From: Peter Xu <peterx@redhat.com>
To: linux-kernel@vger.kernel.org, linux-mm@kvack.org
Cc: Anish Moorthy <amoorthy@google.com>,
Axel Rasmussen <axelrasmussen@google.com>,
Alexander Viro <viro@zeniv.linux.org.uk>,
Mike Kravetz <mike.kravetz@oracle.com>,
Peter Zijlstra <peterz@infradead.org>,
Andrew Morton <akpm@linux-foundation.org>,
Mike Rapoport <rppt@kernel.org>,
Christian Brauner <brauner@kernel.org>,
peterx@redhat.com, linux-fsdevel@vger.kernel.org,
Andrea Arcangeli <aarcange@redhat.com>,
Ingo Molnar <mingo@redhat.com>,
James Houghton <jthoughton@google.com>,
Nadav Amit <nadav.amit@gmail.com>
Subject: [PATCH 5/7] selftests/mm: Replace uffd_read_mutex with a semaphore
Date: Tue, 5 Sep 2023 17:42:33 -0400 [thread overview]
Message-ID: <20230905214235.320571-6-peterx@redhat.com> (raw)
In-Reply-To: <20230905214235.320571-1-peterx@redhat.com>
Each uffd read threads unlocks the read mutex first, probably just to make
sure the thread is reaching a stage where pthread_cancel() can always work
before the main thread moves on.
However keeping the mutex locked always and unlock in the thread is a bit
hacky. Replacing it with a semaphore which should be much clearer, where
the main thread will wait() and the thread will just post(). Move it to
uffd-common.* to be reused later.
Signed-off-by: Peter Xu <peterx@redhat.com>
---
tools/testing/selftests/mm/uffd-common.c | 1 +
tools/testing/selftests/mm/uffd-common.h | 2 ++
tools/testing/selftests/mm/uffd-stress.c | 8 +++-----
3 files changed, 6 insertions(+), 5 deletions(-)
diff --git a/tools/testing/selftests/mm/uffd-common.c b/tools/testing/selftests/mm/uffd-common.c
index 02b89860e193..aded06cab285 100644
--- a/tools/testing/selftests/mm/uffd-common.c
+++ b/tools/testing/selftests/mm/uffd-common.c
@@ -17,6 +17,7 @@ bool map_shared;
bool test_uffdio_wp = true;
unsigned long long *count_verify;
uffd_test_ops_t *uffd_test_ops;
+sem_t uffd_read_sem;
static int uffd_mem_fd_create(off_t mem_size, bool hugetlb)
{
diff --git a/tools/testing/selftests/mm/uffd-common.h b/tools/testing/selftests/mm/uffd-common.h
index 7c4fa964c3b0..521523baded1 100644
--- a/tools/testing/selftests/mm/uffd-common.h
+++ b/tools/testing/selftests/mm/uffd-common.h
@@ -32,6 +32,7 @@
#include <inttypes.h>
#include <stdint.h>
#include <sys/random.h>
+#include <semaphore.h>
#include "../kselftest.h"
#include "vm_util.h"
@@ -97,6 +98,7 @@ extern bool map_shared;
extern bool test_uffdio_wp;
extern unsigned long long *count_verify;
extern volatile bool test_uffdio_copy_eexist;
+extern sem_t uffd_read_sem;
extern uffd_test_ops_t anon_uffd_test_ops;
extern uffd_test_ops_t shmem_uffd_test_ops;
diff --git a/tools/testing/selftests/mm/uffd-stress.c b/tools/testing/selftests/mm/uffd-stress.c
index 469e0476af26..7219f55ae794 100644
--- a/tools/testing/selftests/mm/uffd-stress.c
+++ b/tools/testing/selftests/mm/uffd-stress.c
@@ -125,14 +125,12 @@ static int copy_page_retry(int ufd, unsigned long offset)
return __copy_page(ufd, offset, true, test_uffdio_wp);
}
-pthread_mutex_t uffd_read_mutex = PTHREAD_MUTEX_INITIALIZER;
-
static void *uffd_read_thread(void *arg)
{
struct uffd_args *args = (struct uffd_args *)arg;
struct uffd_msg msg;
- pthread_mutex_unlock(&uffd_read_mutex);
+ sem_post(&uffd_read_sem);
/* from here cancellation is ok */
for (;;) {
@@ -196,7 +194,7 @@ static int stress(struct uffd_args *args)
uffd_read_thread,
(void *)&args[cpu]))
return 1;
- pthread_mutex_lock(&uffd_read_mutex);
+ sem_wait(&uffd_read_sem);
}
if (pthread_create(&background_threads[cpu], &attr,
background_thread, (void *)cpu))
@@ -258,7 +256,7 @@ static int userfaultfd_stress(void)
zeropage = area;
bzero(zeropage, page_size);
- pthread_mutex_lock(&uffd_read_mutex);
+ sem_init(&uffd_read_sem, 0, 0);
pthread_attr_init(&attr);
pthread_attr_setstacksize(&attr, 16*1024*1024);
--
2.41.0
next prev parent reply other threads:[~2023-09-05 21:44 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-09-05 21:42 [PATCH 0/7] mm/userfaultfd/poll: Scale userfaultfd wakeups Peter Xu
2023-09-05 21:42 ` [PATCH 1/7] mm/userfaultfd: Make uffd read() wait event exclusive Peter Xu
2023-09-05 21:42 ` [PATCH 2/7] poll: Add a poll_flags for poll_queue_proc() Peter Xu
2023-09-05 23:21 ` kernel test robot
2023-09-06 17:31 ` kernel test robot
2023-09-06 20:53 ` kernel test robot
2023-09-11 20:00 ` Peter Xu
2023-09-05 21:42 ` [PATCH 3/7] poll: POLL_ENQUEUE_EXCLUSIVE Peter Xu
2023-09-05 21:42 ` [PATCH 4/7] fs/userfaultfd: Use exclusive waitqueue for poll() Peter Xu
2023-09-05 21:42 ` Peter Xu [this message]
2023-09-05 21:42 ` [PATCH 6/7] selftests/mm: Create uffd_fault_thread_create|join() Peter Xu
2023-09-05 21:42 ` [PATCH 7/7] selftests/mm: uffd perf test Peter Xu
2023-09-07 19:18 ` [PATCH 0/7] mm/userfaultfd/poll: Scale userfaultfd wakeups Axel Rasmussen
2023-09-08 22:01 ` Peter Xu
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230905214235.320571-6-peterx@redhat.com \
--to=peterx@redhat.com \
--cc=aarcange@redhat.com \
--cc=akpm@linux-foundation.org \
--cc=amoorthy@google.com \
--cc=axelrasmussen@google.com \
--cc=brauner@kernel.org \
--cc=jthoughton@google.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mike.kravetz@oracle.com \
--cc=mingo@redhat.com \
--cc=nadav.amit@gmail.com \
--cc=peterz@infradead.org \
--cc=rppt@kernel.org \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).