[Devel] [PATCH vz8 20/42] mm/userfaultfd: honor FAULT_FLAG_KILLABLE in fault path

[Andrey Ryabinin]

From: Peter Xu <peterx at redhat.com>

Userfaultfd fault path was by default killable even if the caller does not
have FAULT_FLAG_KILLABLE.  That makes sense before in that when with gup
we don't have FAULT_FLAG_KILLABLE properly set before.  Now after previous
patch we've got FAULT_FLAG_KILLABLE applied even for gup code so it should
also make sense to let userfaultfd to honor the FAULT_FLAG_KILLABLE.

Because we're unconditionally setting FAULT_FLAG_KILLABLE in gup code
right now, this patch should have no functional change.  It also cleaned
the code a little bit by introducing some helpers.

Signed-off-by: Peter Xu <peterx at redhat.com>
Signed-off-by: Andrew Morton <akpm at linux-foundation.org>
Tested-by: Brian Geffon <bgeffon at google.com>
Cc: Andrea Arcangeli <aarcange at redhat.com>
Cc: Bobby Powers <bobbypowers at gmail.com>
Cc: David Hildenbrand <david at redhat.com>
Cc: Denis Plotnikov <dplotnikov at virtuozzo.com>
Cc: "Dr . David Alan Gilbert" <dgilbert at redhat.com>
Cc: Hugh Dickins <hughd at google.com>
Cc: Jerome Glisse <jglisse at redhat.com>
Cc: Johannes Weiner <hannes at cmpxchg.org>
Cc: "Kirill A . Shutemov" <kirill at shutemov.name>
Cc: Martin Cracauer <cracauer at cons.org>
Cc: Marty McFadden <mcfadden8 at llnl.gov>
Cc: Matthew Wilcox <willy at infradead.org>
Cc: Maya Gokhale <gokhale2 at llnl.gov>
Cc: Mel Gorman <mgorman at suse.de>
Cc: Mike Kravetz <mike.kravetz at oracle.com>
Cc: Mike Rapoport <rppt at linux.vnet.ibm.com>
Cc: Pavel Emelyanov <xemul at openvz.org>
Link: http://lkml.kernel.org/r/20200220160300.9941-1-peterx@redhat.com
Signed-off-by: Linus Torvalds <torvalds at linux-foundation.org>

https://jira.sw.ru/browse/PSBM-102938
(cherry picked from commit 3e69ad081c18d138fc7fd0f1ceef3b055ab10549)
Signed-off-by: Andrey Ryabinin <aryabinin at virtuozzo.com>
---
 fs/userfaultfd.c | 36 ++++++++++++++++++++++++++++--------
 1 file changed, 28 insertions(+), 8 deletions(-)

diff --git a/fs/userfaultfd.c b/fs/userfaultfd.c
index ebabf5983830..0bfbe5182773 100644
--- a/fs/userfaultfd.c
+++ b/fs/userfaultfd.c
@@ -325,6 +325,30 @@ static inline bool userfaultfd_must_wait(struct userfaultfd_ctx *ctx,
 	return ret;
 }
 
+/* Should pair with userfaultfd_signal_pending() */
+static inline long userfaultfd_get_blocking_state(unsigned int flags)
+{
+	if (flags & FAULT_FLAG_INTERRUPTIBLE)
+		return TASK_INTERRUPTIBLE;
+
+	if (flags & FAULT_FLAG_KILLABLE)
+		return TASK_KILLABLE;
+
+	return TASK_UNINTERRUPTIBLE;
+}
+
+/* Should pair with userfaultfd_get_blocking_state() */
+static inline bool userfaultfd_signal_pending(unsigned int flags)
+{
+	if (flags & FAULT_FLAG_INTERRUPTIBLE)
+		return signal_pending(current);
+
+	if (flags & FAULT_FLAG_KILLABLE)
+		return fatal_signal_pending(current);
+
+	return false;
+}
+
 /*
  * The locking rules involved in returning VM_FAULT_RETRY depending on
  * FAULT_FLAG_ALLOW_RETRY, FAULT_FLAG_RETRY_NOWAIT and
@@ -346,7 +370,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 	struct userfaultfd_ctx *ctx;
 	struct userfaultfd_wait_queue uwq;
 	int ret;
-	bool must_wait, return_to_userland;
+	bool must_wait;
 	long blocking_state;
 
 	ret = VM_FAULT_SIGBUS;
@@ -455,9 +479,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 	uwq.ctx = ctx;
 	uwq.waken = false;
 
-	return_to_userland = vmf->flags & FAULT_FLAG_INTERRUPTIBLE;
-	blocking_state = return_to_userland ? TASK_INTERRUPTIBLE :
-			 TASK_KILLABLE;
+	blocking_state = userfaultfd_get_blocking_state(vmf->flags);
 
 	spin_lock(&ctx->fault_pending_wqh.lock);
 	/*
@@ -483,8 +505,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 	up_read(&mm->mmap_sem);
 
 	if (likely(must_wait && !READ_ONCE(ctx->released) &&
-		   (return_to_userland ? !signal_pending(current) :
-		    !fatal_signal_pending(current)))) {
+		   !userfaultfd_signal_pending(vmf->flags))) {
 		wake_up_poll(&ctx->fd_wqh, EPOLLIN);
 		schedule();
 		ret |= VM_FAULT_MAJOR;
@@ -506,8 +527,7 @@ int handle_userfault(struct vm_fault *vmf, unsigned long reason)
 			set_current_state(blocking_state);
 			if (READ_ONCE(uwq.waken) ||
 			    READ_ONCE(ctx->released) ||
-			    (return_to_userland ? signal_pending(current) :
-			     fatal_signal_pending(current)))
+			    userfaultfd_signal_pending(vmf->flags))
 				break;
 			schedule();
 		}
-- 
2.25.3