Merge branch 'x86/urgent' into x86/mm, to pick up dependent fix

Signed-off-by: Ingo Molnar <mingo@kernel.org>

4 files changed, 119 insertions, 48 deletions

diff --git a/arch/x86/mm/dump_pagetables.c b/arch/x86/mm/dump_pagetables.c
index fc37bbd23eb8..abcb8d00b014 100644
--- a/arch/x86/mm/dump_pagetables.c
+++ b/arch/x86/mm/dump_pagetables.c
@@ -55,10 +55,10 @@ struct addr_marker {
 enum address_markers_idx {
 	USER_SPACE_NR = 0,
 	KERNEL_SPACE_NR,
-	LOW_KERNEL_NR,
-#if defined(CONFIG_MODIFY_LDT_SYSCALL) && defined(CONFIG_X86_5LEVEL)
+#ifdef CONFIG_MODIFY_LDT_SYSCALL
 	LDT_NR,
 #endif
+	LOW_KERNEL_NR,
 	VMALLOC_START_NR,
 	VMEMMAP_START_NR,
 #ifdef CONFIG_KASAN
@@ -66,9 +66,6 @@ enum address_markers_idx {
 	KASAN_SHADOW_END_NR,
 #endif
 	CPU_ENTRY_AREA_NR,
-#if defined(CONFIG_MODIFY_LDT_SYSCALL) && !defined(CONFIG_X86_5LEVEL)
-	LDT_NR,
-#endif
 #ifdef CONFIG_X86_ESPFIX64
 	ESPFIX_START_NR,
 #endif
@@ -512,11 +509,11 @@ static inline bool is_hypervisor_range(int idx)
 {
 #ifdef CONFIG_X86_64
 	/*
-	 * ffff800000000000 - ffff87ffffffffff is reserved for
-	 * the hypervisor.
+	 * A hole in the beginning of kernel address space reserved
+	 * for a hypervisor.
 	 */
-	return	(idx >= pgd_index(__PAGE_OFFSET) - 16) &&
-		(idx <  pgd_index(__PAGE_OFFSET));
+	return	(idx >= pgd_index(GUARD_HOLE_BASE_ADDR)) &&
+		(idx <  pgd_index(GUARD_HOLE_END_ADDR));
 #else
 	return false;
 #endif
diff --git a/arch/x86/mm/pageattr.c b/arch/x86/mm/pageattr.c
index db7a10082238..a1bcde35db4c 100644
--- a/arch/x86/mm/pageattr.c
+++ b/arch/x86/mm/pageattr.c
@@ -285,20 +285,16 @@ static void cpa_flush_all(unsigned long cache)
 	on_each_cpu(__cpa_flush_all, (void *) cache, 1);
 }
 
-static bool __cpa_flush_range(unsigned long start, int numpages, int cache)
+static bool __inv_flush_all(int cache)
 {
 	BUG_ON(irqs_disabled() && !early_boot_irqs_disabled);
 
-	WARN_ON(PAGE_ALIGN(start) != start);
-
 	if (cache && !static_cpu_has(X86_FEATURE_CLFLUSH)) {
 		cpa_flush_all(cache);
 		return true;
 	}
 
-	flush_tlb_kernel_range(start, start + PAGE_SIZE * numpages);
-
-	return !cache;
+	return false;
 }
 
 static void cpa_flush_range(unsigned long start, int numpages, int cache)
@@ -306,7 +302,14 @@ static void cpa_flush_range(unsigned long start, int numpages, int cache)
 	unsigned int i, level;
 	unsigned long addr;
 
-	if (__cpa_flush_range(start, numpages, cache))
+	WARN_ON(PAGE_ALIGN(start) != start);
+
+	if (__inv_flush_all(cache))
+		return;
+
+	flush_tlb_kernel_range(start, start + PAGE_SIZE * numpages);
+
+	if (!cache)
 		return;
 
 	/*
@@ -332,7 +335,12 @@ static void cpa_flush_array(unsigned long baddr, unsigned long *start,
 {
 	unsigned int i, level;
 
-	if (__cpa_flush_range(baddr, numpages, cache))
+	if (__inv_flush_all(cache))
+		return;
+
+	flush_tlb_all();
+
+	if (!cache)
 		return;
 
 	/*
diff --git a/arch/x86/mm/pat.c b/arch/x86/mm/pat.c
index 08013524fba1..4fe956a63b25 100644
--- a/arch/x86/mm/pat.c
+++ b/arch/x86/mm/pat.c
@@ -519,8 +519,13 @@ static u64 sanitize_phys(u64 address)
 	 * for a "decoy" virtual address (bit 63 clear) passed to
 	 * set_memory_X(). __pa() on a "decoy" address results in a
 	 * physical address with bit 63 set.
+	 *
+	 * Decoy addresses are not present for 32-bit builds, see
+	 * set_mce_nospec().
 	 */
-	return address & __PHYSICAL_MASK;
+	if (IS_ENABLED(CONFIG_X86_64))
+		return address & __PHYSICAL_MASK;
+	return address;
 }
 
 /*
@@ -546,7 +551,11 @@ int reserve_memtype(u64 start, u64 end, enum page_cache_mode req_type,
 
 	start = sanitize_phys(start);
 	end = sanitize_phys(end);
-	BUG_ON(start >= end); /* end is exclusive */
+	if (start >= end) {
+		WARN(1, "%s failed: [mem %#010Lx-%#010Lx], req %s\n", __func__,
+				start, end - 1, cattr_name(req_type));
+		return -EINVAL;
+	}
 
 	if (!pat_enabled()) {
 		/* This is identical to page table setting without PAT */
diff --git a/arch/x86/mm/tlb.c b/arch/x86/mm/tlb.c
index bddd6b3cee1d..03b6b4c2238d 100644
--- a/arch/x86/mm/tlb.c
+++ b/arch/x86/mm/tlb.c
@@ -7,7 +7,6 @@
 #include <linux/export.h>
 #include <linux/cpu.h>
 #include <linux/debugfs.h>
-#include <linux/ptrace.h>
 
 #include <asm/tlbflush.h>
 #include <asm/mmu_context.h>
@@ -31,6 +30,12 @@
  */
 
 /*
+ * Use bit 0 to mangle the TIF_SPEC_IB state into the mm pointer which is
+ * stored in cpu_tlb_state.last_user_mm_ibpb.
+ */
+#define LAST_USER_MM_IBPB	0x1UL
+
+/*
  * We get here when we do something requiring a TLB invalidation
  * but could not go invalidate all of the contexts.  We do the
  * necessary invalidation by clearing out the 'ctx_id' which
@@ -181,17 +186,87 @@ static void sync_current_stack_to_mm(struct mm_struct *mm)
 	}
 }
 
-static bool ibpb_needed(struct task_struct *tsk, u64 last_ctx_id)
+static inline unsigned long mm_mangle_tif_spec_ib(struct task_struct *next)
+{
+	unsigned long next_tif = task_thread_info(next)->flags;
+	unsigned long ibpb = (next_tif >> TIF_SPEC_IB) & LAST_USER_MM_IBPB;
+
+	return (unsigned long)next->mm | ibpb;
+}
+
+static void cond_ibpb(struct task_struct *next)
 {
+	if (!next || !next->mm)
+		return;
+
 	/*
-	 * Check if the current (previous) task has access to the memory
-	 * of the @tsk (next) task. If access is denied, make sure to
-	 * issue a IBPB to stop user->user Spectre-v2 attacks.
-	 *
-	 * Note: __ptrace_may_access() returns 0 or -ERRNO.
+	 * Both, the conditional and the always IBPB mode use the mm
+	 * pointer to avoid the IBPB when switching between tasks of the
+	 * same process. Using the mm pointer instead of mm->context.ctx_id
+	 * opens a hypothetical hole vs. mm_struct reuse, which is more or
+	 * less impossible to control by an attacker. Aside of that it
+	 * would only affect the first schedule so the theoretically
+	 * exposed data is not really interesting.
 	 */
-	return (tsk && tsk->mm && tsk->mm->context.ctx_id != last_ctx_id &&
-		ptrace_may_access_sched(tsk, PTRACE_MODE_SPEC_IBPB));
+	if (static_branch_likely(&switch_mm_cond_ibpb)) {
+		unsigned long prev_mm, next_mm;
+
+		/*
+		 * This is a bit more complex than the always mode because
+		 * it has to handle two cases:
+		 *
+		 * 1) Switch from a user space task (potential attacker)
+		 *    which has TIF_SPEC_IB set to a user space task
+		 *    (potential victim) which has TIF_SPEC_IB not set.
+		 *
+		 * 2) Switch from a user space task (potential attacker)
+		 *    which has TIF_SPEC_IB not set to a user space task
+		 *    (potential victim) which has TIF_SPEC_IB set.
+		 *
+		 * This could be done by unconditionally issuing IBPB when
+		 * a task which has TIF_SPEC_IB set is either scheduled in
+		 * or out. Though that results in two flushes when:
+		 *
+		 * - the same user space task is scheduled out and later
+		 *   scheduled in again and only a kernel thread ran in
+		 *   between.
+		 *
+		 * - a user space task belonging to the same process is
+		 *   scheduled in after a kernel thread ran in between
+		 *
+		 * - a user space task belonging to the same process is
+		 *   scheduled in immediately.
+		 *
+		 * Optimize this with reasonably small overhead for the
+		 * above cases. Mangle the TIF_SPEC_IB bit into the mm
+		 * pointer of the incoming task which is stored in
+		 * cpu_tlbstate.last_user_mm_ibpb for comparison.
+		 */
+		next_mm = mm_mangle_tif_spec_ib(next);
+		prev_mm = this_cpu_read(cpu_tlbstate.last_user_mm_ibpb);
+
+		/*
+		 * Issue IBPB only if the mm's are different and one or
+		 * both have the IBPB bit set.
+		 */
+		if (next_mm != prev_mm &&
+		    (next_mm | prev_mm) & LAST_USER_MM_IBPB)
+			indirect_branch_prediction_barrier();
+
+		this_cpu_write(cpu_tlbstate.last_user_mm_ibpb, next_mm);
+	}
+
+	if (static_branch_unlikely(&switch_mm_always_ibpb)) {
+		/*
+		 * Only flush when switching to a user space task with a
+		 * different context than the user space task which ran
+		 * last on this CPU.
+		 */
+		if (this_cpu_read(cpu_tlbstate.last_user_mm) != next->mm) {
+			indirect_branch_prediction_barrier();
+			this_cpu_write(cpu_tlbstate.last_user_mm, next->mm);
+		}
+	}
 }
 
 void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next,
@@ -292,22 +367,12 @@ void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next,
 		new_asid = prev_asid;
 		need_flush = true;
 	} else {
-		u64 last_ctx_id = this_cpu_read(cpu_tlbstate.last_ctx_id);
-
 		/*
 		 * 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.
-		 *
-		 * As an optimization, flush indirect branches only when
-		 * switching into a processes that can't be ptrace by the
-		 * current one (as in such case, attacker has much more
-		 * convenient way how to tamper with the next process than
-		 * branch buffer poisoning).
 		 */
-		if (static_cpu_has(X86_FEATURE_USE_IBPB) &&
-				ibpb_needed(tsk, last_ctx_id))
-			indirect_branch_prediction_barrier();
+		cond_ibpb(tsk);
 
 		if (IS_ENABLED(CONFIG_VMAP_STACK)) {
 			/*
@@ -365,14 +430,6 @@ void switch_mm_irqs_off(struct mm_struct *prev, struct mm_struct *next,
 		trace_tlb_flush_rcuidle(TLB_FLUSH_ON_TASK_SWITCH, 0);
 	}
 
-	/*
-	 * Record last user mm's context id, so we can avoid
-	 * flushing branch buffer with IBPB if we switch back
-	 * to the same user.
-	 */
-	if (next != &init_mm)
-		this_cpu_write(cpu_tlbstate.last_ctx_id, next->context.ctx_id);
-
 	/* Make sure we write CR3 before loaded_mm. */
 	barrier();
 
@@ -441,7 +498,7 @@ void initialize_tlbstate_and_flush(void)
 	write_cr3(build_cr3(mm->pgd, 0));
 
 	/* Reinitialize tlbstate. */
-	this_cpu_write(cpu_tlbstate.last_ctx_id, mm->context.ctx_id);
+	this_cpu_write(cpu_tlbstate.last_user_mm_ibpb, LAST_USER_MM_IBPB);
 	this_cpu_write(cpu_tlbstate.loaded_mm_asid, 0);
 	this_cpu_write(cpu_tlbstate.next_asid, 1);
 	this_cpu_write(cpu_tlbstate.ctxs[0].ctx_id, mm->context.ctx_id);