[Qemu-devel] [PATCH 0/2] Fix ppc64-linux-user

[Qemu-devel] [PATCH 0/2] Fix ppc64-linux-user

[Richard Henderson]

At the moment, no ppc64-linux-user tests work at all.  With these
two patches, I can at least run the linux-user-test-0.3 binary
and a statically linked empty program (int main() { return 0; }).

Given that linux-user-test-0.3 is in fact busybox, this is rather
more comprehensive testing than it would otherwise sound.


r~


Richard Henderson (2):
  ppc64-linux-user: Properly interpret the entry function descriptor.
  ppc64-linux-user: Fix syscall return type.

 linux-user/elfload.c |    8 ++++----
 linux-user/main.c    |   13 +++----------
 2 files changed, 7 insertions(+), 14 deletions(-)

-- 
1.7.6.4

[Qemu-devel] [PATCH 1/2] ppc64-linux-user: Properly interpret the entry function descriptor.

[Richard Henderson]

Don't confuse the load address with the load bias.  They're equal
for ET_DYN objects (i.e. ld.so) but different for ET_EXEC objects
(i.e. statically linked).

Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 linux-user/elfload.c |    8 ++++----
 1 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/linux-user/elfload.c b/linux-user/elfload.c
index 8677bba..a413976 100644
--- a/linux-user/elfload.c
+++ b/linux-user/elfload.c
@@ -618,8 +618,8 @@ static inline void init_thread(struct target_pt_regs *_regs, struct image_info *
 {
     _regs->gpr[1] = infop->start_stack;
 #if defined(TARGET_PPC64) && !defined(TARGET_ABI32)
-    _regs->gpr[2] = ldq_raw(infop->entry + 8) + infop->load_addr;
-    infop->entry = ldq_raw(infop->entry) + infop->load_addr;
+    _regs->gpr[2] = ldq_raw(infop->entry + 8) + infop->load_bias;
+    infop->entry = ldq_raw(infop->entry) + infop->load_bias;
 #endif
     _regs->nip = infop->entry;
 }
@@ -1884,11 +1884,11 @@ int load_elf_binary(struct linux_binprm * bprm, struct target_pt_regs * regs,
     info->start_stack = bprm->p;
 
     /* If we have an interpreter, set that as the program's entry point.
-       Copy the load_addr as well, to help PPC64 interpret the entry
+       Copy the load_bias as well, to help PPC64 interpret the entry
        point as a function descriptor.  Do this after creating elf tables
        so that we copy the original program entry point into the AUXV.  */
     if (elf_interpreter) {
-        info->load_addr = interp_info.load_addr;
+        info->load_bias = interp_info.load_bias;
         info->entry = interp_info.entry;
         free(elf_interpreter);
     }
-- 
1.7.6.4

Re: [Qemu-devel] [PATCH 1/2] ppc64-linux-user: Properly interpret the entry function descriptor.

[Alexander Graf]

On 26.10.2011, at 18:59, Richard Henderson wrote:

> Don't confuse the load address with the load bias.  They're equal
> for ET_DYN objects (i.e. ld.so) but different for ET_EXEC objects
> (i.e. statically linked).
> 
> Signed-off-by: Richard Henderson <rth@twiddle.net>

Acked-by: Alexander Graf <agraf@suse.de>


Alex

[Qemu-devel] [PATCH 2/2] ppc64-linux-user: Fix syscall return type.

[Richard Henderson]

Use target_ulong instead of hard-coded uint32_t.
Remove the disabled printf's that are redundant with -strace.

Signed-off-by: Richard Henderson <rth@twiddle.net>
---
 linux-user/main.c |   13 +++----------
 1 files changed, 3 insertions(+), 10 deletions(-)

diff --git a/linux-user/main.c b/linux-user/main.c
index e7dad54..dd55648 100644
--- a/linux-user/main.c
+++ b/linux-user/main.c
@@ -1332,7 +1332,7 @@ void cpu_loop(CPUPPCState *env)
 {
     target_siginfo_t info;
     int trapnr;
-    uint32_t ret;
+    target_ulong ret;
 
     for(;;) {
         cpu_exec_start(env);
@@ -1695,27 +1695,20 @@ void cpu_loop(CPUPPCState *env)
              * PPC ABI uses overflow flag in cr0 to signal an error
              * in syscalls.
              */
-#if 0
-            printf("syscall %d 0x%08x 0x%08x 0x%08x 0x%08x\n", env->gpr[0],
-                   env->gpr[3], env->gpr[4], env->gpr[5], env->gpr[6]);
-#endif
             env->crf[0] &= ~0x1;
             ret = do_syscall(env, env->gpr[0], env->gpr[3], env->gpr[4],
                              env->gpr[5], env->gpr[6], env->gpr[7],
                              env->gpr[8], 0, 0);
-            if (ret == (uint32_t)(-TARGET_QEMU_ESIGRETURN)) {
+            if (ret == (target_ulong)(-TARGET_QEMU_ESIGRETURN)) {
                 /* Returning from a successful sigreturn syscall.
                    Avoid corrupting register state.  */
                 break;
             }
-            if (ret > (uint32_t)(-515)) {
+            if (ret > (target_ulong)(-515)) {
                 env->crf[0] |= 0x1;
                 ret = -ret;
             }
             env->gpr[3] = ret;
-#if 0
-            printf("syscall returned 0x%08x (%d)\n", ret, ret);
-#endif
             break;
         case POWERPC_EXCP_STCX:
             if (do_store_exclusive(env)) {
-- 
1.7.6.4

Re: [Qemu-devel] [PATCH 2/2] ppc64-linux-user: Fix syscall return type.

[Alexander Graf]

On 26.10.2011, at 18:59, Richard Henderson wrote:

> Use target_ulong instead of hard-coded uint32_t.
> Remove the disabled printf's that are redundant with -strace.
> 
> Signed-off-by: Richard Henderson <rth@twiddle.net>

Acked-by: Alexander Graf <agraf@suse.de>

Riku, do you want to take these in or should I?


Alex