From mboxrd@z Thu Jan 1 00:00:00 1970 From: Mike Turquette Subject: Re: sdhci_s3c_consider_clock scheduling while atomic - clk_round_rate Date: Mon, 18 Aug 2014 08:46:25 -0700 Message-ID: References: <53A99EB5.6010907@samsung.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Return-path: Received: from mail-wg0-f45.google.com ([74.125.82.45]:64229 "EHLO mail-wg0-f45.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751633AbaHRPqq (ORCPT ); Mon, 18 Aug 2014 11:46:46 -0400 Received: by mail-wg0-f45.google.com with SMTP id x12so5078204wgg.4 for ; Mon, 18 Aug 2014 08:46:45 -0700 (PDT) In-Reply-To: <53A99EB5.6010907@samsung.com> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Tomasz Figa Cc: Daniel Drake , linux-samsung-soc , linux-mmc@vger.kernel.org, Ben Dooks , Tomasz Figa , Ulf Hansson , Chris Ball On Tue, Jun 24, 2014 at 8:52 AM, Tomasz Figa wrote: > Hi Daniel, > > [adding Ulf, Chris and Mike to the discussion] > > On 24.06.2014 11:48, Daniel Drake wrote: >> sdhci_s3c_set_clock is called from sdhci_do_set_ios with interrupts >> disabled, and this calls into sdhci_s3c_consider_clock(). >> >> The patch "mmc: sdhci-s3c: Cache bus clock rates" addressed some >> scheduling while atomic in this function, but there are more issues >> here, seen while testing 3.16-rc2 on exynos4412: >> >> BUG: sleeping function called from invalid context at kernel/locking/mutex.c:103 >> in_atomic(): 1, irqs_disabled(): 128, pid: 75, name: mmcqd/0 >> Preemption disabled at:[< (null)>] (null) >> >> CPU: 0 PID: 75 Comm: mmcqd/0 Not tainted 3.16.0-rc2-00028-ge9fe7eb-dirty #77 >> [] (unwind_backtrace) from [] (show_stack+0x10/0x14) >> [] (show_stack) from [] (dump_stack+0x84/0xc4) >> [] (dump_stack) from [] (mutex_lock+0x1c/0x3c) >> [] (mutex_lock) from [] (clk_prepare_lock+0x6c/0xf4) >> [] (clk_prepare_lock) from [] (clk_round_rate+0x10/0x2c) >> [] (clk_round_rate) from [] (sdhci_s3c_set_clock+0x4c/0x1e8) >> [] (sdhci_s3c_set_clock) from [] >> (sdhci_cmu_set_clock+0x54/0x140) >> [] (sdhci_cmu_set_clock) from [] >> (sdhci_do_set_ios+0x138/0x58c) >> [] (sdhci_do_set_ios) from [] (sdhci_set_ios+0x28/0x34) >> >> clk_round_rate cannot be called here because it takes a mutex. >> >> sdhci_s3c_set_clock() also calls into clk_prepare_enable() which looks >> like it could trigger this problem too. Perhaps this clk_prepare_enable could be converted simply to clk_enable? Of course a call to clk_prepare call would still be required before entering into the section where clk_enable is called. Regards, Mike > > Apparently this is related to generic sdhci code calling .set_clock() > under spin_lock_irqsave(). For reference, see sdhci_do_set_ios() and > sdhci_tasklet_finish(). > > Chris, Ulf, any suggestions? > > Best regards, > Tomasz