gmane.comp.emulators.xen.devel

[xen-unstable test] 12975: tolerable FAIL

xen.org — 2012-05-26T06:32:33

flight 12975 xen-unstable real [real]
http://www.chiark.greenend.org.uk/~xensrcts/logs/12975/

Failures :-/ but no regressions.

Regressions which are regarded as allowable (not blocking):
 test-amd64-amd64-xl-qemuu-win7-amd64 12 guest-localmigrate/x10 fail like 12972

Tests which did not succeed, but are not blocking:
 test-amd64-amd64-xl-pcipt-intel  9 guest-start                 fail never pass
 test-amd64-i386-rhel6hvm-intel 11 leak-check/check             fail never pass
 test-amd64-i386-qemuu-rhel6hvm-amd 11 leak-check/check         fail never pass
 test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check       fail never pass
 test-amd64-i386-xend-winxpsp3 16 leak-check/check             fail  never pass
 test-amd64-amd64-win         16 leak-check/check             fail   never pass
 test-amd64-i386-xl-win-vcpus1 13 guest-stop                   fail  never pass
 test-i386-i386-xl-qemuu-winxpsp3 13 guest-stop                 fail never pass
 test-amd64-i386-win-vcpus1   16 leak-check/check

[xen-unstable test] 12972: tolerable FAIL - PUSHED

xen.org — 2012-05-26T00:23:06

flight 12972 xen-unstable real [real]
http://www.chiark.greenend.org.uk/~xensrcts/logs/12972/

Failures :-/ but no regressions.

Tests which are failing intermittently (not blocking):
 test-amd64-i386-rhel6hvm-amd  7 redhat-install              fail pass in 12971
 test-amd64-amd64-pv           9 guest-start        fail in 12971 pass in 12972

Regressions which are regarded as allowable (not blocking):
 test-amd64-amd64-xl-qemuu-win7-amd64 12 guest-localmigrate/x10 fail like 12968

Tests which did not succeed, but are not blocking:
 test-amd64-amd64-xl-pcipt-intel  9 guest-start                 fail never pass
 test-amd64-i386-rhel6hvm-intel 11 leak-check/check             fail never pass
 test-amd64-i386-qemuu-rhel6hvm-amd 11 leak-check/check         fail never pass
 test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check       fail never pass
 test-amd64-i386-xend-winxpsp3 16 leak-check/check             fail  never pass
 test-amd64-amd64-win         16 leak-check/check             fail   never pass
 test-

Xen Document Day: May 28th

Lars Kurth — 2012-05-25T21:13:23

Hi,

everybody. A quick reminder that the next Xen DocumentDayis happening 
next Monday. More info on document days at 
http://wiki.xen.org/wiki/Xen_Document_Days

Hope to see you on IRC! Feel free to add stuff to the TODO list 
(http://wiki.xen.org/wiki/Xen_Document_Days/TODO) or put your name 
besides an item if you intend to work on it.

Best Regards
Lars

*********************
* Xen Document Days *
*********************

We have another Xen document day come up next Monday. Xen Document Days 
are for people who care about Xen Documentation and want to improve it. 
We introduced Documentation Days, because working on documentation in 
parallel with like minded-people, is just more fun than working alone! 
Everybody who can contribute is welcome to join!

For a list of items that need work, check out the community maintained 
TODO list (http://wiki.xen.org/wiki/Xen_Document_Days/TODO 
<http://wiki.xen.org/wiki/Xen_Document_Days/TODO>). Of course, you can 
work on anything you like: the list just provides s

Xen Document Day: May 28th

Lars Kurth — 2012-05-25T20:36:05

Hi,

everybody. A quick reminder that the next Xen Document Day is happening
next Monday. More info on document days at
http://wiki.xen.org/wiki/Xen_Document_Days

Hope to see you on IRC! Feel free to add stuff to the TODO list (
http://wiki.xen.org/wiki/Xen_Document_Days/TODO) or put your name besides
an item if you intend to work on it.

Best Regards
Lars

*********************
* Xen Document Days *
*********************

We have another Xen document day come up next Monday. Xen Document Days are
for people who care about Xen Documentation and want to improve it. We
introduced Documentation Days, because working on documentation in parallel
with like minded-people, is just more fun than working alone! Everybody who
can contribute is welcome to join!

For a list of items that need work, check out the community maintained TODO
list (http://wiki.xen.org/wiki/Xen_Document_Days/TODO<http://wiki.xen.org/wiki/Xen_Document_Days/TODO>).
Of course, you can work on anything you like: the list just provides
suggestion

Compile error for libxl on centos 5

Andres Lagar-Cavilla — 2012-05-25T20:27:31

I am getting libxenlight.so and xl link failures, complaining about
missing login_tty and opentty symbols. The patch below fixes this for me,
in a quick & dirty manner.

FYI, not proposing this for the tree. I'm not sure to which extent you
want to support an oldish distro, whether libutil is universally necessary
and/or available.

Andres

diff -r 69c3ae25bb1d tools/libxl/Makefile
--- a/tools/libxl/Makefile
+++ b/tools/libxl/Makefile
< at >< at > -60,7 +60,7 < at >< at > ifeq ($(BISON),)
          scanners, please install it an rerun configure)
 endif

-LIBXL_LIBS += -lyajl
+LIBXL_LIBS += -lyajl -lutil

 LIBXL_OBJS = flexarray.o libxl.o libxl_create.o libxl_dm.o libxl_pci.o \
            libxl_dom.o libxl_exec.o libxl_xshelp.o libxl_device.o \
< at >< at > -154,7 +154,7 < at >< at > libxlutil.a: $(LIBXLU_OBJS)
    $(AR) rcs libxlutil.a $^

 xl: $(XL_OBJS) libxlutil.so libxenlight.so
-   $(CC) $(LDFLAGS) -o $< at > $(XL_OBJS) libxlutil.so $(LDLIBS_libxenlight)
$(LDLIBS_libxenctrl) -lyajl $(APPEND_LDFLAGS)
+   $(CC) $(LDFLAGS) -o $< at > $(XL_OBJS) libxlutil

[xen-unstable test] 12971: regressions - FAIL

xen.org — 2012-05-25T19:30:39

flight 12971 xen-unstable real [real]
http://www.chiark.greenend.org.uk/~xensrcts/logs/12971/

Regressions :-(

Tests which did not succeed and are blocking,
including tests which could not be run:
 test-amd64-amd64-pv           9 guest-start               fail REGR. vs. 12969

Tests which did not succeed, but are not blocking:
 test-amd64-amd64-xl-pcipt-intel  9 guest-start                 fail never pass
 test-amd64-i386-rhel6hvm-intel 11 leak-check/check             fail never pass
 test-amd64-i386-qemuu-rhel6hvm-amd 11 leak-check/check         fail never pass
 test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check       fail never pass
 test-amd64-i386-xend-winxpsp3 16 leak-check/check             fail  never pass
 test-amd64-amd64-win         16 leak-check/check             fail   never pass
 test-amd64-i386-xl-win-vcpus1 13 guest-stop                   fail  never pass
 test-i386-i386-xl-qemuu-winxpsp3 13 guest-stop                 fail never pass
 test-amd64-i386-win-vcpus1   16 leak-check/check

[PATCH] libxl: When checking BDF of existing slots, function should be decimal, not hex

George Dunlap — 2012-05-25T16:34:17

# HG changeset patch
# User George Dunlap <george.dunlap< at >eu.citrix.com>
# Date 1337961666 0
# Node ID 1c28051020488782f1277dd60a2418324580297e
# Parent  69c3ae25bb1ddcb0ea44b7566d36d34e9d6a70aa
libxl: When checking BDF of existing slots, function should be decimal, not hex

Spotted-by: Konrad Wilk <konrad.wilk< at >oracle.com>
Signed-off-by: George Dunlap <george.dunlap< at >eu.citrix.com>

diff --git a/tools/libxl/libxl_pci.c b/tools/libxl/libxl_pci.c
--- a/tools/libxl/libxl_pci.c
+++ b/tools/libxl/libxl_pci.c
< at >< at > -480,7 +480,7 < at >< at > static int pciback_dev_has_slot(libxl__g
         return ERROR_FAIL;
     }
 
-    while(fscanf(f, "%x:%x:%x.%x\n", &dom, &bus, &dev, &func)==4) {
+    while(fscanf(f, "%x:%x:%x.%d\n", &dom, &bus, &dev, &func)==4) {
         if(dom == pcidev->domain
            && bus == pcidev->bus
            && dev == pcidev->dev

[PATCH v4 0/6] xen/arm: event channels and shared_infopage

Stefano Stabellini — 2012-05-25T16:21:43

Hi all,
this patch series implements support for injecting event channels into
the guest and enables a wider range of hypercalls for ARM guests.

In order to allow more flexibility I modified the hypercall protocol, in
particular the hypercall number is not passed as imm to hvc anymore,
because we might not always know it at compile time.
The hypercall number is now passed on the r12 register.

With this patch series and using the following Linux tree:

git://xenbits.xen.org/people/sstabellini/linux-pvhvm.git vexpress-dt-privcmd-2

I am able to boot dom0, start xenstored and run basic xl commands, like
"xl list" and "xl uptime".
I have added at the beginning of this series few patches that were
previously sent separately (in particular "arm: support fewer LR
registers than virtual irqs").


Changes in v4:

- drop all the patches that have already been committed;

- rebase on "amd iommu: improve parse_event_log_entry()";

- move the initialization of gic.lr_pending and gic.lr_mask to gic_init;

- pass 0 as me

[xen-4.1-testing test] 12970: tolerable FAIL - PUSHED

xen.org — 2012-05-25T15:27:42

flight 12970 xen-4.1-testing real [real]
http://www.chiark.greenend.org.uk/~xensrcts/logs/12970/

Failures :-/ but no regressions.

Regressions which are regarded as allowable (not blocking):
 test-amd64-amd64-xl-pcipt-intel  8 debian-fixup           fail REGR. vs. 12886
 test-amd64-amd64-xl-sedf      5 xen-boot                     fail   like 12886
 test-amd64-i386-xl-winxpsp3-vcpus1  7 windows-install     fail REGR. vs. 12886

Tests which did not succeed, but are not blocking:
 test-amd64-i386-qemuu-rhel6hvm-amd 11 leak-check/check         fail never pass
 test-amd64-i386-rhel6hvm-amd 11 leak-check/check             fail   never pass
 test-amd64-i386-rhel6hvm-intel 11 leak-check/check             fail never pass
 test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check       fail never pass
 test-amd64-amd64-xl-qemuu-winxpsp3 13 guest-stop               fail never pass
 test-amd64-amd64-win         16 leak-check/check             fail   never pass
 test-amd64-amd64-xl-win      13 guest-stop

kernel 3.4.0 dom0 oops

Andrei Popa — 2012-05-23T17:15:43

Hello,

With kernel 3.4.0 as a dom0 and a 3.3.6 as a domU, i get the following
oops on the dom0 kernel when I rsync  / to a filesystem mounted on a
loop device:
http://77.36.72.222/P5230845.JPG

The domU config:
ierdnac-hp ~ # cat /mnt/gentoo/xen 
kernel = "/boot/vmlinuz-3.3.6-pf-c241.old"
name="gentoo"
memory = 2048
vcpus = 1
pae  = 1
acpi = 1
apic = 1
stdvga=0
serial='pty'
usbdevice='tablet'
on_poweroff = 'destroy'
on_reboot   = 'restart'
on_crash    = 'restart'
vnc=1
vncviewer=1
vncconsole=1
sdl=0
vfb = [ 'type=vnc' ]
disk   = ['file:/mnt/gentoo/ierdnac-hp.bak,sdd1,w']
root   = "root=/dev/xvdd1 ro"

If you are interested in the kernels .config file I can provide them.

The Xenoprof results

suixiufeng — 2012-05-25T13:31:31

Hi, All:
The dom0 kernel is centos5.5, and the Hypervisor is xen-4.1.2. The domU
kernel is centos5.5, which is installed through the virt-install command (I
think it belongs to HVM).
Then I use xenoprof to collect the cycle, retired instruction number, LLC
access number and LLC miss number information when running spec cpu 2006
benchmarks.

*I collected data for the whole system, and I only compared the speccpu
process. For xenoprof, I use active domain method, and collect data on both
dom0 and domU. But only domU has the performance information about speccpu
in VM.*
*Furthermore, there are a 30x difference with regard to the  retired
instruction number which are shown in the following. I wonder what is the
reason? I think it is due to xen virtulization. Thank you!*

PM

Cycles

Insts

MISSES

REFS

CPI

MissRate

gromacs_base.i386

6198401

4458151

106

8865

1.390352413

0.011957135

sjeng_base.i386

2848717

3031079

911

2724

0.939835946

33.4%

VM

gromacs_base.i386

295398087

151265683

33934

32949

[PATCH 3/3] gnttab: cleanup

Jan Beulich — 2012-05-25T12:52:17

- introduce local variables (shortcuts for frequently used <dom>->grant_table)
- adjust first parameter of mapcount()
- drop lock acquisition from gnttab_get_version()
- remove hard tabs and adjust formatting

Signed-off-by: Jan Beulich <jbeulich< at >suse.com>
Tested-by: Andrew Thomas <andrew.thomas< at >oracle.com>

--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
< at >< at > -444,18 +444,17 < at >< at > static int _set_status(unsigned gt_versi
 }
 
 static void mapcount(
-    struct domain *ld, struct domain *rd, unsigned long mfn,
+    struct grant_table *lgt, struct domain *rd, unsigned long mfn,
     unsigned int *wrc, unsigned int *rdc)
 {
-    struct grant_table *gt = ld->grant_table;
     struct grant_mapping *map;
     grant_handle_t handle;
 
     *wrc = *rdc = 0;
 
-    for ( handle = 0; handle < gt->maptrack_limit; handle++ )
+    for ( handle = 0; handle < lgt->maptrack_limit; handle++ )
     {
-        map = &maptrack_entry(gt, handle);
+        map = &maptrack_entry(lgt, handle);
         if ( !(map->flags &

[PATCH 2/3] gnttab: mark maptrack free list tail with an explicit terminator

Jan Beulich — 2012-05-25T12:51:30

... instead of using the mutable current limit.

This also addresses an apparent off-by-one mistake when checking for
exhaustion of the maptrack table.

Signed-off-by: Jan Beulich <jbeulich< at >suse.com>
Tested-by: Andrew Thomas <andrew.thomas< at >oracle.com>

--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
< at >< at > -107,6 +107,8 < at >< at > static unsigned inline int max_nr_maptra
     return (max_nr_grant_frames * MAX_MAPTRACK_TO_GRANTS_RATIO);
 }
 
+#define MAPTRACK_TAIL (~0u)
+
 #define SHGNT_PER_PAGE_V1 (PAGE_SIZE / sizeof(grant_entry_v1_t))
 #define shared_entry_v1(t, e) \
     ((t)->shared_v1[(e)/SHGNT_PER_PAGE_V1][(e)%SHGNT_PER_PAGE_V1])
< at >< at > -198,7 +200,7 < at >< at > __get_maptrack_handle(
     struct grant_table *t)
 {
     unsigned int h;
-    if ( unlikely((h = t->maptrack_head) == (t->maptrack_limit - 1)) )
+    if ( unlikely((h = t->maptrack_head) == MAPTRACK_TAIL) )
         return -1;
     t->maptrack_head = maptrack_entry(t, h).ref;
     return h;
< at >< at > -239,8 +241,10 < at >< at > get_maptrack_handle(
 
         new_mt_limi

[PATCH 1/3] gnttab: don't use domain lock forserialization

Jan Beulich — 2012-05-25T12:50:41

Instead use the affected domain's grant table lock, at once reducing
the scopes during which locks are being held and hence allowing
significantly better parallelism.

Signed-off-by: Jan Beulich <jbeulich< at >suse.com>
Tested-by: Andrew Thomas <andrew.thomas< at >oracle.com>

--- a/xen/arch/x86/mm.c
+++ b/xen/arch/x86/mm.c
< at >< at > -3727,8 +3727,6 < at >< at > static int create_grant_pte_mapping(
     l1_pgentry_t ol1e;
     struct domain *d = v->domain;
 
-    ASSERT(domain_is_locked(d));
-
     adjust_guest_l1e(nl1e, d);
 
     gmfn = pte_addr >> PAGE_SHIFT;
< at >< at > -3855,8 +3853,6 < at >< at > static int create_grant_va_mapping(
     struct page_info *l1pg;
     int okay;
     
-    ASSERT(domain_is_locked(d));
-
     adjust_guest_l1e(nl1e, d);
 
     pl1e = guest_map_l1e(v, va, &gl1mfn);
--- a/xen/common/grant_table.c
+++ b/xen/common/grant_table.c
< at >< at > -169,6 +169,30 < at >< at > static int __get_paged_frame(unsigned lo
     return rc;
 }
 
+static inline void
+double_gt_lock(struct grant_table *lgt, struct grant_table *rgt)
+{
+    if ( lgt < rgt )
+

[PATCH 0/3] gnttab: drop use of domain lock and cleanup

Jan Beulich — 2012-05-25T12:49:55

These three patches aren't really related to one another, but need
to be applied in order.

1: gnttab: don't use domain lock for serialization
2: gnttab: mark maptrack free list tail with an explicit terminator
3: gnttab: cleanup

Signed-off-by: Jan Beulich <jbeulich< at >suse.com>
Tested-by: Andrew Thomas <andrew.thomas< at >oracle.com>

[PATCH] libxl: introduce libxl_vcpuinfo_list_free

Dario Faggioli — 2012-05-25T09:44:29

And fix a leak due to it being missing.

Signed-off-by: Dario Faggioli <dario.faggioli< at >citrix.com>

diff --git a/tools/libxl/libxl.h b/tools/libxl/libxl.h
--- a/tools/libxl/libxl.h
+++ b/tools/libxl/libxl.h
< at >< at > -797,6 +797,7 < at >< at > libxl_cputopology *libxl_get_cpu_topolog
 void libxl_cputopology_list_free(libxl_cputopology *, int nr);
 libxl_vcpuinfo *libxl_list_vcpu(libxl_ctx *ctx, uint32_t domid,
                                        int *nb_vcpu, int *nrcpus);
+void libxl_vcpuinfo_list_free(libxl_vcpuinfo *, int nr);
 int libxl_set_vcpuaffinity(libxl_ctx *ctx, uint32_t domid, uint32_t vcpuid,
                            libxl_cpumap *cpumap);
 int libxl_set_vcpuaffinity_all(libxl_ctx *ctx, uint32_t domid,
diff --git a/tools/libxl/libxl_utils.c b/tools/libxl/libxl_utils.c
--- a/tools/libxl/libxl_utils.c
+++ b/tools/libxl/libxl_utils.c
< at >< at > -558,6 +558,14 < at >< at > void libxl_cputopology_list_free(libxl_c
     free(list);
 }
 
+void libxl_vcpuinfo_list_free(libxl_vcpuinfo *list, int nr)
+{
+    int i;
+    for (i = 0;

[PATCH] libxc: do not "panic" if a kernel is not abzImage

Ian Campbell — 2012-05-25T09:24:52

# HG changeset patch
# User Ian Campbell <ian.campbell< at >citrix.com>
# Date 1337937866 -3600
# Node ID 42b4ca522057ecf3bf48250bf964d12e1a9e865c
# Parent  b68bb5eb0503bcec053de4422212a33fa6be9698
libxc: do not "panic" if a kernel is not a bzImage.

Up until the point where we think this is a bzImage there is no point in
printing panicy messages -- some other loader will have a go (probably the
compressed ELF one)

Signed-off-by: Ian Campbell <ian.campbell< at >citrix.com>

diff -r b68bb5eb0503 -r 42b4ca522057 tools/libxc/xc_dom_bzimageloader.c
--- a/tools/libxc/xc_dom_bzimageloader.cFri May 25 10:19:45 2012 +0100
+++ b/tools/libxc/xc_dom_bzimageloader.cFri May 25 10:24:26 2012 +0100
< at >< at > -575,8 +575,7 < at >< at > static int xc_dom_probe_bzimage_kernel(s
 
     if ( dom->kernel_size < sizeof(struct setup_header) )
     {
-        xc_dom_panic(dom->xch, XC_INTERNAL_ERROR,
-                     "%s: kernel image too small", __FUNCTION__);
+        xc_dom_printf(dom->xch, "%s: kernel image too small", __FUNCTION__);
         ret

[xen-unstable test] 12969: tolerable FAIL

xen.org — 2012-05-25T06:31:09

flight 12969 xen-unstable real [real]
http://www.chiark.greenend.org.uk/~xensrcts/logs/12969/

Failures :-/ but no regressions.

Tests which are failing intermittently (not blocking):
 test-amd64-amd64-xl-qemuu-win7-amd64  7 windows-install     fail pass in 12968
 test-amd64-amd64-xl-sedf     10 guest-saverestore  fail in 12968 pass in 12969
 test-amd64-amd64-xl-qemuu-win7-amd64 12 guest-localmigrate/x10 fail in 12968 pass in 12967

Tests which did not succeed, but are not blocking:
 test-amd64-amd64-xl-pcipt-intel  9 guest-start                 fail never pass
 test-amd64-i386-rhel6hvm-intel 11 leak-check/check             fail never pass
 test-amd64-i386-qemuu-rhel6hvm-amd 11 leak-check/check         fail never pass
 test-amd64-i386-qemuu-rhel6hvm-intel 11 leak-check/check       fail never pass
 test-amd64-i386-xend-winxpsp3 16 leak-check/check             fail  never pass
 test-i386-i386-xl-qemuu-winxpsp3 13 guest-stop                 fail never pass
 test-amd64-amd64-win         16 leak-check/check

[PATCH 3/3] libxc: add instruction length for injecttrap in userspace

Xudong Hao — 2012-05-24T19:07:34

Add a parameter to represent instruction length in function
xc_hvm_inject_trap(), user should set this value when this
function is called.

Signed-off-by: Xudong Hao <xudong.hao< at >intel.com>
Signed-off-by: Xiantao Zhang <xiantao.zhang< at >intel.com>
---
 tools/libxc/xc_misc.c               |    5 +++--
 tools/libxc/xenctrl.h               |    4 ++--
 tools/tests/xen-access/xen-access.c |    2 +-
 3 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/tools/libxc/xc_misc.c b/tools/libxc/xc_misc.c
index a029d79..2989a62 100644
--- a/tools/libxc/xc_misc.c
+++ b/tools/libxc/xc_misc.c
< at >< at > -602,8 +602,8 < at >< at > int xc_hvm_get_mem_access(
 }
 
 int xc_hvm_inject_trap(
-    xc_interface *xch, domid_t dom, int vcpu, uint32_t trap, uint32_t error_code, 
-    uint64_t cr2)
+    xc_interface *xch, domid_t dom, int vcpu, uint32_t trap, uint32_t inslen, 
+    uint32_t error_code, uint64_t cr2)
 {
     DECLARE_HYPERCALL;
     DECLARE_HYPERCALL_BUFFER(struct xen_hvm_inject_trap, arg);
< at >< at > -619,6 +619,7 < at >< at > int xc_hvm_inject_trap

[PATCH 0/3] XEN: fix vmx exception mistake

Xudong Hao — 2012-05-24T19:06:23

This series of patches fix the mistake for debug exception(#DB), overflow
exception(#OF) and INT3(#BP), INTn instruction emulation. 

Introduce new function vmx_inject_sw_exception() which deliver the software
excetion, software interrupt and privileged software exception. Split hardware
exception as a seperate function(old function vmx_inject_hw_exception()).

Also Passed down intruction length to exception emulation handler function.

And supply a interface for userspace to inject trap.


PATCH 1: Pass the instruction length field down to exception emulation,
  by changing hypercall parameter and adding a parameter in fucntion
  hvm_inject_exception(), so that exception emulation can set correct
  instruction length.

PATCH 2: Fix the mistake for debug exception(#DB), overflow exception(#OF) and
  INT3(#BP), INTn instruction emulation.

  Introduce new function vmx_inject_sw_exception() which deliver the software
  excetion, software interrupt and privileged software exception. Split hardware
  exception a

PV multiconsole bug during resume.

Konrad Rzeszutek Wilk — 2012-05-24T19:37:14

So .. we used to have in the event.c a spin_lock to protect the
irq_mapping_update_lock, but with git commit  773659483685d652970583384a0294948e57f8b3
"xen/irq: Alter the locking to use a mutex instead of a spinlock."
I changed it to a mutex b/c we keept on getting WARNs.

But now I get this when I resume a PVHVM guest:

Grant tables using version 2 layout.
BUG: sleeping function called from invalid context at /home/konrad/ssd/linux/kernel/mutex.c:85
in_atomic(): 1, irqs_disabled(): 1, pid: 6, name: migration/0
Pid: 6, comm: migration/0 Tainted: G           O 3.4.0upstream-00113-g598ff45-dirty #1
Call Trace:
 [<ffffffff8109830a>] __might_sleep+0xda/0x100
 [<ffffffff815a47f7>] mutex_lock+0x27/0x50
 [<ffffffff81311ea6>] rebind_evtchn_irq+0x36/0x90
 [<ffffffff81341bfc>] xen_console_resume+0x5c/0x60
 [<ffffffff81313fea>] xen_suspend+0x8a/0xb0
 [<ffffffff810d42f3>] stop_machine_cpu_stop+0xa3/0xf0
 [<ffffffff810d4250>] ? stop_one_cpu_nowait+0x50/0x50
 [<ffffffff810d3f81>] cpu_stopper_thread+0xf1/0x1c0
 [<ffffffff8