From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jaehoon Chung Subject: Re: [PATCH] sdhci: put regulator if probe fails Date: Sat, 13 Oct 2012 16:01:14 +0900 Message-ID: <507911BA.3080204@samsung.com> References: <1350089717-4050-1-git-send-email-cascardo@holoscopio.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: Received: from mailout3.samsung.com ([203.254.224.33]:28668 "EHLO mailout3.samsung.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750975Ab2JMHBe (ORCPT ); Sat, 13 Oct 2012 03:01:34 -0400 In-reply-to: <1350089717-4050-1-git-send-email-cascardo@holoscopio.com> Sender: linux-mmc-owner@vger.kernel.org List-Id: linux-mmc@vger.kernel.org To: Thadeu Lima de Souza Cascardo Cc: linux-kernel@vger.kernel.org, Chris Ball , linux-mmc@vger.kernel.org On 10/13/2012 09:55 AM, Thadeu Lima de Souza Cascardo wrote: > When using the dummy regulator, SDHCI may fail its probing because the > regulator does not support any voltages. > > When reloading the driver, you will get a warning about a duplicate > sysfs link. > > [72211.963386] 0000:03:00.0 supply vmmc not found, using dummy regulator > [72211.963409] ------------[ cut here ]------------ > [72211.963420] WARNING: at fs/sysfs/dir.c:536 sysfs_add_one+0x99/0xad() > [72211.963424] Hardware name: > [72211.963429] sysfs: cannot create duplicate filename '/devices/platform/reg-dummy/regulator/regulator.0/0000:03:00.0-vmmc' > > Avoid this by properly cleaning up when the probe fails calling > regulator_put. Other fail paths get fixed as well. > > Signed-off-by: Thadeu Lima de Souza Cascardo > --- > drivers/mmc/host/sdhci.c | 6 +++++- > 1 file changed, 5 insertions(+), 1 deletion(-) > > diff --git a/drivers/mmc/host/sdhci.c b/drivers/mmc/host/sdhci.c > index 7922adb..dd6bc26 100644 > --- a/drivers/mmc/host/sdhci.c > +++ b/drivers/mmc/host/sdhci.c > @@ -2992,7 +2992,8 @@ int sdhci_add_host(struct sdhci_host *host) > if (mmc->ocr_avail == 0) { > pr_err("%s: Hardware doesn't report any " > "support voltages.\n", mmc_hostname(mmc)); > - return -ENODEV; > + ret = -ENODEV; > + goto out_vmmc; > } > > spin_lock_init(&host->lock); > @@ -3121,6 +3122,9 @@ reset: > untasklet: > tasklet_kill(&host->card_tasklet); > tasklet_kill(&host->finish_tasklet); > +out_vmmc: > + if (host->vmmc) > + regulator_put(host->vmmc); If fall down to untasklet, need not to disable regulator for vmmc? Best Regards, Jaehoon Chung > > return ret; > } >