# HG changeset patch
# User kaf24@xxxxxxxxxxxxxxxxxxxx
# Node ID 48aed1403fe35f80aaed8f03c354cbb8c1a0c4ab
# Parent 7627476544b58da7fd4ba909f2991dfe1f2f4109
Port kallsyms to Xen, as 'symbols'.
Signed-off-by: Keir Fraser <keir@xxxxxxxxxxxxx>
diff -r 7627476544b5 -r 48aed1403fe3 xen/arch/ia64/Makefile
--- a/xen/arch/ia64/Makefile Fri Jul 22 14:25:10 2005
+++ b/xen/arch/ia64/Makefile Fri Jul 22 16:44:33 2005
@@ -62,10 +62,4 @@
rm -f *.o *~ core xen.lds.s
$(BASEDIR)/include/asm-ia64/.offsets.h.stamp asm-offsets.s
rm -f lib/*.o
-# setup.o contains bits of compile.h so it must be blown away
-delete-unfresh-files:
- echo any unfresh-files to delete for ia64\?
-# rm -f setup.o
-
-.PHONY: default clean delete-unfresh-files
-
+.PHONY: default clean
diff -r 7627476544b5 -r 48aed1403fe3 xen/Makefile
--- a/xen/Makefile Fri Jul 22 14:25:10 2005
+++ b/xen/Makefile Fri Jul 22 16:44:33 2005
@@ -50,10 +50,10 @@
$(MAKE) -C arch/$(TARGET_ARCH) clean
rm -f include/asm *.o $(TARGET)* *~ core
rm -f include/asm-*/asm-offsets.h
- rm -f tools/figlet/*.o tools/figlet/figlet
rm -f include/xen/acm_policy.h
$(TARGET): delete-unfresh-files
+ $(MAKE) -C tools
$(MAKE) include/xen/compile.h
$(MAKE) include/xen/acm_policy.h
[ -e include/asm ] || ln -sf asm-$(TARGET_ARCH) include/asm
@@ -71,7 +71,6 @@
delete-unfresh-files:
@if [ ! -r include/xen/compile.h -o -O include/xen/compile.h ]; then \
rm -f include/xen/{banner,compile}.h; \
- $(MAKE) -C arch/$(TARGET_ARCH) delete-unfresh-files; \
fi
# acm_policy.h contains security policy for Xen
@@ -105,12 +104,7 @@
@cat include/xen/banner.h >> $@.new
@mv -f $@.new $@
-tools/figlet/figlet: tools/figlet/figlet.o
- $(HOSTCC) -o $@ $<
-tools/figlet/figlet.o: tools/figlet/figlet.c
- $(HOSTCC) -o $@ -c $<
-
-include/xen/banner.h: tools/figlet/figlet tools/figlet/xen.flf
+include/xen/banner.h:
tools/figlet/figlet -d tools/figlet Xen $(XEN_FULLVERSION) > $@.new
@mv -f $@.new $@
diff -r 7627476544b5 -r 48aed1403fe3 xen/arch/x86/Makefile
--- a/xen/arch/x86/Makefile Fri Jul 22 14:25:10 2005
+++ b/xen/arch/x86/Makefile Fri Jul 22 16:44:33 2005
@@ -37,6 +37,15 @@
$(TARGET)-syms: boot/$(TARGET_SUBARCH).o $(ALL_OBJS) $(TARGET_SUBARCH)/xen.lds
$(LD) $(LDFLAGS) -T $(TARGET_SUBARCH)/xen.lds -N \
boot/$(TARGET_SUBARCH).o $(ALL_OBJS) -o $@
+ nm -n $@ | $(BASEDIR)/tools/symbols >$(BASEDIR)/xen-syms.S
+ $(MAKE) $(BASEDIR)/xen-syms.o
+ $(LD) $(LDFLAGS) -T $(TARGET_SUBARCH)/xen.lds -N \
+ boot/$(TARGET_SUBARCH).o $(ALL_OBJS) $(BASEDIR)/xen-syms.o -o $@
+ nm -n $@ | $(BASEDIR)/tools/symbols >$(BASEDIR)/xen-syms.S
+ $(MAKE) $(BASEDIR)/xen-syms.o
+ $(LD) $(LDFLAGS) -T $(TARGET_SUBARCH)/xen.lds -N \
+ boot/$(TARGET_SUBARCH).o $(ALL_OBJS) $(BASEDIR)/xen-syms.o -o $@
+ rm -f $(BASEDIR)/xen-syms.S $(BASEDIR)/xen-syms.o
asm-offsets.s: $(TARGET_SUBARCH)/asm-offsets.c $(HDRS)
$(CC) $(CFLAGS) -S -o $@ $<
@@ -53,7 +62,4 @@
rm -f genapic/*.o genapic/*~ genapic/core
rm -f cpu/*.o cpu/*~ cpu/core
-delete-unfresh-files:
- # nothing
-
-.PHONY: default clean delete-unfresh-files
+.PHONY: default clean
diff -r 7627476544b5 -r 48aed1403fe3 .hgignore
--- a/.hgignore Fri Jul 22 14:25:10 2005
+++ b/.hgignore Fri Jul 22 16:44:33 2005
@@ -162,10 +162,6 @@
^xen/arch/x86/asm-offsets\.s$
^xen/arch/x86/boot/mkelf32$
^xen/ddb/.*$
-^xen/drivers/pci/classlist\.h$
-^xen/drivers/pci/devlist\.h$
-^xen/drivers/pci/gen-devlist$
-^xen/figlet/figlet$
^xen/include/asm$
^xen/include/asm-.*/asm-offsets\.h$
^xen/include/hypervisor-ifs/arch$
@@ -175,8 +171,8 @@
^xen/include/xen/banner\.h$
^xen/include/xen/compile\.h$
^xen/tags$
-^xen/tools/elf-reloc$
^xen/tools/figlet/figlet$
+^xen/tools/symbols$
^xen/xen$
^xen/xen-syms$
^xen/xen\..*$
diff -r 7627476544b5 -r 48aed1403fe3 xen/arch/x86/x86_32/traps.c
--- a/xen/arch/x86/x86_32/traps.c Fri Jul 22 14:25:10 2005
+++ b/xen/arch/x86/x86_32/traps.c Fri Jul 22 16:44:33 2005
@@ -6,6 +6,7 @@
#include <xen/console.h>
#include <xen/mm.h>
#include <xen/irq.h>
+#include <xen/symbols.h>
#include <asm/current.h>
#include <asm/flushtlb.h>
#include <asm/vmx.h>
@@ -63,10 +64,10 @@
}
}
- printk("CPU: %d\nEIP: %04lx:[<%08lx>] \nEFLAGS: %08lx "
- "CONTEXT: %s\n",
- smp_processor_id(), (unsigned long)0xffff & regs->cs,
- eip, eflags, context);
+ printk("CPU: %d\nEIP: %04lx:[<%08lx>]",
+ smp_processor_id(), (unsigned long)0xffff & regs->cs, eip);
+ print_symbol(" %s\n", eip);
+ printk("EFLAGS: %08lx CONTEXT: %s\n", eflags, context);
printk("eax: %08x ebx: %08x ecx: %08x edx: %08x\n",
regs->eax, regs->ebx, regs->ecx, regs->edx);
printk("esi: %08x edi: %08x ebp: %08x esp: %08lx\n",
@@ -119,8 +120,10 @@
/* Find information saved during fault and dump it to the console. */
tss = &init_tss[cpu];
- printk("CPU: %d\nEIP: %04x:[<%08x>] \nEFLAGS: %08x\n",
- cpu, tss->cs, tss->eip, tss->eflags);
+ printk("CPU: %d\nEIP: %04x:[<%08x>]",
+ cpu, tss->cs, tss->eip);
+ print_symbol(" %s\n", tss->eip);
+ printk("EFLAGS: %08x\n", tss->eflags);
printk("CR3: %08x\n", tss->__cr3);
printk("eax: %08x ebx: %08x ecx: %08x edx: %08x\n",
tss->eax, tss->ebx, tss->ecx, tss->edx);
diff -r 7627476544b5 -r 48aed1403fe3 xen/include/asm-x86/types.h
--- a/xen/include/asm-x86/types.h Fri Jul 22 14:25:10 2005
+++ b/xen/include/asm-x86/types.h Fri Jul 22 16:44:33 2005
@@ -1,10 +1,9 @@
#ifndef __X86_TYPES_H__
#define __X86_TYPES_H__
-/*
- * __xx is ok: it doesn't pollute the POSIX namespace. Use these in the
- * header files exported to user space
- */
+#ifndef __ASSEMBLY__
+
+#include <xen/config.h>
typedef __signed__ char __s8;
typedef unsigned char __u8;
@@ -25,8 +24,6 @@
#endif
#endif
-#include <xen/config.h>
-
typedef signed char s8;
typedef unsigned char u8;
@@ -39,9 +36,6 @@
#if defined(__i386__)
typedef signed long long s64;
typedef unsigned long long u64;
-#define BITS_PER_LONG 32
-#define BYTES_PER_LONG 4
-#define LONG_BYTEORDER 2
#if defined(CONFIG_X86_PAE)
typedef u64 physaddr_t;
#else
@@ -50,12 +44,21 @@
#elif defined(__x86_64__)
typedef signed long s64;
typedef unsigned long u64;
-#define BITS_PER_LONG 64
-#define BYTES_PER_LONG 8
-#define LONG_BYTEORDER 3
typedef u64 physaddr_t;
#endif
typedef unsigned long size_t;
+#endif /* __ASSEMBLY__ */
+
+#if defined(__i386__)
+#define BITS_PER_LONG 32
+#define BYTES_PER_LONG 4
+#define LONG_BYTEORDER 2
+#elif defined(__x86_64__)
+#define BITS_PER_LONG 64
+#define BYTES_PER_LONG 8
+#define LONG_BYTEORDER 3
+#endif
+
#endif /* __X86_TYPES_H__ */
diff -r 7627476544b5 -r 48aed1403fe3 xen/arch/x86/x86_64/traps.c
--- a/xen/arch/x86/x86_64/traps.c Fri Jul 22 14:25:10 2005
+++ b/xen/arch/x86/x86_64/traps.c Fri Jul 22 16:44:33 2005
@@ -6,6 +6,7 @@
#include <xen/errno.h>
#include <xen/mm.h>
#include <xen/irq.h>
+#include <xen/symbols.h>
#include <xen/console.h>
#include <xen/sched.h>
#include <asm/current.h>
@@ -14,8 +15,10 @@
void show_registers(struct cpu_user_regs *regs)
{
- printk("CPU: %d\nEIP: %04x:[<%016lx>] \nEFLAGS: %016lx\n",
- smp_processor_id(), 0xffff & regs->cs, regs->rip, regs->eflags);
+ printk("CPU: %d\nEIP: %04x:[<%016lx>]",
+ smp_processor_id(), 0xffff & regs->cs, regs->rip);
+ print_symbol(" %s\n", regs->rip);
+ printk("EFLAGS: %016lx\n", regs->eflags);
printk("rax: %016lx rbx: %016lx rcx: %016lx rdx: %016lx\n",
regs->rax, regs->rbx, regs->rcx, regs->rdx);
printk("rsi: %016lx rdi: %016lx rbp: %016lx rsp: %016lx\n",
diff -r 7627476544b5 -r 48aed1403fe3 xen/arch/x86/traps.c
--- a/xen/arch/x86/traps.c Fri Jul 22 14:25:10 2005
+++ b/xen/arch/x86/traps.c Fri Jul 22 16:44:33 2005
@@ -40,6 +40,7 @@
#include <xen/perfc.h>
#include <xen/softirq.h>
#include <xen/domain_page.h>
+#include <xen/symbols.h>
#include <asm/shadow.h>
#include <asm/system.h>
#include <asm/io.h>
@@ -100,7 +101,7 @@
static int debug_stack_lines = 20;
integer_param("debug_stack_lines", debug_stack_lines);
-static inline int kernel_text_address(unsigned long addr)
+int is_kernel_text(unsigned long addr)
{
extern char _stext, _etext;
if (addr >= (unsigned long) &_stext &&
@@ -110,6 +111,12 @@
}
+unsigned long kernel_text_end(void)
+{
+ extern char _etext;
+ return (unsigned long) &_etext;
+}
+
void show_guest_stack(void)
{
int i;
@@ -150,11 +157,12 @@
while ( ((long) stack & (STACK_SIZE-1)) != 0 )
{
addr = *stack++;
- if ( kernel_text_address(addr) )
+ if ( is_kernel_text(addr) )
{
if ( (i != 0) && ((i % 6) == 0) )
printk("\n ");
- printk("[<%p>] ", _p(addr));
+ printk("[<%p>]", _p(addr));
+ print_symbol(" %s\n", addr);
i++;
}
}
@@ -177,10 +185,7 @@
if ( (i != 0) && ((i % 8) == 0) )
printk("\n ");
addr = *stack++;
- if ( kernel_text_address(addr) )
- printk("[%p] ", _p(addr));
- else
- printk("%p ", _p(addr));
+ printk("%p ", _p(addr));
}
if ( i == 0 )
printk("Stack empty.");
diff -r 7627476544b5 -r 48aed1403fe3 xen/tools/Makefile
--- a/xen/tools/Makefile Fri Jul 22 14:25:10 2005
+++ b/xen/tools/Makefile Fri Jul 22 16:44:33 2005
@@ -1,6 +1,13 @@
+
+include $(BASEDIR)/../Config.mk
default:
$(MAKE) -C figlet
+ $(MAKE) symbols
clean:
- $(MAKE) -C figlet clean
\ No newline at end of file
+ $(MAKE) -C figlet clean
+ rm -f *.o symbols
+
+symbols: symbols.c
+ $(HOSTCC) -o $@ $<
diff -r 7627476544b5 -r 48aed1403fe3 xen/include/xen/symbols.h
--- /dev/null Fri Jul 22 14:25:10 2005
+++ b/xen/include/xen/symbols.h Fri Jul 22 16:44:33 2005
@@ -0,0 +1,45 @@
+#ifndef _XEN_SYMBOLS_H
+#define _XEN_SYMBOLS_H
+
+#include <xen/config.h>
+#include <xen/types.h>
+
+#define KSYM_NAME_LEN 127
+
+extern int is_kernel_text(unsigned long addr);
+extern unsigned long kernel_text_end(void);
+
+/* Lookup an address. */
+const char *symbols_lookup(unsigned long addr,
+ unsigned long *symbolsize,
+ unsigned long *offset,
+ char *namebuf);
+
+/* Replace "%s" in format with address, if found */
+extern void __print_symbol(const char *fmt, unsigned long address);
+
+/* This macro allows us to keep printk typechecking */
+static void __check_printsym_format(const char *fmt, ...)
+ __attribute__((format(printf,1,2)));
+ static inline void __check_printsym_format(const char *fmt, ...)
+{
+}
+
+/* ia64 and ppc64 use function descriptors, which contain the real address */
+#if defined(CONFIG_IA64) || defined(CONFIG_PPC64)
+#define print_fn_descriptor_symbol(fmt, addr) \
+do { \
+ unsigned long *__faddr = (unsigned long*) addr; \
+ print_symbol(fmt, __faddr[0]); \
+} while (0)
+#else
+#define print_fn_descriptor_symbol(fmt, addr) print_symbol(fmt, addr)
+#endif
+
+#define print_symbol(fmt, addr) \
+do { \
+ __check_printsym_format(fmt, ""); \
+ __print_symbol(fmt, addr); \
+} while(0)
+
+#endif /*_XEN_SYMBOLS_H*/
diff -r 7627476544b5 -r 48aed1403fe3 xen/tools/symbols.c
--- /dev/null Fri Jul 22 14:25:10 2005
+++ b/xen/tools/symbols.c Fri Jul 22 16:44:33 2005
@@ -0,0 +1,716 @@
+/* Generate assembler source containing symbol information
+ *
+ * Copyright 2002 by Kai Germaschewski
+ *
+ * This software may be used and distributed according to the terms
+ * of the GNU General Public License, incorporated herein by reference.
+ *
+ * Usage: nm -n <object-file> | scripts/symbols [--all-symbols] > symbols.S
+ *
+ * ChangeLog:
+ *
+ * (25/Aug/2004) Paulo Marques <pmarques@xxxxxxxxxxxx>
+ * Changed the compression method from stem compression to "table lookup"
+ * compression
+ *
+ * Table compression uses all the unused char codes on the symbols and
+ * maps these to the most used substrings (tokens). For instance, it might
+ * map char code 0xF7 to represent "write_" and then in every symbol where
+ * "write_" appears it can be replaced by 0xF7, saving 5 bytes.
+ * The used codes themselves are also placed in the table so that the
+ * decompresion can work without "special cases".
+ * Applied to kernel symbols, this usually produces a compression ratio
+ * of about 50%.
+ *
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+
+/* maximum token length used. It doesn't pay to increase it a lot, because
+ * very long substrings probably don't repeat themselves too often. */
+#define MAX_TOK_SIZE 11
+#define KSYM_NAME_LEN 127
+
+/* we use only a subset of the complete symbol table to gather the token count,
+ * to speed up compression, at the expense of a little compression ratio */
+#define WORKING_SET 1024
+
+/* first find the best token only on the list of tokens that would profit more
+ * than GOOD_BAD_THRESHOLD. Only if this list is empty go to the "bad" list.
+ * Increasing this value will put less tokens on the "good" list, so the search
+ * is faster. However, if the good list runs out of tokens, we must painfully
+ * search the bad list. */
+#define GOOD_BAD_THRESHOLD 10
+
+/* token hash parameters */
+#define HASH_BITS 18
+#define HASH_TABLE_SIZE (1 << HASH_BITS)
+#define HASH_MASK (HASH_TABLE_SIZE - 1)
+#define HASH_BASE_OFFSET 2166136261U
+#define HASH_FOLD(a) ((a)&(HASH_MASK))
+
+/* flags to mark symbols */
+#define SYM_FLAG_VALID 1
+#define SYM_FLAG_SAMPLED 2
+
+struct sym_entry {
+ unsigned long long addr;
+ char type;
+ unsigned char flags;
+ unsigned char len;
+ unsigned char *sym;
+};
+
+
+static struct sym_entry *table;
+static int size, cnt;
+static unsigned long long _stext, _etext;
+static int all_symbols = 0;
+static char symbol_prefix_char = '\0';
+
+struct token {
+ unsigned char data[MAX_TOK_SIZE];
+ unsigned char len;
+ /* profit: the number of bytes that could be saved by inserting this
+ * token into the table */
+ int profit;
+ struct token *next; /* next token on the hash list */
+ struct token *right; /* next token on the good/bad list */
+ struct token *left; /* previous token on the good/bad list */
+ struct token *smaller; /* token that is less one letter than this one */
+ };
+
+struct token bad_head, good_head;
+struct token *hash_table[HASH_TABLE_SIZE];
+
+/* the table that holds the result of the compression */
+unsigned char best_table[256][MAX_TOK_SIZE+1];
+unsigned char best_table_len[256];
+
+
+static void
+usage(void)
+{
+ fprintf(stderr, "Usage: symbols [--all-symbols]
[--symbol-prefix=<prefix char>] < in.map > out.S\n");
+ exit(1);
+}
+
+/*
+ * This ignores the intensely annoying "mapping symbols" found
+ * in ARM ELF files: $a, $t and $d.
+ */
+static inline int
+is_arm_mapping_symbol(const char *str)
+{
+ return str[0] == '$' && strchr("atd", str[1])
+ && (str[2] == '\0' || str[2] == '.');
+}
+
+static int
+read_symbol(FILE *in, struct sym_entry *s)
+{
+ char str[500];
+ char *sym;
+ int rc;
+
+ rc = fscanf(in, "%llx %c %499s\n", &s->addr, &s->type, str);
+ if (rc != 3) {
+ if (rc != EOF) {
+ /* skip line */
+ fgets(str, 500, in);
+ }
+ return -1;
+ }
+
+ sym = str;
+ /* skip prefix char */
+ if (symbol_prefix_char && str[0] == symbol_prefix_char)
+ sym++;
+
+ /* Ignore most absolute/undefined (?) symbols. */
+ if (strcmp(sym, "_stext") == 0)
+ _stext = s->addr;
+ else if (strcmp(sym, "_etext") == 0)
+ _etext = s->addr;
+ else if (toupper(s->type) == '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(s->type) == 'U' ||
+ is_arm_mapping_symbol(sym))
+ return -1;
+
+ /* include the type field in the symbol name, so that it gets
+ * compressed together */
+ s->len = strlen(str) + 1;
+ s->sym = (char *) malloc(s->len + 1);
+ strcpy(s->sym + 1, str);
+ s->sym[0] = s->type;
+
+ return 0;
+}
+
+static int
+symbol_valid(struct sym_entry *s)
+{
+ /* Symbols which vary between passes. Passes 1 and 2 must have
+ * identical symbol lists. The symbols_* 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[] = {
+ "symbols_addresses",
+ "symbols_num_syms",
+ "symbols_names",
+ "symbols_markers",
+ "symbols_token_table",
+ "symbols_token_index",
+
+ /* Exclude linker generated symbols which vary between passes */
+ "_SDA_BASE_", /* ppc */
+ "_SDA2_BASE_", /* ppc */
+ NULL };
+ int i;
+ int offset = 1;
+
+ /* skip prefix char */
+ if (symbol_prefix_char && *(s->sym + 1) == symbol_prefix_char)
+ offset++;
+
+ /* if --all-symbols is not specified, then symbols outside the text
+ * and inittext sections are discarded */
+ if (!all_symbols) {
+ if (s->addr < _stext || s->addr > _etext)
+ return 0;
+ /* Corner case. Discard any symbols with the same value as
+ * _etext _einittext or _eextratext; they can move between pass
+ * 1 and 2 when the symbols data are added. If these symbols
+ * move then they may get dropped in pass 2, which breaks the
+ * symbols rules.
+ */
+ if (s->addr == _etext && strcmp(s->sym + offset, "_etext"))
+ return 0;
+ }
+
+ /* Exclude symbols which vary between passes. */
+ if (strstr(s->sym + offset, "_compiled."))
+ return 0;
+
+ for (i = 0; special_symbols[i]; i++)
+ if( strcmp(s->sym + offset, special_symbols[i]) == 0 )
+ return 0;
+
+ return 1;
+}
+
+static void
+read_map(FILE *in)
+{
+ while (!feof(in)) {
+ if (cnt >= size) {
+ size += 10000;
+ table = realloc(table, sizeof(*table) * size);
+ if (!table) {
+ fprintf(stderr, "out of memory\n");
+ exit (1);
+ }
+ }
+ if (read_symbol(in, &table[cnt]) == 0)
+ cnt++;
+ }
+}
+
+static void output_label(char *label)
+{
+ if (symbol_prefix_char)
+ printf(".globl %c%s\n", symbol_prefix_char, label);
+ else
+ printf(".globl %s\n", label);
+ printf("\tALGN\n");
+ if (symbol_prefix_char)
+ printf("%c%s:\n", symbol_prefix_char, label);
+ else
+ printf("%s:\n", 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)
+{
+ int c, rlen, total=0;
+
+ while (len) {
+ c = *data;
+ /* if the table holds a single char that is the same as the one
+ * we are looking for, then end the search */
+ if (best_table[c][0]==c && best_table_len[c]==1) {
+ *result++ = c;
+ total++;
+ } else {
+ /* if not, recurse and expand */
+ rlen = expand_symbol(best_table[c], best_table_len[c],
result);
+ total += rlen;
+ result += rlen;
+ }
+ data++;
+ len--;
+ }
+ *result=0;
+
+ return total;
+}
+
+static void
+write_src(void)
+{
+ int i, k, off, valid;
+ unsigned int best_idx[256];
+ unsigned int *markers;
+ char buf[KSYM_NAME_LEN+1];
+
+ printf("#include <asm/types.h>\n");
+ printf("#if BITS_PER_LONG == 64\n");
+ printf("#define PTR .quad\n");
+ printf("#define ALGN .align 8\n");
+ printf("#else\n");
+ printf("#define PTR .long\n");
+ printf("#define ALGN .align 4\n");
+ printf("#endif\n");
+
+ printf(".data\n");
+
+ output_label("symbols_addresses");
+ valid = 0;
+ for (i = 0; i < cnt; i++) {
+ if (table[i].flags & SYM_FLAG_VALID) {
+ printf("\tPTR\t%#llx\n", table[i].addr);
+ valid++;
+ }
+ }
+ printf("\n");
+
+ output_label("symbols_num_syms");
+ printf("\tPTR\t%d\n", valid);
+ printf("\n");
+
+ /* table of offset markers, that give the offset in the compressed
stream
+ * every 256 symbols */
+ markers = (unsigned int *) malloc(sizeof(unsigned int)*((valid + 255) /
256));
+
+ output_label("symbols_names");
+ valid = 0;
+ off = 0;
+ for (i = 0; i < cnt; i++) {
+
+ if (!table[i].flags & SYM_FLAG_VALID)
+ continue;
+
+ if ((valid & 0xFF) == 0)
+ markers[valid >> 8] = off;
+
+ printf("\t.byte 0x%02x", table[i].len);
+ for (k = 0; k < table[i].len; k++)
+ printf(", 0x%02x", table[i].sym[k]);
+ printf("\n");
+
+ off += table[i].len + 1;
+ valid++;
+ }
+ printf("\n");
+
+ output_label("symbols_markers");
+ for (i = 0; i < ((valid + 255) >> 8); i++)
+ printf("\tPTR\t%d\n", markers[i]);
+ printf("\n");
+
+ free(markers);
+
+ output_label("symbols_token_table");
+ off = 0;
+ for (i = 0; i < 256; i++) {
+ best_idx[i] = off;
+ expand_symbol(best_table[i],best_table_len[i],buf);
+ printf("\t.asciz\t\"%s\"\n", buf);
+ off += strlen(buf) + 1;
+ }
+ printf("\n");
+
+ output_label("symbols_token_index");
+ for (i = 0; i < 256; i++)
+ printf("\t.short\t%d\n", best_idx[i]);
+ printf("\n");
+}
+
+
+/* table lookup compression functions */
+
+static inline unsigned int rehash_token(unsigned int hash, unsigned char data)
+{
+ return ((hash * 16777619) ^ data);
+}
+
+static unsigned int hash_token(unsigned char *data, int len)
+{
+ unsigned int hash=HASH_BASE_OFFSET;
+ int i;
+
+ for (i = 0; i < len; i++)
+ hash = rehash_token(hash, data[i]);
+
+ return HASH_FOLD(hash);
+}
+
+/* find a token given its data and hash value */
+static struct token *find_token_hash(unsigned char *data, int len, unsigned
int hash)
+{
+ struct token *ptr;
+
+ ptr = hash_table[hash];
+
+ while (ptr) {
+ if ((ptr->len == len) && (memcmp(ptr->data, data, len) == 0))
+ return ptr;
+ ptr=ptr->next;
+ }
+
+ return NULL;
+}
+
+static inline void insert_token_in_group(struct token *head, struct token *ptr)
+{
+ ptr->right = head->right;
+ ptr->right->left = ptr;
+ head->right = ptr;
+ ptr->left = head;
+}
+
+static inline void remove_token_from_group(struct token *ptr)
+{
+ ptr->left->right = ptr->right;
+ ptr->right->left = ptr->left;
+}
+
+
+/* build the counts for all the tokens that start with "data", and have lenghts
+ * from 2 to "len" */
+static void learn_token(unsigned char *data, int len)
+{
+ struct token *ptr,*last_ptr;
+ int i, newprofit;
+ unsigned int hash = HASH_BASE_OFFSET;
+ unsigned int hashes[MAX_TOK_SIZE + 1];
+
+ if (len > MAX_TOK_SIZE)
+ len = MAX_TOK_SIZE;
+
+ /* calculate and store the hash values for all the sub-tokens */
+ hash = rehash_token(hash, data[0]);
+ for (i = 2; i <= len; i++) {
+ hash = rehash_token(hash, data[i-1]);
+ hashes[i] = HASH_FOLD(hash);
+ }
+
+ last_ptr = NULL;
+ ptr = NULL;
+
+ for (i = len; i >= 2; i--) {
+ hash = hashes[i];
+
+ if (!ptr) ptr = find_token_hash(data, i, hash);
+
+ if (!ptr) {
+ /* create a new token entry */
+ ptr = (struct token *) malloc(sizeof(*ptr));
+
+ memcpy(ptr->data, data, i);
+ ptr->len = i;
+
+ /* when we create an entry, it's profit is 0 because
+ * we also take into account the size of the token on
+ * the compressed table. We then subtract
GOOD_BAD_THRESHOLD
+ * so that the test to see if this token belongs to
+ * the good or bad list, is a comparison to zero */
+ ptr->profit = -GOOD_BAD_THRESHOLD;
+
+ ptr->next = hash_table[hash];
+ hash_table[hash] = ptr;
+
+ insert_token_in_group(&bad_head, ptr);
+
+ ptr->smaller = NULL;
+ } else {
+ newprofit = ptr->profit + (ptr->len - 1);
+ /* check to see if this token needs to be moved to a
+ * different list */
+ if((ptr->profit < 0) && (newprofit >= 0)) {
+ remove_token_from_group(ptr);
+ insert_token_in_group(&good_head,ptr);
+ }
+ ptr->profit = newprofit;
+ }
+
+ if (last_ptr) last_ptr->smaller = ptr;
+ last_ptr = ptr;
+
+ ptr = ptr->smaller;
+ }
+}
+
+/* decrease the counts for all the tokens that start with "data", and have
lenghts
+ * from 2 to "len". This function is much simpler than learn_token because we
have
+ * more guarantees (tho tokens exist, the ->smaller pointer is set, etc.)
+ * The two separate functions exist only because of compression performance */
+static void forget_token(unsigned char *data, int len)
+{
+ struct token *ptr;
+ int i, newprofit;
+ unsigned int hash=0;
+
+ if (len > MAX_TOK_SIZE) len = MAX_TOK_SIZE;
+
+ hash = hash_token(data, len);
+ ptr = find_token_hash(data, len, hash);
+
+ for (i = len; i >= 2; i--) {
+
+ newprofit = ptr->profit - (ptr->len - 1);
+ if ((ptr->profit >= 0) && (newprofit < 0)) {
+ remove_token_from_group(ptr);
+ insert_token_in_group(&bad_head, ptr);
+ }
+ ptr->profit=newprofit;
+
+ ptr=ptr->smaller;
+ }
+}
+
+/* count all the possible tokens in a symbol */
+static void learn_symbol(unsigned char *symbol, int len)
+{
+ int i;
+
+ for (i = 0; i < len - 1; i++)
+ learn_token(symbol + i, len - i);
+}
+
+/* decrease the count for all the possible tokens in a symbol */
+static void forget_symbol(unsigned char *symbol, int len)
+{
+ int i;
+
+ for (i = 0; i < len - 1; i++)
+ forget_token(symbol + i, len - i);
+}
+
+/* set all the symbol flags and do the initial token count */
+static void build_initial_tok_table(void)
+{
+ int i, use_it, valid;
+
+ valid = 0;
+ for (i = 0; i < cnt; i++) {
+ table[i].flags = 0;
+ if ( symbol_valid(&table[i]) ) {
+ table[i].flags |= SYM_FLAG_VALID;
+ valid++;
+ }
+ }
+
+ use_it = 0;
+ for (i = 0; i < cnt; i++) {
+
+ /* subsample the available symbols. This method is almost like
+ * a Bresenham's algorithm to get uniformly distributed samples
+ * across the symbol table */
+ if (table[i].flags & SYM_FLAG_VALID) {
+
+ use_it += WORKING_SET;
+
+ if (use_it >= valid) {
+ table[i].flags |= SYM_FLAG_SAMPLED;
+ use_it -= valid;
+ }
+ }
+ if (table[i].flags & SYM_FLAG_SAMPLED)
+ learn_symbol(table[i].sym, table[i].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 tlen, int idx)
+{
+ int i, len, learn, size;
+ unsigned char *p;
+
+ for (i = 0; i < cnt; i++) {
+
+ if (!(table[i].flags & SYM_FLAG_VALID)) continue;
+
+ len = table[i].len;
+ learn = 0;
+ p = table[i].sym;
+
+ do {
+ /* find the token on the symbol */
+ p = (unsigned char *) strstr((char *) p, (char *) str);
+ if (!p) break;
+
+ if (!learn) {
+ /* if this symbol was used to count, decrease
it */
+ if (table[i].flags & SYM_FLAG_SAMPLED)
+ forget_symbol(table[i].sym, len);
+ learn = 1;
+ }
+
+ *p = idx;
+ size = (len - (p - table[i].sym)) - tlen + 1;
+ memmove(p + 1, p + tlen, size);
+ p++;
+ len -= tlen - 1;
+
+ } while (size >= tlen);
+
+ if(learn) {
+ table[i].len = len;
+ /* if this symbol was used to count, learn it again */
+ if(table[i].flags & SYM_FLAG_SAMPLED)
+ learn_symbol(table[i].sym, len);
+ }
+ }
+}
+
+/* search the token with the maximum profit */
+static struct token *find_best_token(void)
+{
+ struct token *ptr,*best,*head;
+ int bestprofit;
+
+ bestprofit=-10000;
+
+ /* failsafe: if the "good" list is empty search from the "bad" list */
+ if(good_head.right == &good_head) head = &bad_head;
+ else head = &good_head;
+
+ ptr = head->right;
+ best = NULL;
+ while (ptr != head) {
+ if (ptr->profit > bestprofit) {
+ bestprofit = ptr->profit;
+ best = ptr;
+ }
+ ptr = ptr->right;
+ }
+
+ return best;
+}
+
+/* this is the core of the algorithm: calculate the "best" table */
+static void optimize_result(void)
+{
+ struct token *best;
+ int i;
+
+ /* using the '\0' symbol last allows compress_symbols to use standard
+ * fast string functions */
+ for (i = 255; i >= 0; i--) {
+
+ /* if this table slot is empty (it is not used by an actual
+ * original char code */
+ if (!best_table_len[i]) {
+
+ /* find the token with the breates profit value */
+ best = find_best_token();
+
+ /* place it in the "best" table */
+ best_table_len[i] = best->len;
+ memcpy(best_table[i], best->data, best_table_len[i]);
+ /* zero terminate the token so that we can use strstr
+ in compress_symbols */
+ best_table[i][best_table_len[i]]='\0';
+
+ /* replace this token in all the valid symbols */
+ compress_symbols(best_table[i], best_table_len[i], i);
+ }
+ }
+}
+
+/* start by placing the symbols that are actually used on the table */
+static void insert_real_symbols_in_table(void)
+{
+ int i, j, c;
+
+ memset(best_table, 0, sizeof(best_table));
+ memset(best_table_len, 0, sizeof(best_table_len));
+
+ for (i = 0; i < cnt; i++) {
+ if (table[i].flags & SYM_FLAG_VALID) {
+ for (j = 0; j < table[i].len; j++) {
+ c = table[i].sym[j];
+ best_table[c][0]=c;
+ best_table_len[c]=1;
+ }
+ }
+ }
+}
+
+static void optimize_token_table(void)
+{
+ memset(hash_table, 0, sizeof(hash_table));
+
+ good_head.left = &good_head;
+ good_head.right = &good_head;
+
+ bad_head.left = &bad_head;
+ bad_head.right = &bad_head;
+
+ build_initial_tok_table();
+
+ insert_real_symbols_in_table();
+
+ /* When valid symbol is not registered, exit to error */
+ if (good_head.left == good_head.right &&
+ bad_head.left == bad_head.right) {
+ fprintf(stderr, "No valid symbol.\n");
+ exit(1);
+ }
+
+ optimize_result();
+}
+
+
+int
+main(int argc, char **argv)
+{
+ if (argc >= 2) {
+ int i;
+ for (i = 1; i < argc; i++) {
+ if(strcmp(argv[i], "--all-symbols") == 0)
+ all_symbols = 1;
+ else if (strncmp(argv[i], "--symbol-prefix=", 16) == 0)
{
+ char *p = &argv[i][16];
+ /* skip quote */
+ if ((*p == '"' && *(p+2) == '"') || (*p == '\''
&& *(p+2) == '\''))
+ p++;
+ symbol_prefix_char = *p;
+ } else
+ usage();
+ }
+ } else if (argc != 1)
+ usage();
+
+ read_map(stdin);
+ optimize_token_table();
+ write_src();
+
+ return 0;
+}
diff -r 7627476544b5 -r 48aed1403fe3 xen/common/symbols.c
--- /dev/null Fri Jul 22 14:25:10 2005
+++ b/xen/common/symbols.c Fri Jul 22 16:44:33 2005
@@ -0,0 +1,158 @@
+/*
+ * symbols.c: in-kernel printing of symbolic oopses and stack traces.
+ *
+ * Copyright 2002 Rusty Russell <rusty@xxxxxxxxxxxxxxx> IBM Corporation
+ *
+ * ChangeLog:
+ *
+ * (25/Aug/2004) Paulo Marques <pmarques@xxxxxxxxxxxx>
+ * Changed the compression method from stem compression to "table lookup"
+ * compression (see tools/symbols.c for a more complete description)
+ */
+
+#include <xen/config.h>
+#include <xen/symbols.h>
+#include <xen/init.h>
+#include <xen/lib.h>
+#include <xen/string.h>
+
+/* These will be re-linked against their real values during the second link
stage */
+extern unsigned long symbols_addresses[] __attribute__((weak));
+extern unsigned long symbols_num_syms __attribute__((weak,section("data")));
+extern u8 symbols_names[] __attribute__((weak));
+
+extern u8 symbols_token_table[] __attribute__((weak));
+extern u16 symbols_token_index[] __attribute__((weak));
+
+extern unsigned long symbols_markers[] __attribute__((weak));
+
+/* expand a compressed symbol data into the resulting uncompressed string,
+ given the offset to where the symbol is in the compressed stream */
+static unsigned int symbols_expand_symbol(unsigned int off, char *result)
+{
+ int len, skipped_first = 0;
+ u8 *tptr, *data;
+
+ /* get the compressed symbol length from the first symbol byte */
+ data = &symbols_names[off];
+ len = *data;
+ data++;
+
+ /* update the offset to return the offset for the next symbol on
+ * the compressed stream */
+ off += len + 1;
+
+ /* for every byte on the compressed symbol data, copy the table
+ entry for that byte */
+ while(len) {
+ tptr = &symbols_token_table[ symbols_token_index[*data] ];
+ data++;
+ len--;
+
+ while (*tptr) {
+ if(skipped_first) {
+ *result = *tptr;
+ result++;
+ } else
+ skipped_first = 1;
+ tptr++;
+ }
+ }
+
+ *result = '\0';
+
+ /* return to offset to the next symbol */
+ return off;
+}
+
+/* find the offset on the compressed stream given and index in the
+ * symbols array */
+static unsigned int get_symbol_offset(unsigned long pos)
+{
+ u8 *name;
+ int i;
+
+ /* use the closest marker we have. We have markers every 256 positions,
+ * so that should be close enough */
+ name = &symbols_names[ symbols_markers[pos>>8] ];
+
+ /* sequentially scan all the symbols up to the point we're searching for.
+ * Every symbol is stored in a [<len>][<len> bytes of data] format, so we
+ * just need to add the len to the current pointer for every symbol we
+ * wish to skip */
+ for(i = 0; i < (pos&0xFF); i++)
+ name = name + (*name) + 1;
+
+ return name - symbols_names;
+}
+
+const char *symbols_lookup(unsigned long addr,
+ unsigned long *symbolsize,
+ unsigned long *offset,
+ char *namebuf)
+{
+ unsigned long i, low, high, mid;
+ unsigned long symbol_end = 0;
+
+ /* This kernel should never had been booted. */
+ BUG_ON(!symbols_addresses);
+
+ namebuf[KSYM_NAME_LEN] = 0;
+ namebuf[0] = 0;
+
+ if (!is_kernel_text(addr))
+ return NULL;
+
+ /* do a binary search on the sorted symbols_addresses array */
+ low = 0;
+ high = symbols_num_syms;
+
+ while (high-low > 1) {
+ mid = (low + high) / 2;
+ if (symbols_addresses[mid] <= addr) low = mid;
+ else high = mid;
+ }
+
+ /* search for the first aliased symbol. Aliased symbols are
+ symbols with the same address */
+ while (low && symbols_addresses[low - 1] == symbols_addresses[low])
+ --low;
+
+ /* Grab name */
+ symbols_expand_symbol(get_symbol_offset(low), namebuf);
+
+ /* Search for next non-aliased symbol */
+ for (i = low + 1; i < symbols_num_syms; i++) {
+ if (symbols_addresses[i] > symbols_addresses[low]) {
+ symbol_end = symbols_addresses[i];
+ break;
+ }
+ }
+
+ /* if we found no next symbol, we use the end of the section */
+ if (!symbol_end)
+ symbol_end = kernel_text_end();
+
+ *symbolsize = symbol_end - symbols_addresses[low];
+ *offset = addr - symbols_addresses[low];
+ return namebuf;
+}
+
+/* Replace "%s" in format with address, or returns -errno. */
+void __print_symbol(const char *fmt, unsigned long address)
+{
+ const char *name;
+ unsigned long offset, size;
+ char namebuf[KSYM_NAME_LEN+1];
+ char buffer[sizeof("%s+%#lx/%#lx [%s]") + KSYM_NAME_LEN +
+ 2*(BITS_PER_LONG*3/10) + 1];
+
+ name = symbols_lookup(address, &size, &offset, namebuf);
+
+ if (!name)
+ sprintf(buffer, "???");
+ else
+ sprintf(buffer, "%s+%#lx/%#lx", name, offset, size);
+
+ printk(fmt, buffer);
+}
_______________________________________________
Xen-changelog mailing list
Xen-changelog@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-changelog
|
Home