I'm working on a simple operating system, and I've run into some issues with flickering in the output (see this clip). I'm pretty at a loss as to what could be causing this apart from perhaps an issue with memory.
kernel.cpp
typedef unsigned char uint8_t;
typedef unsigned short uint16_t;
#define FPS 30
#define PIT_HERTZ 1193131.666
#define CLOCK_HIT (int)(PIT_HERTZ/FPS)
const unsigned char font[128-32][8] = {
{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}, // U+0020 (space)
{ 0x18, 0x3C, 0x3C, 0x18, 0x18, 0x00, 0x18, 0x00}, // U+0021 (!)
{ 0x36, 0x36, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}, // U+0022 (")
{ 0x36, 0x36, 0x7F, 0x36, 0x7F, 0x36, 0x36, 0x00}, // U+0023 (#)
{ 0x0C, 0x3E, 0x03, 0x1E, 0x30, 0x1F, 0x0C, 0x00}, // U+0024 ($)
{ 0x00, 0x63, 0x33, 0x18, 0x0C, 0x66, 0x63, 0x00}, // U+0025 (%)
{ 0x1C, 0x36, 0x1C, 0x6E, 0x3B, 0x33, 0x6E, 0x00}, // U+0026 (&)
{ 0x06, 0x06, 0x03, 0x00, 0x00, 0x00, 0x00, 0x00}, // U+0027 (')
{ 0x18, 0x0C, 0x06, 0x06, 0x06, 0x0C, 0x18, 0x00}, // U+0028 (()
{ 0x06, 0x0C, 0x18, 0x18, 0x18, 0x0C, 0x06, 0x00}, // U+0029 ())
{ 0x00, 0x66, 0x3C, 0xFF, 0x3C, 0x66, 0x00, 0x00}, // U+002A (*)
{ 0x00, 0x0C, 0x0C, 0x3F, 0x0C, 0x0C, 0x00, 0x00}, // U+002B (+)
{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x0C, 0x0C, 0x06}, // U+002C (,)
{ 0x00, 0x00, 0x00, 0x3F, 0x00, 0x00, 0x00, 0x00}, // U+002D (-)
{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x0C, 0x0C, 0x00}, // U+002E (.)
{ 0x60, 0x30, 0x18, 0x0C, 0x06, 0x03, 0x01, 0x00}, // U+002F (/)
{ 0x3E, 0x63, 0x73, 0x7B, 0x6F, 0x67, 0x3E, 0x00}, // U+0030 (0)
{ 0x0C, 0x0E, 0x0C, 0x0C, 0x0C, 0x0C, 0x3F, 0x00}, // U+0031 (1)
{ 0x1E, 0x33, 0x30, 0x1C, 0x06, 0x33, 0x3F, 0x00}, // U+0032 (2)
{ 0x1E, 0x33, 0x30, 0x1C, 0x30, 0x33, 0x1E, 0x00}, // U+0033 (3)
{ 0x38, 0x3C, 0x36, 0x33, 0x7F, 0x30, 0x78, 0x00}, // U+0034 (4)
{ 0x3F, 0x03, 0x1F, 0x30, 0x30, 0x33, 0x1E, 0x00}, // U+0035 (5)
{ 0x1C, 0x06, 0x03, 0x1F, 0x33, 0x33, 0x1E, 0x00}, // U+0036 (6)
{ 0x3F, 0x33, 0x30, 0x18, 0x0C, 0x0C, 0x0C, 0x00}, // U+0037 (7)
{ 0x1E, 0x33, 0x33, 0x1E, 0x33, 0x33, 0x1E, 0x00}, // U+0038 (8)
{ 0x1E, 0x33, 0x33, 0x3E, 0x30, 0x18, 0x0E, 0x00}, // U+0039 (9)
{ 0x00, 0x0C, 0x0C, 0x00, 0x00, 0x0C, 0x0C, 0x00}, // U+003A (:)
{ 0x00, 0x0C, 0x0C, 0x00, 0x00, 0x0C, 0x0C, 0x06}, // U+003B (;)
{ 0x18, 0x0C, 0x06, 0x03, 0x06, 0x0C, 0x18, 0x00}, // U+003C (<)
{ 0x00, 0x00, 0x3F, 0x00, 0x00, 0x3F, 0x00, 0x00}, // U+003D (=)
{ 0x06, 0x0C, 0x18, 0x30, 0x18, 0x0C, 0x06, 0x00}, // U+003E (>)
{ 0x1E, 0x33, 0x30, 0x18, 0x0C, 0x00, 0x0C, 0x00}, // U+003F (?)
{ 0x3E, 0x63, 0x7B, 0x7B, 0x7B, 0x03, 0x1E, 0x00}, // U+0040 (@)
{ 0x0C, 0x1E, 0x33, 0x33, 0x3F, 0x33, 0x33, 0x00}, // U+0041 (A)
{ 0x3F, 0x66, 0x66, 0x3E, 0x66, 0x66, 0x3F, 0x00}, // U+0042 (B)
{ 0x3C, 0x66, 0x03, 0x03, 0x03, 0x66, 0x3C, 0x00}, // U+0043 (C)
{ 0x1F, 0x36, 0x66, 0x66, 0x66, 0x36, 0x1F, 0x00}, // U+0044 (D)
{ 0x7F, 0x46, 0x16, 0x1E, 0x16, 0x46, 0x7F, 0x00}, // U+0045 (E)
{ 0x7F, 0x46, 0x16, 0x1E, 0x16, 0x06, 0x0F, 0x00}, // U+0046 (F)
{ 0x3C, 0x66, 0x03, 0x03, 0x73, 0x66, 0x7C, 0x00}, // U+0047 (G)
{ 0x33, 0x33, 0x33, 0x3F, 0x33, 0x33, 0x33, 0x00}, // U+0048 (H)
{ 0x1E, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x1E, 0x00}, // U+0049 (I)
{ 0x78, 0x30, 0x30, 0x30, 0x33, 0x33, 0x1E, 0x00}, // U+004A (J)
{ 0x67, 0x66, 0x36, 0x1E, 0x36, 0x66, 0x67, 0x00}, // U+004B (K)
{ 0x0F, 0x06, 0x06, 0x06, 0x46, 0x66, 0x7F, 0x00}, // U+004C (L)
{ 0x63, 0x77, 0x7F, 0x7F, 0x6B, 0x63, 0x63, 0x00}, // U+004D (M)
{ 0x63, 0x67, 0x6F, 0x7B, 0x73, 0x63, 0x63, 0x00}, // U+004E (N)
{ 0x1C, 0x36, 0x63, 0x63, 0x63, 0x36, 0x1C, 0x00}, // U+004F (O)
{ 0x3F, 0x66, 0x66, 0x3E, 0x06, 0x06, 0x0F, 0x00}, // U+0050 (P)
{ 0x1E, 0x33, 0x33, 0x33, 0x3B, 0x1E, 0x38, 0x00}, // U+0051 (Q)
{ 0x3F, 0x66, 0x66, 0x3E, 0x36, 0x66, 0x67, 0x00}, // U+0052 (R)
{ 0x1E, 0x33, 0x07, 0x0E, 0x38, 0x33, 0x1E, 0x00}, // U+0053 (S)
{ 0x3F, 0x2D, 0x0C, 0x0C, 0x0C, 0x0C, 0x1E, 0x00}, // U+0054 (T)
{ 0x33, 0x33, 0x33, 0x33, 0x33, 0x33, 0x3F, 0x00}, // U+0055 (U)
{ 0x33, 0x33, 0x33, 0x33, 0x33, 0x1E, 0x0C, 0x00}, // U+0056 (V)
{ 0x63, 0x63, 0x63, 0x6B, 0x7F, 0x77, 0x63, 0x00}, // U+0057 (W)
{ 0x63, 0x63, 0x36, 0x1C, 0x1C, 0x36, 0x63, 0x00}, // U+0058 (X)
{ 0x33, 0x33, 0x33, 0x1E, 0x0C, 0x0C, 0x1E, 0x00}, // U+0059 (Y)
{ 0x7F, 0x63, 0x31, 0x18, 0x4C, 0x66, 0x7F, 0x00}, // U+005A (Z)
{ 0x1E, 0x06, 0x06, 0x06, 0x06, 0x06, 0x1E, 0x00}, // U+005B ([)
{ 0x03, 0x06, 0x0C, 0x18, 0x30, 0x60, 0x40, 0x00}, // U+005C (\)
{ 0x1E, 0x18, 0x18, 0x18, 0x18, 0x18, 0x1E, 0x00}, // U+005D (])
{ 0x08, 0x1C, 0x36, 0x63, 0x00, 0x00, 0x00, 0x00}, // U+005E (^)
{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xFF}, // U+005F (_)
{ 0x0C, 0x0C, 0x18, 0x00, 0x00, 0x00, 0x00, 0x00}, // U+0060 (`)
{ 0x00, 0x00, 0x1E, 0x30, 0x3E, 0x33, 0x6E, 0x00}, // U+0061 (a)
{ 0x07, 0x06, 0x06, 0x3E, 0x66, 0x66, 0x3B, 0x00}, // U+0062 (b)
{ 0x00, 0x00, 0x1E, 0x33, 0x03, 0x33, 0x1E, 0x00}, // U+0063 (c)
{ 0x38, 0x30, 0x30, 0x3e, 0x33, 0x33, 0x6E, 0x00}, // U+0064 (d)
{ 0x00, 0x00, 0x1E, 0x33, 0x3f, 0x03, 0x1E, 0x00}, // U+0065 (e)
{ 0x1C, 0x36, 0x06, 0x0f, 0x06, 0x06, 0x0F, 0x00}, // U+0066 (f)
{ 0x00, 0x00, 0x6E, 0x33, 0x33, 0x3E, 0x30, 0x1F}, // U+0067 (g)
{ 0x07, 0x06, 0x36, 0x6E, 0x66, 0x66, 0x67, 0x00}, // U+0068 (h)
{ 0x0C, 0x00, 0x0E, 0x0C, 0x0C, 0x0C, 0x1E, 0x00}, // U+0069 (i)
{ 0x30, 0x00, 0x30, 0x30, 0x30, 0x33, 0x33, 0x1E}, // U+006A (j)
{ 0x07, 0x06, 0x66, 0x36, 0x1E, 0x36, 0x67, 0x00}, // U+006B (k)
{ 0x0E, 0x0C, 0x0C, 0x0C, 0x0C, 0x0C, 0x1E, 0x00}, // U+006C (l)
{ 0x00, 0x00, 0x33, 0x7F, 0x7F, 0x6B, 0x63, 0x00}, // U+006D (m)
{ 0x00, 0x00, 0x1F, 0x33, 0x33, 0x33, 0x33, 0x00}, // U+006E (n)
{ 0x00, 0x00, 0x1E, 0x33, 0x33, 0x33, 0x1E, 0x00}, // U+006F (o)
{ 0x00, 0x00, 0x3B, 0x66, 0x66, 0x3E, 0x06, 0x0F}, // U+0070 (p)
{ 0x00, 0x00, 0x6E, 0x33, 0x33, 0x3E, 0x30, 0x78}, // U+0071 (q)
{ 0x00, 0x00, 0x3B, 0x6E, 0x66, 0x06, 0x0F, 0x00}, // U+0072 (r)
{ 0x00, 0x00, 0x3E, 0x03, 0x1E, 0x30, 0x1F, 0x00}, // U+0073 (s)
{ 0x08, 0x0C, 0x3E, 0x0C, 0x0C, 0x2C, 0x18, 0x00}, // U+0074 (t)
{ 0x00, 0x00, 0x33, 0x33, 0x33, 0x33, 0x6E, 0x00}, // U+0075 (u)
{ 0x00, 0x00, 0x33, 0x33, 0x33, 0x1E, 0x0C, 0x00}, // U+0076 (v)
{ 0x00, 0x00, 0x63, 0x6B, 0x7F, 0x7F, 0x36, 0x00}, // U+0077 (w)
{ 0x00, 0x00, 0x63, 0x36, 0x1C, 0x36, 0x63, 0x00}, // U+0078 (x)
{ 0x00, 0x00, 0x33, 0x33, 0x33, 0x3E, 0x30, 0x1F}, // U+0079 (y)
{ 0x00, 0x00, 0x3F, 0x19, 0x0C, 0x26, 0x3F, 0x00}, // U+007A (z)
{ 0x38, 0x0C, 0x0C, 0x07, 0x0C, 0x0C, 0x38, 0x00}, // U+007B ({)
{ 0x18, 0x18, 0x18, 0x00, 0x18, 0x18, 0x18, 0x00}, // U+007C (|)
{ 0x07, 0x0C, 0x0C, 0x38, 0x0C, 0x0C, 0x07, 0x00}, // U+007D (})
{ 0x6E, 0x3B, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00}, // U+007E (~)
{ 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00} // U+007F
};
static inline void outb(uint16_t port, uint8_t val)
{
asm volatile ( "outb %0, %1" : : "a"(val), "Nd"(port) );
}
static inline uint8_t inb(uint16_t port)
{
uint8_t ret;
asm volatile ( "inb %1, %0"
: "=a"(ret)
: "Nd"(port) );
return ret;
}
unsigned read_pit(void) {
unsigned count = 0;
// al = channel in bits 6 and 7, remaining bits clear
outb(0x43,0b0000000);
count = inb(0x40); // Low byte
count |= inb(0x40)<<8; // High byte
return count;
}
void draw_pixel(int pos_x, int pos_y, unsigned char color)
{
unsigned char* location = (unsigned char*)0xA0000 + 320 * pos_y + pos_x;
*location = color;
}
void draw_char(char c, int x, int y, unsigned char color)
{
const unsigned char *glyph = font[(int)c-32];
for(int cy=0;cy<8;cy++){
for(int cx=0;cx<8;cx++){
if(((int)glyph[cy]&(1<<cx))==(1<<cx)){
draw_pixel(x+cx,y+cy,color);
}
}
}
}
void draw_string(const char * s, int x, int y, unsigned char color) {
int i = 0;
while(s[i] != false) {
draw_char(s[i],x+(i*8),y,color);
i++;
}
}
void draw_rect(int pos_x, int pos_y, int w, int h, unsigned char color) {
for(int y = 0; y<h; y++) {
for(int x = 0; x<w; x++) {
draw_pixel(pos_x+x,pos_y+y,color);
}
}
}
extern "C" void main(){
int clock = 0;
int textposadd = 0;
bool bh = false;
while(true) {
clock++;
if(read_pit()!= 0 && clock == CLOCK_HIT) {
int ten_plus = 10+textposadd;
if(ten_plus==208){bh = true;}
else if(ten_plus==10){bh = false;}
if(bh){textposadd--;}
else{textposadd++;}
clock = 0;
draw_rect(0,0,320,200,0);
draw_string("Hello, reader. This is written text.", 10, 10+textposadd, 15);
draw_string("If this is displayed, my code works.", 18, 10, 15);
}
}
return;
}
boot.asm
[org 0x7c00]
KERNEL_LOCATION equ 0x1000
mov [BOOT_DISK], dl
xor ax, ax
mov es, ax
mov ds, ax
mov bp, 0x8000
mov sp, bp
mov bx, KERNEL_LOCATION
mov dh, 16
mov ah, 0x02
mov al, dh
mov ch, 0x00
mov dh, 0x00
mov cl, 0x02
mov dl, [BOOT_DISK]
int 0x13 ; no error management, do your homework!
mov ah, 0x00
mov al, 0x13
int 0x10 ; text mode
CODE_SEG equ GDT_code - GDT_start
DATA_SEG equ GDT_data - GDT_start
cli
lgdt [GDT_descriptor]
mov eax, cr0
or eax, 1
mov cr0, eax
jmp CODE_SEG:start_protected_mode
jmp $
BOOT_DISK: db 0
GDT_start:
GDT_null:
dd 0x0
dd 0x0
GDT_code:
dw 0xffff
dw 0x0
db 0x0
db 0b10011010
db 0b11001111
db 0x0
GDT_data:
dw 0xffff
dw 0x0
db 0x0
db 0b10010010
db 0b11001111
db 0x0
GDT_end:
GDT_descriptor:
dw GDT_end - GDT_start - 1
dd GDT_start
[bits 32]
start_protected_mode:
mov ax, DATA_SEG
mov ds, ax
mov ss, ax
mov es, ax
mov fs, ax
mov gs, ax
mov ebp, 0x90000 ; 32 bit stack base pointer
mov esp, ebp
jmp KERNEL_LOCATION
times 510-($-$$) db 0
dw 0xaa55
Your call draw_rect(0,0,320,200,0); clears the screen. Even though you subsequently write new pixels, you're at the mercy of your monitor's vsync. If the screen redraws while you're currently drawing your text, then you'll get whatever partial state it's in.
Many moons ago, the approach to fix this without using double buffering was to introduce some assembly that waits for the next vsync event. Essentially, the timing loop would be driven off that instead of a timer. You wait for vsync and then immediately do your drawing.
As to how you do this on your hardware, it's gonna vary. It looks like you're just using old x86. I really can't recall. I think the last time I did this I was writing games in MS-DOS in the 90s some time. Maybe start here.
Here is essentially the same approach but written in C with inline assembly, from a post in 2000 from gamedev.net:
void wait_for_retrace(void)
{
_DX = 0x03DA;
l1:
asm {
in al, dx
and al, 0x08
jnz l1
}
l2:
asm {
in al, dx
and al, 0x08
jz l2
}
}
Other than this, the standard approach is to at least use another buffer. You do all your drawing onto that buffer, and then you copy the buffer into the video memory when it's fully drawn. You can still get tearing if you do this while the screen is drawing, but it won't flicker as badly.
There's a third option, but it's complex and relates to erasing specific pixels before redrawing. It's more reserved for doing things like masking/moving around sprites. Basically, that approach utilizes XOR to "undo" a draw operation.
Bottom line is that if you really want smooth graphics, you must begin and end your entire screen update between screen update cycles.