From 58b73e220885887fa590bb343e7b6aa22c60195c Mon Sep 17 00:00:00 2001 From: Stephen Marz Date: Sat, 25 Apr 2020 22:25:32 -0400 Subject: [PATCH] Added tons of comments. --- risc_v/src/test.rs | 105 +++++++++++++++++++++++++++++++++------------ 1 file changed, 78 insertions(+), 27 deletions(-) diff --git a/risc_v/src/test.rs b/risc_v/src/test.rs index 237435c..684aea6 100644 --- a/risc_v/src/test.rs +++ b/risc_v/src/test.rs @@ -1,21 +1,34 @@ // test.rs -use crate::{kmem::{kfree, kmalloc}, +use crate::{cpu::{build_satp, + memcpy, + satp_fence_asid, + CpuMode, + SatpMode, + TrapFrame}, + kmem::{kfree, kmalloc}, + page::{map, zalloc, EntryBits, Table, PAGE_SIZE}, process::{Process, + ProcessData, + ProcessState, NEXT_PID, PROCESS_LIST, PROCESS_STARTING_ADDR, STACK_ADDR, - STACK_PAGES, ProcessState, ProcessData}, - syscall::syscall_fs_read}; -use crate::page::{zalloc, Table, map, EntryBits}; -use crate::cpu::{memcpy, TrapFrame, CpuMode, satp_fence_asid, SatpMode, build_satp}; + STACK_PAGES}, + syscall::syscall_fs_read}; pub fn test_block() { - // Let's test the block driver! + // The bytes to read would usually come from the inode, but we are in an + // interrupt context right now, so we cannot pause. Usually, this would be done + // by an exec system call. let bytes_to_read = 1024 * 50; let buffer = kmalloc(bytes_to_read); + // Read the file from the disk. let bytes_read = syscall_fs_read(8, 8, buffer, bytes_to_read as u32, 0); + // After compiling our program, I manually looked and saw it was 12,288 + // bytes. So, to make sure we got the right one, I do a manual check + // here. if bytes_read != 12288 { println!( "Unable to load program at inode 8, which should be \ @@ -25,50 +38,80 @@ pub fn test_block() { } else { // Let's get this program running! - let program_pages = (bytes_read / 4096) + 1; + // Everything is "page" based since we're going to map pages to + // user space. So, we need to know how many program pages we + // need. Each page is 4096 bytes. + let program_pages = (bytes_read / PAGE_SIZE) + 1; let my_pid = unsafe { NEXT_PID + 1 }; unsafe { NEXT_PID += 1; } - satp_fence_asid(my_pid as usize); - let mut my_proc= + let mut my_proc = Process { frame: zalloc(1) as *mut TrapFrame, stack: zalloc(STACK_PAGES), pid: my_pid, root: zalloc(1) as *mut Table, state: ProcessState::Running, data: ProcessData::zero(), - sleep_until: 0, - program: zalloc(program_pages) - }; + sleep_until: 0, + program: zalloc(program_pages), }; // Map the program in the MMU. let ptr = my_proc.program; unsafe { memcpy(ptr, buffer, bytes_read); } let table = unsafe { my_proc.root.as_mut().unwrap() }; + // This will map all of the program pages. Notice that in linker.lds in userspace + // we set the entry point address to 0x2000_0000. This is the same address as + // PROCESS_STARTING_ADDR, and they must match. for i in 0..program_pages { - let vaddr = PROCESS_STARTING_ADDR + (i << 12); - let paddr = ptr as usize + (i << 12); - map(table, vaddr, paddr, EntryBits::UserReadWriteExecute.val(), 0); + let vaddr = PROCESS_STARTING_ADDR + i * PAGE_SIZE; + let paddr = ptr as usize + i * PAGE_SIZE; + // We don't have an ELF loader yet, so we're loading raw binaries into memory. Since + // it is a flat binary, all .data, .rodata, and .bss sections get wrapped into + // the .text section. Normally, we don't want the .text section to be writeable, + // however because of this "flattening", we don't have a choice. + // Notice that USER shows up here. Since we're running in user mode, this bit MUST + // BE SET! Otherwise, we'll get a page fault from the beginning. + map( + table, + vaddr, + paddr, + EntryBits::UserReadWriteExecute.val(), + 0, + ); } // Map the stack let ptr = my_proc.stack as *mut u8; for i in 0..STACK_PAGES { - let vaddr = STACK_ADDR + (i << 12); - let paddr = ptr as usize + (i << 12); - map(table, vaddr, paddr, EntryBits::UserReadWrite.val(), 0); + let vaddr = STACK_ADDR + i * PAGE_SIZE; + let paddr = ptr as usize + i * PAGE_SIZE; + // We create the stack. We don't load a stack from the disk. This is why I don't + // need to make the stack executable. + map( + table, + vaddr, + paddr, + EntryBits::UserReadWrite.val(), + 0, + ); } // Set everything up in the trap frame unsafe { + // The program counter is a virtual memory address and is loaded into mepc + // when we execute mret. (*my_proc.frame).pc = PROCESS_STARTING_ADDR; - // Stack pointer + // Stack pointer. The stack starts at the bottom and works its way up, so we have to + // set the stack pointer to the bottom. (*my_proc.frame).regs[2] = - STACK_ADDR as usize + STACK_PAGES * 4096; + STACK_ADDR as usize + STACK_PAGES * PAGE_SIZE; + // USER MODE! This is how we set what'll go into mstatus when we run the process. (*my_proc.frame).mode = CpuMode::User as usize; (*my_proc.frame).pid = my_proc.pid as usize; - } - unsafe { + // The SATP register is used for the MMU, so we need to + // map our table into that register. The switch_to_user + // function will load .satp into the actual register + // when the time comes. (*my_proc.frame).satp = build_satp( SatpMode::Sv39, @@ -76,8 +119,17 @@ pub fn test_block() { my_proc.root as usize, ); } + // We don't reuse PIDs, so this really shouldn't matter. + satp_fence_asid(my_pid as usize); + // I took a different tact here than in process.rs. In there I created the process + // while holding onto the process list. It doesn't really matter since this is synchronous, + // but it might get dicey if let Some(mut pl) = unsafe { PROCESS_LIST.take() } { - println!("Added user process to the scheduler...get ready for take-off!"); + println!( + "Added user process to the scheduler...get \ + ready for take-off!" + ); + // As soon as we push this process on the list, it'll be schedule-able. pl.push_back(my_proc); unsafe { PROCESS_LIST.replace(pl); @@ -85,10 +137,9 @@ pub fn test_block() { } else { println!("Unable to spawn process."); - // Since my_proc couldn't enter the process list, it will - // be dropped and all of the associated allocations will - // be deallocated. - + // Since my_proc couldn't enter the process list, it + // will be dropped and all of the associated allocations + // will be deallocated. } } println!();