[Crash-utility] Fwd: [PATCH] fix s390x stack dump

Subject: [PATCH] fix s390x stack dump Date: Wed, 31 Jan 2007 14:51:55 -0800 From: "Kurtis D. Rader" <krader(a)us.ibm.com&gt; To: crash-utility(a)redhat.com The s390x_back_trace_cmd() function treats the stack as a sequence of 32-bit words. That's obviously wrong and causes confusion because the two halves of the 64-bit word gets dumped "backwards"; that is, at the wrong offset. This patch fixes that. I created it against crash 4.0-3.9 but it applies cleanly to 4.0-3.17. diff -U3 -p --recursive crash-4.0-3.9/s390x.c crash-4.0-3.9.fixed/s390x.c --- crash-4.0-3.9/s390x.c 2006-11-02 10:26:50.000000000 -0800 +++ crash-4.0-3.9.fixed/s390x.c 2006-11-02 18:50:57.000000000 -0800 @@ -730,9 +730,9 @@ s390x_back_trace_cmd(struct bt_info *bt) backchain = ULONG(&stack[backchain - stack_base + bc_offset]); /* print stack content if -f is specified */ - if((bt->flags & BT_FULL) && !BT_REFERENCE_CHECK(bt)){ + if ((bt->flags & BT_FULL) && !BT_REFERENCE_CHECK(bt)) { int frame_size; - if(backchain == 0){ + if (backchain == 0) { frame_size = stack_base - old_backchain + KERNEL_STACK_SIZE; } else { @@ -740,14 +740,15 @@ s390x_back_trace_cmd(struct bt_info *bt) (stack_base - old_backchain + KERNEL_STACK_SIZE)); } - for(j=0; j< frame_size; j+=4){ + for (j = 0; j < frame_size; j += 8) { if(j % 16 == 0){ - fprintf(fp,"\n%08lx: ",old_backchain+j); + fprintf(fp, "%s %016lx: ", + j ? "\n" : "", old_backchain + j); } - fprintf(fp," %08lx",ULONG(&stack[old_backchain - - stack_base + j])); + fprintf(fp," %016lx", + ULONG(&stack[old_backchain - stack_base + j])); } - fprintf(fp,"\n\n"); + fprintf(fp, "\n"); } /* Check for interrupt stackframe */ -- Kurtis D. Rader, Linux level 3 support email: krader(a)us.ibm.com IBM Integrated Technology Services DID: +1 503-578-3714 15350 SW Koll Pkwy, MS DES1-01 service: 800-IBM-SERV Beaverton, OR 97006-6063 http://www.ibm.com