Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On Mon, Jun 08, 2015 at 10:14:49PM -0700, Vinson Lee wrote: On Tue, May 26, 2015 at 7:53 PM, Xiao Guangrong guangrong.x...@linux.intel.com wrote: On 05/26/2015 10:48 PM, Paolo Bonzini wrote: On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) It has been fixed by Andrew: From: Andrew Morton a...@linux-foundation.org Subject: arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug arch/x86/kvm/mmu.c: In function 'kvm_mmu_pte_write': arch/x86/kvm/mmu.c:4256: error: unknown field 'cr0_wp' specified in initializer arch/x86/kvm/mmu.c:4257: error: unknown field 'cr4_pae' specified in initializer arch/x86/kvm/mmu.c:4257: warning: excess elements in union initializer ... gcc-4.4.4 (at least) has issues when using anonymous unions in initializers. Fixes: edc90b7dc4ceef6 (KVM: MMU: fix SMAP virtualization) Cc: Xiao Guangrong guangrong.x...@linux.intel.com Cc: Paolo Bonzini pbonz...@redhat.com Signed-off-by: Andrew Morton a...@linux-foundation.org Should be found at -mm tree. -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ Hi. Please backport arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug to stable trees that have backported KVM: MMU: fix SMAP virtualization. Currently, the Linux 4.0.5 build is broken with GCC 4.4. KVM: MMU: fix SMAP virtualization is currently queued for 3.16.y-ckt and 3.18.y. Cheers, Vinson -- To unsubscribe from this list: send the line unsubscribe stable in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html I believe you're referring to the following commit: commit 5ec45a192fe6e287f0fc06d5ca4f3bd446d94803 Author: Andrew Morton a...@linux-foundation.org Date: Wed Jun 10 11:15:02 2015 -0700 arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug I am queuing it for the 3.16 kernel. Thanks! Cheers, -- Luís -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On Wed, 2015-05-27 at 10:53 +0800, Xiao Guangrong wrote: On 05/26/2015 10:48 PM, Paolo Bonzini wrote: On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) It has been fixed by Andrew: From: Andrew Morton a...@linux-foundation.org Subject: arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug Folks, this fix is missing in both -tip and Linus' (4.1-rc7) tree. At least I'm running into it. Thanks, Davidlohr -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On Wed, 10 Jun 2015 11:02:46 -0700 Davidlohr Bueso d...@stgolabs.net wrote: On Wed, 2015-05-27 at 10:53 +0800, Xiao Guangrong wrote: On 05/26/2015 10:48 PM, Paolo Bonzini wrote: On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) It has been fixed by Andrew: From: Andrew Morton a...@linux-foundation.org Subject: arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug Folks, this fix is missing in both -tip and Linus' (4.1-rc7) tree. At least I'm running into it. hm, yes, it's in linux-next so I dropped my copy. I'll send it in to Linus today. -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On Tue, May 26, 2015 at 7:53 PM, Xiao Guangrong guangrong.x...@linux.intel.com wrote: On 05/26/2015 10:48 PM, Paolo Bonzini wrote: On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) It has been fixed by Andrew: From: Andrew Morton a...@linux-foundation.org Subject: arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug arch/x86/kvm/mmu.c: In function 'kvm_mmu_pte_write': arch/x86/kvm/mmu.c:4256: error: unknown field 'cr0_wp' specified in initializer arch/x86/kvm/mmu.c:4257: error: unknown field 'cr4_pae' specified in initializer arch/x86/kvm/mmu.c:4257: warning: excess elements in union initializer ... gcc-4.4.4 (at least) has issues when using anonymous unions in initializers. Fixes: edc90b7dc4ceef6 (KVM: MMU: fix SMAP virtualization) Cc: Xiao Guangrong guangrong.x...@linux.intel.com Cc: Paolo Bonzini pbonz...@redhat.com Signed-off-by: Andrew Morton a...@linux-foundation.org Should be found at -mm tree. -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/ Hi. Please backport arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug to stable trees that have backported KVM: MMU: fix SMAP virtualization. Currently, the Linux 4.0.5 build is broken with GCC 4.4. KVM: MMU: fix SMAP virtualization is currently queued for 3.16.y-ckt and 3.18.y. Cheers, Vinson -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On 05/26/2015 10:48 PM, Paolo Bonzini wrote: On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) It has been fixed by Andrew: From: Andrew Morton a...@linux-foundation.org Subject: arch/x86/kvm/mmu.c: work around gcc-4.4.4 bug arch/x86/kvm/mmu.c: In function 'kvm_mmu_pte_write': arch/x86/kvm/mmu.c:4256: error: unknown field 'cr0_wp' specified in initializer arch/x86/kvm/mmu.c:4257: error: unknown field 'cr4_pae' specified in initializer arch/x86/kvm/mmu.c:4257: warning: excess elements in union initializer ... gcc-4.4.4 (at least) has issues when using anonymous unions in initializers. Fixes: edc90b7dc4ceef6 (KVM: MMU: fix SMAP virtualization) Cc: Xiao Guangrong guangrong.x...@linux.intel.com Cc: Paolo Bonzini pbonz...@redhat.com Signed-off-by: Andrew Morton a...@linux-foundation.org Should be found at -mm tree. -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. So I think the kernel code has to change, probably by naming the structure. The information contained in this message is confidential and is intended for the addressee(s) only. If you have received this message in error, please notify the sender immediately and delete the message. Unless you are an addressee (or authorized to receive for an addressee), you may not use, copy or disclose to anyone this message or any information contained in this message. The unauthorized use, disclosure, copying or alteration of this message is strictly prohibited. -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On 26/05/2015 16:45, Edward Cree wrote: This breaks older compilers that can't initialize anon structures. How old ? Even gcc 3.1 says you can use unnamed struct/union fields and 3.2 is the minimum version required to compile the kernel as mentioned in the README. We could simply just name the structure, but I doubt this is the only place in the kernel code where it's being used this way :) This appears to be GCC bug #10676, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=10676 Says it was fixed in 4.6, but I believe the kernel supports GCCs much older than that (back to 3.2). I personally hit it on 4.4.7, the version shipped with RHEL6.6. Yes, it will be fixed soon(ish). Probably before you can get rid of the obnoxious disclaimer... :) Paolo The information contained in this message is confidential and is intended for the addressee(s) only. If you have received this message in error, please notify the sender immediately and delete the message. Unless you are an addressee (or authorized to receive for an addressee), you may not use, copy or disclose to anyone this message or any information contained in this message. The unauthorized use, disclosure, copying or alteration of this message is strictly prohibited. -- To unsubscribe from this list: send the line unsubscribe kvm in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On 05/11/2015 10:55 AM, Xiao Guangrong wrote:
KVM may turn a user page to a kernel page when kernel writes a readonly
user page if CR0.WP = 1. This shadow page entry will be reused after
SMAP is enabled so that kernel is allowed to access this user page
Fix it by setting SMAP !CR0.WP into shadow page's role and reset mmu
once CR4.SMAP is updated
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
@@ -4208,12 +4211,18 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
const u8 *new, int bytes)
{
gfn_t gfn = gpa PAGE_SHIFT;
- union kvm_mmu_page_role mask = { .word = 0 };
struct kvm_mmu_page *sp;
LIST_HEAD(invalid_list);
u64 entry, gentry, *spte;
int npte;
bool remote_flush, local_flush, zap_page;
+ union kvm_mmu_page_role mask = (union kvm_mmu_page_role) {
+ .cr0_wp = 1,
+ .cr4_pae = 1,
+ .nxe = 1,
+ .smep_andnot_wp = 1,
+ .smap_andnot_wp = 1,
+ };
This breaks older compilers that can't initialize anon structures.
-boris
--
To unsubscribe from this list: send the line unsubscribe kvm in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
Boris Ostrovsky boris.ostrov...@oracle.com writes:
On 05/11/2015 10:55 AM, Xiao Guangrong wrote:
KVM may turn a user page to a kernel page when kernel writes a readonly
user page if CR0.WP = 1. This shadow page entry will be reused after
SMAP is enabled so that kernel is allowed to access this user page
Fix it by setting SMAP !CR0.WP into shadow page's role and reset mmu
once CR4.SMAP is updated
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
@@ -4208,12 +4211,18 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t
gpa,
const u8 *new, int bytes)
{
gfn_t gfn = gpa PAGE_SHIFT;
-union kvm_mmu_page_role mask = { .word = 0 };
struct kvm_mmu_page *sp;
LIST_HEAD(invalid_list);
u64 entry, gentry, *spte;
int npte;
bool remote_flush, local_flush, zap_page;
+union kvm_mmu_page_role mask = (union kvm_mmu_page_role) {
+.cr0_wp = 1,
+.cr4_pae = 1,
+.nxe = 1,
+.smep_andnot_wp = 1,
+.smap_andnot_wp = 1,
+};
This breaks older compilers that can't initialize anon structures.
How old ? Even gcc 3.1 says you can use unnamed struct/union fields and
3.2 is the minimum version required to compile the kernel as mentioned
in the README.
We could simply just name the structure, but I doubt this is the
only place in the kernel code where it's being used this way :)
Bandan
-boris
--
To unsubscribe from this list: send the line unsubscribe kvm in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
To unsubscribe from this list: send the line unsubscribe kvm in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH 1/2] KVM: MMU: fix SMAP virtualization
On 05/22/2015 07:54 PM, Bandan Das wrote:
Boris Ostrovsky boris.ostrov...@oracle.com writes:
On 05/11/2015 10:55 AM, Xiao Guangrong wrote:
KVM may turn a user page to a kernel page when kernel writes a readonly
user page if CR0.WP = 1. This shadow page entry will be reused after
SMAP is enabled so that kernel is allowed to access this user page
Fix it by setting SMAP !CR0.WP into shadow page's role and reset mmu
once CR4.SMAP is updated
Signed-off-by: Xiao Guangrong guangrong.x...@linux.intel.com
---
@@ -4208,12 +4211,18 @@ void kvm_mmu_pte_write(struct kvm_vcpu *vcpu, gpa_t gpa,
const u8 *new, int bytes)
{
gfn_t gfn = gpa PAGE_SHIFT;
- union kvm_mmu_page_role mask = { .word = 0 };
struct kvm_mmu_page *sp;
LIST_HEAD(invalid_list);
u64 entry, gentry, *spte;
int npte;
bool remote_flush, local_flush, zap_page;
+ union kvm_mmu_page_role mask = (union kvm_mmu_page_role) {
+ .cr0_wp = 1,
+ .cr4_pae = 1,
+ .nxe = 1,
+ .smep_andnot_wp = 1,
+ .smap_andnot_wp = 1,
+ };
This breaks older compilers that can't initialize anon structures.
How old ? Even gcc 3.1 says you can use unnamed struct/union fields and
3.2 is the minimum version required to compile the kernel as mentioned
in the README.
We could simply just name the structure, but I doubt this is the
only place in the kernel code where it's being used this way :)
You can use them but you can't use initializers. Unfortunately my build
system (F13) conveniently went down but this is an example from an old
email:
FC-64 build@build-mk2:~/xtt-x86_64/bootstrap cat anon.c
struct bar {
struct {
int i;
};
};
main()
{
struct bar a = {.i = 0};
}
FC-64 build@build-mk2:~/xtt-x86_64/bootstrap gcc --version|head -1
gcc (GCC) 4.4.4 20100503 (Red Hat 4.4.4-2)
FC-64 build@build-mk2:~/xtt-x86_64/bootstrap gcc anon.c
anon.c: In function ‘main’:
anon.c:9: error: unknown field ‘i’ specified in initializer
FC-64 build@build-mk2:~/xtt-x86_64/bootstrap
but
build@build-mk2 bootstrap]$ gcc --version|head -1
gcc (GCC) 4.6.3 20120306 (Red Hat 4.6.3-2)
[build@build-mk2 bootstrap]$ gcc anon.c
[build@build-mk2 bootstrap]$
-boris
--
To unsubscribe from this list: send the line unsubscribe kvm in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
