Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751202AbbD3Pgg (ORCPT ); Thu, 30 Apr 2015 11:36:36 -0400 Received: from exprod5og113.obsmtp.com ([64.18.0.26]:33468 "EHLO mail-lb0-f175.google.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S1750746AbbD3Pgd (ORCPT ); Thu, 30 Apr 2015 11:36:33 -0400 From: Semen Protsenko To: Greg Kroah-Hartman Cc: Jiri Slaby , Grygorii Strashko , Nishanth Menon , linux-serial@vger.kernel.org, linux-kernel@vger.kernel.org, linux-omap@vger.kernel.org Subject: [PATCH v2] serial: omap: Fix error handling in probe Date: Thu, 30 Apr 2015 18:35:27 +0300 Message-Id: <1430408127-5237-1-git-send-email-semen.protsenko@globallogic.com> X-Mailer: git-send-email 1.7.9.5 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2276 Lines: 54 There is pm_qos_add_request() being executed on serial_omap_probe(), which stores "&up->pm_qos_request" from omap-serial driver to "pm_qos_array[PM_QOS_CPU_DMA_LATENCY]->constraints". If serial_omap_probe() fails after pm_qos_add_request() (e.g. on uart_add_one_port() call), pm_qos_array still keeping pm_qos_request struct from omap-serial driver, which is not valid anymore (since driver failed). This leads further to kernel crash on pm_qos_update_target(), executing from some completely different driver. We were observing this while trying to run audio playback while having one of omap-serial driver instances failed on uart_add_one_port() call: Unable to handle kernel paging request at virtual address fffffffc Backtrace: (plist_add) from (pm_qos_update_target) (pm_qos_update_target) from (pm_qos_add_request) (pm_qos_add_request) from (snd_pcm_hw_params) (snd_pcm_hw_params) from (snd_pcm_common_ioctl1) (snd_pcm_common_ioctl1) from (snd_pcm_playback_ioctl1) (snd_pcm_playback_ioctl1) from (snd_pcm_playback_ioctl) (snd_pcm_playback_ioctl) from (do_vfs_ioctl) (do_vfs_ioctl) from (SyS_ioctl) (SyS_ioctl) from (ret_fast_syscall) This patch adds pm_qos_remove_request() on fail path in serial_omap_probe() in order to fix this issue. While at it, free the wakeup settings on fail path as well, just like it's done in serial_omap_remove(). Signed-off-by: Semen Protsenko --- drivers/tty/serial/omap-serial.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/drivers/tty/serial/omap-serial.c b/drivers/tty/serial/omap-serial.c index 211479a..7f49172 100644 --- a/drivers/tty/serial/omap-serial.c +++ b/drivers/tty/serial/omap-serial.c @@ -1735,6 +1735,8 @@ static int serial_omap_probe(struct platform_device *pdev) err_add_port: pm_runtime_put(&pdev->dev); pm_runtime_disable(&pdev->dev); + pm_qos_remove_request(&up->pm_qos_request); + device_init_wakeup(up->dev, false); err_rs485: err_port_line: return ret; -- 1.7.9.5 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/