[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Kernel crash on qemu-m68k-system while building r-cran packages



Le 30/05/2020 à 17:27, John Paul Adrian Glaubitz a écrit :
> Hi Laurent!
> 
> On 5/30/20 5:12 PM, Laurent Vivier wrote:
>> As the package is available, I've been able to reproduce the problem
>> with logs enabled. Apparently the instruction is not decoded correctly:
>>
>> ----------------
>> IN:
>> 0x00016a2c:  fmovel #0,%fpsr
>> Disassembler disagrees with translator over instruction decoding
>> Please report this to qemu-devel@nongnu.org
>>
>> OP:
>>  ld_i32 tmp0,env,$0xfffffffffffffff0
>>  movi_i32 tmp1,$0x0
>>  brcond_i32 tmp0,tmp1,lt,$L0
>>
>>  ---- 00016a2c 00000000
>>  movi_i32 PC,$0x16a2c
>>  movi_i32 tmp0,$0x3
>>  call raise_exception,$0x0,$0,env,tmp0
>>  set_label $L0
>>  exit_tb $0x7fc2a0da5643
>>
>> INT  49331: Address Error(0xc) pc=00016a2c sp=3cc91f30 sr=2004
>>
>> I try to fix that...
> 
> Ah, cool. Thanks a lot. And sorry for not replying earlier, I missed your
> previous mail.

No problem.

I've investigated and the fmovel with immediate value is not implemented...

The following patch fixes the crash but then I have a soft lockup. IT
needs more investigation.

diff --git a/target/m68k/translate.c b/target/m68k/translate.c
index 0f80888203d3..c093f6c683e8 100644
--- a/target/m68k/translate.c
+++ b/target/m68k/translate.c
@@ -4936,6 +4936,20 @@ static void gen_op_fmove_fcr(CPUM68KState *env,
DisasContext *s,
             gen_store_fcr(s, AREG(insn, 0), mask);
         }
         return;
+    case 7: /* Immediate */
+        if (REG(insn, 0) == 4) {
+            if (is_write ||
+                (mask != M68K_FPIAR && mask != M68K_FPSR &&
+                 mask != M68K_FPCR)) {
+                gen_exception(s, s->base.pc_next, EXCP_ILLEGAL);
+                return;
+            }
+            tmp = tcg_const_i32(read_im32(env, s));
+            gen_store_fcr(s, tmp, mask);
+            tcg_temp_free(tmp);
+            return;
+        }
+        break;
     default:
         break;
     }

Thanks,
Laurent


Reply to: