[scripts] Add symbol table building script

Create a script that builds a simple-to-read symbol table from the
output of `nm`. Include running that script over the kernel in the
build, and including that output in the initrd.

Tags: callstack debugging

assets/initrd.toml

@@ -12,6 +12,10 @@
 dest = "screenfont.psf"
 source = "../assets/fonts/tamsyn8x16r.psf"
 
+[[files]]
+dest = "symbol_table.dat"
+source = "symbol_table.dat"
+
 [[files]]
 dest = "nulldrv1"
 source = "user/nulldrv"

scripts/build_symbol_table.py

@@ -0,0 +1,71 @@
+#!/usr/bin/env python3
+#
+# Generate the jsix style symbol table. The format in memory of this table
+# is as follows:
+#
+# <num_entires> : 8 bytes
+# <index>       : 16 * N bytes
+# <name data>   : variable
+#
+# Each index entry has the format
+# <symbol address> : 8 bytes
+# <offset of name> : 8 bytes
+#
+# Name offsets are from the start of the symbol table as a whole. (ie,
+# where <num_entries> is located.)
+
+def parse_syms(infile):
+    """Take the output of the `nm` command, and parse it into a tuple
+    representing the symbols in the text segment of the binary. Returns
+    a list of (address, symbol_name)."""
+
+    from cxxfilt import demangle
+
+    syms = []
+    for line in sys.stdin:
+        addr, t, mangled = line.split()
+        if t not in "tTvVwW": continue
+
+        addr = int(addr, base=16)
+        name = demangle(mangled)
+        syms.append((addr, name))
+
+    return sorted(syms)
+
+
+def write_table(syms, outfile):
+    """Write the given symbol table as generated by parse_syms()
+    to the outfile, index first, and then name character data."""
+
+    import struct
+
+    outfile.write(struct.pack("@Q", len(syms)))
+    index_pos = outfile.tell()
+
+    outfile.seek(struct.calcsize("@QQ") * len(syms), 1)
+    nul = b'\0'
+
+    positions = {}
+    for s in syms:
+        addr, name = s
+        positions[addr] = outfile.tell()
+
+        data = name.encode('utf-8')
+        outfile.write(name.encode('utf-8'))
+        outfile.write(nul)
+
+    outfile.seek(index_pos)
+    for s in syms:
+        addr = s[0]
+        pos = positions[addr]
+        outfile.write(struct.pack("@QQ", addr, pos))
+
+
+if __name__ == "__main__":
+    import sys
+    if len(sys.argv) != 2:
+        print(f"Usage: {sys.argv[0]} <output>")
+        sys.exit(1)
+
+    outfile = open(sys.argv[1], "wb")
+    write_table(parse_syms(sys.stdin), outfile)

scripts/templates/build.ninja.j2

@@ -122,6 +122,10 @@ rule makerd
     description = Making init ramdisk
     command = $builddir/native/makerd $in $out
 
+rule makest
+    description = Making symbol table
+    command = nm $in | ${srcroot}/scripts/build_symbol_table.py $out
+
 rule makeefi
     description = Converting $name
     command = objcopy $
@@ -186,9 +190,13 @@ build $builddir/fatroot/jsix.elf : cp $builddir/jsix.elf
 build $builddir/fatroot/efi/boot/bootx64.efi : cp $builddir/boot/boot.efi
     name = bootloader to FAT image
 
+build ${builddir}/symbol_table.dat : makest ${builddir}/jsix.elf | $
+    ${builddir}/native/makest
+
 build $builddir/fatroot/initrd.img : makerd ${srcroot}/assets/initrd.toml | $
     ${builddir}/native/makerd $
-    ${builddir}/user/nulldrv
+    ${builddir}/user/nulldrv $
+    ${builddir}/symbol_table.dat
 
 build $builddir/jsix.img : makefat | $
     $builddir/fatroot/initrd.img $