linux-kernel - [PATCH AUTOSEL 6.1 05/14] KVM: s390: fix setting of fpc register

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Mon, 15 Jan 2024 20:06:07 -0500
From: Sasha Levin <sashal@...nel.org>
To: linux-kernel@...r.kernel.org,
	stable@...r.kernel.org
Cc: Heiko Carstens <hca@...ux.ibm.com>,
	Christian Borntraeger <borntraeger@...ux.ibm.com>,
	Claudio Imbrenda <imbrenda@...ux.ibm.com>,
	Alexander Gordeev <agordeev@...ux.ibm.com>,
	Sasha Levin <sashal@...nel.org>,
	frankja@...ux.ibm.com,
	gor@...ux.ibm.com,
	kvm@...r.kernel.org,
	linux-s390@...r.kernel.org
Subject: [PATCH AUTOSEL 6.1 05/14] KVM: s390: fix setting of fpc register

From: Heiko Carstens <hca@...ux.ibm.com>

[ Upstream commit b988b1bb0053c0dcd26187d29ef07566a565cf55 ]

kvm_arch_vcpu_ioctl_set_fpu() allows to set the floating point control
(fpc) register of a guest cpu. The new value is tested for validity by
temporarily loading it into the fpc register.

This may lead to corruption of the fpc register of the host process:
if an interrupt happens while the value is temporarily loaded into the fpc
register, and within interrupt context floating point or vector registers
are used, the current fp/vx registers are saved with save_fpu_regs()
assuming they belong to user space and will be loaded into fp/vx registers
when returning to user space.

test_fp_ctl() restores the original user space / host process fpc register
value, however it will be discarded, when returning to user space.

In result the host process will incorrectly continue to run with the value
that was supposed to be used for a guest cpu.

Fix this by simply removing the test. There is another test right before
the SIE context is entered which will handles invalid values.

This results in a change of behaviour: invalid values will now be accepted
instead of that the ioctl fails with -EINVAL. This seems to be acceptable,
given that this interface is most likely not used anymore, and this is in
addition the same behaviour implemented with the memory mapped interface
(replace invalid values with zero) - see sync_regs() in kvm-s390.c.

Reviewed-by: Christian Borntraeger <borntraeger@...ux.ibm.com>
Reviewed-by: Claudio Imbrenda <imbrenda@...ux.ibm.com>
Signed-off-by: Heiko Carstens <hca@...ux.ibm.com>
Signed-off-by: Alexander Gordeev <agordeev@...ux.ibm.com>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---
 arch/s390/kvm/kvm-s390.c | 5 -----
 1 file changed, 5 deletions(-)

diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
index 3775363471f0..f604946ab2c8 100644
--- a/arch/s390/kvm/kvm-s390.c
+++ b/arch/s390/kvm/kvm-s390.c
@@ -4138,10 +4138,6 @@ int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)

 	vcpu_load(vcpu);

-	if (test_fp_ctl(fpu->fpc)) {
-		ret = -EINVAL;
-		goto out;
-	}
 	vcpu->run->s.regs.fpc = fpu->fpc;
 	if (MACHINE_HAS_VX)
 		convert_fp_to_vx((__vector128 *) vcpu->run->s.regs.vrs,
@@ -4149,7 +4145,6 @@ int kvm_arch_vcpu_ioctl_set_fpu(struct kvm_vcpu *vcpu, struct kvm_fpu *fpu)
 	else
 		memcpy(vcpu->run->s.regs.fprs, &fpu->fprs, sizeof(fpu->fprs));

-out:
 	vcpu_put(vcpu);
 	return ret;
 }
-- 
2.43.0