| 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
| ||
|
Message-Id: <20230504225351.10765-6-decui@microsoft.com>
Date: Thu, 4 May 2023 15:53:50 -0700
From: Dexuan Cui <decui@...rosoft.com>
To: ak@...ux.intel.com, arnd@...db.de, bp@...en8.de,
brijesh.singh@....com, dan.j.williams@...el.com,
dave.hansen@...ux.intel.com, haiyangz@...rosoft.com, hpa@...or.com,
jane.chu@...cle.com, kirill.shutemov@...ux.intel.com,
kys@...rosoft.com, linux-arch@...r.kernel.org,
linux-hyperv@...r.kernel.org, luto@...nel.org, mingo@...hat.com,
peterz@...radead.org, rostedt@...dmis.org,
sathyanarayanan.kuppuswamy@...ux.intel.com, seanjc@...gle.com,
tglx@...utronix.de, tony.luck@...el.com, wei.liu@...nel.org,
x86@...nel.org, mikelley@...rosoft.com
Cc: linux-kernel@...r.kernel.org, Tianyu.Lan@...rosoft.com,
Dexuan Cui <decui@...rosoft.com>
Subject: [PATCH v6 5/6] Drivers: hv: vmbus: Support TDX guests
Add Hyper-V specific code so that a TDX guest can run on Hyper-V:
No need to use hv_vp_assist_page.
Don't use the unsafe Hyper-V TSC page.
Don't try to use HV_REGISTER_CRASH_CTL.
Don't trust Hyper-V's TLB-flushing hypercalls.
Don't use lazy EOI.
Signed-off-by: Dexuan Cui <decui@...rosoft.com>
---
Changes in v2:
Used a new function hv_set_memory_enc_dec_needed() in
__set_memory_enc_pgtable().
Added the missing set_memory_encrypted() in hv_synic_free().
Changes in v3:
Use pgprot_decrypted(PAGE_KERNEL)in hv_ringbuffer_init().
(Do not use PAGE_KERNEL_NOENC, which doesn't exist for ARM64).
Used cc_mkdec() in hv_synic_enable_regs().
ms_hyperv_init_platform():
Explicitly do not use HV_X64_REMOTE_TLB_FLUSH_RECOMMENDED.
Explicitly do not use HV_X64_APIC_ACCESS_RECOMMENDED.
Enabled __send_ipi_mask() and __send_ipi_one() for TDX guests.
Changes in v4:
A minor rebase to Michael's v7 DDA patchset. I'm very happy that
I can drop my v3 change to arch/x86/mm/pat/set_memory.c due to
Michael's work.
Changes in v5:
Added memset() to clear synic_message_page and synic_event_page()
after set_memory_decrypted().
Rebased the patch since "post_msg_page" has been removed in
hyperv-next.
Improved the error handling in hv_synic_alloc()/free() [Michael
Kelley]
Changes in v6:
Adressed Michael Kelley's comments on patch 5:
Removed 2 unnecessary lines of messages from the commit log.
Fixed the error handling path for hv_synic_alloc()/free().
Printed the 'ret' in hv_synic_alloc()/free().
arch/x86/hyperv/hv_apic.c | 6 ++--
arch/x86/hyperv/hv_init.c | 19 +++++++---
arch/x86/kernel/cpu/mshyperv.c | 21 ++++++++++-
drivers/hv/hv.c | 65 ++++++++++++++++++++++++++++++++--
4 files changed, 101 insertions(+), 10 deletions(-)
diff --git a/arch/x86/hyperv/hv_apic.c b/arch/x86/hyperv/hv_apic.c
index 1fbda2f94184..b28da8b41b45 100644
--- a/arch/x86/hyperv/hv_apic.c
+++ b/arch/x86/hyperv/hv_apic.c
@@ -177,7 +177,8 @@ static bool __send_ipi_mask(const struct cpumask *mask, int vector,
(exclude_self && weight == 1 && cpumask_test_cpu(this_cpu, mask)))
return true;
- if (!hv_hypercall_pg)
+ /* A TDX guest doesn't use hv_hypercall_pg. */
+ if (!hv_isolation_type_tdx() && !hv_hypercall_pg)
return false;
if ((vector < HV_IPI_LOW_VECTOR) || (vector > HV_IPI_HIGH_VECTOR))
@@ -231,7 +232,8 @@ static bool __send_ipi_one(int cpu, int vector)
trace_hyperv_send_ipi_one(cpu, vector);
- if (!hv_hypercall_pg || (vp == VP_INVAL))
+ /* A TDX guest doesn't use hv_hypercall_pg. */
+ if ((!hv_isolation_type_tdx() && !hv_hypercall_pg) || (vp == VP_INVAL))
return false;
if ((vector < HV_IPI_LOW_VECTOR) || (vector > HV_IPI_HIGH_VECTOR))
diff --git a/arch/x86/hyperv/hv_init.c b/arch/x86/hyperv/hv_init.c
index f175e0de821c..f28357ecad7d 100644
--- a/arch/x86/hyperv/hv_init.c
+++ b/arch/x86/hyperv/hv_init.c
@@ -79,7 +79,7 @@ static int hyperv_init_ghcb(void)
static int hv_cpu_init(unsigned int cpu)
{
union hv_vp_assist_msr_contents msr = { 0 };
- struct hv_vp_assist_page **hvp = &hv_vp_assist_page[cpu];
+ struct hv_vp_assist_page **hvp;
int ret;
ret = hv_common_cpu_init(cpu);
@@ -89,6 +89,7 @@ static int hv_cpu_init(unsigned int cpu)
if (!hv_vp_assist_page)
return 0;
+ hvp = &hv_vp_assist_page[cpu];
if (hv_root_partition) {
/*
* For root partition we get the hypervisor provided VP assist
@@ -398,11 +399,21 @@ void __init hyperv_init(void)
if (hv_common_init())
return;
- hv_vp_assist_page = kcalloc(num_possible_cpus(),
- sizeof(*hv_vp_assist_page), GFP_KERNEL);
+ /*
+ * The VP assist page is useless to a TDX guest: the only use we
+ * would have for it is lazy EOI, which can not be used with TDX.
+ */
+ if (hv_isolation_type_tdx())
+ hv_vp_assist_page = NULL;
+ else
+ hv_vp_assist_page = kcalloc(num_possible_cpus(),
+ sizeof(*hv_vp_assist_page),
+ GFP_KERNEL);
if (!hv_vp_assist_page) {
ms_hyperv.hints &= ~HV_X64_ENLIGHTENED_VMCS_RECOMMENDED;
- goto common_free;
+
+ if (!hv_isolation_type_tdx())
+ goto common_free;
}
if (hv_isolation_type_snp()) {
diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
index 2fd687a80033..b95b689efa07 100644
--- a/arch/x86/kernel/cpu/mshyperv.c
+++ b/arch/x86/kernel/cpu/mshyperv.c
@@ -404,8 +404,27 @@ static void __init ms_hyperv_init_platform(void)
if (hv_get_isolation_type() == HV_ISOLATION_TYPE_SNP)
static_branch_enable(&isolation_type_snp);
- else if (hv_get_isolation_type() == HV_ISOLATION_TYPE_TDX)
+ else if (hv_get_isolation_type() == HV_ISOLATION_TYPE_TDX) {
static_branch_enable(&isolation_type_tdx);
+
+ /*
+ * The GPAs of SynIC Event/Message pages and VMBus
+ * Moniter pages need to be added by this offset.
+ */
+ ms_hyperv.shared_gpa_boundary = cc_mkdec(0);
+
+ /* Don't use the unsafe Hyper-V TSC page */
+ ms_hyperv.features &= ~HV_MSR_REFERENCE_TSC_AVAILABLE;
+
+ /* HV_REGISTER_CRASH_CTL is unsupported */
+ ms_hyperv.misc_features &= ~HV_FEATURE_GUEST_CRASH_MSR_AVAILABLE;
+
+ /* Don't trust Hyper-V's TLB-flushing hypercalls */
+ ms_hyperv.hints &= ~HV_X64_REMOTE_TLB_FLUSH_RECOMMENDED;
+
+ /* A TDX VM must use x2APIC and doesn't use lazy EOI */
+ ms_hyperv.hints &= ~HV_X64_APIC_ACCESS_RECOMMENDED;
+ }
}
if (hv_max_functions_eax >= HYPERV_CPUID_NESTED_FEATURES) {
diff --git a/drivers/hv/hv.c b/drivers/hv/hv.c
index de6708dbe0df..af959e87b6e7 100644
--- a/drivers/hv/hv.c
+++ b/drivers/hv/hv.c
@@ -18,6 +18,7 @@
#include <linux/clockchips.h>
#include <linux/delay.h>
#include <linux/interrupt.h>
+#include <linux/set_memory.h>
#include <clocksource/hyperv_timer.h>
#include <asm/mshyperv.h>
#include "hyperv_vmbus.h"
@@ -80,6 +81,7 @@ int hv_synic_alloc(void)
{
int cpu;
struct hv_per_cpu_context *hv_cpu;
+ int ret = -ENOMEM;
/*
* First, zero all per-cpu memory areas so hv_synic_free() can
@@ -120,9 +122,42 @@ int hv_synic_alloc(void)
(void *)get_zeroed_page(GFP_ATOMIC);
if (hv_cpu->synic_event_page == NULL) {
pr_err("Unable to allocate SYNIC event page\n");
+
+ free_page((unsigned long)hv_cpu->synic_message_page);
+ hv_cpu->synic_message_page = NULL;
+
goto err;
}
}
+
+ /* It's better to leak the page if the decryption fails. */
+ if (hv_isolation_type_tdx()) {
+ ret = set_memory_decrypted(
+ (unsigned long)hv_cpu->synic_message_page, 1);
+ if (ret) {
+ pr_err("Failed to decrypt SYNIC msg page: %d\n", ret);
+ hv_cpu->synic_message_page = NULL;
+
+ /*
+ * Free the event page so that a TDX VM won't
+ * try to encrypt the page in hv_synic_free().
+ */
+ free_page((unsigned long)hv_cpu->synic_event_page);
+ hv_cpu->synic_event_page = NULL;
+ goto err;
+ }
+
+ ret = set_memory_decrypted(
+ (unsigned long)hv_cpu->synic_event_page, 1);
+ if (ret) {
+ pr_err("Failed to decrypt SYNIC event page: %d\n", ret);
+ hv_cpu->synic_event_page = NULL;
+ goto err;
+ }
+
+ memset(hv_cpu->synic_message_page, 0, PAGE_SIZE);
+ memset(hv_cpu->synic_event_page, 0, PAGE_SIZE);
+ }
}
return 0;
@@ -131,18 +166,40 @@ int hv_synic_alloc(void)
* Any memory allocations that succeeded will be freed when
* the caller cleans up by calling hv_synic_free()
*/
- return -ENOMEM;
+ return ret;
}
void hv_synic_free(void)
{
int cpu;
+ int ret;
for_each_present_cpu(cpu) {
struct hv_per_cpu_context *hv_cpu
= per_cpu_ptr(hv_context.cpu_context, cpu);
+ /* It's better to leak the page if the encryption fails. */
+ if (hv_isolation_type_tdx()) {
+ if (hv_cpu->synic_message_page) {
+ ret = set_memory_encrypted((unsigned long)
+ hv_cpu->synic_message_page, 1);
+ if (ret) {
+ pr_err("Failed to encrypt SYNIC msg page: %d\n", ret);
+ hv_cpu->synic_message_page = NULL;
+ }
+ }
+
+ if (hv_cpu->synic_event_page) {
+ ret = set_memory_encrypted((unsigned long)
+ hv_cpu->synic_event_page, 1);
+ if (ret) {
+ pr_err("Failed to encrypt SYNIC event page: %d\n", ret);
+ hv_cpu->synic_event_page = NULL;
+ }
+ }
+ }
+
free_page((unsigned long)hv_cpu->synic_event_page);
free_page((unsigned long)hv_cpu->synic_message_page);
}
@@ -179,7 +236,8 @@ void hv_synic_enable_regs(unsigned int cpu)
if (!hv_cpu->synic_message_page)
pr_err("Fail to map synic message page.\n");
} else {
- simp.base_simp_gpa = virt_to_phys(hv_cpu->synic_message_page)
+ simp.base_simp_gpa =
+ cc_mkdec(virt_to_phys(hv_cpu->synic_message_page))
>> HV_HYP_PAGE_SHIFT;
}
@@ -198,7 +256,8 @@ void hv_synic_enable_regs(unsigned int cpu)
if (!hv_cpu->synic_event_page)
pr_err("Fail to map synic event page.\n");
} else {
- siefp.base_siefp_gpa = virt_to_phys(hv_cpu->synic_event_page)
+ siefp.base_siefp_gpa =
+ cc_mkdec(virt_to_phys(hv_cpu->synic_event_page))
>> HV_HYP_PAGE_SHIFT;
}
--
2.25.1
Powered by blists - more mailing lists