[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date: Tue, 4 Apr 2017 09:28:08 -0500
From: Thor Thayer <thor.thayer@...ux.intel.com>
To: Borislav Petkov <bp@...en8.de>
Cc: mchehab@...nel.org, linux-edac@...r.kernel.org,
linux-kernel@...r.kernel.org, thor.thayer@...ux.intel.com.com
Subject: Re: [PATCH] EDAC, altera: Fix peripheral warnings for Cyclone5
Hi Boris,
On 04/04/2017 04:59 AM, Borislav Petkov wrote:
> On Mon, Apr 03, 2017 at 02:01:06PM -0500, thor.thayer@...ux.intel.com wrote:
>> From: Thor Thayer <thor.thayer@...ux.intel.com>
>>
>> The peripherals EDACs only exist on the Arria10 SoCFPGA. The Cyclone5
>> initialization has EDAC warnings when the peripherals aren't found
>> in the device tree. Fix by checking for Arria10 in the init functions.
>>
>> Signed-off-by: Thor Thayer <thor.thayer@...ux.intel.com>
>> ---
>> drivers/edac/altera_edac.c | 29 +++++++++++++++++++++++++----
>> 1 file changed, 25 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/edac/altera_edac.c b/drivers/edac/altera_edac.c
>> index 6421cc3..98e83f5 100644
>> --- a/drivers/edac/altera_edac.c
>> +++ b/drivers/edac/altera_edac.c
>> @@ -1024,13 +1024,30 @@ static int __maybe_unused altr_init_memory_port(void __iomem *ioaddr, int port)
>> return ret;
>> }
>>
>> +static int invalid_model(void)
>> +{
>> + struct device_node *np = of_find_node_by_path("/");
>
> That needs to have its return value checked, of course.
>
Actually, some of the of_ functions have validity checking in them. In
this case, of_get_property() calls of_find_property() which returns NULL
if np is 0 which is propagated back up to model so this code should be safe.
In the case of the of_node_put() below, the node is also checked for
non-NULL so I should be OK there.
>> + const char *model = of_get_property(np, "model", NULL);
>> +
>> + of_node_put(np);
>> + if (!model || strncmp(model, "Altera SOCFPGA Arria 10", 23) != 0)
>
> No need for the "!= 0"
>
OK. Thanks.
>> + return -ENODEV;
>> +
>> + return 0;
>> +}
>
> That function name "invalid_model" sounds like a boolean: is the model
> invalid. So you can simply return bools and not -ENODEV as the context
> you're using it is boolean.
>
Good point, thanks. I'll fix and re-submit.
Thanks for reviewing!
Powered by blists - more mailing lists