[build] Warn on zero-length symbols when building symbol table

Make build_symbol_table.py output statistics on the symbol table it
builds, and emit warnings for zero-length symbols. Also added lengths to
several functions defined in asm that this uncovered.

scripts/build_symbol_table.py

@@ -17,7 +17,7 @@
 
 import re
 
-sym_re = re.compile(r'([0-9a-fA-F]{16}) ([0-9a-fA-F]{16}) [tTvVwW] (.*)')
+sym_re = re.compile(r'([0-9a-fA-F]{16}) ([0-9a-fA-F]{16} )?[tTvVwW] (.*)')
 
 def parse_syms(infile):
     """Take the output of the `nm` command, and parse it into a tuple
@@ -30,8 +30,14 @@ def parse_syms(infile):
         if not match: continue
 
         addr = int(match.group(1), base=16)
-        size = int(match.group(2), base=16)
+        size = int(match.group(2) or "0", base=16)
         name = match.group(3)
+
+        if size == 0:
+            if not "." in name:
+                print(f"SYMBOL WARNING: zero size for symbol {name}")
+            continue
+
         syms.append([addr, size, name, 0])
 
     return syms
@@ -61,6 +67,8 @@ def write_table(syms, outfile):
         pos = s[3]
         outfile.write(struct.pack("@QQQ", addr, size, pos))
 
+    return len(syms)
+
 
 if __name__ == "__main__":
     import sys
@@ -68,5 +76,11 @@ if __name__ == "__main__":
         print(f"Usage: nm -n -S --demangle | {sys.argv[0]} <output>")
         sys.exit(1)
 
-    outfile = open(sys.argv[1], "wb")
-    write_table(parse_syms(sys.stdin), outfile)
+    syms = 0
+    size = 0
+    with open(sys.argv[1], "wb") as outfile:
+        syms = write_table(parse_syms(sys.stdin), outfile)
+        outfile.seek(0, 2)
+        size = outfile.tell()
+
+    print(f"Wrote {syms} symbols ({size/1024:.1f} KiB) to {sys.argv[1]}.")

src/kernel/boot.s

@@ -42,15 +42,17 @@ bsp_idle:
 	jmp bsp_idle
 .end:
 
-global interrupts_enable: function hidden
+global interrupts_enable: function hidden (interrupts_enable.end - interrupts_enable)
 interrupts_enable:
 	sti
 	ret
+.end:
 
-global interrupts_disable: function hidden
+global interrupts_disable: function hidden (interrupts_disable.end - interrupts_disable)
 interrupts_disable:
 	cli
 	ret
+.end:
 
 section .bss
 align 0x100

src/kernel/debug.s

@@ -1,33 +1,39 @@
-global get_rsp: function hidden
+global get_rsp: function hidden (get_rsp.end - get_rsp)
 get_rsp:
 	mov rax, rsp
 	ret
+.end:
 
-global get_rip: function hidden
+global get_rip: function hidden (get_rip.end - get_rip)
 get_rip:
 	pop rax ; do the same thing as 'ret', except with 'jmp'
 	jmp rax ; with the return address still in rax
+.end:
 
-global get_caller: function hidden
+global get_caller: function hidden (get_caller.end - get_caller)
 get_caller:
 	; No prelude - don't touch rsp or rbp
 	mov rax, [rbp+8]
 	ret
+.end:
 
-global get_grandcaller: function hidden
+global get_grandcaller: function hidden (get_grandcaller.end - get_grandcaller)
 get_grandcaller:
 	; No prelude - don't touch rsp or rbp
 	mov rax, [rbp]
 	mov rax, [rax+8]
 	ret
+.end:
 
-global get_gsbase: function hidden
+global get_gsbase: function hidden (get_gsbase.end - get_gsbase)
 get_gsbase:
 	rdgsbase rax
 	ret
+.end:
 
-global _halt: function hidden
+global _halt: function hidden (_halt.end - _halt)
 _halt:
 	hlt
 	jmp _halt
+.end:
 

src/kernel/gdtidt.s

@@ -1,15 +1,17 @@
 
-global idt_write: function hidden
+global idt_write: function hidden (idt_write.end - idt_write)
 idt_write:
 	lidt [rdi] ; first arg is the IDT pointer location
 	ret
+.end:
 
-global idt_load: function hidden
+global idt_load: function hidden (idt_load.end - idt_load)
 idt_load:
 	sidt [rdi] ; first arg is where to write the idtr value
 	ret
+.end:
 
-global gdt_write: function hidden
+global gdt_write: function hidden (gdt_write.end - gdt_write)
 gdt_write:
 	lgdt [rdi] ; first arg is the GDT pointer location
 
@@ -27,9 +29,11 @@ gdt_write:
 .next:
 	ltr cx ; fourth arg is the TSS
 	ret
+.end:
 
-global gdt_load: function hidden
+global gdt_load: function hidden (gdt_load.end - gdt_load)
 gdt_load:
 	sgdt [rdi] ; first arg is where to write the gdtr value
 	ret
+.end:
 

src/kernel/smp.s

@@ -26,7 +26,7 @@ bits 16
 default rel
 align 8
 
-global ap_startup: function hidden
+global ap_startup: function hidden (ap_startup.end - ap_startup)
 ap_startup:
 	jmp .start_real
 
@@ -102,15 +102,16 @@ align 8
 
 	mov rax, [BASE + (.ret - ap_startup)]
 	jmp rax
+.end:
 
 
-global ap_startup_code_size: function hidden
+global ap_startup_code_size: data hidden
 ap_startup_code_size:
 	dq ($ - ap_startup)
 
 
 section .text
-global init_ap_trampoline: function hidden
+global init_ap_trampoline: function hidden (init_ap_trampoline.end - init_ap_trampoline)
 init_ap_trampoline:
 	push rbp
 	mov rbp, rsp
@@ -129,6 +130,7 @@ init_ap_trampoline:
 
 	pop rbp
 	ret
+.end:
 
 extern long_ap_startup
 global ap_idle:function hidden (ap_idle.end - ap_idle)

src/kernel/task.s

@@ -1,6 +1,6 @@
 %include "tasking.inc"
 
-global task_switch: function hidden
+global task_switch: function hidden (task_switch.end - task_switch)
 task_switch:
 	push rbp
 	mov rbp, rsp
@@ -60,8 +60,10 @@ task_switch:
 
 	pop rbp
 	ret
+.end:
 
-global _current_gsbase: function hidden
+global _current_gsbase: function hidden (_current_gsbase.end - _current_gsbase)
 _current_gsbase:
 	mov rax, [gs:CPU_DATA.self]
 	ret
+.end: