mu - Soul of a tiny new machine. More thorough tests → More comprehensible and rewrite-friendly software → More resilient society.

	Commit message (Expand)	Author	Age	Files	Lines
*	3431	Kartik K. Agaram	2016-09-30	1	-28/+30
*	3395	Kartik K. Agaram	2016-09-17	1	-50/+50
*	3355	Kartik K. Agaram	2016-09-15	1	-99/+215
*	3315	Kartik K. Agaram	2016-09-10	1	-3/+8
*	3266	Kartik K. Agaram	2016-08-27	1	-19/+125
*	3219	Kartik K. Agaram	2016-08-17	1	-0/+46
*	3158	Kartik K. Agaram	2016-07-27	1	-0/+367

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01//EN" "http://www.w3.org/TR/html4/strict.dtd"> <html> <head> <meta http-equiv="content-type" content="text/html; charset=UTF-8"> <title>Mu - 302stack_allocate.subx</title> <meta name="Generator" content="Vim/8.1"> <meta name="plugin-version" content="vim8.1_v1"> <meta name="syntax" content="none"> <meta name="settings" content="number_lines,use_css,pre_wrap,no_foldcolumn,expand_tabs,line_ids,prevent_copy="> <meta name="colorscheme" content="minimal-light"> <style type="text/css">  </style> <script type='text/javascript'>  </script> </head> <body onload='JumpToLine();'> <a href='https://github.com/akkartik/mu/blob/main/302stack_allocate.subx'>https://github.com/akkartik/mu/blob/main/302stack_allocate.subx</a> <pre id='vimCodeElement'> 1 # A function which pushes n zeros on the stack. 2 # Really only intended to be called from code generated by mu.subx (for array 3 # vars on the stack). 4 5 == code 6 7 #? Entry: 8 #? # . prologue 9 #? 89/<- %ebp 4/r32/esp 10 #? # 11 #? 68/push 0xfcfdfeff/imm32 12 #? b8/copy-to-eax 0x34353637/imm32 13 #? $dump-stack0: 14 #? (push-n-zero-bytes 4) 15 #? 68/push 0x20/imm32 16 #? $dump-stack9: 17 #? b8/copy-to-eax 1/imm32/exit 18 #? cd/syscall 0x80/imm8 19 20 # This is not a regular function, so it won't be idiomatic. 21 # Registers must be properly restored. 22 # Registers can be spilled, but that modifies the stack and needs to be 23 # cleaned up. 24 25 # Overhead: 26 # 62 + n*6 instructions to push n bytes. 27 # If we just emitted code to push n zeroes, it would be: 28 # 5 bytes for 4 zero bytes, so 1.25 bytes per zero. And that's not even 29 # instructions. 30 # But on the other hand it would destroy the instruction cache, where this 31 # approach requires 15 instructions, fixed. 32 33 # n must be positive 34 push-n-zero-bytes: # n: int 35 $push-n-zero-bytes:prologue: 36 89/<- *<a href='302stack_allocate.subx.html#L56'>Push-n-zero-bytes-ebp</a> 5/r32/ebp # spill ebp without affecting stack 37 89/<- %ebp 4/r32/esp 38 $push-n-zero-bytes:copy-ra: 39 # -- esp = ebp 40 89/<- *<a href='302stack_allocate.subx.html#L60'>Push-n-zero-bytes-eax</a> 0/r32/eax 41 8b/-> *esp 0/r32/eax 42 2b/subtract *(ebp+4) 4/r32/esp 43 # -- esp+n = ebp 44 89/<- *esp 0/r32/eax 45 8b/-> *<a href='302stack_allocate.subx.html#L60'>Push-n-zero-bytes-eax</a> 0/r32/eax 46 $push-n-zero-bytes:bulk-cleaning: 47 89/<- *<a href='302stack_allocate.subx.html#L58'>Push-n-zero-bytes-esp</a> 4/r32/esp 48 81 0/subop/add *<a href='302stack_allocate.subx.html#L58'>Push-n-zero-bytes-esp</a> 4/imm32 49 81 0/subop/add *(ebp+4) 4/imm32 50 (<a href='120allocate.subx.html#L869'>zero-out</a> *<a href='302stack_allocate.subx.html#L58'>Push-n-zero-bytes-esp</a> *(ebp+4)) # n+4 51 $push-n-zero-bytes:epilogue: 52 8b/-> *<a href='302stack_allocate.subx.html#L56'>Push-n-zero-bytes-ebp</a> 5/r32/ebp # restore spill 53 c3/return 54 55 == data 56 Push-n-zero-bytes-ebp: # (addr int) 57 0/imm32 58 Push-n-zero-bytes-esp: # (addr int) 59 0/imm32 60 Push-n-zero-bytes-eax: 61 0/imm32 </pre> </body> </html>