[PATCH RESEND v6 15/16] mm/gup: Allow to react to fatal signals

From: Peter Xu <peterx@redhat.com>
To: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Cc: Peter Xu <peterx@redhat.com>, Martin Cracauer <cracauer@cons.org>,
	Mike Rapoport <rppt@linux.vnet.ibm.com>,
	Hugh Dickins <hughd@google.com>,
	Jerome Glisse <jglisse@redhat.com>,
	"Kirill A . Shutemov" <kirill@shutemov.name>,
	Matthew Wilcox <willy@infradead.org>,
	Pavel Emelyanov <xemul@virtuozzo.com>,
	Brian Geffon <bgeffon@google.com>,
	Maya Gokhale <gokhale2@llnl.gov>,
	Denis Plotnikov <dplotnikov@virtuozzo.com>,
	Andrea Arcangeli <aarcange@redhat.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	"Dr . David Alan Gilbert" <dgilbert@redhat.com>,
	Linus Torvalds <torvalds@linux-foundation.org>,
	Mike Kravetz <mike.kravetz@oracle.com>,
	Marty McFadden <mcfadden8@llnl.gov>,
	David Hildenbrand <david@redhat.com>,
	Bobby Powers <bobbypowers@gmail.com>,
	Mel Gorman <mgorman@suse.de>
Subject: [PATCH RESEND v6 15/16] mm/gup: Allow to react to fatal signals
Date: Thu, 20 Feb 2020 11:02:56 -0500	[thread overview]
Message-ID: <20200220160256.9887-1-peterx@redhat.com> (raw)
In-Reply-To: <20200220155353.8676-1-peterx@redhat.com>

The existing gup code does not react to the fatal signals in many code
paths.  For example, in one retry path of gup we're still using
down_read() rather than down_read_killable().  Also, when doing page
faults we don't pass in FAULT_FLAG_KILLABLE as well, which means that
within the faulting process we'll wait in non-killable way as well.
These were spotted by Linus during the code review of some other
patches.

Let's allow the gup code to react to fatal signals to improve the
responsiveness of threads when during gup and being killed.

Signed-off-by: Peter Xu <peterx@redhat.com>
---
 mm/gup.c     | 12 +++++++++---
 mm/hugetlb.c |  3 ++-
 2 files changed, 11 insertions(+), 4 deletions(-)

diff --git a/mm/gup.c b/mm/gup.c
index ec2b76f44a01..3f0cb14334ac 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -648,7 +648,7 @@ static int faultin_page(struct task_struct *tsk, struct vm_area_struct *vma,
 	if (*flags & FOLL_REMOTE)
 		fault_flags |= FAULT_FLAG_REMOTE;
 	if (locked)
-		fault_flags |= FAULT_FLAG_ALLOW_RETRY;
+		fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
 	if (*flags & FOLL_NOWAIT)
 		fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_RETRY_NOWAIT;
 	if (*flags & FOLL_TRIED) {
@@ -991,7 +991,7 @@ int fixup_user_fault(struct task_struct *tsk, struct mm_struct *mm,
 	address = untagged_addr(address);
 
 	if (unlocked)
-		fault_flags |= FAULT_FLAG_ALLOW_RETRY;
+		fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE;
 
 retry:
 	vma = find_extend_vma(mm, address);
@@ -1113,7 +1113,13 @@ static __always_inline long __get_user_pages_locked(struct task_struct *tsk,
 			break;
 
 		*locked = 1;
-		down_read(&mm->mmap_sem);
+		ret = down_read_killable(&mm->mmap_sem);
+		if (ret) {
+			BUG_ON(ret > 0);
+			if (!pages_done)
+				pages_done = ret;
+			break;
+		}
 
 		ret = __get_user_pages(tsk, mm, start, 1, flags | FOLL_TRIED,
 				       pages, NULL, locked);
diff --git a/mm/hugetlb.c b/mm/hugetlb.c
index ac9a28d51674..c342b091a7a4 100644
--- a/mm/hugetlb.c
+++ b/mm/hugetlb.c
@@ -4338,7 +4338,8 @@ long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma,
 			if (flags & FOLL_WRITE)
 				fault_flags |= FAULT_FLAG_WRITE;
 			if (locked)
-				fault_flags |= FAULT_FLAG_ALLOW_RETRY;
+				fault_flags |= FAULT_FLAG_ALLOW_RETRY |
+					FAULT_FLAG_KILLABLE;
 			if (flags & FOLL_NOWAIT)
 				fault_flags |= FAULT_FLAG_ALLOW_RETRY |
 					FAULT_FLAG_RETRY_NOWAIT;
-- 
2.24.1


