KVM: s390: Add gmap_helper_set_unused()

Add gmap_helper_set_unused() to mark userspace ptes as unused.

Core mm code will use that information to discard unused pages instead
of attempting to swap them.

Reviewed-by: Nico Boehr <nrb@linux.ibm.com>
Tested-by: Nico Boehr <nrb@linux.ibm.com>
Acked-by: Christoph Schlameuss <schlameuss@linux.ibm.com>
Acked-by: Heiko Carstens <hca@linux.ibm.com>
Signed-off-by: Claudio Imbrenda <imbrenda@linux.ibm.com>

arch/s390/include/asm/gmap_helpers.h

@@ -11,5 +11,6 @@
 void gmap_helper_zap_one_page(struct mm_struct *mm, unsigned long vmaddr);
 void gmap_helper_discard(struct mm_struct *mm, unsigned long vmaddr, unsigned long end);
 int gmap_helper_disable_cow_sharing(void);
+void gmap_helper_try_set_pte_unused(struct mm_struct *mm, unsigned long vmaddr);
 
 #endif /* _ASM_S390_GMAP_HELPERS_H */

arch/s390/mm/gmap_helpers.c

@@ -129,6 +129,85 @@ void gmap_helper_discard(struct mm_struct *mm, unsigned long vmaddr, unsigned lo
 }
 EXPORT_SYMBOL_GPL(gmap_helper_discard);
 
+/**
+ * gmap_helper_try_set_pte_unused() - mark a pte entry as unused
+ * @mm: the mm
+ * @vmaddr: the userspace address whose pte is to be marked
+ *
+ * Mark the pte corresponding the given address as unused. This will cause
+ * core mm code to just drop this page instead of swapping it.
+ *
+ * This function needs to be called with interrupts disabled (for example
+ * while holding a spinlock), or while holding the mmap lock. Normally this
+ * function is called as a result of an unmap operation, and thus KVM common
+ * code will already hold kvm->mmu_lock in write mode.
+ *
+ * Context: Needs to be called while holding the mmap lock or with interrupts
+ *          disabled.
+ */
+void gmap_helper_try_set_pte_unused(struct mm_struct *mm, unsigned long vmaddr)
+{
+	pmd_t *pmdp, pmd, pmdval;
+	pud_t *pudp, pud;
+	p4d_t *p4dp, p4d;
+	pgd_t *pgdp, pgd;
+	spinlock_t *ptl;	/* Lock for the host (userspace) page table */
+	pte_t *ptep;
+
+	pgdp = pgd_offset(mm, vmaddr);
+	pgd = pgdp_get(pgdp);
+	if (pgd_none(pgd) || !pgd_present(pgd))
+		return;
+
+	p4dp = p4d_offset(pgdp, vmaddr);
+	p4d = p4dp_get(p4dp);
+	if (p4d_none(p4d) || !p4d_present(p4d))
+		return;
+
+	pudp = pud_offset(p4dp, vmaddr);
+	pud = pudp_get(pudp);
+	if (pud_none(pud) || pud_leaf(pud) || !pud_present(pud))
+		return;
+
+	pmdp = pmd_offset(pudp, vmaddr);
+	pmd = pmdp_get_lockless(pmdp);
+	if (pmd_none(pmd) || pmd_leaf(pmd) || !pmd_present(pmd))
+		return;
+
+	ptep = pte_offset_map_rw_nolock(mm, pmdp, vmaddr, &pmdval, &ptl);
+	if (!ptep)
+		return;
+
+	/*
+	 * Several paths exists that takes the ptl lock and then call the
+	 * mmu_notifier, which takes the mmu_lock. The unmap path, instead,
+	 * takes the mmu_lock in write mode first, and then potentially
+	 * calls this function, which takes the ptl lock. This can lead to a
+	 * deadlock.
+	 * The unused page mechanism is only an optimization, if the
+	 * _PAGE_UNUSED bit is not set, the unused page is swapped as normal
+	 * instead of being discarded.
+	 * If the lock is contended the bit is not set and the deadlock is
+	 * avoided.
+	 */
+	if (spin_trylock(ptl)) {
+		/*
+		 * Make sure the pte we are touching is still the correct
+		 * one. In theory this check should not be needed, but
+		 * better safe than sorry.
+		 * Disabling interrupts or holding the mmap lock is enough to
+		 * guarantee that no concurrent updates to the page tables
+		 * are possible.
+		 */
+		if (likely(pmd_same(pmdval, pmdp_get_lockless(pmdp))))
+			__atomic64_or(_PAGE_UNUSED, (long *)ptep);
+		spin_unlock(ptl);
+	}
+
+	pte_unmap(ptep);
+}
+EXPORT_SYMBOL_GPL(gmap_helper_try_set_pte_unused);
+
 static int find_zeropage_pte_entry(pte_t *pte, unsigned long addr,
 				   unsigned long end, struct mm_walk *walk)
 {