[<prev] [next>] [day] [month] [year] [list]
Message-ID: <1448593294.3216.11.camel@amd.com>
Date: Fri, 27 Nov 2015 03:01:42 +0000
From: "Xue, Ken" <Ken.Xue@....com>
To: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"mlin@...nel.org" <mlin@...nel.org>,
"stern@...land.harvard.edu" <stern@...land.harvard.edu>
CC: "linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
"Huang, Shane" <Shane.Huang@....com>,
SPG_Linux_Kernel <SPG_Linux_Kernel@....com>,
"aaron.lu@...el.com" <aaron.lu@...el.com>,
"stable@...r.kernel.org" <stable@...r.kernel.org>,
"JBottomley@...n.com" <JBottomley@...n.com>,
"Yu, Xiangliang" <Xiangliang.Yu@....com>
Subject: Re: [PATCH 1/2]Revert "SCSI: Fix NULL pointer dereference in runtime
PM"
On 四, 2015-09-10 at 10:23 +0800, Ken Xue wrote:
Can someone help to apply this patch series? Thanks.
http://marc.info/?l=linux-scsi&m=144185206825609&w=2
http://marc.info/?l=linux-scsi&m=144185208525611&w=2
> Revert "SCSI: Fix NULL pointer dereference in runtime PM"
>
> This reverts commit 49718f0fb8c9 ("SCSI: Fix NULL pointer dereference
> in
> runtime PM")
>
> The old commit may lead to a issue that
> blk_{pre|post}_runtime_suspend and
> blk_{pre|post}_runtime_resume can not be called in pairs.
>
> Take sr device as example, when sr device goes to runtime suspend,
> blk_{pre|post}_runtime_suspend will be called since sr device defined
> pm->runtime_suspend. But blk_{pre|post}_runtime_resume will not be
> called
> since sr device doesn't have pm->runtime_resume. Then sr device can
> not
> resume correctly anymore.
>
> Signed-off-by: Ken Xue <Ken.Xue@....com>
> Acked-by: Alan Stern <stern@...land.harvard.edu>
> Cc: stable@...r.kernel.org
>
> ---
> drivers/scsi/scsi_pm.c | 20 ++++++++++----------
> 1 file changed, 10 insertions(+), 10 deletions(-)
>
> diff --git a/drivers/scsi/scsi_pm.c b/drivers/scsi/scsi_pm.c
> index e4b7998..459abe1 100644
> --- a/drivers/scsi/scsi_pm.c
> +++ b/drivers/scsi/scsi_pm.c
> @@ -219,13 +219,13 @@ static int sdev_runtime_suspend(struct device
> *dev)
> struct scsi_device *sdev = to_scsi_device(dev);
> int err = 0;
>
> - if (pm && pm->runtime_suspend) {
> - err = blk_pre_runtime_suspend(sdev->request_queue);
> - if (err)
> - return err;
> + err = blk_pre_runtime_suspend(sdev->request_queue);
> + if (err)
> + return err;
> + if (pm && pm->runtime_suspend)
> err = pm->runtime_suspend(dev);
> - blk_post_runtime_suspend(sdev->request_queue, err);
> - }
> + blk_post_runtime_suspend(sdev->request_queue, err);
> +
> return err;
> }
>
> @@ -248,11 +248,11 @@ static int sdev_runtime_resume(struct device
> *dev)
> const struct dev_pm_ops *pm = dev->driver ? dev->driver->pm
> : NULL;
> int err = 0;
>
> - if (pm && pm->runtime_resume) {
> - blk_pre_runtime_resume(sdev->request_queue);
> + blk_pre_runtime_resume(sdev->request_queue);
> + if (pm && pm->runtime_resume)
> err = pm->runtime_resume(dev);
> - blk_post_runtime_resume(sdev->request_queue, err);
> - }
> + blk_post_runtime_resume(sdev->request_queue, err);
> +
> return err;
> }
>
Powered by blists - more mailing lists