[v8,4/9] kallsyms: Reduce the memory occupied by kallsyms_seqs_of_names[]

Message ID 20221102084921.1615-5-thunder.leizhen@huawei.com
State New
Headers
Series kallsyms: Optimizes the performance of lookup symbols |

Commit Message

Zhen Lei Nov. 2, 2022, 8:49 a.m. UTC
  kallsyms_seqs_of_names[] records the symbol index sorted by address, the
maximum value in kallsyms_seqs_of_names[] is the number of symbols. And
2^24 = 16777216, which means that three bytes are enough to store the
index. This can help us save (1 * kallsyms_num_syms) bytes of memory.

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
---
 kernel/kallsyms.c          | 18 ++++++++++++++----
 kernel/kallsyms_internal.h |  2 +-
 scripts/kallsyms.c         |  5 ++++-
 3 files changed, 19 insertions(+), 6 deletions(-)
  

Comments

David Laight Nov. 2, 2022, noon UTC | #1
From: Zhen Lei
> Sent: 02 November 2022 08:49
> 
> kallsyms_seqs_of_names[] records the symbol index sorted by address, the
> maximum value in kallsyms_seqs_of_names[] is the number of symbols. And
> 2^24 = 16777216, which means that three bytes are enough to store the
> index. This can help us save (1 * kallsyms_num_syms) bytes of memory.

You can get the compiler to do the 'heavy lifting' for you.

struct uint24 {
    unsigned int val24:24;
} __attribute__((packed));

struct uint24 table[1024];

works fine.

	David

-
Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
Registration No: 1397386 (Wales)
  
Zhen Lei Nov. 7, 2022, 8:01 a.m. UTC | #2
On 2022/11/2 20:00, David Laight wrote:
> From: Zhen Lei
>> Sent: 02 November 2022 08:49
>>
>> kallsyms_seqs_of_names[] records the symbol index sorted by address, the
>> maximum value in kallsyms_seqs_of_names[] is the number of symbols. And
>> 2^24 = 16777216, which means that three bytes are enough to store the
>> index. This can help us save (1 * kallsyms_num_syms) bytes of memory.
> 
> You can get the compiler to do the 'heavy lifting' for you.
> 
> struct uint24 {
>     unsigned int val24:24;
> } __attribute__((packed));

This method depends on byte order. If the byte order of the tool is
different from that of the kernel, it's a problem. For example,
cross-compile PowerPC kernel on x86.

> 
> struct uint24 table[1024];
> 
> works fine.
> 
> 	David
> 
> -
> Registered Address Lakeside, Bramley Road, Mount Farm, Milton Keynes, MK1 1PT, UK
> Registration No: 1397386 (Wales)
> 
> 
> .
>
  

Patch

diff --git a/kernel/kallsyms.c b/kernel/kallsyms.c
index ba351dfa109b6ac..48f36fd7e10b95e 100644
--- a/kernel/kallsyms.c
+++ b/kernel/kallsyms.c
@@ -201,6 +201,16 @@  static int compare_symbol_name(const char *name, char *namebuf)
 	return ret;
 }
 
+static unsigned int get_symbol_seq(int index)
+{
+	unsigned int i, seq = 0;
+
+	for (i = 0; i < 3; i++)
+		seq = (seq << 8) | kallsyms_seqs_of_names[3 * index + i];
+
+	return seq;
+}
+
 static int kallsyms_lookup_names(const char *name,
 				 unsigned int *start,
 				 unsigned int *end)
@@ -215,7 +225,7 @@  static int kallsyms_lookup_names(const char *name,
 
 	while (low <= high) {
 		mid = low + (high - low) / 2;
-		seq = kallsyms_seqs_of_names[mid];
+		seq = get_symbol_seq(mid);
 		off = get_symbol_offset(seq);
 		kallsyms_expand_symbol(off, namebuf, ARRAY_SIZE(namebuf));
 		ret = compare_symbol_name(name, namebuf);
@@ -232,7 +242,7 @@  static int kallsyms_lookup_names(const char *name,
 
 	low = mid;
 	while (low) {
-		seq = kallsyms_seqs_of_names[low - 1];
+		seq = get_symbol_seq(low - 1);
 		off = get_symbol_offset(seq);
 		kallsyms_expand_symbol(off, namebuf, ARRAY_SIZE(namebuf));
 		if (compare_symbol_name(name, namebuf))
@@ -244,7 +254,7 @@  static int kallsyms_lookup_names(const char *name,
 	if (end) {
 		high = mid;
 		while (high < kallsyms_num_syms - 1) {
-			seq = kallsyms_seqs_of_names[high + 1];
+			seq = get_symbol_seq(high + 1);
 			off = get_symbol_offset(seq);
 			kallsyms_expand_symbol(off, namebuf, ARRAY_SIZE(namebuf));
 			if (compare_symbol_name(name, namebuf))
@@ -269,7 +279,7 @@  unsigned long kallsyms_lookup_name(const char *name)
 
 	ret = kallsyms_lookup_names(name, &i, NULL);
 	if (!ret)
-		return kallsyms_sym_address(kallsyms_seqs_of_names[i]);
+		return kallsyms_sym_address(get_symbol_seq(i));
 
 	return module_kallsyms_lookup_name(name);
 }
diff --git a/kernel/kallsyms_internal.h b/kernel/kallsyms_internal.h
index a04b7a5cb1e3eaf..27fabdcc40f5793 100644
--- a/kernel/kallsyms_internal.h
+++ b/kernel/kallsyms_internal.h
@@ -26,6 +26,6 @@  extern const char kallsyms_token_table[] __weak;
 extern const u16 kallsyms_token_index[] __weak;
 
 extern const unsigned int kallsyms_markers[] __weak;
-extern const unsigned int kallsyms_seqs_of_names[] __weak;
+extern const u8 kallsyms_seqs_of_names[] __weak;
 
 #endif // LINUX_KALLSYMS_INTERNAL_H_
diff --git a/scripts/kallsyms.c b/scripts/kallsyms.c
index 07ecf7e5c49f616..04e04fbd9625caf 100644
--- a/scripts/kallsyms.c
+++ b/scripts/kallsyms.c
@@ -600,7 +600,10 @@  static void write_src(void)
 	sort_symbols_by_name();
 	output_label("kallsyms_seqs_of_names");
 	for (i = 0; i < table_cnt; i++)
-		printf("\t.long\t%u\n", table[i]->seq);
+		printf("\t.byte 0x%02x, 0x%02x, 0x%02x\n",
+			(unsigned char)(table[i]->seq >> 16),
+			(unsigned char)(table[i]->seq >> 8),
+			(unsigned char)(table[i]->seq >> 0));
 	printf("\n");
 
 	output_label("kallsyms_token_table");