mbox series

[SRU,focal/linux-hwe-5.8,0/4] CVE fixes for 5.8 kernels

Message ID 20210923235101.49134-1-cascardo@canonical.com
Headers show
Series CVE fixes for 5.8 kernels | expand

Message

Thadeu Lima de Souza Cascardo Sept. 23, 2021, 11:50 p.m. UTC
[CVEs]
These are fixes that are only needed for 5.8 kernels (though currently that
is only azure-5.8), which are already pending or released for other
kernels.

[Test]
This has been built and booted as part of hwe-5.8 and azure-5.8.

The one I had a test for was CVE-2021-3612, and I can confirm I don't see a
crash anymore.

[Potential regression]
Two of the fixes affect KVM, so regressions would be observed on KVM guests
running on Azure instances. There is one for a Xilinx network device, so
any systems with that device could be impacted. Finally, the one that was
tested against affects the /dev/input/js* interface to some input devices.

Alexander Larkin (1):
  Input: joydev - prevent use of not validated data in JSIOCSBTNMAP
    ioctl

David Rientjes (1):
  KVM: SVM: Periodically schedule when unregistering regions on destroy

Esben Haabendal (1):
  net: ll_temac: Fix TX BD buffer overwrite

Nicholas Piggin (1):
  KVM: do not allow mapping valid but non-reference-counted pages

 arch/x86/kvm/svm/sev.c                      |  1 +
 drivers/input/joydev.c                      |  2 +-
 drivers/net/ethernet/xilinx/ll_temac_main.c |  2 +-
 virt/kvm/kvm_main.c                         | 19 +++++++++++++++++--
 4 files changed, 20 insertions(+), 4 deletions(-)

Comments

Stefan Bader Sept. 24, 2021, 7:29 a.m. UTC | #1
On 24.09.21 01:50, Thadeu Lima de Souza Cascardo wrote:
> [CVEs]
> These are fixes that are only needed for 5.8 kernels (though currently that
> is only azure-5.8), which are already pending or released for other
> kernels.
> 
> [Test]
> This has been built and booted as part of hwe-5.8 and azure-5.8.
> 
> The one I had a test for was CVE-2021-3612, and I can confirm I don't see a
> crash anymore.
> 
> [Potential regression]
> Two of the fixes affect KVM, so regressions would be observed on KVM guests
> running on Azure instances. There is one for a Xilinx network device, so
> any systems with that device could be impacted. Finally, the one that was
> tested against affects the /dev/input/js* interface to some input devices.
> 
> Alexander Larkin (1):
>    Input: joydev - prevent use of not validated data in JSIOCSBTNMAP
>      ioctl
> 
> David Rientjes (1):
>    KVM: SVM: Periodically schedule when unregistering regions on destroy
> 
> Esben Haabendal (1):
>    net: ll_temac: Fix TX BD buffer overwrite
> 
> Nicholas Piggin (1):
>    KVM: do not allow mapping valid but non-reference-counted pages
> 
>   arch/x86/kvm/svm/sev.c                      |  1 +
>   drivers/input/joydev.c                      |  2 +-
>   drivers/net/ethernet/xilinx/ll_temac_main.c |  2 +-
>   virt/kvm/kvm_main.c                         | 19 +++++++++++++++++--
>   4 files changed, 20 insertions(+), 4 deletions(-)
> 
Acked-by: Stefan Bader <stefan.bader@canonical.com>
Kleber Sacilotto de Souza Sept. 24, 2021, 7:45 a.m. UTC | #2
On 24.09.21 01:50, Thadeu Lima de Souza Cascardo wrote:
> [CVEs]
> These are fixes that are only needed for 5.8 kernels (though currently that
> is only azure-5.8), which are already pending or released for other
> kernels.
> 
> [Test]
> This has been built and booted as part of hwe-5.8 and azure-5.8.
> 
> The one I had a test for was CVE-2021-3612, and I can confirm I don't see a
> crash anymore.
> 
> [Potential regression]
> Two of the fixes affect KVM, so regressions would be observed on KVM guests
> running on Azure instances. There is one for a Xilinx network device, so
> any systems with that device could be impacted. Finally, the one that was
> tested against affects the /dev/input/js* interface to some input devices.
> 
> Alexander Larkin (1):
>    Input: joydev - prevent use of not validated data in JSIOCSBTNMAP
>      ioctl
> 
> David Rientjes (1):
>    KVM: SVM: Periodically schedule when unregistering regions on destroy
> 
> Esben Haabendal (1):
>    net: ll_temac: Fix TX BD buffer overwrite
> 
> Nicholas Piggin (1):
>    KVM: do not allow mapping valid but non-reference-counted pages
> 
>   arch/x86/kvm/svm/sev.c                      |  1 +
>   drivers/input/joydev.c                      |  2 +-
>   drivers/net/ethernet/xilinx/ll_temac_main.c |  2 +-
>   virt/kvm/kvm_main.c                         | 19 +++++++++++++++++--
>   4 files changed, 20 insertions(+), 4 deletions(-)
> 

Acked-by: Kleber Sacilotto de Souza <kleber.souza@canonical.com>

Thanks
Stefan Bader Sept. 24, 2021, 9:52 a.m. UTC | #3
On 24.09.21 01:50, Thadeu Lima de Souza Cascardo wrote:
> [CVEs]
> These are fixes that are only needed for 5.8 kernels (though currently that
> is only azure-5.8), which are already pending or released for other
> kernels.
> 
> [Test]
> This has been built and booted as part of hwe-5.8 and azure-5.8.
> 
> The one I had a test for was CVE-2021-3612, and I can confirm I don't see a
> crash anymore.
> 
> [Potential regression]
> Two of the fixes affect KVM, so regressions would be observed on KVM guests
> running on Azure instances. There is one for a Xilinx network device, so
> any systems with that device could be impacted. Finally, the one that was
> tested against affects the /dev/input/js* interface to some input devices.
> 
> Alexander Larkin (1):
>    Input: joydev - prevent use of not validated data in JSIOCSBTNMAP
>      ioctl
> 
> David Rientjes (1):
>    KVM: SVM: Periodically schedule when unregistering regions on destroy
> 
> Esben Haabendal (1):
>    net: ll_temac: Fix TX BD buffer overwrite
> 
> Nicholas Piggin (1):
>    KVM: do not allow mapping valid but non-reference-counted pages
> 
>   arch/x86/kvm/svm/sev.c                      |  1 +
>   drivers/input/joydev.c                      |  2 +-
>   drivers/net/ethernet/xilinx/ll_temac_main.c |  2 +-
>   virt/kvm/kvm_main.c                         | 19 +++++++++++++++++--
>   4 files changed, 20 insertions(+), 4 deletions(-)
> 

Applied to focal:linux-hwe-5.8/hwe-5.8. Thanks.

-Stefan