[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <f7d2c4e6-101d-5f52-f3df-6adb0c8b8396@linux.intel.com>
Date: Tue, 12 Nov 2019 15:18:07 +0800
From: Dilip Kota <eswara.kota@...ux.intel.com>
To: Andy Shevchenko <andriy.shevchenko@...el.com>
Cc: gustavo.pimentel@...opsys.com, lorenzo.pieralisi@....com,
andrew.murray@....com, helgaas@...nel.org, jingoohan1@...il.com,
robh@...nel.org, martin.blumenstingl@...glemail.com,
linux-pci@...r.kernel.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, cheol.yong.kim@...el.com,
chuanhua.lei@...ux.intel.com, qi-ming.wu@...el.com
Subject: Re: [PATCH v5 2/3] dwc: PCI: intel: PCIe RC controller driver
On 11/11/2019 4:08 PM, Dilip Kota wrote:
>
> On 11/6/2019 8:24 PM, Andy Shevchenko wrote:
>> On Wed, Nov 06, 2019 at 11:44:02AM +0800, Dilip Kota wrote:
[...]
>>
>>> + return ret;
>>> +}
>>> + platform_set_drvdata(pdev, lpp);
>> I think it makes sense to setup at the end of the function (before
>> dev_info()
>> call).
> I have done it immediately after the memory allocation.
> Ok, i will move it before dev_info().
>
I ran test with all the changes and kernel panic is hit due to NULL
pointer access. It is because of platform_set_drvdata() moved before
dev_info, which resulted in intel_pcie_get_resources() doing
platform_get_drvdata and end up accessing NULL pointer. I will keep
'platform_set_drvdata()' remain unchanged.
Regards,
Dilip
Powered by blists - more mailing lists