A simple hello world does not display in Qemu

[Pius Onobhayedo]

I followed the instructions on https://github.com/rust-osdev/bootloader/blob/main/README.md and https://github.com/rust-osdev/bootloader/blob/main/docs/create-disk-image.md and got all successfully setup

A simple code hello world code like that shown below in my test kernel, builds but when I run, Qemu opens quite alright and shows a series of info. Unfortunately, it never ends up showing my Hello World. I am not using any custom linker. Am I missing something?

#![no_std]
#![no_main]
#![feature(alloc_error_handler)]

bootloader_api::entry_point!(my_entry_point);

fn my_entry_point(bootinfo: &'static mut bootloader_api::BootInfo) -> ! {

 let display_cursor_position = 0xb8000 as *mut u8;
 let hello = b"Hello World";
    for (i, &byte) in hello.iter().enumerate() {
        unsafe {
            *display_cursor_position.offset(i as isize * 2) = byte;
            *display_cursor_position.offset(i as isize * 2 + 1) = 0x0f;
        }
    }
loop {}
}

#[panic_handler]
fn panic(_info: &core::panic::PanicInfo) -> ! {
    loop {}
}

/// The global allocator type.
#[derive(Default)]
pub struct Allocator;

unsafe impl GlobalAlloc for Allocator {
    unsafe fn alloc(&self, layout: Layout) -> *mut u8 {
        alloc(layout) as *mut u8
    }
    unsafe fn dealloc(&self, ptr: *mut u8, _layout: Layout) {
        dealloc(ptr, _layout);
    }
}

/// If there is an out of memory error, just panic.

#[alloc_error_handler]
fn my_allocator_error(_layout: Layout) -> ! {
    panic!("out of memory");
}

/// The static global allocator.
#[global_allocator]
static GLOBAL_ALLOCATOR: Allocator = Allocator;

[Philipp Oppermann]

We don't use the VGA text mode because it is not compatible with UEFI booting. Instead, we set up a pixel-based framebuffer, which you can access through the bootinfo.framebuffer field.

As this buffer is pixel-based instead of text-based, you have to do the font rendering yourself. I recommend to use the noto_sans_mono_bitmap for that as a start. See our internal framebuffer submodule as an example.

Hope this helps!

[Pius Onobhayedo]

We don't use the VGA text mode because it is not compatible with UEFI booting. Instead, we set up a pixel-based framebuffer, which you can access through the bootinfo.framebuffer field.

As this buffer is pixel-based instead of text-based, you have to do the font rendering yourself. I recommend to use the noto_sans_mono_bitmap for that as a start. See our internal framebuffer submodule as an example.

Hope this helps!

This approach worked. I was able to see the "Hello, World". Thank you.