[coreboot-gerrit] New patch to review for coreboot: d791a9e ARMV7: threading support for cooperative multitasking

[Isaac Christensen]

Isaac Christensen ([e-mail]) just uploaded a new patch set to gerrit, which you can find at http://review.coreboot.org/6517

-gerrit

commit d791a9ecdd17320c53d154b8e5eaa7ceada7807a
Author: Ronald G. Minnich 
Date:   Fri Aug 23 11:22:42 2013 -0700

ARMV7: threading support for cooperative multitasking

These functions add support for cooperative multitasking.
    Currently, since we only have one ARM SOC that uses or supports multitasking,
    arch_get_thread_stackbase returns CONFIG_STACK_BOTTOM for the thread stack.
    We may end up having to make a cpu-specific function that arch_get_thread_stackbase calls,
    but let's avoid adding complexity until we're sure we need to. We also wish to avoid
    creating Yet Another Config Variable but will do so if pressed.

The switch code only saves r4-r11 and lr, which is consistent with the standard.

Change-Id: I0338a9c11127351e1f3a190bc51a7a558420b141
    Signed-off-by: Ronald G. Minnich 
    Reviewed-on: https://gerrit.chromium.org/gerrit/66845
    Reviewed-by: Aaron Durbin 
    Commit-Queue: Ronald G. Minnich 
    Tested-by: Ronald G. Minnich 
    (cherry picked from commit 22b62af3c26b6b504498b434d29a56a8932f3061)
    Signed-off-by: Isaac Christensen 
---
 src/arch/armv7/thread.c | 127 ++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 127 insertions(+)

diff --git a/src/arch/armv7/thread.c b/src/arch/armv7/thread.c
new file mode 100644
index 0000000..d0c23ff
--- /dev/null
+++ b/src/arch/armv7/thread.c
@@ -0,0 +1,127 @@
+/*
+ * This file is part of the coreboot project.
+ *
+ * Copyright (C) 2013 Google, Inc.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; version 2 of the License.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
+ */
+#include <console/console.h>
+#include <thread.h>
+
+/* The stack frame looks like the following. */
+struct pushed_regs {
+	u32 r4;
+	u32 r5;
+	u32 r6;
+	u32 r7;
+	u32 r8;
+	u32 r9;
+	u32 r10;
+	u32 r11;
+	u32 lr;
+};
+
+static inline uintptr_t push_stack(uintptr_t cur_stack, uintptr_t value)
+{
+	uintptr_t *addr;
+
+	cur_stack -= sizeof(value);
+	addr = (uintptr_t *)cur_stack;
+	*addr = value;
+	return cur_stack;
+}
+
+void arch_prepare_thread(struct thread *t,
+                         void asmlinkage (*thread_entry)(void *), void *arg)
+{
+	uintptr_t stack = t->stack_current;
+	int i;
+	uintptr_t poison = 0xdeadbeef;
+
+	/* Push the LR. thread_entry()
+	 * is assumed to never return.
+	 */
+	stack = push_stack(stack, (uintptr_t)thread_entry);
+	/* Make room for the registers.
+	 * Poison the initial stack. This is good hygiene and finds bugs.
+	 * Poisoning the stack with different values helps when you're
+	 * hunting for (e.g.) misaligned stacks or other such
+	 * weirdness. The -1 is because we already pushed lr.
+	 */
+	for(i = 0; i < sizeof(struct pushed_regs)/sizeof(u32)-1; i++)
+		stack = push_stack(stack, poison++);
+
+	t->stack_current = stack;
+}
+
+/* We could write this as a .S and the first time around that's how we
+ * did it. But there's always the question of matching our ARM
+ * directives in the .S with how gcc is doing things.  It seems best
+ * to follow the pattern of the rest of the ARM port and just use
+ * inline assembly and let gcc get all the ELF magic right.
+ */
+void __attribute__((naked))
+switch_to_thread(uintptr_t new_stack, uintptr_t *saved_stack)
+{
+	/* Defintions for those of us not totally familiar with ARM:
+	 * R15 -- PC, R14 -- LR, R13 -- SP
+	 * R0-R3 need not be saved, nor R12.
+	 * on entry, the only saved state is in LR -- the old PC.
+	 * The args are in R0,R1.
+	 * R0 is the new stack
+	 * R1 is a pointer to the old stack save location
+	 * Push R4-R11 and LR
+	 * then switch stacks
+	 * then pop R0-R12 and LR
+	 * then mov PC,LR
+	 *
+	 * stack layout
+	 * +------------+
+	 * |    LR      | <-- sp + 0x20
+	 * +------------+
+	 * |    R11     | <-- sp + 0x1c
+	 * +------------+
+	 * |    R10     | <-- sp + 0x18
+	 * +------------+
+	 * |    R9      | <-- sp + 0x14
+	 * +------------+
+	 * |    R8      | <-- sp + 0x10
+	 * +------------+
+	 * |    R7      | <-- sp + 0x0c
+	 * +------------+
+	 * |    R6      | <-- sp + 0x08
+	 * +------------+
+	 * |    R5      | <-- sp + 0x04
+	 * +------------+
+	 * |    R4      | <-- sp + 0x00
+	 * +------------+
+	 */
+        asm volatile (
+	/* save context. */
+	"push {r4-r11,lr}\n\t"
+	/* Save the current stack */
+	"str sp,[r1]\n\t"
+	/* switch to the new stack */
+	"mov sp,r0\n\t"
+	/* restore the registers */
+	"pop {r4-r11,lr}\n\t"
+	/* resume other thread. */
+	"mov pc,lr\n\t"
+	);
+}
+
+void *arch_get_thread_stackbase(void)
+{
+	return (void *)CONFIG_STACK_BOTTOM;
+}