Skip to contentSkip to conversation

Optionally flush L1D on context switch

PATCH Applied Balbir Singh 3 participants 26 messages May 10, 2020 – Sep 16, 2020

Subsystems (5)
Documentation x86 KVM x86 x86 MM +1 more subsystems

v6 of 7 Applied Latest is v7 →

lore
Share & tools
lore
#0 lore raw

v6 cover Optionally flush L1D on context switch

Provide a mechanism to flush the L1D cache on context switch. The goal is to allow tasks that are paranoid due to the recent snoop assisted data sampling vulnerabilites, to flush their L1D on being switched out. This protects their data from being snooped or leaked via side channels after the task has context switched out.

Changelog v6:
- Fix the complaint about variable shadowing (Reported-by: kbuild test robot <[email protected]>)
Changelog v5:
- Based on Tom's recommendation, restrict the patches to Intel CPUs only ([email protected])
- Update reviewed-by tags based on v4.
Changelog v4:
- Refactor the L1D flushing code even further, pages are now allocated once and never freed. Simplify the exported functions.
- Change the name prefixs to be more consistent (l1d_flush_*)
- Refactoring of the code done in the spirit of the comments, prctl still requires arch bits for get/set L1D flush and ofcourse in the arch switch_mm bits flushing the L1D cache.
Changelog v3:
- Refactor the return value of what flush_l1d_cache_hw() returns
- Refactor the code, so that the generic setup bits come first (patch 3 from previous posting is now patches 3 and 4)
- Move from arch_prctl() to the prctl() interface as recommend in the reviews.
Changelog v2:
- Fix a miss of mutex_unlock (caught by Borislav Petkov <[email protected]>)
- Add documentation about the changes (Josh Poimboeuf <[email protected]>)

Changelog:
- Refactor the code and reuse cond_ibpb() - code bits provided by tglx
- Merge mm state tracking for ibpb and l1d flush
- Rename TIF_L1D_FLUSH to TIF_SPEC_FLUSH_L1D

Changelog RFC:
- Reuse existing code for allocation and flush
- Simplify the goto logic in the actual l1d_flush function
- Optimize the code path with jump labels/static functions

The previous version of these patches are posted at:

https://lore.kernel.org/lkml/[email protected]/

Balbir Singh (6):
arch/x86/kvm: Refactor l1d flush lifecycle management arch/x86/kvm: Refactor tlbflush and l1d flush arch/x86/mm: Refactor cond_ibpb() to support other use cases arch/x86/kvm: Refactor L1D flushing
Optionally flush L1D on context switch
Documentation: Add L1D flushing Documentation

 Documentation/admin-guide/hw-vuln/index.rst   |   1 +
 .../admin-guide/hw-vuln/l1d_flush.rst         |  40 ++++++
 arch/x86/include/asm/cacheflush.h             |   8 ++
 arch/x86/include/asm/thread_info.h            |   7 +-
 arch/x86/include/asm/tlbflush.h               |   2 +-
 arch/x86/kernel/Makefile                      |   1 +
 arch/x86/kernel/l1d_flush.c                   | 120 ++++++++++++++++++
 arch/x86/kvm/vmx/vmx.c                        |  62 +--------
 arch/x86/mm/tlb.c                             |  83 +++++++++---
 include/uapi/linux/prctl.h                    |   4 +
 kernel/sys.c                                  |  20 +++
 11 files changed, 266 insertions(+), 82 deletions(-)
 create mode 100644 Documentation/admin-guide/hw-vuln/l1d_flush.rst
 create mode 100644 arch/x86/kernel/l1d_flush.c
Signature · 1 line
2.17.1
#1 lore raw

PATCH v6 1/6 arch/x86/kvm: Refactor l1d flush lifecycle management

Split out the allocation and free routines to be used in a follow up set of patches (to reuse for L1D flushing).

---
 arch/x86/include/asm/cacheflush.h |  3 +++
 arch/x86/kernel/Makefile          |  1 +
 arch/x86/kernel/l1d_flush.c       | 36 +++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.c            | 25 +++------------------
 4 files changed, 43 insertions(+), 22 deletions(-)
 create mode 100644 arch/x86/kernel/l1d_flush.c
arch/x86/include/asm/cacheflush.h +3 -0
index 63feaf2a5f93..bac56fcd9790 100644
@@ -6,6 +6,9 @@
 #include <asm-generic/cacheflush.h>
 #include <asm/special_insns.h>
 
+#define L1D_CACHE_ORDER 4
 void clflush_cache_range(void *addr, unsigned int size);
+void *l1d_flush_alloc_pages(void);
+void l1d_flush_cleanup_pages(void *l1d_flush_pages);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/Makefile +1 -0
index e77261db2391..c17c1e3c1a0b 100644
@@ -157,3 +157,4 @@ ifeq ($(CONFIG_X86_64),y)
 endif
 
 obj-$(CONFIG_IMA_SECURE_AND_OR_TRUSTED_BOOT)	+= ima_arch.o
+obj-y						+= l1d_flush.o
arch/x86/kernel/l1d_flush.c new file +36 -0
new file mode 100644
index 000000000000..d605878c8f28
@@ -0,0 +1,36 @@
+#include <linux/mm.h>
+#include <asm/cacheflush.h>
+
+void *l1d_flush_alloc_pages(void)
+{
+	struct page *page;
+	void *l1d_flush_pages = NULL;
+	int i;
+
+	/*
+	 * This allocation for l1d_flush_pages is not tied to a VM/task's
+	 * lifetime and so should not be charged to a memcg.
+	 */
+	page = alloc_pages(GFP_KERNEL, L1D_CACHE_ORDER);
+	if (!page)
+		return NULL;
+	l1d_flush_pages = page_address(page);
+
+	/*
+	 * Initialize each page with a different pattern in
+	 * order to protect against KSM in the nested
+	 * virtualization case.
+	 */
+	for (i = 0; i < 1u << L1D_CACHE_ORDER; ++i) {
+		memset(l1d_flush_pages + i * PAGE_SIZE, i + 1,
+				PAGE_SIZE);
+	}
+	return l1d_flush_pages;
+}
+EXPORT_SYMBOL_GPL(l1d_flush_alloc_pages);
+
+void l1d_flush_cleanup_pages(void *l1d_flush_pages)
+{
+	free_pages((unsigned long)l1d_flush_pages, L1D_CACHE_ORDER);
+}
+EXPORT_SYMBOL_GPL(l1d_flush_cleanup_pages);
arch/x86/kvm/vmx/vmx.c +3 -22
index 00d31a5e0089..f35654db904a 100644
@@ -203,14 +203,10 @@ static const struct {
 	[VMENTER_L1D_FLUSH_NOT_REQUIRED] = {"not required", false},
 };
 
-#define L1D_CACHE_ORDER 4
 static void *vmx_l1d_flush_pages;
 
 static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 {
-	struct page *page;
-	unsigned int i;
-
 	if (!boot_cpu_has_bug(X86_BUG_L1TF)) {
 		l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_NOT_REQUIRED;
 		return 0;
@@ -253,24 +249,9 @@ static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 
 	if (l1tf != VMENTER_L1D_FLUSH_NEVER && !vmx_l1d_flush_pages &&
 	    !boot_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		/*
-		 * This allocation for vmx_l1d_flush_pages is not tied to a VM
-		 * lifetime and so should not be charged to a memcg.
-		 */
-		page = alloc_pages(GFP_KERNEL, L1D_CACHE_ORDER);
-		if (!page)
+		vmx_l1d_flush_pages = l1d_flush_alloc_pages();
+		if (!vmx_l1d_flush_pages)
 			return -ENOMEM;
-		vmx_l1d_flush_pages = page_address(page);
-
-		/*
-		 * Initialize each page with a different pattern in
-		 * order to protect against KSM in the nested
-		 * virtualization case.
-		 */
-		for (i = 0; i < 1u << L1D_CACHE_ORDER; ++i) {
-			memset(vmx_l1d_flush_pages + i * PAGE_SIZE, i + 1,
-			       PAGE_SIZE);
-		}
 	}
 
 	l1tf_vmx_mitigation = l1tf;
@@ -8099,7 +8080,7 @@ static struct kvm_x86_init_ops vmx_init_ops __initdata = {
 static void vmx_cleanup_l1d_flush(void)
 {
 	if (vmx_l1d_flush_pages) {
-		free_pages((unsigned long)vmx_l1d_flush_pages, L1D_CACHE_ORDER);
+		l1d_flush_cleanup_pages(vmx_l1d_flush_pages);
 		vmx_l1d_flush_pages = NULL;
 	}
 	/* Restore state so sysfs ignores VMX */
Signature · 1 line
2.17.1
#7 lore raw

PATCH v6 2/6 arch/x86/kvm: Refactor tlbflush and l1d flush

In reply to Balbir Singh

Refactor the existing assembly bits into smaller helper functions and also abstract L1D_FLUSH into a helper function. Use these functions in kvm for L1D flushing.

---
 arch/x86/include/asm/cacheflush.h |  3 ++
 arch/x86/kernel/l1d_flush.c       | 54 +++++++++++++++++++++++++++++++
 arch/x86/kvm/vmx/vmx.c            | 29 ++---------------
 3 files changed, 60 insertions(+), 26 deletions(-)
arch/x86/include/asm/cacheflush.h +3 -0
index bac56fcd9790..21cc3b28fa63 100644
@@ -8,7 +8,10 @@
 
 #define L1D_CACHE_ORDER 4
 void clflush_cache_range(void *addr, unsigned int size);
+void l1d_flush_populate_tlb(void *l1d_flush_pages);
 void *l1d_flush_alloc_pages(void);
 void l1d_flush_cleanup_pages(void *l1d_flush_pages);
+void l1d_flush_sw(void *l1d_flush_pages);
+int l1d_flush_hw(void);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/l1d_flush.c +54 -0

57 lines not on this page · raw message

arch/x86/kvm/vmx/vmx.c +3 -26
index f35654db904a..4f95927aad4c 100644
@@ -6031,8 +6031,6 @@ static int vmx_handle_exit(struct kvm_vcpu *vcpu,
  */
 static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 {
-	int size = PAGE_SIZE << L1D_CACHE_ORDER;
-
 	/*
 	 * This code is only executed when the the flush mode is 'cond' or
 	 * 'always'
@@ -6061,32 +6059,11 @@ static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 
 	vcpu->stat.l1d_flush++;
 
-	if (static_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		wrmsrl(MSR_IA32_FLUSH_CMD, L1D_FLUSH);
+	if (!l1d_flush_hw())
 		return;
-	}
 
-	asm volatile(
-		/* First ensure the pages are in the TLB */
-		"xorl	%%eax, %%eax\n"
-		".Lpopulate_tlb:\n\t"
-		"movzbl	(%[flush_pages], %%" _ASM_AX "), %%ecx\n\t"
-		"addl	$4096, %%eax\n\t"
-		"cmpl	%%eax, %[size]\n\t"
-		"jne	.Lpopulate_tlb\n\t"
-		"xorl	%%eax, %%eax\n\t"
-		"cpuid\n\t"
-		/* Now fill the cache */
-		"xorl	%%eax, %%eax\n"
-		".Lfill_cache:\n"
-		"movzbl	(%[flush_pages], %%" _ASM_AX "), %%ecx\n\t"
-		"addl	$64, %%eax\n\t"
-		"cmpl	%%eax, %[size]\n\t"
-		"jne	.Lfill_cache\n\t"
-		"lfence\n"
-		:: [flush_pages] "r" (vmx_l1d_flush_pages),
-		    [size] "r" (size)
-		: "eax", "ebx", "ecx", "edx");
+	l1d_flush_populate_tlb(vmx_l1d_flush_pages);
+	l1d_flush_sw(vmx_l1d_flush_pages);
 }
 
 static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
Signature · 1 line
2.17.1
#9 lore raw

PATCH v6 3/6 arch/x86/mm: Refactor cond_ibpb() to support other use cases

In reply to Balbir Singh

cond_ibpb() has the necessary bits required to track the previous mm in switch_mm_irqs_off(). This can be reused for other use cases like L1D flushing (on context switch out).

---
 arch/x86/include/asm/tlbflush.h |  2 +-
 arch/x86/mm/tlb.c               | 43 +++++++++++++++++----------------
 2 files changed, 23 insertions(+), 22 deletions(-)
arch/x86/include/asm/tlbflush.h +1 -1
index 8c87a2e0b660..a927d40664df 100644
@@ -83,7 +83,7 @@ struct tlb_state {
 	/* Last user mm for optimizing IBPB */
 	union {
 		struct mm_struct	*last_user_mm;
-		unsigned long		last_user_mm_ibpb;
+		unsigned long		last_user_mm_spec;
 	};
 
 	u16 loaded_mm_asid;
arch/x86/mm/tlb.c +22 -21

105 lines not on this page · raw message

Signature · 1 line
2.17.1
#15 lore raw

PATCH v6 5/6 Optionally flush L1D on context switch

In reply to Balbir Singh

Implement a mechanism to selectively flush the L1D cache. The goal is to allow tasks that are paranoid due to the recent snoop assisted data sampling vulnerabilites, to flush their L1D on being switched out. This protects their data from being snooped or leaked via side channels after the task has context switched out.

There are two scenarios we might want to protect against, a task leaving the CPU with data still in L1D (which is the main concern of this patch), the second scenario is a malicious task coming in (not so well trusted) for which we want to clean up the cache before it starts. Only the case for the former is addressed.

A new thread_info flag TIF_SPEC_FLUSH_L1D is added to track tasks which opt-into L1D flushing. cpu_tlbstate.last_user_mm_spec is used to convert the TIF flags into mm state (per cpu via last_user_mm_spec) in cond_mitigation(), which then used to do decide when to call flush_l1d().

Add prctl()'s to opt-in to the L1D cache on context switch out, the existing mechanisms of tracking prev_mm via cpu_tlbstate is reused to track state of the tasks and to flush the L1D cache. The prctl interface is generic and can be ported over to other architectures.

---
 arch/x86/include/asm/thread_info.h |  7 ++++-
 arch/x86/mm/tlb.c                  | 44 ++++++++++++++++++++++++++++--
 include/uapi/linux/prctl.h         |  4 +++
 kernel/sys.c                       | 20 ++++++++++++++
 4 files changed, 72 insertions(+), 3 deletions(-)
arch/x86/include/asm/thread_info.h +6 -1
index 8de8ceccb8bc..67de693d9ba1 100644
@@ -84,7 +84,7 @@ struct thread_info {
 #define TIF_SYSCALL_AUDIT	7	/* syscall auditing active */
 #define TIF_SECCOMP		8	/* secure computing */
 #define TIF_SPEC_IB		9	/* Indirect branch speculation mitigation */
-#define TIF_SPEC_FORCE_UPDATE	10	/* Force speculation MSR update in context switch */
+#define TIF_SPEC_FLUSH_L1D	10	/* Flush L1D on mm switches (processes) */
 #define TIF_USER_RETURN_NOTIFY	11	/* notify kernel of userspace return */
 #define TIF_UPROBE		12	/* breakpointed or singlestepping */
 #define TIF_PATCH_PENDING	13	/* pending live patching update */
@@ -96,6 +96,7 @@ struct thread_info {
 #define TIF_MEMDIE		20	/* is terminating due to OOM killer */
 #define TIF_POLLING_NRFLAG	21	/* idle is polling for TIF_NEED_RESCHED */
 #define TIF_IO_BITMAP		22	/* uses I/O bitmap */
+#define TIF_SPEC_FORCE_UPDATE	23	/* Force speculation MSR update in context switch */
 #define TIF_FORCED_TF		24	/* true if TF in eflags artificially */
 #define TIF_BLOCKSTEP		25	/* set when we want DEBUGCTLMSR_BTF */
 #define TIF_LAZY_MMU_UPDATES	27	/* task is updating the mmu lazily */
@@ -132,6 +133,7 @@ struct thread_info {
 #define _TIF_ADDR32		(1 << TIF_ADDR32)
 #define _TIF_X32		(1 << TIF_X32)
 #define _TIF_FSCHECK		(1 << TIF_FSCHECK)
+#define _TIF_SPEC_FLUSH_L1D	(1 << TIF_SPEC_FLUSH_L1D)
 
 /* Work to do before invoking the actual syscall. */
 #define _TIF_WORK_SYSCALL_ENTRY	\
@@ -235,6 +237,9 @@ static inline int arch_within_stack_frames(const void * const stack,
 			   current_thread_info()->status & TS_COMPAT)
 #endif
 
+extern int arch_prctl_l1d_flush_set(struct task_struct *tsk, unsigned long enable);
+extern int arch_prctl_l1d_flush_get(struct task_struct *tsk);
+
 extern void arch_task_cache_init(void);
 extern int arch_dup_task_struct(struct task_struct *dst, struct task_struct *src);
 extern void arch_release_task_struct(struct task_struct *tsk);
arch/x86/mm/tlb.c +42 -2

77 lines not on this page · raw message

include/uapi/linux/prctl.h +4 -0
index 07b4f8131e36..42cb3038c81a 100644
@@ -238,4 +238,8 @@ struct prctl_mm_map {
 #define PR_SET_IO_FLUSHER		57
 #define PR_GET_IO_FLUSHER		58
 
+/* Flush L1D on context switch (mm) */
+#define PR_SET_L1D_FLUSH		59
+#define PR_GET_L1D_FLUSH		60
+
 #endif /* _LINUX_PRCTL_H */
kernel/sys.c +20 -0
index d325f3ab624a..578aa8b6d87e 100644
@@ -2262,6 +2262,16 @@ int __weak arch_prctl_spec_ctrl_set(struct task_struct *t, unsigned long which,
 	return -EINVAL;
 }
 
+int __weak arch_prctl_l1d_flush_set(struct task_struct *tsk, unsigned long enable)
+{
+	return -EINVAL;
+}
+
+int __weak arch_prctl_l1d_flush_get(struct task_struct *t)
+{
+	return -EINVAL;
+}
+
 #define PR_IO_FLUSHER (PF_MEMALLOC_NOIO | PF_LESS_THROTTLE)
 
 SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3,
@@ -2514,6 +2524,16 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3,
 
 		error = (current->flags & PR_IO_FLUSHER) == PR_IO_FLUSHER;
 		break;
+	case PR_SET_L1D_FLUSH:
+		if (arg3 || arg4 || arg5)
+			return -EINVAL;
+		error = arch_prctl_l1d_flush_set(me, arg2);
+		break;
+	case PR_GET_L1D_FLUSH:
+		if (arg2 || arg3 || arg4 || arg5)
+			return -EINVAL;
+		error = arch_prctl_l1d_flush_get(me);
+		break;
 	default:
 		error = -EINVAL;
 		break;
Signature · 1 line
2.17.1
#23 lore raw

PATCH v6 6/6 Documentation: Add L1D flushing Documentation

In reply to Balbir Singh

Add documentation of l1d flushing, explain the need for the feature and how it can be used.

---
 Documentation/admin-guide/hw-vuln/index.rst   |  1 +
 .../admin-guide/hw-vuln/l1d_flush.rst         | 40 +++++++++++++++++++
 2 files changed, 41 insertions(+)
 create mode 100644 Documentation/admin-guide/hw-vuln/l1d_flush.rst
Documentation/admin-guide/hw-vuln/index.rst +1 -0
index 0795e3c2643f..35633b299d45 100644
@@ -14,3 +14,4 @@ are configurable at compile, boot or run time.
    mds
    tsx_async_abort
    multihit.rst
+   l1d_flush
Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +40 -0
new file mode 100644
index 000000000000..7d515b8c29f1
@@ -0,0 +1,40 @@
+L1D Flushing for the paranoid
+=============================
+
+With an increasing number of vulnerabilities being reported around data
+leaks from L1D, a new user space mechanism to flush the L1D cache on
+context switch is added to the kernel. This should help address
+CVE-2020-0550 and for paranoid applications, keep them safe from any
+yet to be discovered vulnerabilities, related to leaks from the L1D
+cache.
+
+Tasks can opt in to this mechanism by using a prctl (implemented only
+for x86 at the moment).
+
+Related CVES
+------------
+At the present moment, the following CVEs can be addressed by this
+mechanism
+
+    =============       ========================     ==================
+    CVE-2020-0550       Improper Data Forwarding     OS related aspects
+    =============       ========================     ==================
+
+Usage Guidelines
+----------------
+Applications can call ``prctl(2)`` with one of these two arguments
+
+1. PR_SET_L1D_FLUSH - flush the L1D cache on context switch (out)
+2. PR_GET_L1D_FLUSH - get the current state of the L1D cache flush, returns 1
+   if set and 0 if not set.
+
+**NOTE**: The feature is disabled by default, applications to need to specifically
+opt into the feature to enable it.
+
+Mitigation
+----------
+When PR_SET_L1D_FLUSH is enabled for a task, on switching tasks (when
+the address space changes), a flush of the L1D cache is performed for
+the task when it leaves the CPU. If the underlying CPU supports L1D
+flushing in hardware, the hardware mechanism is used, otherwise a software
+fallback, similar to the mechanism used by L1TF is used.
Signature · 1 line
2.17.1
#13 lore raw

PATCH v6 4/6 arch/x86/kvm: Refactor L1D flushing

In reply to Balbir Singh

Move out the initialization function to l1d_flush_init_once() so that it can be reused for subsequent patches. The side-effect of this patch is that the memory allocated for l1d flush pages is no longer freed up and the memory allocated once is shared amongst callers.

l1d_flush_sw/hw() are now abstracted under arch_l1d_flush(). vmx_l1d_flush_mutex however continues to exist as it also used from other code paths.

---
 arch/x86/include/asm/cacheflush.h | 12 +++---
 arch/x86/kernel/l1d_flush.c       | 64 +++++++++++++++++++++++--------
 arch/x86/kvm/vmx/vmx.c            | 20 ++--------
 3 files changed, 57 insertions(+), 39 deletions(-)
arch/x86/include/asm/cacheflush.h +7 -5
index 21cc3b28fa63..851d8f1ab827 100644
@@ -7,11 +7,13 @@
 #include <asm/special_insns.h>
 
 #define L1D_CACHE_ORDER 4
+
+enum l1d_flush_options {
+	L1D_FLUSH_POPULATE_TLB = 0x1,
+};
+
 void clflush_cache_range(void *addr, unsigned int size);
-void l1d_flush_populate_tlb(void *l1d_flush_pages);
-void *l1d_flush_alloc_pages(void);
-void l1d_flush_cleanup_pages(void *l1d_flush_pages);
-void l1d_flush_sw(void *l1d_flush_pages);
-int l1d_flush_hw(void);
+int l1d_flush_init_once(void);
+void arch_l1d_flush(enum l1d_flush_options options);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/l1d_flush.c +47 -17

111 lines not on this page · raw message

arch/x86/kvm/vmx/vmx.c +3 -17
index 4f95927aad4c..d56702578588 100644
@@ -203,8 +203,6 @@ static const struct {
 	[VMENTER_L1D_FLUSH_NOT_REQUIRED] = {"not required", false},
 };
 
-static void *vmx_l1d_flush_pages;
-
 static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 {
 	if (!boot_cpu_has_bug(X86_BUG_L1TF)) {
@@ -247,12 +245,9 @@ static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 		l1tf = VMENTER_L1D_FLUSH_ALWAYS;
 	}
 
-	if (l1tf != VMENTER_L1D_FLUSH_NEVER && !vmx_l1d_flush_pages &&
-	    !boot_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		vmx_l1d_flush_pages = l1d_flush_alloc_pages();
-		if (!vmx_l1d_flush_pages)
+	if (l1tf != VMENTER_L1D_FLUSH_NEVER)
+		if (l1d_flush_init_once())
 			return -ENOMEM;
-	}
 
 	l1tf_vmx_mitigation = l1tf;
 
@@ -6058,12 +6053,7 @@ static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 	}
 
 	vcpu->stat.l1d_flush++;
-
-	if (!l1d_flush_hw())
-		return;
-
-	l1d_flush_populate_tlb(vmx_l1d_flush_pages);
-	l1d_flush_sw(vmx_l1d_flush_pages);
+	arch_l1d_flush(L1D_FLUSH_POPULATE_TLB);
 }
 
 static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
@@ -8056,10 +8046,6 @@ static struct kvm_x86_init_ops vmx_init_ops __initdata = {
 
 static void vmx_cleanup_l1d_flush(void)
 {
-	if (vmx_l1d_flush_pages) {
-		l1d_flush_cleanup_pages(vmx_l1d_flush_pages);
-		vmx_l1d_flush_pages = NULL;
-	}
 	/* Restore state so sysfs ignores VMX */
 	l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_AUTO;
 }
Signature · 1 line
2.17.1
#24 lore raw

Re: [PATCH v6 6/6] Documentation: Add L1D flushing Documentation

In reply to Balbir Singh

Review

Quoted 3 lines · Balbir Singh
(quoted diff) +3 -0
+With an increasing number of vulnerabilities being reported around data
+leaks from L1D, a new user space mechanism to flush the L1D cache on
+context switch is added to the kernel. This should help address

is added to the kernel? This is documentation of an existing feature...

Review

Quoted 7 lines
(quoted diff) +7 -0
+Mitigation
+----------
+When PR_SET_L1D_FLUSH is enabled for a task, on switching tasks (when
+the address space changes), a flush of the L1D cache is performed for
+the task when it leaves the CPU. If the underlying CPU supports L1D
+flushing in hardware, the hardware mechanism is used, otherwise a software
+fallback, similar to the mechanism used by L1TF is used.

This lacks documentation of the limitations, especially that this does not help against cross Hyperthread attacks.

I've massaged the whole thing a bit. See below.

Thanks,

tglx
8<-----------------

Documentation/admin-guide/hw-vuln/index.rst +1 -0
@@ -14,3 +14,4 @@ are configurable at compile, boot or run
    mds
    tsx_async_abort
    multihit.rst
+   l1d_flush
Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +53 -0

53 lines not on this page · raw message

#2 lore raw

Re: [PATCH v6 1/6] arch/x86/kvm: Refactor l1d flush lifecycle management

In reply to Balbir Singh
Quoted 1 line · Balbir Singh

Subject: [PATCH v6 1/6] arch/x86/kvm: Refactor....

arch/x86/kvm: is really not the correct subsystem prefix...

I'll fix it up this time.

#4 lore raw

Re: [PATCH v6 1/6] arch/x86/kvm: Refactor l1d flush lifecycle management

In reply to Balbir Singh

Review

Quoted 2 lines · Balbir Singh

+++ b/arch/x86/kernel/l1d_flush.c

(quoted diff) +0 -0
@@ -0,0 +1,36 @@

Lacks

+// SPDX-License-Identifier: GPL-2.0-only

#10 lore raw

Re: [PATCH v6 3/6] arch/x86/mm: Refactor cond_ibpb() to support other use cases

In reply to Balbir Singh

arch/x86/mm/tlb.c void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next, · line 553

Quoted 6 lines · Balbir Singh
arch/x86/mm/tlb.c +2 -0
@@ -550,8 +549,10 @@ void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next,
 		 * Avoid user/user BTB poisoning by flushing the branch
 		 * predictor when switching between processes. This stops
 		 * one process from doing Spectre-v2 attacks on another.
+		 * The hook can also be used for mitigations that rely
+		 * on switch_mm for hooks.

The new function name has absolutely nothing to do with IBPB and is clearly talking about mitigations. So the IBPB comment wants to move and that extra sentence you bolted on can go away. It's nonsensical word salad anyway.

arch/x86/mm/tlb.c void initialize_tlbstate_and_flush(void) · line 681

Quoted 4 lines
arch/x86/mm/tlb.c +1 -1
 	/* Reinitialize tlbstate. */
-	this_cpu_write(cpu_tlbstate.last_user_mm_ibpb, LAST_USER_MM_IBPB);
+	this_cpu_write(cpu_tlbstate.last_user_mm_spec, LAST_USER_MM_IBPB);

There is still no comment why this only needs MM_IBPB. I'll change this to LAST_USER_MM_INIT and put that define close to the others.

Thanks,

tglx

#16 lore raw
In reply to Balbir Singh

arch/x86/mm/tlb.c static void cond_mitigation(struct task_struct *next) · line 459

Quoted 3 lines · Balbir Singh
arch/x86/mm/tlb.c +2 -0
+	if (prev_mm & LAST_USER_MM_L1D_FLUSH)
+		arch_l1d_flush(0); /* Just flush, don't populate the TLB */

Bah. I fundamentally hate tail comments. They are just disturbing the reading flow. Aside of that, this states the WHAT but not the WHY. And if you add that explanation then you need more than 20 characters and end up with

	if (prev_mm & LAST_USER_MM_L1D_FLUSH) {
        	/*
                 * Proper comment explaining why this is flushing
        	 * without prepopulating the TLB.
                 */
 		arch_l1d_flush(0);
        }

anyway. And even for a short comment which fits after the function call it's way better to have:

	if (prev_mm & LAST_USER_MM_L1D_FLUSH) {
        	/* Short explanation */
 		arch_l1d_flush(0);
        }

Hmm?

arch/x86/mm/tlb.c void initialize_tlbstate_and_flush(void) · line 719

Quoted 4 lines
arch/x86/mm/tlb.c +4 -0
+	/*
+	 * Leave last_user_mm_spec at LAST_USER_MM_IBPB, we don't
+	 * want to set LAST_USER_MM_L1D_FLUSH and force a flush before
+	 * we've allocated the flush pages.

Ah here is the comment. I still like the explicit define for the (re) init.

#18 lore raw
In reply to Balbir Singh
Quoted 55 lines · Balbir Singh

Implement a mechanism to selectively flush the L1D cache. The goal is to allow tasks that are paranoid due to the recent snoop assisted data sampling vulnerabilites, to flush their L1D on being switched out. This protects their data from being snooped or leaked via side channels after the task has context switched out.

There are two scenarios we might want to protect against, a task leaving the CPU with data still in L1D (which is the main concern of this patch), the second scenario is a malicious task coming in (not so well trusted) for which we want to clean up the cache before it starts. Only the case for the former is addressed.

A new thread_info flag TIF_SPEC_FLUSH_L1D is added to track tasks which opt-into L1D flushing. cpu_tlbstate.last_user_mm_spec is used to convert the TIF flags into mm state (per cpu via last_user_mm_spec) in cond_mitigation(), which then used to do decide when to call flush_l1d().

Add prctl()'s to opt-in to the L1D cache on context switch out, the existing mechanisms of tracking prev_mm via cpu_tlbstate is reused to track state of the tasks and to flush the L1D cache. The prctl interface is generic and can be ported over to other architectures.

---
 arch/x86/include/asm/thread_info.h |  7 ++++-
 arch/x86/mm/tlb.c                  | 44 ++++++++++++++++++++++++++++--
 include/uapi/linux/prctl.h         |  4 +++
 kernel/sys.c                       | 20 ++++++++++++++
 4 files changed, 72 insertions(+), 3 deletions(-)
arch/x86/include/asm/thread_info.h +2 -1
index 8de8ceccb8bc..67de693d9ba1 100644
@@ -84,7 +84,7 @@ struct thread_info {
 #define TIF_SYSCALL_AUDIT	7	/* syscall auditing active */
 #define TIF_SECCOMP		8	/* secure computing */
 #define TIF_SPEC_IB		9	/* Indirect branch speculation mitigation */
-#define TIF_SPEC_FORCE_UPDATE	10	/* Force speculation MSR update in context switch */
+#define TIF_SPEC_FLUSH_L1D	10	/* Flush L1D on mm switches (processes) */
 #define TIF_USER_RETURN_NOTIFY	11	/* notify kernel of userspace return */
 #define TIF_UPROBE		12	/* breakpointed or singlestepping */
 #define TIF_PATCH_PENDING	13	/* pending live patching update */
@@ -96,6 +96,7 @@ struct thread_info {
 #define TIF_MEMDIE		20	/* is terminating due to OOM killer */
 #define TIF_POLLING_NRFLAG	21	/* idle is polling for TIF_NEED_RESCHED */
 #define TIF_IO_BITMAP		22	/* uses I/O bitmap */
+#define TIF_SPEC_FORCE_UPDATE	23	/* Force speculation MSR update in context switch */
 #define TIF_FORCED_TF		24	/* true if TF in eflags artificially */
 #define TIF_BLOCKSTEP		25	/* set when we want DEBUGCTLMSR_BTF */
 #define TIF_LAZY_MMU_UPDATES	27	/* task is updating the mmu lazily */

arch/x86/include/asm/thread_info.h struct thread_info { · line 136

Quoted 4 lines
arch/x86/include/asm/thread_info.h +1 -0
@@ -132,6 +133,7 @@ struct thread_info {
 #define _TIF_ADDR32		(1 << TIF_ADDR32)
 #define _TIF_X32		(1 << TIF_X32)
 #define _TIF_FSCHECK		(1 << TIF_FSCHECK)
+#define _TIF_SPEC_FLUSH_L1D	(1 << TIF_SPEC_FLUSH_L1D)

Bah. These defines are ordered in the same way as the TIF defines....

arch/x86/include/asm/thread_info.h

Quoted 9 lines
arch/x86/include/asm/thread_info.h +1 -1
 /*
- * Bits to mangle the TIF_SPEC_IB state into the mm pointer which is
+ * Bits to mangle the TIF_SPEC_* state into the mm pointer which is
  * stored in cpu_tlb_state.last_user_mm_spec.

*/
#define LAST_USER_MM_IBPB 0x1UL

arch/x86/include/asm/thread_info.h +2 -1
-#define LAST_USER_MM_SPEC_MASK	(LAST_USER_MM_IBPB)
+#define LAST_USER_MM_L1D_FLUSH	0x2UL
+#define LAST_USER_MM_SPEC_MASK	(LAST_USER_MM_IBPB | LAST_USER_MM_L1D_FLUSH)

You lost

+ BUILD_BUG_ON(TIF_SPEC_FLUSH_L1D != TIF_SPEC_IB + 1);

from patch I gave you.

#20 lore raw
In reply to Balbir Singh

Balbir Singh <[email protected]> writes:

This part:

Quoted 9 lines
include/uapi/linux/prctl.h +3 -0
@@ -238,4 +238,8 @@ struct prctl_mm_map {
 #define PR_SET_IO_FLUSHER		57
 #define PR_GET_IO_FLUSHER		58
 
+/* Flush L1D on context switch (mm) */
+#define PR_SET_L1D_FLUSH		59
+#define PR_GET_L1D_FLUSH		60

...

Quoted 7 lines
include/uapi/linux/prctl.h +3 -0
@@ -2514,6 +2524,16 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3,
 
 		error = (current->flags & PR_IO_FLUSHER) == PR_IO_FLUSHER;
 		break;
+	case PR_SET_L1D_FLUSH:
+		if (arg3 || arg4 || arg5)
+			return -EINVAL;

include/uapi/linux/prctl.h SYSCALL_DEFINE5(prctl, int, option, unsigned long, arg2, unsigned long, arg3, · line 2539

Quoted 10 lines
include/uapi/linux/prctl.h +7 -0
+		error = arch_prctl_l1d_flush_set(me, arg2);
+		break;
+	case PR_GET_L1D_FLUSH:
+		if (arg2 || arg3 || arg4 || arg5)
+			return -EINVAL;
+		error = arch_prctl_l1d_flush_get(me);
+		break;
 	default:
 		error = -EINVAL;
 		break;

wants to be split into a separate patch, really. Then we get a proper subject lines with proper subsystem prefixes. This part also lacks a description in Documentation/userspace-api/ and function prototypes for the arch_prctl* functions.

But looking at this deeper (yes I should have noticed earlier):

Why do we need yet another PRCTL?

We already have PR_SET_SPECULATION_CTRL/PR_GET_SPECULATION_CTRL. That L1D flush thingy fits into this category, right?

This makes even more sense if you think about the second use case for L1D flush, i.e. the flush when a untrusted task comes in. If we ever want to support that case then this will be imposed by seccomp and then we'd need yet another interface there.

And for this reason we should also name that current opt-in thingy:
L1D_FLUSH_OUT in the prctl and also for the TIF bits.

Hmm? Kees?

I've applied the first 4 patches to:

git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git x86/mm

so the polishing I did gets preserved and you don't have to resend the whole pile.

Thanks,

tglx

#25 lore raw

Re: [PATCH v6 6/6] Documentation: Add L1D flushing Documentation

In reply to Thomas Gleixner
Quoted 12 lines · Thomas Gleixner
Quoted 5 lines · Balbir Singh

+With an increasing number of vulnerabilities being reported around data
+leaks from L1D, a new user space mechanism to flush the L1D cache on
+context switch is added to the kernel. This should help address

is added to the kernel? This is documentation of an existing feature...

Good catch! Thanks

Quoted 5 lines
Quoted 5 lines
Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +3 -0
+Mitigation
+----------
+When PR_SET_L1D_FLUSH is enabled for a task, on switching tasks

(when
+the address space changes), a flush of the L1D cache is performed

Quoted 9 lines
Quoted 5 lines

for

(quoted diff) +2 -0
+the task when it leaves the CPU. If the underlying CPU supports L1D
+flushing in hardware, the hardware mechanism is used, otherwise a

software
+fallback, similar to the mechanism used by L1TF is used.

This lacks documentation of the limitations, especially that this does not help against cross Hyperthread attacks.

Yes, true

Quoted 79 lines

I've massaged the whole thing a bit. See below.

Thanks,

tglx
8<-----------------

Documentation/admin-guide/hw-vuln/index.rst +1 -0
@@ -14,3 +14,4 @@ are configurable at compile, boot or run
    mds
    tsx_async_abort
    multihit.rst
+   l1d_flush
Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +4 -0
@@ -0,0 +1,53 @@
+L1D Flushing for the paranoid
+=============================
+
+With an increasing number of vulnerabilities being reported around

data

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +4 -0
+leaks from the Level 1 Data cache (L1D) the kernel provides an opt-in
+mechanism to flush the L1D cache on context switch.
+
+This mechanism can be used to address e.g. CVE-2020-0550. For

paranoid
+applications the mechanism keeps them safe from any yet to be discovered

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +9 -0
+vulnerabilities, related to leaks from the L1D cache.
+
+
+Related CVEs
+------------
+At the present moment, the following CVEs can be addressed by this
+mechanism
+
+    =============       ========================     ================

==
+ CVE-2020-0550 Improper Data Forwarding OS related
aspects
+ ============= ======================== ================
==

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +7 -0
+
+Usage Guidelines
+----------------
+Applications can call ``prctl(2)`` with one of these two arguments
+
+1. PR_SET_L1D_FLUSH - flush the L1D cache on context switch (out)
+2. PR_GET_L1D_FLUSH - get the current state of the L1D cache flush,

returns 1

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +9 -0
+   if set and 0 if not set.
+
+**NOTE**: The feature is disabled by default, applications need to
+specifically opt into the feature to enable it.
+
+Mitigation
+----------
+
+When PR_SET_L1D_FLUSH is enabled for a task a flush of the L1D cache

is
+performed when the task is scheduled out and the incoming task belongs to a

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +10 -0
+different process and therefore to a different address space.
+
+If the underlying CPU supports L1D flushing in hardware, the hardware
+mechanism is used, otherwise a software fallback, similar to the L1TF
+mitigation, is invoked.
+
+Limitations
+-----------
+
+The mechanism does not mitigate L1D data leaks between tasks

belonging to
+different processes which are concurrently executing on sibling threads of

Documentation/admin-guide/hw-vuln/l1d_flush.rst new file +3 -0
+a physical CPU core when SMT is enabled on the system.
+
+This can be addressed by controlled placement of processes on
Quoted 5 lines

physical CPU
+cores or by disabling SMT. See the relevant chapter in the L1TF mitigation
+document: :ref:Documentation/admin-guide/hw-vuln/l1tf.rst <smt_control>.

I like your addition above

Thanks,
Balbir Singh.

#21 lore raw
In reply to Thomas Gleixner
Quoted 42 lines · Thomas Gleixner

Balbir Singh <[email protected]> writes:

This part:

Quoted 9 lines
include/uapi/linux/prctl.h +3 -0
@@ -238,4 +238,8 @@ struct prctl_mm_map {
 #define PR_SET_IO_FLUSHER            57
 #define PR_GET_IO_FLUSHER            58
 
+/* Flush L1D on context switch (mm) */
+#define PR_SET_L1D_FLUSH             59
+#define PR_GET_L1D_FLUSH             60

...

Quoted 17 lines
include/uapi/linux/prctl.h +10 -0
@@ -2514,6 +2524,16 @@ SYSCALL_DEFINE5(prctl, int, option, unsigned
 long, arg2, unsigned long, arg3,
 
              error = (current->flags & PR_IO_FLUSHER) ==
 PR_IO_FLUSHER;
              break;
+     case PR_SET_L1D_FLUSH:
+             if (arg3 || arg4 || arg5)
+                     return -EINVAL;
+             error = arch_prctl_l1d_flush_set(me, arg2);
+             break;
+     case PR_GET_L1D_FLUSH:
+             if (arg2 || arg3 || arg4 || arg5)
+                     return -EINVAL;
+             error = arch_prctl_l1d_flush_get(me);
+             break;
      default:
Quoted 2 lines
              error = -EINVAL;
              break;

wants to be split into a separate patch, really. Then we get a proper subject lines with proper subsystem prefixes. This part also lacks a description in Documentation/userspace-api/ and function prototypes for
the arch_prctl* functions.

Quoted 6 lines

But looking at this deeper (yes I should have noticed earlier):

Why do we need yet another PRCTL?

We already have PR_SET_SPECULATION_CTRL/PR_GET_SPECULATION_CTRL. That L1D flush thingy fits into this category, right?

It does, I thought about it for a while when I was changing the code and left it aside because, looking at the definition

1 PR_SPEC_ENABLE The speculation feature is enabled,
mitigation is disabled.
2 PR_SPEC_DISABLE The speculation feature is disabled,
mitigation is enabled.

With L1D flush, there is no overriding of the feature as such (as in enable when the mitigation is disabled and vice-versa). I am happy to reconsider my initial thought though.

Quoted 7 lines

This makes even more sense if you think about the second use case for L1D flush, i.e. the flush when a untrusted task comes in. If we ever want to support that case then this will be imposed by seccomp and then
we'd need yet another interface there.

Yep, I see what you mean

Quoted 12 lines

And for this reason we should also name that current opt-in thingy:
L1D_FLUSH_OUT in the prctl and also for the TIF bits.

Hmm? Kees?

I've applied the first 4 patches to:

git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git x86/mm

so the polishing I did gets preserved and you don't have to resend the whole pile.

Thanks, I think your change to patch 6 makes sense as well. Let me respin this based on what you think of the argument above

Balbir

#17 lore raw
In reply to Thomas Gleixner
Quoted 13 lines · Thomas Gleixner
Quoted 4 lines · Balbir Singh
(quoted diff) +2 -0
+     if (prev_mm & LAST_USER_MM_L1D_FLUSH)
+             arch_l1d_flush(0); /* Just flush, don't populate the

TLB */

Bah. I fundamentally hate tail comments. They are just disturbing the reading flow. Aside of that, this states the WHAT but not the WHY. And if you add that explanation then you need more than 20 characters and end up with

Quoted 8 lines
        if (prev_mm & LAST_USER_MM_L1D_FLUSH) {
                /*
                 * Proper comment explaining why this is flushing
                 * without prepopulating the TLB.
                 */
                arch_l1d_flush(0);
        }

I added a comment due to the use of 0, 0 is usually seen as true or false and I wanted to add some comments in there to indicate we don't populate the TLB, the reason we don't do it is, I don't think we need to. I am happy to revisit the placement of the comment.

Quoted 11 lines

anyway. And even for a short comment which fits after the function call
it's way better to have:

        if (prev_mm & LAST_USER_MM_L1D_FLUSH) {
                /* Short explanation */
                arch_l1d_flush(0);
        }

Hmm?

Quoted 8 lines

arch/x86/mm/tlb.c void initialize_tlbstate_and_flush(void) · line 719

Quoted 4 lines
arch/x86/mm/tlb.c +4 -0
+     /*
+      * Leave last_user_mm_spec at LAST_USER_MM_IBPB, we don't
+      * want to set LAST_USER_MM_L1D_FLUSH and force a flush before
+      * we've allocated the flush pages.

Ah here is the comment. I still like the explicit define for the (re) init.

I saw your tree and it sounds like you fixed it up in there in patch 3.

Balbir Singh.

#3 lore raw

Re: [PATCH v6 1/6] arch/x86/kvm: Refactor l1d flush lifecycle management

In reply to Thomas Gleixner
Quoted 6 lines · Thomas Gleixner

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.

Quoted 7 lines
Quoted 1 line · Balbir Singh

Subject: [PATCH v6 1/6] arch/x86/kvm: Refactor....

arch/x86/kvm: is really not the correct subsystem prefix...

I'll fix it up this time.

Thanks, noted!

Balbir Singh.

#5 lore raw

Re: [PATCH v6 1/6] arch/x86/kvm: Refactor l1d flush lifecycle management

In reply to Thomas Gleixner
Quoted 10 lines · Thomas Gleixner

Review

Quoted 2 lines · Balbir Singh

+++ b/arch/x86/kernel/l1d_flush.c

(quoted diff) +0 -0
@@ -0,0 +1,36 @@

Lacks

+// SPDX-License-Identifier: GPL-2.0-only

Agreed, it should match the license in arch/x86/kvm/vmx/vmx.c

Thanks,
Balbir

#22 lore raw
In reply to Singh, Balbir

Balbir,

Quoted 12 lines · Balbir
Quoted 7 lines · Thomas Gleixner

Balbir Singh <[email protected]> writes:
But looking at this deeper (yes I should have noticed earlier):

Why do we need yet another PRCTL?

We already have PR_SET_SPECULATION_CTRL/PR_GET_SPECULATION_CTRL. That L1D flush thingy fits into this category, right?

It does, I thought about it for a while when I was changing the code and left it aside because, looking at the definition

Quoted 8 lines

1 PR_SPEC_ENABLE The speculation feature is enabled,
mitigation is disabled.
2 PR_SPEC_DISABLE The speculation feature is disabled,
mitigation is enabled.

With L1D flush, there is no overriding of the feature as such (as in enable when the mitigation is disabled and vice-versa). I am happy to reconsider my initial thought though.

L1D is always enabled as L1D will be a source of trouble forever :)

Thanks,

tglx

#19 lore raw
In reply to Thomas Gleixner
Quoted 109 lines · Thomas Gleixner

CAUTION: This email originated from outside of the organization. Do not click links or open attachments unless you can confirm the sender and know the content is safe.

Quoted 80 lines · Balbir Singh

Implement a mechanism to selectively flush the L1D cache. The goal is to
allow tasks that are paranoid due to the recent snoop assisted data sampling
vulnerabilites, to flush their L1D on being switched out. This protects
their data from being snooped or leaked via side channels after the task
has context switched out.

There are two scenarios we might want to protect against, a task leaving
the CPU with data still in L1D (which is the main concern of this patch),
the second scenario is a malicious task coming in (not so well trusted)
for which we want to clean up the cache before it starts. Only the case
for the former is addressed.

A new thread_info flag TIF_SPEC_FLUSH_L1D is added to track tasks which
opt-into L1D flushing. cpu_tlbstate.last_user_mm_spec is used to convert
the TIF flags into mm state (per cpu via last_user_mm_spec) in cond_mitigation(), which then used to do decide when to call flush_l1d().

Add prctl()'s to opt-in to the L1D cache on context switch out, the existing mechanisms of tracking prev_mm via cpu_tlbstate is reused to track state of the tasks and to flush the L1D cache. The prctl interface is generic and can be ported over to other architectures.

---
 arch/x86/include/asm/thread_info.h |  7 ++++-
 arch/x86/mm/tlb.c                  | 44
++++++++++++++++++++++++++++--
 include/uapi/linux/prctl.h         |  4 +++
 kernel/sys.c                       | 20 ++++++++++++++
 4 files changed, 72 insertions(+), 3 deletions(-)
arch/x86/include/asm/thread_info.h +0 -0
index 8de8ceccb8bc..67de693d9ba1 100644
@@ -84,7 +84,7 @@ struct thread_info {
 #define TIF_SYSCALL_AUDIT    7       /* syscall auditing active */
 #define TIF_SECCOMP          8       /* secure computing */
 #define TIF_SPEC_IB          9       /* Indirect branch speculation

mitigation */
-#define TIF_SPEC_FORCE_UPDATE 10 /* Force speculation
MSR update in context switch */
+#define TIF_SPEC_FLUSH_L1D 10 /* Flush L1D on mm switches
(processes) */
#define TIF_USER_RETURN_NOTIFY 11 /* notify kernel of
userspace return */
#define TIF_UPROBE 12 /* breakpointed or
singlestepping */
#define TIF_PATCH_PENDING 13 /* pending live patching
update */

arch/x86/include/asm/thread_info.h +0 -0
@@ -96,6 +96,7 @@ struct thread_info {
 #define TIF_MEMDIE           20      /* is terminating due to OOM

killer */
#define TIF_POLLING_NRFLAG 21 /* idle is polling for
TIF_NEED_RESCHED */

 #define TIF_IO_BITMAP                22      /* uses I/O bitmap */
+#define TIF_SPEC_FORCE_UPDATE        23      /* Force speculation

MSR update in context switch */
#define TIF_FORCED_TF 24 /* true if TF in
eflags artificially */
#define TIF_BLOCKSTEP 25 /* set when we want
DEBUGCTLMSR_BTF */
#define TIF_LAZY_MMU_UPDATES 27 /* task is updating the mmu
lazily */

arch/x86/include/asm/thread_info.h struct thread_info { · line 136

Quoted 4 lines
arch/x86/include/asm/thread_info.h +1 -0
@@ -132,6 +133,7 @@ struct thread_info {
 #define _TIF_ADDR32          (1 << TIF_ADDR32)
 #define _TIF_X32             (1 << TIF_X32)
 #define _TIF_FSCHECK         (1 << TIF_FSCHECK)
+#define _TIF_SPEC_FLUSH_L1D  (1 << TIF_SPEC_FLUSH_L1D)

Bah. These defines are ordered in the same way as the TIF defines....

Quoted 11 lines
arch/x86/include/asm/thread_info.h +0 -1
 /*
- * Bits to mangle the TIF_SPEC_IB state into the mm pointer which

is
+ * Bits to mangle the TIF_SPEC_* state into the mm pointer which is
* stored in cpu_tlb_state.last_user_mm_spec.
*/
#define LAST_USER_MM_IBPB 0x1UL

arch/x86/include/asm/thread_info.h +2 -1
-#define LAST_USER_MM_SPEC_MASK       (LAST_USER_MM_IBPB)
+#define LAST_USER_MM_L1D_FLUSH       0x2UL
+#define LAST_USER_MM_SPEC_MASK       (LAST_USER_MM_IBPB |

LAST_USER_MM_L1D_FLUSH)

You lost

Quoted 3 lines

+ BUILD_BUG_ON(TIF_SPEC_FLUSH_L1D != TIF_SPEC_IB + 1);

from patch I gave you.

Oops.. I'll fix up both and redo patch 5/6, by splitting it up, into interface vs flush bits

Thanks,
Balbir Singh.

merge bot Applied
#8 lore raw

[tip: x86/mm] x86/kvm: Refactor L1D flush operations

In reply to Balbir Singh
The following commit has been merged into the x86/mm branch of tip:

The following commit has been merged into the x86/mm branch of tip:

Commit-ID:     e3efae20ec69e9a8c9db1ad81b37de629219bbc4
Gitweb:        https://git.kernel.org/tip/e3efae20ec69e9a8c9db1ad81b37de629219bbc4
Author:        Balbir Singh <[email protected]>
AuthorDate:    Sun, 10 May 2020 11:47:59 +10:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Wed, 13 May 2020 18:12:19 +02:00

x86/kvm: Refactor L1D flush operations

Move the L1D flush functions into builtin code so they can be reused for L1D flush on context switch.

Split them up into:
- Hardware L1D flush
- TLB pre-populating of L1D pages for software based flushing
- Software based L1D flush

Adjust the KVM code accordingly.

[ tglx: Massaged changelog ]

---
 arch/x86/include/asm/cacheflush.h |  3 ++-
 arch/x86/kernel/l1d_flush.c       | 49 ++++++++++++++++++++++++++++++-
 arch/x86/kvm/vmx/vmx.c            | 29 +-----------------
 3 files changed, 55 insertions(+), 26 deletions(-)
arch/x86/include/asm/cacheflush.h +3 -0
index bac56fc..21cc3b2 100644
@@ -8,7 +8,10 @@
 
 #define L1D_CACHE_ORDER 4
 void clflush_cache_range(void *addr, unsigned int size);
+void l1d_flush_populate_tlb(void *l1d_flush_pages);
 void *l1d_flush_alloc_pages(void);
 void l1d_flush_cleanup_pages(void *l1d_flush_pages);
+void l1d_flush_sw(void *l1d_flush_pages);
+int l1d_flush_hw(void);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/l1d_flush.c +49 -0

52 lines not on this page · raw message

arch/x86/kvm/vmx/vmx.c +3 -26
index 225aa82..786d161 100644
@@ -5983,8 +5983,6 @@ unexpected_vmexit:
  */
 static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 {
-	int size = PAGE_SIZE << L1D_CACHE_ORDER;
-
 	/*
 	 * This code is only executed when the the flush mode is 'cond' or
 	 * 'always'
@@ -6013,32 +6011,11 @@ static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 
 	vcpu->stat.l1d_flush++;
 
-	if (static_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		wrmsrl(MSR_IA32_FLUSH_CMD, L1D_FLUSH);
+	if (!l1d_flush_hw())
 		return;
-	}
 
-	asm volatile(
-		/* First ensure the pages are in the TLB */
-		"xorl	%%eax, %%eax\n"
-		".Lpopulate_tlb:\n\t"
-		"movzbl	(%[flush_pages], %%" _ASM_AX "), %%ecx\n\t"
-		"addl	$4096, %%eax\n\t"
-		"cmpl	%%eax, %[size]\n\t"
-		"jne	.Lpopulate_tlb\n\t"
-		"xorl	%%eax, %%eax\n\t"
-		"cpuid\n\t"
-		/* Now fill the cache */
-		"xorl	%%eax, %%eax\n"
-		".Lfill_cache:\n"
-		"movzbl	(%[flush_pages], %%" _ASM_AX "), %%ecx\n\t"
-		"addl	$64, %%eax\n\t"
-		"cmpl	%%eax, %[size]\n\t"
-		"jne	.Lfill_cache\n\t"
-		"lfence\n"
-		:: [flush_pages] "r" (vmx_l1d_flush_pages),
-		    [size] "r" (size)
-		: "eax", "ebx", "ecx", "edx");
+	l1d_flush_populate_tlb(vmx_l1d_flush_pages);
+	l1d_flush_sw(vmx_l1d_flush_pages);
 }
 
 static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
merge bot Applied
#6 lore raw

[tip: x86/mm] x86/kvm: Refactor L1D flush page management

In reply to Balbir Singh
The following commit has been merged into the x86/mm branch of tip:

The following commit has been merged into the x86/mm branch of tip:

Commit-ID:     b9b3bc1c30be1f056c1c0564bc7268820ea8bf70
Gitweb:        https://git.kernel.org/tip/b9b3bc1c30be1f056c1c0564bc7268820ea8bf70
Author:        Balbir Singh <[email protected]>
AuthorDate:    Sun, 10 May 2020 11:47:58 +10:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Wed, 13 May 2020 18:12:18 +02:00

x86/kvm: Refactor L1D flush page management

Split out the allocation and free routines and move them into builtin code so they can be reused for the upcoming paranoid L1D flush on context switch mitigation.

[ tglx: Add missing SPDX identifier and massage subject and changelog ]

---
 arch/x86/include/asm/cacheflush.h |  3 ++-
 arch/x86/kernel/Makefile          |  1 +-
 arch/x86/kernel/l1d_flush.c       | 39 ++++++++++++++++++++++++++++++-
 arch/x86/kvm/vmx/vmx.c            | 25 ++-----------------
 4 files changed, 46 insertions(+), 22 deletions(-)
 create mode 100644 arch/x86/kernel/l1d_flush.c
arch/x86/include/asm/cacheflush.h +3 -0
index 63feaf2..bac56fc 100644
@@ -6,6 +6,9 @@
 #include <asm-generic/cacheflush.h>
 #include <asm/special_insns.h>
 
+#define L1D_CACHE_ORDER 4
 void clflush_cache_range(void *addr, unsigned int size);
+void *l1d_flush_alloc_pages(void);
+void l1d_flush_cleanup_pages(void *l1d_flush_pages);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/Makefile +1 -0
index ba89cab..c04d218 100644
@@ -156,3 +156,4 @@ ifeq ($(CONFIG_X86_64),y)
 endif
 
 obj-$(CONFIG_IMA_SECURE_AND_OR_TRUSTED_BOOT)	+= ima_arch.o
+obj-y						+= l1d_flush.o
arch/x86/kernel/l1d_flush.c new file +39 -0
new file mode 100644
index 0000000..4f298b7
@@ -0,0 +1,39 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#include <linux/mm.h>
+
+#include <asm/cacheflush.h>
+
+void *l1d_flush_alloc_pages(void)
+{
+	struct page *page;
+	void *l1d_flush_pages = NULL;
+	int i;
+
+	/*
+	 * This allocation for l1d_flush_pages is not tied to a VM/task's
+	 * lifetime and so should not be charged to a memcg.
+	 */
+	page = alloc_pages(GFP_KERNEL, L1D_CACHE_ORDER);
+	if (!page)
+		return NULL;
+	l1d_flush_pages = page_address(page);
+
+	/*
+	 * Initialize each page with a different pattern in
+	 * order to protect against KSM in the nested
+	 * virtualization case.
+	 */
+	for (i = 0; i < 1u << L1D_CACHE_ORDER; ++i) {
+		memset(l1d_flush_pages + i * PAGE_SIZE, i + 1,
+				PAGE_SIZE);
+	}
+	return l1d_flush_pages;
+}
+EXPORT_SYMBOL_GPL(l1d_flush_alloc_pages);
+
+void l1d_flush_cleanup_pages(void *l1d_flush_pages)
+{
+	free_pages((unsigned long)l1d_flush_pages, L1D_CACHE_ORDER);
+}
+EXPORT_SYMBOL_GPL(l1d_flush_cleanup_pages);
arch/x86/kvm/vmx/vmx.c +3 -22
index 8305097..225aa82 100644
@@ -203,14 +203,10 @@ static const struct {
 	[VMENTER_L1D_FLUSH_NOT_REQUIRED] = {"not required", false},
 };
 
-#define L1D_CACHE_ORDER 4
 static void *vmx_l1d_flush_pages;
 
 static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 {
-	struct page *page;
-	unsigned int i;
-
 	if (!boot_cpu_has_bug(X86_BUG_L1TF)) {
 		l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_NOT_REQUIRED;
 		return 0;
@@ -253,24 +249,9 @@ static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 
 	if (l1tf != VMENTER_L1D_FLUSH_NEVER && !vmx_l1d_flush_pages &&
 	    !boot_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		/*
-		 * This allocation for vmx_l1d_flush_pages is not tied to a VM
-		 * lifetime and so should not be charged to a memcg.
-		 */
-		page = alloc_pages(GFP_KERNEL, L1D_CACHE_ORDER);
-		if (!page)
+		vmx_l1d_flush_pages = l1d_flush_alloc_pages();
+		if (!vmx_l1d_flush_pages)
 			return -ENOMEM;
-		vmx_l1d_flush_pages = page_address(page);
-
-		/*
-		 * Initialize each page with a different pattern in
-		 * order to protect against KSM in the nested
-		 * virtualization case.
-		 */
-		for (i = 0; i < 1u << L1D_CACHE_ORDER; ++i) {
-			memset(vmx_l1d_flush_pages + i * PAGE_SIZE, i + 1,
-			       PAGE_SIZE);
-		}
 	}
 
 	l1tf_vmx_mitigation = l1tf;
@@ -8026,7 +8007,7 @@ static struct kvm_x86_init_ops vmx_init_ops __initdata = {
 static void vmx_cleanup_l1d_flush(void)
 {
 	if (vmx_l1d_flush_pages) {
-		free_pages((unsigned long)vmx_l1d_flush_pages, L1D_CACHE_ORDER);
+		l1d_flush_cleanup_pages(vmx_l1d_flush_pages);
 		vmx_l1d_flush_pages = NULL;
 	}
 	/* Restore state so sysfs ignores VMX */
merge bot Applied
#14 lore raw

[tip: x86/mm] x86/kvm: Refactor L1D flushing

In reply to Balbir Singh
The following commit has been merged into the x86/mm branch of tip:

The following commit has been merged into the x86/mm branch of tip:

Commit-ID:     3f768f0032dbc0657ed7e48f4735a3c4e49e25d7
Gitweb:        https://git.kernel.org/tip/3f768f0032dbc0657ed7e48f4735a3c4e49e25d7
Author:        Balbir Singh <[email protected]>
AuthorDate:    Sun, 10 May 2020 11:48:01 +10:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Wed, 13 May 2020 18:12:20 +02:00

x86/kvm: Refactor L1D flushing

Move more L1D flush related code out of KVM/VMX into builtin code to allow reuse for L1D flushing:

- Move the initialization to l1d_flush_init_once() and remove the deallocation of the L1D flush pages.

This avoids adding complex refcounting of users (VMX or tasks which opt into a L1D flush on context switch) for the price of a few pages potentially wasted when no users are left.

- Unify the flush invocations as arch_l1d_flush() which attempts hardware flushing and falls back to the software implementation with the option of prepopulating the TLB entries first.

[ tglx: Massage changelog and add a paranoid check of the flush pages
pointer ]

---
 arch/x86/include/asm/cacheflush.h | 12 ++---
 arch/x86/kernel/l1d_flush.c       | 68 ++++++++++++++++++++++--------
 arch/x86/kvm/vmx/vmx.c            | 20 +--------
 3 files changed, 60 insertions(+), 40 deletions(-)
arch/x86/include/asm/cacheflush.h +7 -5
index 21cc3b2..851d8f1 100644
@@ -7,11 +7,13 @@
 #include <asm/special_insns.h>
 
 #define L1D_CACHE_ORDER 4
+
+enum l1d_flush_options {
+	L1D_FLUSH_POPULATE_TLB = 0x1,
+};
+
 void clflush_cache_range(void *addr, unsigned int size);
-void l1d_flush_populate_tlb(void *l1d_flush_pages);
-void *l1d_flush_alloc_pages(void);
-void l1d_flush_cleanup_pages(void *l1d_flush_pages);
-void l1d_flush_sw(void *l1d_flush_pages);
-int l1d_flush_hw(void);
+int l1d_flush_init_once(void);
+void arch_l1d_flush(enum l1d_flush_options options);
 
 #endif /* _ASM_X86_CACHEFLUSH_H */
arch/x86/kernel/l1d_flush.c +50 -18

109 lines not on this page · raw message

arch/x86/kvm/vmx/vmx.c +3 -17
index 786d161..d489234 100644
@@ -203,8 +203,6 @@ static const struct {
 	[VMENTER_L1D_FLUSH_NOT_REQUIRED] = {"not required", false},
 };
 
-static void *vmx_l1d_flush_pages;
-
 static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 {
 	if (!boot_cpu_has_bug(X86_BUG_L1TF)) {
@@ -247,12 +245,9 @@ static int vmx_setup_l1d_flush(enum vmx_l1d_flush_state l1tf)
 		l1tf = VMENTER_L1D_FLUSH_ALWAYS;
 	}
 
-	if (l1tf != VMENTER_L1D_FLUSH_NEVER && !vmx_l1d_flush_pages &&
-	    !boot_cpu_has(X86_FEATURE_FLUSH_L1D)) {
-		vmx_l1d_flush_pages = l1d_flush_alloc_pages();
-		if (!vmx_l1d_flush_pages)
+	if (l1tf != VMENTER_L1D_FLUSH_NEVER)
+		if (l1d_flush_init_once())
 			return -ENOMEM;
-	}
 
 	l1tf_vmx_mitigation = l1tf;
 
@@ -6010,12 +6005,7 @@ static void vmx_l1d_flush(struct kvm_vcpu *vcpu)
 	}
 
 	vcpu->stat.l1d_flush++;
-
-	if (!l1d_flush_hw())
-		return;
-
-	l1d_flush_populate_tlb(vmx_l1d_flush_pages);
-	l1d_flush_sw(vmx_l1d_flush_pages);
+	arch_l1d_flush(L1D_FLUSH_POPULATE_TLB);
 }
 
 static void update_cr8_intercept(struct kvm_vcpu *vcpu, int tpr, int irr)
@@ -7983,10 +7973,6 @@ static struct kvm_x86_init_ops vmx_init_ops __initdata = {
 
 static void vmx_cleanup_l1d_flush(void)
 {
-	if (vmx_l1d_flush_pages) {
-		l1d_flush_cleanup_pages(vmx_l1d_flush_pages);
-		vmx_l1d_flush_pages = NULL;
-	}
 	/* Restore state so sysfs ignores VMX */
 	l1tf_vmx_mitigation = VMENTER_L1D_FLUSH_AUTO;
 }
merge bot Applied
#11 lore raw

[tip: x86/mm] x86/mm: Refactor cond_ibpb() to support other use cases

In reply to Balbir Singh
The following commit has been merged into the x86/mm branch of tip:

The following commit has been merged into the x86/mm branch of tip:

Commit-ID:     83ce56f712af79eac5f761e6b058359336803500
Gitweb:        https://git.kernel.org/tip/83ce56f712af79eac5f761e6b058359336803500
Author:        Balbir Singh <[email protected]>
AuthorDate:    Sun, 10 May 2020 11:48:00 +10:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Wed, 13 May 2020 18:12:20 +02:00

x86/mm: Refactor cond_ibpb() to support other use cases

cond_ibpb() has the necessary bits required to track the previous mm in switch_mm_irqs_off(). This can be reused for other use cases like L1D flushing on context switch.

[ tglx: Moved comment, added a separate define for state (re)initialization ]

---
 arch/x86/include/asm/tlbflush.h |  2 +-
 arch/x86/mm/tlb.c               | 53 +++++++++++++++++---------------
 2 files changed, 30 insertions(+), 25 deletions(-)
arch/x86/include/asm/tlbflush.h +1 -1
index 8c87a2e..a927d40 100644
@@ -83,7 +83,7 @@ struct tlb_state {
 	/* Last user mm for optimizing IBPB */
 	union {
 		struct mm_struct	*last_user_mm;
-		unsigned long		last_user_mm_ibpb;
+		unsigned long		last_user_mm_spec;
 	};
 
 	u16 loaded_mm_asid;
arch/x86/mm/tlb.c +29 -24

116 lines not on this page · raw message

merge bot Applied
#12 lore raw

[tip: x86/pti] x86/mm: Refactor cond_ibpb() to support other use cases

In reply to Balbir Singh
The following commit has been merged into the x86/pti branch of tip:

The following commit has been merged into the x86/pti branch of tip:

Commit-ID:     81f449985c12b83b91849d94724b803ebf856301
Gitweb:        https://git.kernel.org/tip/81f449985c12b83b91849d94724b803ebf856301
Author:        Balbir Singh <[email protected]>
AuthorDate:    Wed, 29 Jul 2020 10:11:00 +10:00
Committer:     Thomas Gleixner <[email protected]>
CommitterDate: Wed, 16 Sep 2020 15:08:02 +02:00

x86/mm: Refactor cond_ibpb() to support other use cases

cond_ibpb() has the necessary bits required to track the previous mm in switch_mm_irqs_off(). This can be reused for other use cases like L1D flushing on context switch.

[ tglx: Moved comment, added a separate define for state (re)initialization ]

---
 arch/x86/include/asm/tlbflush.h |  2 +-
 arch/x86/mm/tlb.c               | 53 +++++++++++++++++---------------
 2 files changed, 30 insertions(+), 25 deletions(-)
arch/x86/include/asm/tlbflush.h +1 -1
index 8c87a2e..a927d40 100644
@@ -83,7 +83,7 @@ struct tlb_state {
 	/* Last user mm for optimizing IBPB */
 	union {
 		struct mm_struct	*last_user_mm;
-		unsigned long		last_user_mm_ibpb;
+		unsigned long		last_user_mm_spec;
 	};
 
 	u16 loaded_mm_asid;
arch/x86/mm/tlb.c +29 -24

116 lines not on this page · raw message