diff mbox series

[4.19,111/120] KVM: do not assume PTE is writable after follow_pfn

Message ID 20210726153836.031515657@linuxfoundation.org
State Superseded
Headers show
Series None | expand

Commit Message

Greg Kroah-Hartman July 26, 2021, 3:39 p.m. UTC 
From: Paolo Bonzini <pbonzini@redhat.com>

commit bd2fae8da794b55bf2ac02632da3a151b10e664c upstream.

In order to convert an HVA to a PFN, KVM usually tries to use
the get_user_pages family of functinso.  This however is not
possible for VM_IO vmas; in that case, KVM instead uses follow_pfn.

In doing this however KVM loses the information on whether the
PFN is writable.  That is usually not a problem because the main
use of VM_IO vmas with KVM is for BARs in PCI device assignment,
however it is a bug.  To fix it, use follow_pte and check pte_write
while under the protection of the PTE lock.  The information can
be used to fail hva_to_pfn_remapped or passed back to the
caller via *writable.

Usage of follow_pfn was introduced in commit add6a0cd1c5b ("KVM: MMU: try to fix
up page faults before giving up", 2016-07-05); however, even older version
have the same issue, all the way back to commit 2e2e3738af33 ("KVM:
Handle vma regions with no backing page", 2008-07-20), as they also did
not check whether the PFN was writable.

Fixes: 2e2e3738af33 ("KVM: Handle vma regions with no backing page")
Reported-by: David Stevens <stevensd@google.com>
Cc: 3pvd@google.com
Cc: Jann Horn <jannh@google.com>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[OP: backport to 4.19, adjust follow_pte() -> follow_pte_pmd()]
Signed-off-by: Ovidiu Panait <ovidiu.panait@windriver.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 virt/kvm/kvm_main.c |   15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

Comments

Ovidiu Panait July 27, 2021, 8:33 a.m. UTC | #1 
Hi,

On 7/27/21 1:02 AM, Paolo Bonzini wrote:

> [Please note: This e-mail is from an EXTERNAL e-mail address]

>

> On 26/07/21 23:17, Sudip Mukherjee wrote:

>> Hi Greg,

>>

>> On Mon, Jul 26, 2021 at 4:58 PM Greg Kroah-Hartman

>> <gregkh@linuxfoundation.org> wrote:

>>>

>>> From: Paolo Bonzini <pbonzini@redhat.com>

>>>

>>> commit bd2fae8da794b55bf2ac02632da3a151b10e664c upstream.

>>

>> The build of mips malta_kvm_defconfig fails with the error:

>> In file included from arch/mips/kvm/../../../virt/kvm/kvm_main.c:21:

>> arch/mips/kvm/../../../virt/kvm/kvm_main.c: In function 

>> 'hva_to_pfn_remapped':

>> ./include/linux/kvm_host.h:70:33: error: conversion from 'long long

>> unsigned int' to 'long unsigned int' changes value from

>> '9218868437227405314' to '2' [-Werror=overflow]

>>     70 | #define KVM_PFN_ERR_RO_FAULT    (KVM_PFN_ERR_MASK + 2)

>>        |                                 ^

>> arch/mips/kvm/../../../virt/kvm/kvm_main.c:1530:23: note: in expansion

>> of macro 'KVM_PFN_ERR_RO_FAULT'

>>   1530 |                 pfn = KVM_PFN_ERR_RO_FAULT;

>>

>> It built fine after reverting this patch.

>> gcc version 11.1.1 20210723

>

> I'll resend a version that works tomorrow (including the second patch

> too, which depends on this one for context).

>

The following upstream commit needed to be backported as well:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=a9545779ee9e9e103648f6f2552e73cfe808d0f4


I have sent v2 for this patchset.


Thanks!

Ovidiu


> Paolo

>
diff mbox series

Patch

--- a/virt/kvm/kvm_main.c
+++ b/virt/kvm/kvm_main.c
@@ -1495,9 +1495,11 @@  static int hva_to_pfn_remapped(struct vm
 			       kvm_pfn_t *p_pfn)
 {
 	unsigned long pfn;
+	pte_t *ptep;
+	spinlock_t *ptl;
 	int r;
 
-	r = follow_pfn(vma, addr, &pfn);
+	r = follow_pte_pmd(vma->vm_mm, addr, NULL, NULL, &ptep, NULL, &ptl);
 	if (r) {
 		/*
 		 * get_user_pages fails for VM_IO and VM_PFNMAP vmas and does
@@ -1512,14 +1514,19 @@  static int hva_to_pfn_remapped(struct vm
 		if (r)
 			return r;
 
-		r = follow_pfn(vma, addr, &pfn);
+		r = follow_pte_pmd(vma->vm_mm, addr, NULL, NULL, &ptep, NULL, &ptl);
 		if (r)
 			return r;
+	}
 
+	if (write_fault && !pte_write(*ptep)) {
+		pfn = KVM_PFN_ERR_RO_FAULT;
+		goto out;
 	}
 
 	if (writable)
-		*writable = true;
+		*writable = pte_write(*ptep);
+	pfn = pte_pfn(*ptep);
 
 	/*
 	 * Get a reference here because callers of *hva_to_pfn* and
@@ -1534,6 +1541,8 @@  static int hva_to_pfn_remapped(struct vm
 	 */ 
 	kvm_get_pfn(pfn);
 
+out:
+	pte_unmap_unlock(ptep, ptl);
 	*p_pfn = pfn;
 	return 0;
 }