KVM/arm64 fixes for 6.3, part #4
Hi Paolo,
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
Please pull,
Oliver
The following changes since commit e81625218bf7986ba1351a98c43d346b15601d26:
KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV2/3 to protected VMs (2023-04-04 15:52:06 +0000)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.git tags/kvmarm-fixes-6.3-4
for you to fetch changes up to a25bc8486f9c01c1af6b6c5657234b2eee2c39d6:
KVM: arm64: Fix buffer overflow in kvm_arm_set_fw_reg() (2023-04-19 15:22:37 +0000)
----------------------------------------------------------------
KVM/arm64 fixes for 6.3, part #4
- Plug a buffer overflow due to the use of the user-provided register width for firmware regs. Outright reject accesses where the user register width does not match the kernel representation.
- Protect non-atomic RMW operations on vCPU flags against preemption, as an update to the flags by an intervening preemption could be lost.
----------------------------------------------------------------
Dan Carpenter (1):
KVM: arm64: Fix buffer overflow in kvm_arm_set_fw_reg()
Marc Zyngier (1):
KVM: arm64: Make vcpu flag updates non-preemptible
arch/arm64/include/asm/kvm_host.h | 19 ++++++++++++++++++- arch/arm64/kvm/hypercalls.c | 2 ++ 2 files changed, 20 insertions(+), 1 deletion(-)
Paolo Bonzini
maintainerQuoted 15 lines · Oliver Upton
Hi Paolo,
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
Please pull,
Oliver
The following changes since commit
e81625218bf7986ba1351a98c43d346b15601d26:KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV2/3 to protected VMs (2023-04-04 15:52:06 +0000)
Quoted 7 lines
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm.gittags/kvmarm-fixes-6.3-4for you to fetch changes up to
a25bc8486f9c01c1af6b6c5657234b2eee2c39d6:KVM: arm64: Fix buffer overflow in kvm_arm_set_fw_reg() (2023-04-19 15:22:37 +0000)
Pulled, thanks.
Paolo
Quoted 22 lines
----------------------------------------------------------------
KVM/arm64 fixes for 6.3, part #4- Plug a buffer overflow due to the use of the user-provided register width for firmware regs. Outright reject accesses where the user register width does not match the kernel representation.
- Protect non-atomic RMW operations on vCPU flags against preemption, as an update to the flags by an intervening preemption could be lost.
----------------------------------------------------------------
Dan Carpenter (1):
KVM: arm64: Fix buffer overflow inkvm_arm_set_fw_reg()Marc Zyngier (1):
KVM: arm64: Make vcpu flag updates non-preemptiblearch/arm64/include/asm/kvm_host.h | 19 ++++++++++++++++++- arch/arm64/kvm/hypercalls.c | 2 ++ 2 files changed, 20 insertions(+), 1 deletion(-)
Paolo Bonzini
maintainerQuoted 6 lines · Oliver Upton
Hi Paolo,
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
At least going by the Fixes tag, I think this one should have been Cc'd to stable as well. Can you send it next week or would you like someone else to handle the backport?
Thanks,
Paolo
Marc Zyngier
maintainerQuoted 11 lines · Paolo Bonzini
Quoted 6 lines · Oliver Upton
Hi Paolo,
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
At least going by the Fixes tag, I think this one should have been Cc'd to stable as well. Can you send it next week or would you like someone else to handle the backport?
Indeed, that's missing. But yes, backports are definitely on the cards, and we'll make sure all stable versions get fixed as soon as the fix hits Linus' tree.
M.
Signature · 1 line
Jazz is not dead. It just smells funny...
Quoted 4 lines · Marc Zyngier
Quoted 3 lines · Paolo Bonzini
Quoted 2 lines · Oliver Upton
Hi Paolo,
Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
At least going by the Fixes tag, I think this one should have been Cc'd to stable as well. Can you send it next week or would you like someone else to handle the backport?
Thanks for spotting that, I had a mental note to do so, but my memory is fleeting at best :)
Quoted 3 lines
Indeed, that's missing. But yes, backports are definitely on the cards, and we'll make sure all stable versions get fixed as soon as the fix hits Linus' tree.
Between this last batch of fixes for 6.3 and the 6.4 pull we've accrued quite a backlog of stable-worthy patches, many of them are likely to be nontrivial backports.
I'll do the config_lock series, and I can pick up the firmware reg fix if nobody else is handling that backport.
Are you going to take a stab at the vCPU flags fix?
Signature · 2 lines
Thanks, Oliver
Marc Zyngier
maintainerQuoted 22 lines · Oliver Upton
Quoted 4 lines · Marc Zyngier
Quoted 3 lines · Paolo Bonzini
Quoted 2 lines · Oliver Upton
Hi Paolo,
Quoted 8 lines
Quoted 8 lines
Quoted 4 lines
Here is v2 of the last batch of fixes for 6.3 (for real this time!)
Details in the tag, but the noteworthy addition is Dan's fix for a rather obvious buffer overflow when writing to a firmware register.
At least going by the Fixes tag, I think this one should have been Cc'd to stable as well. Can you send it next week or would you like someone else to handle the backport?
Thanks for spotting that, I had a mental note to do so, but my memory is fleeting at best :)
Quoted 3 lines
Indeed, that's missing. But yes, backports are definitely on the cards, and we'll make sure all stable versions get fixed as soon as the fix hits Linus' tree.
Quoted 8 lines
Between this last batch of fixes for 6.3 and the 6.4 pull we've accrued quite a backlog of stable-worthy patches, many of them are likely to be nontrivial backports.
I'll do the
config_lockseries, and I can pick up the firmware reg fix if nobody else is handling that backport.Are you going to take a stab at the vCPU flags fix?
Yup, that's on my list. The backport will be easy down to 6.0, but way more creative before that, as we don't have a proper flag management infrastructure in older kernels.
Cheers,
M.
Signature · 1 line
Without deviation from the norm, progress is not possible.