[1/1] target/riscv: fix endless translation loop on big endian systems

Commit Message

Ziqiao Kong April 13, 2025, 5:22 p.m. UTC

On big endian systems, pte and updated_pte holds big endian host data
while pte_pa points to little endiaon target data. This means the branch
at cpu_helper.c:1669 will never be satisfied and thus causes an endless
translation loop.

Signed-off-by: Ziqiao Kong <[e-mail]>
---
 target/riscv/cpu_helper.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Patch

diff --git a/target/riscv/cpu_helper.c b/target/riscv/cpu_helper.c
index 6c4391d96b..578927ec9c 100644
--- a/target/riscv/cpu_helper.c
+++ b/target/riscv/cpu_helper.c
@@ -1662,9 +1662,9 @@  static int get_physical_address(CPURISCVState *env, hwaddr *physical,
             target_ulong *pte_pa = qemu_map_ram_ptr(mr->ram_block, addr1);
             target_ulong old_pte;
             if (riscv_cpu_sxl(env) == MXL_RV32) {
-                old_pte = qatomic_cmpxchg((uint32_t *)pte_pa, pte, updated_pte);
+                old_pte = qatomic_cmpxchg((uint32_t *)pte_pa, cpu_to_le64(pte), cpu_to_le64(updated_pte));
             } else {
-                old_pte = qatomic_cmpxchg(pte_pa, pte, updated_pte);
+                old_pte = qatomic_cmpxchg(pte_pa, cpu_to_le64(pte), cpu_to_le64(updated_pte));
             }
             if (old_pte != pte) {
                 goto restart;