Xenalyze questions

* Xenalyze questions
@ 2013-06-06 11:35 Simon Graham
  2013-06-06 13:46 ` George Dunlap
  0 siblings, 1 reply; 5+ messages in thread
From: Simon Graham @ 2013-06-06 11:35 UTC (permalink / raw)
  To: xen-devel

[-- Attachment #1: Type: text/plain, Size: 2489 bytes --]

I've been trying to use xenalyze to investigate some problems on our Xen 4.2.2 based product and have run into a couple of issues:

First, I ran into problem running xenalyze on 64-bit Linux because some of the trace data structures were not properly packed  - I worked around this by adding the packed attribute where needed but I suspect there is a more elegant solution (see attached patch).

With that out of the way, I am now running it on a trace file and it craps out in the following code because e->pf_case has a large -ve value (presumably uninitialized):
 
void shadow_mmio_postprocess(struct hvm_data *h)
{
    struct pf_xen_extra *e = &h->inflight.pf_xen;
    if ( opt.summary_info )
    {
        if(e->pf_case)
=>          update_summary(&h->summary.pf_xen[e->pf_case],
                           h->arc_cycles);
        else
            fprintf(warn, "Strange, pf_case 0!\n");
 
        hvm_update_short_summary(h, HVM_SHORT_SUMMARY_MMIO);
    }
 
    if(opt.with_mmio_enumeration)
        enumerate_mmio(h);
}
 
The output from running the program is:
 
Starting program: /home/sgraham/sandbox/Community/xenalyze/xenalyze ~/tmp/e.evt
No output defined, using summary.
Using VMX hardware-assisted virtualization.
scan_for_new_pcpu: Activating pcpu 0 at offset 0
Creating vcpu 0 for dom 32768
scan_for_new_pcpu: Activating pcpu 1 at offset 1560
Creating vcpu 1 for dom 32768
scan_for_new_pcpu: Activating pcpu 2 at offset 1772
Creating vcpu 2 for dom 32768
scan_for_new_pcpu: Activating pcpu 3 at offset 68428
Creating vcpu 3 for dom 32768
scan_for_new_pcpu: Activating pcpu 5 at offset 68520
Creating vcpu 5 for dom 32768
scan_for_new_pcpu: Activating pcpu 10 at offset 73632
Creating vcpu 10 for dom 32768
scan_for_new_pcpu: Activating pcpu 11 at offset 90588
Creating vcpu 11 for dom 32768
init_pcpus: through first trace write, done for now.
hvm_generic_postprocess_init: Strange, h->postprocess set!
 
Program received signal SIGSEGV, Segmentation fault.
0x0000000000402888 in update_summary (s=0x7fe8aa769ca8, c=23368) at xenalyze.c:2244
 
This is using Xen 4.2.2 with several VMs running (32-bit Win7, 64-bit HVM Linux, 64-bit Win8). The trace file was generated by a 10s run of 'xentrace -e all' and is 200MB so I haven't attached it!
 
I suspect the problem is related to the specific set of trace records being output but I'm having a hard time grok'ing the code - Any suggestions welcome!

Simon Graham




[-- Attachment #2: xenalyze-packed.patch --]
[-- Type: application/octet-stream, Size: 1416 bytes --]

diff -r 71eeac989efc Makefile

--- a/Makefile
+++ b/Makefile
@@ -1,4 +1,4 @@
-CFLAGS += -g -O2
+CFLAGS += -g -O0
 CFLAGS += -fno-strict-aliasing
 CFLAGS += -std=gnu99
 CFLAGS += -Wall -Wstrict-prototypes
diff -r 71eeac989efc xenalyze.c
--- a/xenalyze.c
+++ b/xenalyze.c
@@ -5628,7 +5628,7 @@
             unsigned long long gl1e, write_val;
             unsigned long long va;
             unsigned flags:29, emulation_count:3;
-        } gpl4;
+        } __attribute__((packed)) gpl4;
     } *r = (typeof(r))ri->d;
  
     union shadow_event sevt = { .event = ri->event };
@@ -5725,7 +5725,7 @@
         struct {
             unsigned long long gfn;
             unsigned int va;
-        } gpl3;
+        } __attribute__((packed)) gpl3;
         struct {
             unsigned long long gfn, va;
         } gpl4;
@@ -5932,7 +5932,7 @@
         struct {
             unsigned long long gl1e, va;
             unsigned int flags;
-        } gpl4;
+        } __attribute__((packed)) gpl4;
     } *r = (typeof(r))ri->d;
     union shadow_event sevt = { .event = ri->event };
     int rec_gpl = sevt.paging_levels + 2;
@@ -6120,7 +6120,7 @@
         struct {
             unsigned long long gl1e, va;
             unsigned int flags;
-        } gpl4;
+        } __attribute__((packed)) gpl4;
     } *r = (typeof(r))ri->d;
     union shadow_event sevt = { .event = ri->event };
     int rec_gpl = sevt.paging_levels + 2;

[-- Attachment #3: Type: text/plain, Size: 126 bytes --]

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

^ permalink raw reply	[flat|nested] 5+ messages in thread