| Message ID | 1408355431-115633-1-git-send-email-wanpeng.li@linux.intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
Il 18/08/2014 11:50, Wanpeng Li ha scritto: > EPT misconfig handler in kvm will check which reason lead to EPT > misconfiguration after vmexit. One of the reasons is that an EPT > paging-structure entry is configured with settings reserved for > future functionality. However, the handler can't identify if > paging-structure entry of reserved bits for 1-GByte page are > configured, since PDPTE which point to 1-GByte page will reserve > bits 29:12 instead of bits 7:3 which are reserved for PDPTE that > references an EPT Page Directory. This patch fix it by reserve > bits 29:12 for 1-GByte page. > > Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com> > --- > arch/x86/kvm/vmx.c | 14 ++++++++++---- > 1 file changed, 10 insertions(+), 4 deletions(-) > > diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c > index bfe11cf..71cbee5 100644 > --- a/arch/x86/kvm/vmx.c > +++ b/arch/x86/kvm/vmx.c > @@ -5521,9 +5521,14 @@ static u64 ept_rsvd_mask(u64 spte, int level) > for (i = 51; i > boot_cpu_data.x86_phys_bits; i--) > mask |= (1ULL << i); > > - if (level > 2) > - /* bits 7:3 reserved */ > - mask |= 0xf8; > + if (level > 2) { level can be 4 here. You have to return 0xf8 for level == 4. The same "if" statement then can cover both 2MB and 1GB pages, like if (spte & (1ULL << 7)) /* 1GB/2MB page, bits 29:12 or 20:12 reserved respectively */ mask |= (PAGE_SIZE << ((level - 1) * 9)) - PAGE_SIZE; else /* bits 6:3 reserved */ mask |= 0x78; > - if (level == 1 || (level == 2 && (spte & (1ULL << 7)))) { > + if (level == 1 || ((level == 3 || level == 2) > + && (spte & (1ULL << 7)))) { This condition can be simplified by checking the return value of ept_rsvd_mask. If it includes 0x38, this is a large page. Otherwise it is a leaf page and you can go down the "if". Paolo > u64 ept_mem_type = (spte & 0x38) >> 3; > > if (ept_mem_type == 2 || ept_mem_type == 3 || > -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
On 08/18/2014 05:50 PM, Wanpeng Li wrote: > EPT misconfig handler in kvm will check which reason lead to EPT > misconfiguration after vmexit. One of the reasons is that an EPT > paging-structure entry is configured with settings reserved for > future functionality. However, the handler can't identify if > paging-structure entry of reserved bits for 1-GByte page are > configured, since PDPTE which point to 1-GByte page will reserve > bits 29:12 instead of bits 7:3 which are reserved for PDPTE that > references an EPT Page Directory. This patch fix it by reserve > bits 29:12 for 1-GByte page. That mask is only set in the lowest pte for 4K page, i think it is not a problem, no? -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Il 18/08/2014 12:52, Xiao Guangrong ha scritto: >> > EPT misconfig handler in kvm will check which reason lead to EPT >> > misconfiguration after vmexit. One of the reasons is that an EPT >> > paging-structure entry is configured with settings reserved for >> > future functionality. However, the handler can't identify if >> > paging-structure entry of reserved bits for 1-GByte page are >> > configured, since PDPTE which point to 1-GByte page will reserve >> > bits 29:12 instead of bits 7:3 which are reserved for PDPTE that >> > references an EPT Page Directory. This patch fix it by reserve >> > bits 29:12 for 1-GByte page. > That mask is only set in the lowest pte for 4K page, i think it > is not a problem, no? It will cause KVM to WARN. The EPT memory type will also be ignored for gigabyte pages. Paolo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi Paolo, On Mon, Aug 18, 2014 at 12:18:59PM +0200, Paolo Bonzini wrote: >Il 18/08/2014 11:50, Wanpeng Li ha scritto: >> EPT misconfig handler in kvm will check which reason lead to EPT >> misconfiguration after vmexit. One of the reasons is that an EPT >> paging-structure entry is configured with settings reserved for >> future functionality. However, the handler can't identify if >> paging-structure entry of reserved bits for 1-GByte page are >> configured, since PDPTE which point to 1-GByte page will reserve >> bits 29:12 instead of bits 7:3 which are reserved for PDPTE that >> references an EPT Page Directory. This patch fix it by reserve >> bits 29:12 for 1-GByte page. >> >> Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com> >> --- >> arch/x86/kvm/vmx.c | 14 ++++++++++---- >> 1 file changed, 10 insertions(+), 4 deletions(-) >> >> diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c >> index bfe11cf..71cbee5 100644 >> --- a/arch/x86/kvm/vmx.c >> +++ b/arch/x86/kvm/vmx.c >> @@ -5521,9 +5521,14 @@ static u64 ept_rsvd_mask(u64 spte, int level) >> for (i = 51; i > boot_cpu_data.x86_phys_bits; i--) >> mask |= (1ULL << i); >> >> - if (level > 2) >> - /* bits 7:3 reserved */ >> - mask |= 0xf8; >> + if (level > 2) { > >level can be 4 here. You have to return 0xf8 for level == 4. > >The same "if" statement then can cover both 2MB and 1GB pages, like > > if (spte & (1ULL << 7)) > /* 1GB/2MB page, bits 29:12 or 20:12 reserved respectively */ > mask |= (PAGE_SIZE << ((level - 1) * 9)) - PAGE_SIZE; > else > /* bits 6:3 reserved */ > mask |= 0x78; > >> - if (level == 1 || (level == 2 && (spte & (1ULL << 7)))) { >> + if (level == 1 || ((level == 3 || level == 2) >> + && (spte & (1ULL << 7)))) { > >This condition can be simplified by checking the return value of ept_rsvd_mask. >If it includes 0x38, this is a large page. Otherwise it is a leaf page and >you can go down the "if". As you know, 5:3 bits which used for EPT MT are not reserved bits, so I fail to understand why check the return value of ept_rsvd_mask and it's a large page if includes 0x38. Could you eplain in more details? ;-) Regards, Wanpeng Li > >Paolo > >> u64 ept_mem_type = (spte & 0x38) >> 3; >> >> if (ept_mem_type == 2 || ept_mem_type == 3 || >> -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Il 19/08/2014 04:17, Wanpeng Li ha scritto: >>> >> - if (level == 1 || (level == 2 && (spte & (1ULL << 7)))) { >>> >> + if (level == 1 || ((level == 3 || level == 2) >>> >> + && (spte & (1ULL << 7)))) { >> > >> >This condition can be simplified by checking the return value of ept_rsvd_mask. >> >If it includes 0x38, this is a large page. Oops, a "not" was missing. If it includes 0x38, this is _not_ a large page (it is a page directory / page directory pointer / PML4). >> Otherwise it is a leaf page and >> you can go down the "if". > As you know, 5:3 bits which used for EPT MT are not reserved bits, so > I fail to understand why check the return value of ept_rsvd_mask and > it's a large page if includes 0x38. Could you eplain in more details? ;-) A non-leaf page will always have 0x38 in the ept_rsvd_mask. A leaf page will never have 0x38 in the ept_rsvd_mask. Paolo -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/arch/x86/kvm/vmx.c b/arch/x86/kvm/vmx.c index bfe11cf..71cbee5 100644 --- a/arch/x86/kvm/vmx.c +++ b/arch/x86/kvm/vmx.c @@ -5521,9 +5521,14 @@ static u64 ept_rsvd_mask(u64 spte, int level) for (i = 51; i > boot_cpu_data.x86_phys_bits; i--) mask |= (1ULL << i); - if (level > 2) - /* bits 7:3 reserved */ - mask |= 0xf8; + if (level > 2) { + if (spte & (1 << 7)) + /* 1GB ref, bits 29:12 */ + mask |= 0x3ffff000; + else + /* bits 7:3 reserved */ + mask |= 0xf8; + } else if (level == 2) { if (spte & (1ULL << 7)) /* 2MB ref, bits 20:12 reserved */ @@ -5561,7 +5566,8 @@ static void ept_misconfig_inspect_spte(struct kvm_vcpu *vcpu, u64 spte, WARN_ON(1); } - if (level == 1 || (level == 2 && (spte & (1ULL << 7)))) { + if (level == 1 || ((level == 3 || level == 2) + && (spte & (1ULL << 7)))) { u64 ept_mem_type = (spte & 0x38) >> 3; if (ept_mem_type == 2 || ept_mem_type == 3 ||
EPT misconfig handler in kvm will check which reason lead to EPT misconfiguration after vmexit. One of the reasons is that an EPT paging-structure entry is configured with settings reserved for future functionality. However, the handler can't identify if paging-structure entry of reserved bits for 1-GByte page are configured, since PDPTE which point to 1-GByte page will reserve bits 29:12 instead of bits 7:3 which are reserved for PDPTE that references an EPT Page Directory. This patch fix it by reserve bits 29:12 for 1-GByte page. Signed-off-by: Wanpeng Li <wanpeng.li@linux.intel.com> --- arch/x86/kvm/vmx.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-)