From mboxrd@z Thu Jan 1 00:00:00 1970 From: Omar Ramirez Luna Subject: [PATCH 0/2] ARM: OMAP: dmtimer: scheduling while atomic fixes Date: Thu, 24 Nov 2011 22:12:48 -0600 Message-ID: <1322194370-8073-1-git-send-email-omar.ramirez@ti.com> Return-path: Received: from na3sys009aog103.obsmtp.com ([74.125.149.71]:32930 "EHLO na3sys009aog103.obsmtp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751458Ab1KYEMx (ORCPT ); Thu, 24 Nov 2011 23:12:53 -0500 Received: by ggnp4 with SMTP id p4so3784753ggn.31 for ; Thu, 24 Nov 2011 20:12:51 -0800 (PST) Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: Tony Lindgren Cc: Russell King , Tarun Kanti DebBarma , Kevin Hilman , Santosh Shilimkar , Benoit Cousson , linux-omap@vger.kernel.org, linux-arm-kernel@lists.infradead.org, Omar Ramirez Luna Changes in dmtimer framework have introduced scheduling-while-atomic and other lock state BUGs, these are related to the usage of clk_get (internal path holds a mutex_lock) while holding a spin_lock_irqsave. The other inconsistent lock state BUGs are caused when calling omap_dm_timer_request* on softirq or hardirq, because code handling source parent clocks is still using clk_get, since there is only one user of those APIs that acquires a lock in a softirq context (tidspbridge) for now it can be changed. Omar Ramirez Luna (2): ARM: OMAP: dmtimer: fix sleeping function called from invalid context ARM: OMAP: dmtimer: reorganize omap_dm_timer_request_* arch/arm/plat-omap/dmtimer.c | 104 +++++++++++------------------ arch/arm/plat-omap/include/plat/dmtimer.h | 6 +- 2 files changed, 44 insertions(+), 66 deletions(-) -- 1.7.5.4