|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] Re: [Xen-devel] xl create crash when using stub domains
On Wed, 2011-09-21 at 02:34 +0100, Jeremy Fitzhardinge wrote:
> On 09/15/2011 05:33 PM, Jeremy Fitzhardinge wrote:
> > When I create an HVM domain with stubdom enabled, it crashes at:
> >
> > (gdb) run create -c /etc/xen/f14hv64 vcpus=4 xen_platform_pci=0 'boot="d"'
> > Starting program: /usr/sbin/xl create -c /etc/xen/f14hv64 vcpus=4
> > xen_platform_pci=0 'boot="d"'
> > [Thread debugging using libthread_db enabled]
> > Parsing config file /etc/xen/f14hv64
> > xc: info: VIRTUAL MEMORY ARRANGEMENT:
> > Loader: 0000000000100000->000000000017b9ec
> > TOTAL: 0000000000000000->000000003f800000
> > ENTRY ADDRESS: 0000000000100000
> > xc: info: PHYSICAL MEMORY ALLOCATION:
> > 4KB PAGES: 0x0000000000000200
> > 2MB PAGES: 0x00000000000001fb
> > 1GB PAGES: 0x0000000000000000
> > xc: error: panic: xc_dom_bzimageloader.c:588: xc_dom_probe_bzimage_kernel:
> > kernel is not a bzImage: Invalid kernel
FWIW I don't get this message. It seems unrelated to the issue here but
makes me curious...
> > Detaching after fork from child process 26888.
> > [New Thread 0x7ffff7342700 (LWP 26889)]
> > [Thread 0x7ffff7342700 (LWP 26889) exited]
> > [New Thread 0x7ffff7342700 (LWP 26921)]
> >
> > Program received signal SIGSEGV, Segmentation fault.
> > 0x00007ffff7bbbec5 in libxl__wait_for_device_model (gc=0x7fffffffdbb0,
> > domid=22, state=0x7ffff7bc1b8c "running", starting=0x623760,
> > check_callback=0, check_callback_userdata=0x0) at libxl_device.c:555
> > 555 if (starting && starting->for_spawn->fd > xs_fileno(xsh))
> > (gdb) bt
>
> This patch seems to fix it, but I don't know if it is a real fix or just
> papering over something else.
I think this is correct because starting->for_spawn is only valid if the
device model was launched with libxl__spawn_spawn which is only the case
for process based stubdom.
libxl__create_device_model heads off into libxl__create_stubdom for this
case and explicitly sets for_spawn == NULL.
Hmm, actually this function never uses starting except to get at
for_spawn perhaps we should just pass in the for_spawn directly. Patch
to that effect follows.
Ian.
ps: can you add this to your ~/.hgrc please:
[diff]
showfunc = True
8<-----------------------------------------------
# HG changeset patch
# User Ian Campbell <ian.campbell@xxxxxxxxxx>
# Date 1316595312 -3600
# Node ID eb9330c89fd3843ff0b1348b0ef21cfeb22d4a76
# Parent 21db7a7dd18483aab5c651f2364c09e8e492d7b1
libxl: make libxl__wait_for_device_model use libxl__spawn_starrting directly
Instead of indirecting via libxl_device_model_starting. This fixes a
segmentation fault using stubdomains where starting->for_spawn is
(validly) NULL because starting a stubdom doesn't need to spawn a
process.
Most callers of libxl__wait_for_device_model already pass NULL for
this variable (because they are not on the starting path) so on
libxl__confirm_device_model_startup needs to change.
Reported-by: Jeremy Fitzhardinge <jeremy@xxxxxxxx>
Signed-off-by: Ian Campbell <ian.campbell@xxxxxxxxxx>
diff -r 21db7a7dd184 -r eb9330c89fd3 tools/libxl/libxl_device.c
--- a/tools/libxl/libxl_device.c Tue Sep 20 16:50:44 2011 +0100
+++ b/tools/libxl/libxl_device.c Wed Sep 21 09:55:12 2011 +0100
@@ -528,7 +528,7 @@ out:
int libxl__wait_for_device_model(libxl__gc *gc,
uint32_t domid, char *state,
- libxl__device_model_starting *starting,
+ libxl__spawn_starting *spawning,
int (*check_callback)(libxl__gc *gc,
uint32_t domid,
const char *state,
@@ -558,12 +558,12 @@ int libxl__wait_for_device_model(libxl__
tv.tv_sec = LIBXL_DEVICE_MODEL_START_TIMEOUT;
tv.tv_usec = 0;
nfds = xs_fileno(xsh) + 1;
- if (starting && starting->for_spawn->fd > xs_fileno(xsh))
- nfds = starting->for_spawn->fd + 1;
+ if (spawning && spawning->fd > xs_fileno(xsh))
+ nfds = spawning->fd + 1;
while (rc > 0 || (!rc && tv.tv_sec > 0)) {
- if ( starting ) {
- rc = libxl__spawn_check(gc, starting->for_spawn);
+ if ( spawning ) {
+ rc = libxl__spawn_check(gc, spawning);
if ( rc ) {
LIBXL__LOG(ctx, LIBXL__LOG_ERROR,
"Device Model died during startup");
@@ -592,8 +592,8 @@ again:
free(p);
FD_ZERO(&rfds);
FD_SET(xs_fileno(xsh), &rfds);
- if (starting)
- FD_SET(starting->for_spawn->fd, &rfds);
+ if (spawning)
+ FD_SET(spawning->fd, &rfds);
rc = select(nfds, &rfds, NULL, NULL, &tv);
if (rc > 0) {
if (FD_ISSET(xs_fileno(xsh), &rfds)) {
@@ -603,9 +603,9 @@ again:
else
goto again;
}
- if (starting && FD_ISSET(starting->for_spawn->fd, &rfds)) {
+ if (spawning && FD_ISSET(spawning->fd, &rfds)) {
unsigned char dummy;
- if (read(starting->for_spawn->fd, &dummy, sizeof(dummy)) != 1)
+ if (read(spawning->fd, &dummy, sizeof(dummy)) != 1)
LIBXL__LOG_ERRNO(ctx, LIBXL__LOG_DEBUG,
"failed to read spawn status pipe");
}
diff -r 21db7a7dd184 -r eb9330c89fd3 tools/libxl/libxl_dm.c
--- a/tools/libxl/libxl_dm.c Tue Sep 20 16:50:44 2011 +0100
+++ b/tools/libxl/libxl_dm.c Wed Sep 21 09:55:12 2011 +0100
@@ -934,7 +934,7 @@ int libxl__confirm_device_model_startup(
{
int detach;
int problem = libxl__wait_for_device_model(gc, starting->domid, "running",
- starting, NULL, NULL);
+ starting->for_spawn, NULL,
NULL);
detach = detach_device_model(gc, starting);
return problem ? problem : detach;
}
diff -r 21db7a7dd184 -r eb9330c89fd3 tools/libxl/libxl_internal.h
--- a/tools/libxl/libxl_internal.h Tue Sep 20 16:50:44 2011 +0100
+++ b/tools/libxl/libxl_internal.h Wed Sep 21 09:55:12 2011 +0100
@@ -288,7 +288,7 @@ _hidden int libxl__confirm_device_model_
_hidden int libxl__detach_device_model(libxl__gc *gc,
libxl__device_model_starting *starting);
_hidden int libxl__wait_for_device_model(libxl__gc *gc,
uint32_t domid, char *state,
- libxl__device_model_starting *starting,
+ libxl__spawn_starting *spawning,
int (*check_callback)(libxl__gc *gc,
uint32_t domid,
const char *state,
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
 |
Lists.xenproject.org is hosted with RackSpace, monitoring our |