[kvm-unit-tests,2/2] powerpc: Add Special Purpose Register persistency test

Submitted by Cédric Le Goater on March 10, 2017, 8:53 a.m.

Details

Message ID 5d28fcc2-8a6d-f975-5f35-d0dc7dc044da@kaod.org
State New
Headers show

Commit Message

Cédric Le Goater March 10, 2017, 8:53 a.m.
On 03/10/2017 09:48 AM, Cédric Le Goater wrote:
> On 03/09/2017 06:27 PM, Thomas Huth wrote:
>> This test has two purposes: First, check whether the hypervisor can be
>> destabilized by writing random values into the SPRs of the PowerPC CPU
>> (this indeed revealed a bug last year, see CVE-2016-3044).
>> Second, this test can be used to check whether the SPRs are synchronized
>> properly between the KVM host CPU and QEMU, e.g. when migrating the VM
>> from one QEMU instance to another.
>> The test first fills the various SPRs with some non-zero value, then reads
>> the values back into a first array. It then either sleeps a short period
>> of time (for testing without migration, in the hope that we're rescheduled
>> on another host CPU), or it waits for a key or NMI (with the '-w' option)
>> so that it is possible to migrate the VM before continuing. The test then
>> finally reads the values from the SPRs back into another array and then
>> compares them with the initial values.
>> Currently the test only supports the SPRs from the PowerISA v2.01
>> (PowerPC 970) and PowerISA v2.07 specification (i.e. POWER8 CPUs),
>> but other versions should be pretty easy to add later.
>>
>> Signed-off-by: Thomas Huth <thuth@redhat.com>
> 
> It looks good to me. I gave it a try and it worked fine with some
> extra tuning in the migration script.  
> 
> LDFLAGS needs fix for binutils 2.28 but that is another issue.

An here's the "fix" for it. I haven't studied deeply the question 
though.  

Thanks,

C.

From 820e2d614d63bf1cd911d13396b44a5ff8fb7bdc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?C=C3=A9dric=20Le=20Goater?= <clg@kaod.org>
Date: Fri, 10 Mar 2017 08:54:26 +0100
Subject: [PATCH] powerpc: add -N to LDFLAGS
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

ld 2.28 complains with :

ld -EL -nostdlib -pie -o powerpc/selftest.elf \
	-T powerpc/flat.lds --build-id=none \
	powerpc/selftest.o powerpc/cstart64.o powerpc/reloc64.o lib/libcflat.a lib/libfdt/libfdt.a powerpc/selftest.aux.o
ld: powerpc/selftest.elf: Not enough room for program headers, try linking with -N
ld: final link failed: Bad value

Signed-off-by: Cédric Le Goater <clg@kaod.org>
---
 powerpc/Makefile.common | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Thomas Huth March 10, 2017, 9:38 a.m.
On 10.03.2017 09:53, Cédric Le Goater wrote:
> On 03/10/2017 09:48 AM, Cédric Le Goater wrote:
>> On 03/09/2017 06:27 PM, Thomas Huth wrote:
>>> This test has two purposes: First, check whether the hypervisor can be
>>> destabilized by writing random values into the SPRs of the PowerPC CPU
>>> (this indeed revealed a bug last year, see CVE-2016-3044).
>>> Second, this test can be used to check whether the SPRs are synchronized
>>> properly between the KVM host CPU and QEMU, e.g. when migrating the VM
>>> from one QEMU instance to another.
>>> The test first fills the various SPRs with some non-zero value, then reads
>>> the values back into a first array. It then either sleeps a short period
>>> of time (for testing without migration, in the hope that we're rescheduled
>>> on another host CPU), or it waits for a key or NMI (with the '-w' option)
>>> so that it is possible to migrate the VM before continuing. The test then
>>> finally reads the values from the SPRs back into another array and then
>>> compares them with the initial values.
>>> Currently the test only supports the SPRs from the PowerISA v2.01
>>> (PowerPC 970) and PowerISA v2.07 specification (i.e. POWER8 CPUs),
>>> but other versions should be pretty easy to add later.
>>>
>>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>>
>> It looks good to me. I gave it a try and it worked fine with some
>> extra tuning in the migration script.  
>>
>> LDFLAGS needs fix for binutils 2.28 but that is another issue.
> 
> An here's the "fix" for it. I haven't studied deeply the question 
> though.  

I think this could be an explanation for this issue:
https://lists.gnu.org/archive/html/bug-gnu-utils/2002-08/msg00242.html

Could you please check whether it works with "-n" instead of "-N", too?
If so, I think that would be the better option.

 Thanks,
  Thomas

> From 820e2d614d63bf1cd911d13396b44a5ff8fb7bdc Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?C=C3=A9dric=20Le=20Goater?= <clg@kaod.org>
> Date: Fri, 10 Mar 2017 08:54:26 +0100
> Subject: [PATCH] powerpc: add -N to LDFLAGS
> MIME-Version: 1.0
> Content-Type: text/plain; charset=UTF-8
> Content-Transfer-Encoding: 8bit
> 
> ld 2.28 complains with :
> 
> ld -EL -nostdlib -pie -o powerpc/selftest.elf \
> 	-T powerpc/flat.lds --build-id=none \
> 	powerpc/selftest.o powerpc/cstart64.o powerpc/reloc64.o lib/libcflat.a lib/libfdt/libfdt.a powerpc/selftest.aux.o
> ld: powerpc/selftest.elf: Not enough room for program headers, try linking with -N
> ld: final link failed: Bad value
> 
> Signed-off-by: Cédric Le Goater <clg@kaod.org>
> ---
>  powerpc/Makefile.common | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/powerpc/Makefile.common b/powerpc/Makefile.common
> index 92809a5468ea..6c62ac85311a 100644
> --- a/powerpc/Makefile.common
> +++ b/powerpc/Makefile.common
> @@ -43,7 +43,7 @@ cflatobjs += lib/powerpc/smp.o
>  
>  FLATLIBS = $(libcflat) $(LIBFDT_archive)
>  %.elf: CFLAGS += $(arch_CFLAGS)
> -%.elf: LDFLAGS = $(arch_LDFLAGS) -nostdlib -pie
> +%.elf: LDFLAGS = $(arch_LDFLAGS) -nostdlib -pie -N
>  %.elf: %.o $(FLATLIBS) powerpc/flat.lds $(cstart.o) $(reloc.o)
>  	$(CC) $(CFLAGS) -c -o $(@:.elf=.aux.o) lib/auxinfo.c -DPROGNAME=\"$@\"
>  	$(LD) $(LDFLAGS) -o $@ \
> 

--
To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Cédric Le Goater March 10, 2017, 10:27 a.m.
On 03/10/2017 10:38 AM, Thomas Huth wrote:
> On 10.03.2017 09:53, Cédric Le Goater wrote:
>> On 03/10/2017 09:48 AM, Cédric Le Goater wrote:
>>> On 03/09/2017 06:27 PM, Thomas Huth wrote:
>>>> This test has two purposes: First, check whether the hypervisor can be
>>>> destabilized by writing random values into the SPRs of the PowerPC CPU
>>>> (this indeed revealed a bug last year, see CVE-2016-3044).
>>>> Second, this test can be used to check whether the SPRs are synchronized
>>>> properly between the KVM host CPU and QEMU, e.g. when migrating the VM
>>>> from one QEMU instance to another.
>>>> The test first fills the various SPRs with some non-zero value, then reads
>>>> the values back into a first array. It then either sleeps a short period
>>>> of time (for testing without migration, in the hope that we're rescheduled
>>>> on another host CPU), or it waits for a key or NMI (with the '-w' option)
>>>> so that it is possible to migrate the VM before continuing. The test then
>>>> finally reads the values from the SPRs back into another array and then
>>>> compares them with the initial values.
>>>> Currently the test only supports the SPRs from the PowerISA v2.01
>>>> (PowerPC 970) and PowerISA v2.07 specification (i.e. POWER8 CPUs),
>>>> but other versions should be pretty easy to add later.
>>>>
>>>> Signed-off-by: Thomas Huth <thuth@redhat.com>
>>>
>>> It looks good to me. I gave it a try and it worked fine with some
>>> extra tuning in the migration script.  
>>>
>>> LDFLAGS needs fix for binutils 2.28 but that is another issue.
>>
>> An here's the "fix" for it. I haven't studied deeply the question 
>> though.  
> 
> I think this could be an explanation for this issue:
> https://lists.gnu.org/archive/html/bug-gnu-utils/2002-08/msg00242.html
> 
> Could you please check whether it works with "-n" instead of "-N", too?
> If so, I think that would be the better option.


It works. I will send the patch

Thanks,

C. 

--
To unsubscribe from this list: send the line "unsubscribe kvm-ppc" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch hide | download patch | download mbox

diff --git a/powerpc/Makefile.common b/powerpc/Makefile.common
index 92809a5468ea..6c62ac85311a 100644
--- a/powerpc/Makefile.common
+++ b/powerpc/Makefile.common
@@ -43,7 +43,7 @@  cflatobjs += lib/powerpc/smp.o
 
 FLATLIBS = $(libcflat) $(LIBFDT_archive)
 %.elf: CFLAGS += $(arch_CFLAGS)
-%.elf: LDFLAGS = $(arch_LDFLAGS) -nostdlib -pie
+%.elf: LDFLAGS = $(arch_LDFLAGS) -nostdlib -pie -N
 %.elf: %.o $(FLATLIBS) powerpc/flat.lds $(cstart.o) $(reloc.o)
 	$(CC) $(CFLAGS) -c -o $(@:.elf=.aux.o) lib/auxinfo.c -DPROGNAME=\"$@\"
 	$(LD) $(LDFLAGS) -o $@ \