR16: 00000001 c0658e58 c058297c 00200002 ffff8cb9 00000000 c072f4a0 c0760000
> GPR24: c0760000 00000000 00000000 c0762f1c c0762f1c 00000000 c001258c 00000000
> NIP [c00aef68] smp_call_function_many+0x318/0x320
> LR [c00af114] smp_call_function+0x34/0x44
> Call Trace:
> [c075dcf8] [ef006320] 0xef006320 (unreliable)
> [c075dd38] [c00af114] smp_call_function+0x34/0x44
> [c075dd48] [c00af158] on_each_cpu+0x1c/0x4c
> [c075dd68] [c0012768] tau_timeout_smp+0x3c/0x4c
> [c075dd78] [c0092c04] call_timer_fn.isra.26+0x20/0x84
> [c075dd88] [c0092d18] expire_timers+0xb0/0xc0
> [c075dda8] [c0093a2c] run_timer_softirq+0xa4/0x1a4
> [c075dde8] [c0580224] __do_softirq+0x11c/0x280
> [c075de48] [c00382ec] irq_exit+0xc0/0xd4
> [c075de58] [c000d2a0] timer_interrupt+0x154/0x260
> [c075de88] [c001353c] ret_from_except+0x0/0x14
> --- interrupt: 901 at arch_cpu_idle+0x24/0x78
>     LR = arch_cpu_idle+0x24/0x78
> [c075df50] [c0723214] 0xc0723214 (unreliable)
> [c075df60] [c057fa20] default_idle_call+0x38/0x58
> [c075df70] [c005de48] do_idle+0xd4/0x17c
> [c075df90] [c005e054] cpu_startup_entry+0x24/0x28
> [c075dfa0] [c0004cc8] rest_init+0xa8/0xbc
> [c075dfb0] [c06f136c] start_kernel+0x40c/0x420
> [c075dff0] [00003500] 0x3500
> Instruction dump:
> 7c0803a6 7fa5eb78 7d808120 7ea6ab78 baa10014 38210040 4bfffbb0 7f64db78
> 7f85e378 484b31b1 7c601b78 4bfffdf4 <0fe00000> 4bfffd60 9421ffe0 7c0802a6
> ---[ end trace 784c7f15ecd23941 ]---
>
> Has anyone else observed these problems (either the WARNING from
> smp_call_function_many() or the missing symbol names)?
>
> What is the best way to fix this? Should I upgrade binutils?



I got a similar report before.

I'd like to know whether or not
this is the same issue as fixed by
7883a14339299773b2ce08dcfd97c63c199a9289


Does your problem happen on the latest kernel?
Which version of binutils are you using?


-- 
Best Regards
Masahiro Yamada

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Finn Thain]

On Mon, 20 Jul 2020, Masahiro Yamada wrote:

> 
> I got a similar report before.
> 
> I'd like to know whether or not
> this is the same issue as fixed by
> 7883a14339299773b2ce08dcfd97c63c199a9289
> 

The problem can be observed with 3d77e6a8804ab ("Linux 5.7").
So it appears that 7883a14339299 ("scripts/kallsyms: fix wrong 
kallsyms_relative_base") is not sufficient to fix it.

> 
> Does your problem happen on the latest kernel?

Unfortunately this cross compiler (gcc 4.6.4) is too old to build 
v5.8-rc1 or later. I will have to upgrade.

> Which version of binutils are you using?
> 

This toolchain uses binutils 2.22.

In case it helps,

$ powerpc-linux-gnu-nm -n vmlinux |head                  
         w mach_chrp
00000005 a LG_CACHELINE_BYTES
00000005 a LG_CACHELINE_BYTES
00000005 a LG_CACHELINE_BYTES
0000000c a Hash_bits
0000001f a CACHELINE_MASK
0000001f a CACHELINE_MASK
0000001f a CACHELINE_MASK
00000020 a CACHELINE_BYTES
00000020 a CACHELINE_BYTES
00000020 a CACHELINE_BYTES
00000020 a reg
0003ffc0 a Hash_msk
c0000000 T _start
c0000000 A _stext
c0000000 A _text
c000000c T __start
c0000054 t __after_mmu_off
c0000090 t turn_on_mmu
c00000c4 T __secondary_hold
c00000dc T __secondary_hold_spinloop
c00000e0 T __secondary_hold_acknowledge
c0000100 t Reset

Re: [PATCH v2 13/16] scripts/kallsyms: move ignored symbol types to is_ignored_symbol()

[Finn Thain]

On Mon, 20 Jul 2020, Masahiro Yamada wrote:

> 
> I got a similar report before.
> 
> I'd like to know whether or not
> this is the same issue as fixed by
> 7883a14339299773b2ce08dcfd97c63c199a9289
> 

The problem can be observed with 3d77e6a8804ab ("Linux 5.7").
So it appears that 7883a14339299 ("scripts/kallsyms: fix wrong 
kallsyms_relative_base") is not sufficient to fix it.

> 
> Does your problem happen on the latest kernel?

Unfortunately this cross compiler (gcc 4.6.4) is too old to build 
v5.8-rc1 or later. I will have to upgrade.

> Which version of binutils are you using?
> 

This toolchain uses binutils 2.22.

In case it helps,

$ powerpc-linux-gnu-nm -n vmlinux |head                  
         w mach_chrp
00000005 a LG_CACHELINE_BYTES
00000005 a LG_CACHELINE_BYTES
00000005 a LG_CACHELINE_BYTES
0000000c a Hash_bits
0000001f a CACHELINE_MASK
0000001f a CACHELINE_MASK
0000001f a CACHELINE_MASK
00000020 a CACHELINE_BYTES
00000020 a CACHELINE_BYTES
00000020 a CACHELINE_BYTES
00000020 a reg
0003ffc0 a Hash_msk
c0000000 T _start
c0000000 A _stext
c0000000 A _text
c000000c T __start
c0000054 t __after_mmu_off
c0000090 t turn_on_mmu
c00000c4 T __secondary_hold
c00000dc T __secondary_hold_spinloop
c00000e0 T __secondary_hold_acknowledge
c0000100 t Reset

[PATCH v2 14/16] scripts/kallsyms: make check_symbol_range() void function

[Masahiro Yamada]

There is no more reason to check the return value of
check_symbol_range().

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 15 ++++++---------
 1 file changed, 6 insertions(+), 9 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index d90a6133d7b8..f4d5f131556d 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -155,8 +155,8 @@ static bool is_ignored_symbol(const char *name, char type)
 	return false;
 }
 
-static int check_symbol_range(const char *sym, unsigned long long addr,
-			      struct addr_range *ranges, int entries)
+static void check_symbol_range(const char *sym, unsigned long long addr,
+			       struct addr_range *ranges, int entries)
 {
 	size_t i;
 	struct addr_range *ar;
@@ -166,14 +166,12 @@ static int check_symbol_range(const char *sym, unsigned long long addr,
 
 		if (strcmp(sym, ar->start_sym) == 0) {
 			ar->start = addr;
-			return 0;
+			return;
 		} else if (strcmp(sym, ar->end_sym) == 0) {
 			ar->end = addr;
-			return 0;
+			return;
 		}
 	}
-
-	return 1;
 }
 
 static int read_symbol(FILE *in, struct sym_entry *s)
@@ -200,9 +198,8 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 	/* Ignore most absolute/undefined (?) symbols. */
 	if (strcmp(sym, "_text") == 0)
 		_text = s->addr;
-	else if (check_symbol_range(sym, s->addr, text_ranges,
-				    ARRAY_SIZE(text_ranges)) == 0)
-		/* nothing to do */;
+
+	check_symbol_range(sym, s->addr, text_ranges, ARRAY_SIZE(text_ranges));
 
 	/* include the type field in the symbol name, so that it gets
 	 * compressed together */
-- 
2.17.1

[PATCH v2 15/16] scripts/kallsyms: put check_symbol_range() calls close together

[Masahiro Yamada]

Put the relevant code close together.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index f4d5f131556d..b9b1a4cf1c65 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -200,6 +200,7 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 		_text = s->addr;
 
 	check_symbol_range(sym, s->addr, text_ranges, ARRAY_SIZE(text_ranges));
+	check_symbol_range(sym, s->addr, &percpu_range, 1);
 
 	/* include the type field in the symbol name, so that it gets
 	 * compressed together */
@@ -215,9 +216,6 @@ static int read_symbol(FILE *in, struct sym_entry *s)
 
 	s->percpu_absolute = 0;
 
-	/* Record if we've found __per_cpu_start/end. */
-	check_symbol_range(sym, s->addr, &percpu_range, 1);
-
 	return 0;
 }
 
-- 
2.17.1

[PATCH v2 16/16] scripts/kallsyms: remove redundant initializers

[Masahiro Yamada]

These are set to zero without the explicit initializers.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
---

Changes in v2: None

 scripts/kallsyms.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index b9b1a4cf1c65..fb55f262f42d 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -57,9 +57,9 @@ static struct addr_range percpu_range = {
 
 static struct sym_entry *table;
 static unsigned int table_size, table_cnt;
-static int all_symbols = 0;
-static int absolute_percpu = 0;
-static int base_relative = 0;
+static int all_symbols;
+static int absolute_percpu;
+static int base_relative;
 
 static int token_profit[0x10000];
 
-- 
2.17.1