From 6132623dbf3076da11cdb85c494fadd7ee0f0a93 Mon Sep 17 00:00:00 2001 From: Miquel Sabaté Solà Date: Mon, 2 Dec 2024 21:02:35 +0100 Subject: Share the same stack everywhere MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit As documented in the code, we are not implementing any kind of memory protection, so in theory any process (regardless if running in user or kernel space) would be able to tamper with other processes' stack. Hence, don't even pretend that we are separating stacks and share the same global stack everywhere. This simplifies things a bit. Signed-off-by: Miquel Sabaté Solà --- kernel/head.S | 8 +++++--- kernel/main.c | 18 ++++++++++++------ 2 files changed, 17 insertions(+), 9 deletions(-) (limited to 'kernel') diff --git a/kernel/head.S b/kernel/head.S index 9a12833..4390cc9 100644 --- a/kernel/head.S +++ b/kernel/head.S @@ -94,11 +94,13 @@ _start_kernel: li t6, 0 csrw sscratch, 0 - // Point tp and sp to the init task. The 'tp' register will always point to - // the current process being executed, and it will be shown on debug when + // Point 'tp' to the init task. The 'tp' register will always point to the + // current process being executed, and it will be shown on debug when // printing out messages. la tp, tasks - la sp, tasks + THREAD_SIZE + + // Point 'sp' the our general stack. + la sp, stack + THREAD_SIZE // The `start_kernel` function requires an argument to be passed, which is // the pointer to the `fdt` blob. The bootloader puts this on the `a1` diff --git a/kernel/main.c b/kernel/main.c index 9c524c4..19220f6 100644 --- a/kernel/main.c +++ b/kernel/main.c @@ -4,16 +4,22 @@ #include #include -// Stacks to be used by our processes. -unsigned long stack[4][THREAD_SIZE / sizeof(unsigned long)]; +// Stack to be used by our processes. "Blasphemy!" I hear you say. "How dare you +// use the same stack for kernel and user space?" It's not like this is some +// sort of utopian system in which everyone shares everything, but since this +// stupidly simple kernel does not even bother to implement paging nor any other +// memory protection of any kind, it's not like separating stacks for each +// process and kernel space would make much of a difference. Hence, let's keep +// it simple and have the same stack everwhere. +uint64_t stack[THREAD_SIZE / sizeof(uint64_t)]; // Initialize the list of structs by providing a fixed stack address and empty // values everywhere else. struct task_struct tasks[4] = { - [TASK_INIT] = { .stack = stack[0], .name = "init", .entry_addr = nullptr, }, - [TASK_FIZZ] = { .stack = stack[1], .name = "fizz", .entry_addr = nullptr, }, - [TASK_BUZZ] = { .stack = stack[2], .name = "buzz", .entry_addr = nullptr, }, - [TASK_FIZZBUZZ] = { .stack = stack[3], .name = "fizzbuzz", .entry_addr = nullptr, }, + [TASK_INIT] = { .stack = stack, .name = "init", .entry_addr = nullptr, }, + [TASK_FIZZ] = { .stack = stack, .name = "fizz", .entry_addr = nullptr, }, + [TASK_BUZZ] = { .stack = stack, .name = "buzz", .entry_addr = nullptr, }, + [TASK_FIZZBUZZ] = { .stack = stack, .name = "fizzbuzz", .entry_addr = nullptr, }, }; // Defined in fbos/init.h. -- cgit v1.2.3