target/i386: Do not re-compute new pc with CF_PCREL

With PCREL, we have a page-relative view of EIP, and an
approximation of PC = EIP+CSBASE that is good enough to
detect page crossings.  If we try to recompute PC after
masking EIP, we will mess up that approximation and write
a corrupt value to EIP.

We already handled masking properly for PCREL, so the
fix in b5e0d5d was only needed for the !PCREL path.

Cc: qemu-stable@nongnu.org
Fixes: b5e0d5d ("target/i386: Fix 32-bit wrapping of pc/eip computation")
Reported-by: Michael Tokarev <mjt@tls.msk.ru>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
Message-ID: <20240101230617.129349-1-richard.henderson@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

target/i386/tcg/translate.c

@@ -2866,10 +2866,6 @@ static void gen_jmp_rel(DisasContext *s, MemOp ot, int diff, int tb_num)
         }
     }
     new_eip &= mask;
-    new_pc = new_eip + s->cs_base;
-    if (!CODE64(s)) {
-        new_pc = (uint32_t)new_pc;
-    }
 
     gen_update_cc_op(s);
     set_cc_op(s, CC_OP_DYNAMIC);
@@ -2885,6 +2881,8 @@ static void gen_jmp_rel(DisasContext *s, MemOp ot, int diff, int tb_num)
             tcg_gen_andi_tl(cpu_eip, cpu_eip, mask);
             use_goto_tb = false;
         }
+    } else if (!CODE64(s)) {
+        new_pc = (uint32_t)(new_eip + s->cs_base);
     }
 
     if (use_goto_tb && translator_use_goto_tb(&s->base, new_pc)) {