From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964916AbbLVVUG (ORCPT ); Tue, 22 Dec 2015 16:20:06 -0500 Received: from mail-wm0-f51.google.com ([74.125.82.51]:37539 "EHLO mail-wm0-f51.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753826AbbLVVUD (ORCPT ); Tue, 22 Dec 2015 16:20:03 -0500 From: Richard Cochran To: Cc: John Stultz , Julia Lawall , Markus Elfring , Thomas Gleixner , kernel-janitors@vger.kernel.org Subject: [PATCH] posix-clock: fix return code on the poll method's error path Date: Tue, 22 Dec 2015 22:19:58 +0100 Message-Id: <1450819198-17420-1-git-send-email-richardcochran@gmail.com> X-Mailer: git-send-email 2.1.4 In-Reply-To: <56766488.9050402@users.sourceforge.net> References: <56766488.9050402@users.sourceforge.net> Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org The posix_clock_poll function is supposed to return a bit mask of POLLxxx values. However, in case the hardware has disappeared (due to hot plugging for example) this code returns -ENODEV in a futile attempt to throw an error at the file descriptor level. The kernel's file_operations interface does not accept such error codes from the poll method. Instead, this function aught to return POLLERR. The value -ENODEV does, in fact, contain the POLLERR bit (and almost all the other POLLxxx bits as well), but only by chance. This patch fixes code to return a proper bit mask. Credit goes to Markus Elfring for pointing out the suspicious signed/unsigned mismatch. Signed-off-by: Richard Cochran --- kernel/time/posix-clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/time/posix-clock.c b/kernel/time/posix-clock.c index ce033c7..9cff0ab 100644 --- a/kernel/time/posix-clock.c +++ b/kernel/time/posix-clock.c @@ -69,10 +69,10 @@ static ssize_t posix_clock_read(struct file *fp, char __user *buf, static unsigned int posix_clock_poll(struct file *fp, poll_table *wait) { struct posix_clock *clk = get_posix_clock(fp); - int result = 0; + unsigned int result = 0; if (!clk) - return -ENODEV; + return POLLERR; if (clk->ops.poll) result = clk->ops.poll(clk, fp, wait); -- 2.1.4 From mboxrd@z Thu Jan 1 00:00:00 1970 From: Richard Cochran Date: Tue, 22 Dec 2015 21:19:58 +0000 Subject: [PATCH] posix-clock: fix return code on the poll method's error path Message-Id: <1450819198-17420-1-git-send-email-richardcochran@gmail.com> List-Id: References: <56766488.9050402@users.sourceforge.net> In-Reply-To: <56766488.9050402@users.sourceforge.net> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-kernel@vger.kernel.org Cc: John Stultz , Julia Lawall , Markus Elfring , Thomas Gleixner , kernel-janitors@vger.kernel.org The posix_clock_poll function is supposed to return a bit mask of POLLxxx values. However, in case the hardware has disappeared (due to hot plugging for example) this code returns -ENODEV in a futile attempt to throw an error at the file descriptor level. The kernel's file_operations interface does not accept such error codes from the poll method. Instead, this function aught to return POLLERR. The value -ENODEV does, in fact, contain the POLLERR bit (and almost all the other POLLxxx bits as well), but only by chance. This patch fixes code to return a proper bit mask. Credit goes to Markus Elfring for pointing out the suspicious signed/unsigned mismatch. Signed-off-by: Richard Cochran --- kernel/time/posix-clock.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/kernel/time/posix-clock.c b/kernel/time/posix-clock.c index ce033c7..9cff0ab 100644 --- a/kernel/time/posix-clock.c +++ b/kernel/time/posix-clock.c @@ -69,10 +69,10 @@ static ssize_t posix_clock_read(struct file *fp, char __user *buf, static unsigned int posix_clock_poll(struct file *fp, poll_table *wait) { struct posix_clock *clk = get_posix_clock(fp); - int result = 0; + unsigned int result = 0; if (!clk) - return -ENODEV; + return POLLERR; if (clk->ops.poll) result = clk->ops.poll(clk, fp, wait); -- 2.1.4