This is a variant of SandyBridge with indirect branch prediction
protection. The only difference between SandyBridge and SandyBridge-IBRS
is the added "spec-ctrl" feature.
The SandyBridge-IBRS model in QEMU is a bit different since SandyBridge
got several additional features since we added it in
The CPU contains the updated microcode for CVE-2017-5715.
Signed-off-by: Jiri Denemark
---
tests/cputest.c| 1 +
...86_64-cpuid-EPYC-7601-32-Core-ibpb-disabled.xml | 7 +
...x86_64-cpuid-EPYC-7601-32-Core-ibpb-enabled.xml | 9 +
This is a variant of Nehalem with indirect branch prediction protection.
The only difference between Nehalem and Nehalem-IBRS is the added
"spec-ctrl" feature.
Thus the diff matches QEMU, but the new CPU model itself is different.
The QEMU's versions of both models contain "vme" feature, while
This is a variant of EPYC with indirect branch prediction protection.
The only difference between EPYC and EPYC-IBPB is the added "ibpb"
feature.
Signed-off-by: Jiri Denemark
---
src/cpu/cpu_map.xml| 72 ++
This is a variant of Westmere with indirect branch prediction
protection. The only difference between Westmere and Westmere-IBRS is
the added "spec-ctrl" feature.
The Westmere-IBRS model in QEMU is a bit different since Westmere got
several additional features since we added it in cpu_map.xml:
This is a variant of Broadwell-noTSX with indirect branch prediction
protection. The only difference between Broadwell-noTSX and
Broadwell-noTSX-IBRS is the added "spec-ctrl" feature.
The Broadwell-noTSX-IBRS model in QEMU is a bit different since
Broadwell-noTSX got several additional features
This is a variant of Skylake-Client with indirect branch prediction
protection. The only difference between Skylake-Client and
Skylake-Client-IBRS is the added "spec-ctrl" feature.
Signed-off-by: Jiri Denemark
---
src/cpu/cpu_map.xml| 69
This is a variant of IvyBridge with indirect branch prediction
protection. The only difference between IvyBridge and IvyBridge-IBRS is
the added "spec-ctrl" feature.
The IvyBridge-IBRS model in QEMU is a bit different since IvyBridge got
several additional features since we added it in
The CPU contains the updated microcode for CVE-2017-5715.
Signed-off-by: Jiri Denemark
---
tests/cputest.c| 1 +
.../x86_64-cpuid-Xeon-Gold-5115-disabled.xml | 8 +
.../x86_64-cpuid-Xeon-Gold-5115-enabled.xml| 8 +
This is a variant of Haswell-noTSX with indirect branch prediction
protection. The only difference between Haswell-noTSX and
Haswell-noTSX-IBRS is the added "spec-ctrl" feature.
The Haswell-noTSX-IBRS model in QEMU is a bit different since
Haswell-noTSX got several additional features since we
This is a variant of Broadwell with indirect branch prediction
protection. The only difference between Broadwell and Broadwell-IBRS is
the added "spec-ctrl" feature.
The Broadwell-IBRS model in QEMU is a bit different since Broadwell got
several additional features since we added it in
This is a variant of Skylake-Server with indirect branch prediction
protection. The only difference between Skylake-Server and
Skylake-Server-IBRS is the added "spec-ctrl" feature.
Signed-off-by: Jiri Denemark
---
src/cpu/cpu_map.xml| 76
This is a variant of Haswell with indirect branch prediction protection.
The only difference between Haswell and Haswell-IBRS is the added
"spec-ctrl" feature.
The Haswell-IBRS model in QEMU is a bit different since Haswell got
several additional features since we added it in cpu_map.xml:
The CPU contains the updated microcode for CVE-2017-5715.
Signed-off-by: Jiri Denemark
---
tests/cputest.c| 1 +
.../x86_64-cpuid-Xeon-E5-2609-v3-disabled.xml | 6 +
.../x86_64-cpuid-Xeon-E5-2609-v3-enabled.xml | 8 +
The CPU contains the updated microcode for CVE-2017-5715.
Signed-off-by: Jiri Denemark
---
tests/cputest.c| 1 +
.../x86_64-cpuid-Core-i7-5600U-ibrs-disabled.xml | 6 +
.../x86_64-cpuid-Core-i7-5600U-ibrs-enabled.xml| 8 +
The CPU contains the updated microcode for CVE-2017-5715.
The *-guest.xml and *-json.xml CPU definitions use Skylake-Client CPU
model rather than Broadwell. This is similar to Xeon-E5-2650-v4 and it
is caused by our CPU model selection code when no model matches the CPU
signature (family +
From: Paolo Bonzini
Added in QEMU commits TBD and TBD.
Signed-off-by: Paolo Bonzini
Signed-off-by: Jiri Denemark
---
src/cpu/cpu_map.xml | 8
1 file changed, 8 insertions(+)
diff --git a/src/cpu/cpu_map.xml
This is the libvirt's part of the changes related to CVE-2017-5715. The
new models can be used to pass the protective CPU features to guests.
But remember, the host CPU microcode, host kernel, QEMU, and libvirt all
need to be updated for this to be any useful.
Based on a patch from Paolo Bonzini.
Prepare for hash table volume lists by creating the object infrastructure
for a Volume Object and Volume Object List
The _virStorageVolObj will contain just a pointer to the "current"
(and live) volume definition.
The _virStorageVolObjList will contain three hash tables, one for
each of the
Alter the logic such that we only add the volume to the pool once
we've filled in all the information and cause failure to go to a
common error: label. Patches to place the @vol into a few hash tables
will soon "require" that at least the keys (name, target.path, and key)
be populated with valid
This series converts the storage volume forward linked list into
an object that contains hash tables of volume objects.
The first two patches resolve issues seen during testing with
the Disk backend which had some special cased code to handle
the Create/Delete paths and interactions with the
Alter the volume logic to use the hash tables instead of forward
linked lists. There are three hash tables to allow for fast lookup
by name, target.path, and key.
Modify the virStoragePoolObjAddVol to place the object in all 3
tables if possible using self locking RWLock on the volumes object.
For a disk backend, the deleteVol code will clear all the
volumes in the pool and perform a pool refresh, thus the
storageVolDeleteInternal should not use access @voldef
after deleteVol succeeds.
Signed-off-by: John Ferlan
---
src/storage/storage_driver.c | 14 ++
On Tue, Jan 09, 2018 at 19:44:18 +0100, Kashyap Chamarthy wrote:
> On Tue, Jan 09, 2018 at 04:37:10PM +0100, Jiri Denemark wrote:
> > On Tue, Jan 09, 2018 at 15:32:54 +0100, Kashyap Chamarthy wrote:
>
> [...]
>
> > > But doesn't tell *what* the default value is. It is check='partial'.
> > >
On Tue, Jan 09, 2018 at 04:37:10PM +0100, Jiri Denemark wrote:
> On Tue, Jan 09, 2018 at 15:32:54 +0100, Kashyap Chamarthy wrote:
[...]
> > But doesn't tell *what* the default value is. It is check='partial'.
> > Mention it so.
[...]
> NACK
>
> As I said on IRC, the default differs with guest
On Mon, Jan 08, 2018 at 07:35:22PM +0800, Wu Zongyong wrote:
> In current implemention, mdev_types caps keep constant all
> the time. But, it is possible that a device capable of
> mdev_types sometime(for example:bind to proper driver) and
> incapable of mdev_types at other times(for example:
On Tue, Jan 9, 2018 at 5:31 PM, Michal Privoznik wrote:
> On 01/09/2018 04:04 PM, Christian Ehrhardt wrote:
>> virSecurityManagerDomainSetPathLabel is used to make a path known
>> to the security modules, but today is used interchangably for
>> - paths to files/dirs to be
On 01/09/2018 04:04 PM, Christian Ehrhardt wrote:
> virSecurityManagerDomainSetPathLabel is used to make a path known
> to the security modules, but today is used interchangably for
> - paths to files/dirs to be accessed directly
> - paths to a dir, but the access will actually be to files
On 01/09/2018 04:04 PM, Christian Ehrhardt wrote:
> Based on a discussion in [1] I found that the AppArmor security
> module lacked some callbacks. Implementing those not only fixes
> the issue I had before but will also cover a few more cases I
> didn't even run into so far.
>
> [1]:
On Tue, Jan 09, 2018 at 15:32:54 +0100, Kashyap Chamarthy wrote:
> The 'check' attribute is referring to this:
>
>
>
> Upstream documentation says, it is:
>
> used to request a specific way of checking whether the virtual CPU
> matches the specification. It is usually safe to omit
Ping
On 2018/1/5 10:53, Jie Wang wrote:
> offset and len can also be equal to 0 on failed if blockjob return
> status:"BLOCK_JOB_COMPLETED" with error:"File descriptor in bad state",
> so we need to check 'error' in this case.
> ---
> src/qemu/qemu_monitor_json.c | 2 +-
> 1 file changed, 1
Based on a discussion in [1] I found that the AppArmor security
module lacked some callbacks. Implementing those not only fixes
the issue I had before but will also cover a few more cases I
didn't even run into so far.
[1]: https://www.redhat.com/archives/libvir-list/2017-December/msg00726.html
virSecurityManagerDomainSetPathLabel is used to make a path known
to the security modules, but today is used interchangably for
- paths to files/dirs to be accessed directly
- paths to a dir, but the access will actually be to files therein
Depending on the security module it is important to
Since 1b4f66e "security: introduce virSecurityManager
(Set|Restore)ChardevLabel" this is a public API of security manager.
Implementing this in apparmor avoids miss any rules that should be
added for devices labeled via these calls.
Signed-off-by: Christian Ehrhardt
This came up in discussions around huge pages, but it will cover
more per guest paths that should be added to the guests apparmor profile:
- keys via qemuDomainWriteMasterKeyFile
- per domain dirs via qemuProcessMakeDir
- memory backing paths via qemuProcessBuildDestroyMemoryPathsImpl
This is now covered by DomainSetPathLabel being implemented in apparmor.
Signed-off-by: Christian Ehrhardt
---
src/security/virt-aa-helper.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/src/security/virt-aa-helper.c b/src/security/virt-aa-helper.c
index
>> +char *in = NULL, *out = NULL;
>> +int ret = -1;
>> +
>> +virSecurityLabelDefPtr secdef = virDomainDefGetSecurityLabelDef(def,
>> +SECURITY_APPARMOR_NAME);
>> +if (!secdef)
>> +return 0;
>
> There shouldn't be empty line in
On Tue, Jan 9, 2018 at 11:02 AM, Michal Privoznik wrote:
> On 01/03/2018 06:00 PM, Christian Ehrhardt wrote:
[...]
>> AppArmorSetPathLabel(virSecurityManagerPtr mgr,
>> virDomainDefPtr def,
>> - const char *path)
>> +
> [libvirt] [PATCH] usb: allow host devices to be specified by port
>
>
>
> From: Thomas Hebb To: libvir-list redhat comCc:
> Thomas Hebb Subject: [libvirt] [PATCH] usb: allow
> host devices to be specified by portDate: Tue, 5 Jul
On Tue, Jan 9, 2018 at 11:02 AM, Michal Privoznik wrote:
> On 01/03/2018 06:00 PM, Christian Ehrhardt wrote:
>> Based on a discussion in [1] I found that the AppArmor security
>> module lacked some callbacks. Implementing those not only fixes
>> the issue I had before but
The 'check' attribute is referring to this:
Upstream documentation says, it is:
used to request a specific way of checking whether the virtual CPU
matches the specification. It is usually safe to omit this attribute
when starting a domain and stick with the default value.
But
On Wed, Dec 20, 2017 at 04:47:50PM +, Daniel P. Berrange wrote:
This introduces a new binary 'libvirt_qemu' which can be used to launch
guests externally from libvirtd. eg
libvirt_qemu -c qemu:///system /path/to/xml/file
This will launch a guest from the requested XML file and then
On Wed, Dec 20, 2017 at 04:47:49PM +, Daniel P. Berrange wrote:
Signed-off-by: Daniel P. Berrange
---
src/qemu/qemu_driver.c | 57 +
src/qemu/qemu_process.c | 31 ++-
src/qemu/qemu_process.h | 1 +
3
I'm so sorry for not getting to this earlier. I though I'll get to this over
the holidays, but they were very busy with no free time for me.
On Wed, Dec 20, 2017 at 04:47:46PM +, Daniel P. Berrange wrote:
Currently the QEMU driver has three ways of setting up cgroups. It either
skips them
model was allocated by asprintf but never freed after usage
and assignment.
Signed-off-by: Adam Majer
---
src/Virt_ComputerSystem.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/src/Virt_ComputerSystem.c b/src/Virt_ComputerSystem.c
index da07f93..b4930ac 100644
---
In some instances, asprintf allocated memory is not freed
resulting in a slow memory leak.
Adam Majer (2):
Fix memory leak in set_other_id_info
Fix memory leak in set_input_props
src/Virt_ComputerSystem.c | 2 ++
src/Virt_SettingsDefineCapabilities.c | 6 --
2 files changed,
When caption is specified, cap was allocated and assigned a new pointer
while the old never freed. Free cap before replacing it.
get_input_dev_caption only returns valid cap pointer on success, so
do not try to free it on failure.
Signed-off-by: Adam Majer
---
On 12/18/2017 02:31 PM, John Ferlan wrote:
> Looks OK; however, there hasn't been a non build related source code
> change to libvirt-cim since June 2014. So while this could be pushed,
> I'm curious to understand why the "sudden interest" in CIM? The last
> release was July 2013!
I'm not sure
On Mon, Jan 08, 2018 at 10:08:59AM -0700, Jim Fehlig wrote:
> Based loosely on a patch from Fei Li .
>
> Commit 8708ca01c added virNetDevSwitchdevFeature() to check if a network
> device has Switchdev capabilities. virNetDevSwitchdevFeature() attempts
> to retrieve the PCI device
On 01/09/2018 04:06 AM, Andrea Bolognani wrote:
> On Mon, 2018-01-08 at 11:19 -0500, John Ferlan wrote:
>> On 01/08/2018 09:50 AM, Peter Krempa wrote:
>>> On Mon, Jan 08, 2018 at 15:10:29 +0100, Andrea Bolognani wrote:
Instead of formatting 'MHz: 0', which can be confusing, skip the
On Tue, Jan 9, 2018 at 11:02 AM, Cedric Bosdonnat wrote:
> On Wed, 2018-01-03 at 17:55 +0100, Christian Ehrhardt wrote:
>> [...]
>>
>> > > To me, 1 feels most correct cause while the other two fix hugepages,
>> > > there seem to be lurking bugs since we aren't implementing
>>
Hello,
I have not got revert back since I sent last patch on 12/22/2017, just want
to know how it is going on?
Regards,
Charles.
On Fri, Dec 22, 2017 at 3:08 PM, Charles Kelimod wrote:
> Hi Peter,
>
> I will remove the comment.
> I did the test and noticed the issue, then
On Mon, Jan 08, 2018 at 19:06:43 +0800, Feng, Shaohe wrote:
> On 2018年01月05日 02:52, John Ferlan wrote:
> >
> > On 12/17/2017 06:02 PM, Shaohe Feng wrote:
> >> We can start qemu with a "cpu,+la57" to set 57-bit vitrual address
> >> space. So VM can be aware that it need to enable 5-level paging.
>
On 01/03/2018 06:00 PM, Christian Ehrhardt wrote:
> Since 1b4f66e "security: introduce virSecurityManager
> (Set|Restore)ChardevLabel" this is a public API of security manager.
>
> Implementing this in apparmor avoids miss any rules that should be
> added for devices labeled via these calls.
I
On Wed, 2018-01-03 at 17:55 +0100, Christian Ehrhardt wrote:
> [...]
>
> > > To me, 1 feels most correct cause while the other two fix hugepages,
> > > there seem to be lurking bugs since we aren't implementing
> > > domainSetPathLabel.
> > >
> >
> > I work on #1 a while and I think we can do a
On 01/03/2018 06:00 PM, Christian Ehrhardt wrote:
> virSecurityManagerDomainSetPathLabel is used to make a path known
> to the security modules, but today is used interchangably for
> - paths to files/dirs to be accessed directly
> - paths to a dir, but the access will actually be to files
On 01/03/2018 06:00 PM, Christian Ehrhardt wrote:
> Based on a discussion in [1] I found that the AppArmor security
> module lacked some callbacks. Implementing those not only fixes
> the issue I had before but will also cover a few more cases I
> didn't even run into so far.
>
> [1]:
John Ferlan [2018-01-08, 07:55AM -0500]:
>
>
> On 01/08/2018 03:39 AM, Bjoern Walk wrote:
> > John Ferlan [2018-01-04, 03:56PM -0500]:
> >> On 12/19/2017 05:08 AM, Bjoern Walk wrote:
> >>> diff --git a/src/util/virsysinfo.c b/src/util/virsysinfo.c
> >>>
On Mon, 2018-01-08 at 11:19 -0500, John Ferlan wrote:
> On 01/08/2018 09:50 AM, Peter Krempa wrote:
> > On Mon, Jan 08, 2018 at 15:10:29 +0100, Andrea Bolognani wrote:
> > > Instead of formatting 'MHz: 0', which can be confusing, skip the
> > > field altogether. This behavior is consistent with
59 matches
Mail list logo