KVM: SEV: Do cache maintenance on the source VM during intra-host migrationgrafted

Manually perform cache maintenance on the source VM during intra-host
migration to ensure no stale data is left in CPU caches after the VM is
destroyed.  Because the source VM is "converted" to a non-SEV VM, KVM's
memory reclaim flows won't trigger cache maintenance, e.g. when all guest
memory is reclaimed in response to detaching from the mmu_notifier.

Note, relying on the destination VM to do cache maintenance isn't an option
as KVM doesn't require identical guest memory configurations, i.e. the
source VM may have access to memory that the destination VM does not.
Enforcing equivalent memory configurations is infeasible, as it would
require a *deep* comparison of memslots, e.g. to verify that not only are
the memslot identical, but what the memslots point at is also identical.

Fixes: b56639318bb2 ("KVM: SEV: Add support for SEV intra host migration")
Cc: stable@vger.kernel.org
Reported-by: Stefan Teodorescu <fane@google.com>
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-ID: <20260923163721.1584779-3-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

1 files changed, 193 insertions, 0 deletions

diff --git a/arch/arc/mm/fault.c b/arch/arc/mm/fault.c
new file mode 100644
index 000000000..95119a5e7
--- /dev/null
+++ b/arch/arc/mm/fault.c
@@ -0,0 +1,193 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/* Page Fault Handling for ARC (TLB Miss / ProtV)
+ *
+ * Copyright (C) 2004, 2007-2010, 2011-2012 Synopsys, Inc. (www.synopsys.com)
+ */
+
+#include <linux/signal.h>
+#include <linux/interrupt.h>
+#include <linux/sched/signal.h>
+#include <linux/errno.h>
+#include <linux/ptrace.h>
+#include <linux/uaccess.h>
+#include <linux/kdebug.h>
+#include <linux/perf_event.h>
+#include <linux/mm_types.h>
+#include <asm/entry.h>
+#include <asm/mmu.h>
+
+/*
+ * kernel virtual address is required to implement vmalloc/pkmap/fixmap
+ * Refer to asm/processor.h for System Memory Map
+ *
+ * It simply copies the PMD entry (pointer to 2nd level page table or hugepage)
+ * from swapper pgdir to task pgdir. The 2nd level table/page is thus shared
+ */
+noinline static int handle_kernel_vaddr_fault(unsigned long address)
+{
+	/*
+	 * Synchronize this task's top level page-table
+	 * with the 'reference' page table.
+	 */
+	pgd_t *pgd, *pgd_k;
+	p4d_t *p4d, *p4d_k;
+	pud_t *pud, *pud_k;
+	pmd_t *pmd, *pmd_k;
+
+	pgd = pgd_offset(current->active_mm, address);
+	pgd_k = pgd_offset_k(address);
+
+	if (pgd_none (*pgd_k))
+		goto bad_area;
+	if (!pgd_present(*pgd))
+		set_pgd(pgd, *pgd_k);
+
+	p4d = p4d_offset(pgd, address);
+	p4d_k = p4d_offset(pgd_k, address);
+	if (p4d_none(*p4d_k))
+		goto bad_area;
+	if (!p4d_present(*p4d))
+		set_p4d(p4d, *p4d_k);
+
+	pud = pud_offset(p4d, address);
+	pud_k = pud_offset(p4d_k, address);
+	if (pud_none(*pud_k))
+		goto bad_area;
+	if (!pud_present(*pud))
+		set_pud(pud, *pud_k);
+
+	pmd = pmd_offset(pud, address);
+	pmd_k = pmd_offset(pud_k, address);
+	if (pmd_none(*pmd_k))
+		goto bad_area;
+	if (!pmd_present(*pmd))
+		set_pmd(pmd, *pmd_k);
+
+	/* XXX: create the TLB entry here */
+	return 0;
+
+bad_area:
+	return 1;
+}
+
+void do_page_fault(unsigned long address, struct pt_regs *regs)
+{
+	struct vm_area_struct *vma = NULL;
+	struct task_struct *tsk = current;
+	struct mm_struct *mm = tsk->mm;
+	int sig, si_code = SEGV_MAPERR;
+	unsigned int write = 0, exec = 0, mask;
+	vm_fault_t fault = VM_FAULT_SIGSEGV;	/* handle_mm_fault() output */
+	unsigned int flags;			/* handle_mm_fault() input */
+
+	/*
+	 * NOTE! We MUST NOT take any locks for this case. We may
+	 * be in an interrupt or a critical region, and should
+	 * only copy the information from the master page table,
+	 * nothing more.
+	 */
+	if (address >= VMALLOC_START && !user_mode(regs)) {
+		if (unlikely(handle_kernel_vaddr_fault(address)))
+			goto no_context;
+		else
+			return;
+	}
+
+	/*
+	 * If we're in an interrupt or have no user
+	 * context, we must not take the fault..
+	 */
+	if (faulthandler_disabled() || !mm)
+		goto no_context;
+
+	if (regs->ecr.cause & ECR_C_PROTV_STORE)	/* ST/EX */
+		write = 1;
+	else if ((regs->ecr.vec == ECR_V_PROTV) &&
+	         (regs->ecr.cause == ECR_C_PROTV_INST_FETCH))
+		exec = 1;
+
+	flags = FAULT_FLAG_DEFAULT;
+	if (user_mode(regs))
+		flags |= FAULT_FLAG_USER;
+	if (write)
+		flags |= FAULT_FLAG_WRITE;
+
+	perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address);
+retry:
+	vma = lock_mm_and_find_vma(mm, address, regs);
+	if (!vma)
+		goto bad_area_nosemaphore;
+
+	/*
+	 * vm_area is good, now check permissions for this memory access
+	 */
+	mask = VM_READ;
+	if (write)
+		mask = VM_WRITE;
+	if (exec)
+		mask = VM_EXEC;
+
+	if (!(vma->vm_flags & mask)) {
+		si_code = SEGV_ACCERR;
+		goto bad_area;
+	}
+
+	fault = handle_mm_fault(vma, address, flags, regs);
+
+	/* Quick path to respond to signals */
+	if (fault_signal_pending(fault, regs)) {
+		if (!user_mode(regs))
+			goto no_context;
+		return;
+	}
+
+	/* The fault is fully completed (including releasing mmap lock) */
+	if (fault & VM_FAULT_COMPLETED)
+		return;
+
+	/*
+	 * Fault retry nuances, mmap_lock already relinquished by core mm
+	 */
+	if (unlikely(fault & VM_FAULT_RETRY)) {
+		flags |= FAULT_FLAG_TRIED;
+		goto retry;
+	}
+
+bad_area:
+	mmap_read_unlock(mm);
+
+bad_area_nosemaphore:
+	/*
+	 * Major/minor page fault accounting
+	 * (in case of retry we only land here once)
+	 */
+	if (likely(!(fault & VM_FAULT_ERROR)))
+		/* Normal return path: fault Handled Gracefully */
+		return;
+
+	if (!user_mode(regs))
+		goto no_context;
+
+	if (fault & VM_FAULT_OOM) {
+		pagefault_out_of_memory();
+		return;
+	}
+
+	if (fault & VM_FAULT_SIGBUS) {
+		sig = SIGBUS;
+		si_code = BUS_ADRERR;
+	}
+	else {
+		sig = SIGSEGV;
+	}
+
+	tsk->thread.fault_address = address;
+	force_sig_fault(sig, si_code, (void __user *)address);
+	return;
+
+no_context:
+	if (fixup_exception(regs))
+		return;
+
+	die("Oops", regs, address);
+}