riscv: Add initial support for 32bit boards

* Adding separate targets for 32bit and 64bit qemu
* Using the riscv64 toolchain for 32bit builds requires setting -m elf32lriscv
* rv32/rv64 is currently configured with ARCH_RISCV_RV32/RV64 and not per stage.
  This should probably be changed later.

TEST=Boots to "Payload not loaded." on 32bit qemu using the following commands:

util/riscv/make-spike-elf.sh build/coreboot.rom build/coreboot.elf
qemu-system-riscv32 -M virt -m 1024M -nographic -kernel build/coreboot.elf

Change-Id: I35e59b459d1770df10b51fe9e77dcc474d7c75a0
Signed-off-by: Philipp Hug <philipp@hug.cx>
Reviewed-on: https://review.coreboot.org/c/31253
Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
Reviewed-by: ron minnich <rminnich@gmail.com>

7 files changed, 46 insertions, 16 deletions

diff --git a/src/arch/riscv/Makefile.inc b/src/arch/riscv/Makefile.inc
index 9d91f0cefe..d5f62954eb 100644
--- a/src/arch/riscv/Makefile.inc
+++ b/src/arch/riscv/Makefile.inc
@@ -86,6 +86,10 @@ $(objcbfs)/bootblock.debug: $$(bootblock-objs)
 bootblock-c-ccopts += $(riscv_flags)
 bootblock-S-ccopts += $(riscv_asm_flags)
 
+ifeq ($(CONFIG_ARCH_RISCV_RV32),y)
+LDFLAGS_bootblock += -m elf32lriscv
+endif #CONFIG_ARCH_RISCV_RV32
+
 endif #CONFIG_ARCH_BOOTBLOCK_RISCV
 
 ################################################################################
@@ -116,6 +120,10 @@ $(objcbfs)/romstage.debug: $$(romstage-objs)
 romstage-c-ccopts += $(riscv_flags)
 romstage-S-ccopts += $(riscv_asm_flags)
 
+ifeq ($(CONFIG_ARCH_RISCV_RV32),y)
+LDFLAGS_romstage += -m elf32lriscv
+endif #CONFIG_ARCH_RISCV_RV32
+
 endif #CONFIG_ARCH_ROMSTAGE_RISCV
 
 ################################################################################
@@ -161,5 +169,9 @@ $(objcbfs)/ramstage.debug: $$(ramstage-objs)
 ramstage-c-ccopts += $(riscv_flags)
 ramstage-S-ccopts += $(riscv_asm_flags)
 
+ifeq ($(CONFIG_ARCH_RISCV_RV32),y)
+LDFLAGS_ramstage += -m elf32lriscv
+endif #CONFIG_ARCH_RISCV_RV32
+
 endif #CONFIG_ARCH_RAMSTAGE_RISCV
 endif #CONFIG_ARCH_RISCV
diff --git a/src/arch/riscv/bootblock.S b/src/arch/riscv/bootblock.S
index d4b8be7c2a..b0796f9fbc 100644
--- a/src/arch/riscv/bootblock.S
+++ b/src/arch/riscv/bootblock.S
@@ -16,6 +16,7 @@
  */
 
 #include <arch/encoding.h>
+#include <bits.h>
 #include <mcall.h>
 
 .section ".text._start", "ax", %progbits
@@ -44,7 +45,7 @@ _start:
 	slli t1, a0, RISCV_PGSHIFT
 	add  t0, t0, t1
 	li   t1, 0xDEADBEEF
-	sd   t1, 0(t0)
+	STORE t1, 0(t0)
 	li   t1, RISCV_PGSIZE - HLS_SIZE
 	add  sp, t0, t1
 
diff --git a/src/arch/riscv/include/arch/smp/spinlock.h b/src/arch/riscv/include/arch/smp/spinlock.h
index dc561d30a1..95e60bfefc 100644
--- a/src/arch/riscv/include/arch/smp/spinlock.h
+++ b/src/arch/riscv/include/arch/smp/spinlock.h
@@ -21,7 +21,7 @@
 #define barrier() { asm volatile ("fence" ::: "memory"); }
 
 typedef struct {
-	volatile atomic_t lock;
+	atomic_t lock;
 } spinlock_t;
 
 static inline void spinlock_lock(spinlock_t *lock)
diff --git a/src/arch/riscv/include/bits.h b/src/arch/riscv/include/bits.h
index f69c7ec0da..d824f3ec98 100644
--- a/src/arch/riscv/include/bits.h
+++ b/src/arch/riscv/include/bits.h
@@ -47,10 +47,19 @@
 #define STR(x) XSTR(x)
 #define XSTR(x) #x
 
-# define SLL32    sllw
-# define STORE    sd
-# define LOAD     ld
-# define LOG_REGBYTES 3
+#if __riscv_xlen == 64
+#define SLL32 sllw
+#define STORE sd
+#define LOAD ld
+#define LWU lwu
+#define LOG_REGBYTES 3
+#else
+#define SLL32 sll
+#define STORE sw
+#define LOAD lw
+#define LWU lw
+#define LOG_REGBYTES 2
+#endif
 
 #define REGBYTES (1 << LOG_REGBYTES)
 
diff --git a/src/arch/riscv/include/mcall.h b/src/arch/riscv/include/mcall.h
index cd1ed6dc18..192d2b4564 100644
--- a/src/arch/riscv/include/mcall.h
+++ b/src/arch/riscv/include/mcall.h
@@ -18,7 +18,13 @@
 
 // NOTE: this is the size of hls_t below. A static_assert would be
 // nice to have.
+#if __riscv_xlen == 64
 #define HLS_SIZE 88
+#endif
+
+#if __riscv_xlen == 32
+#define HLS_SIZE 52
+#endif
 
 /* We save 37 registers, currently. */
 #define MENTRY_FRAME_SIZE (HLS_SIZE + 37 * 8)
@@ -26,6 +32,7 @@
 #ifndef __ASSEMBLER__
 
 #include <arch/encoding.h>
+#include <arch/smp/atomic.h>
 #include <stdint.h>
 
 typedef struct {
@@ -38,8 +45,8 @@ typedef struct {
 struct blocker {
 	void *arg;
 	void (*fn)(void *arg);
-	uint32_t sync_a;
-	uint32_t sync_b;
+	atomic_t sync_a;
+	atomic_t sync_b;
 };
 
 typedef struct {
diff --git a/src/arch/riscv/ramstage.S b/src/arch/riscv/ramstage.S
index 115a55f290..28183e50e0 100644
--- a/src/arch/riscv/ramstage.S
+++ b/src/arch/riscv/ramstage.S
@@ -14,6 +14,7 @@
  */
 
 #include <arch/encoding.h>
+#include <bits.h>
 #include <mcall.h>
 
 .section ".text._start", "ax", %progbits
@@ -27,7 +28,7 @@ _start:
 	slli t1, a0, RISCV_PGSHIFT
 	add  t0, t0, t1
 	li   t1, 0xDEADBEEF
-	sd   t1, 0(t0)
+	STORE t1, 0(t0)
 	li   t1, RISCV_PGSIZE - HLS_SIZE
 	add  sp, t0, t1
 
diff --git a/src/arch/riscv/smp.c b/src/arch/riscv/smp.c
index 8942ec5b93..b32e4b8694 100644
--- a/src/arch/riscv/smp.c
+++ b/src/arch/riscv/smp.c
@@ -32,13 +32,13 @@ void smp_pause(int working_hartid)
 		/* waiting for work hart */
 		do {
 			barrier();
-		} while (SYNCA != 0x01234567);
+		} while (atomic_read(&SYNCA) != 0x01234567);
 
 		clear_csr(mstatus, MSTATUS_MIE);
 		write_csr(mie, MIP_MSIP);
 
 		/* count how many cores enter the halt */
-		__sync_fetch_and_add(&SYNCB, 1);
+		atomic_add(&SYNCB, 1);
 
 		do {
 			barrier();
@@ -49,17 +49,17 @@ void smp_pause(int working_hartid)
 	} else {
 		/* Initialize the counter and
 		 * mark the work hart into smp_pause */
-		SYNCB = 0;
-		SYNCA = 0x01234567;
+		atomic_set(&SYNCB, 0);
+		atomic_set(&SYNCA, 0x01234567);
 
 		/* waiting for other Hart to enter the halt */
 		do {
 			barrier();
-		} while (SYNCB + 1 < CONFIG_MAX_CPUS);
+		} while (atomic_read(&SYNCB) + 1 < CONFIG_MAX_CPUS);
 
 		/* initialize for the next call */
-		SYNCA = 0;
-		SYNCB = 0;
+		atomic_set(&SYNCA, 0);
+		atomic_set(&SYNCB, 0);
 	}
 #undef SYNCA
 #undef SYNCB