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

Re: [Xen-devel] [Fwd: stack overflow "cause" found]


  • To: xen-devel@xxxxxxxxxxxxxxxxxxx
  • From: Rik van Riel <riel@xxxxxxxxxx>
  • Date: Wed, 26 Jul 2006 16:49:21 -0400
  • Delivery-date: Wed, 26 Jul 2006 13:49:52 -0700
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>

Rik van Riel wrote:

Xen really should not be doing this.  We could either uninline some of
these functions, or try gcc -O2 or gcc -Os.

OK, with -Os it doesn't quite build due to the built-ins not being
inlined.  With -O2 the stack usage changes dramatically...

Currently the top stack users are:

$ objdump -dS xen-syms | ./checkstack.pl x86_64
0xffff8300001c7c47 find_unisys_acpi_oem_table:          1296
0xffff83000011182e do_xen_version:                      1280
0xffff83000015bc3e x86_emulate_memop:                   1264
0xffff8300001adbaa vmx_world_save:                      1040
0xffff8300001aeac5 vmx_world_save:                      1040

0xffff8300001af923 vmx_assist:                          224

The function vmx_world_save is inlined into vmx_assist with -O2,
and the combined stack use is only 200 bytes, down from 1264 !

I guess we'll want to build Xen with -O2.  I'm testing a -O2 built
hypervisor right now...

--
"Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are,
by definition, not smart enough to debug it." - Brian W. Kernighan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.