lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Fri, 27 Apr 2018 08:02:49 -0500
From:   Alex Elder <elder@...aro.org>
To:     Arvind Yadav <arvind.yadav.cs@...il.com>, hvaibhav.linux@...il.com,
        johan@...nel.org, elder@...nel.org, gregkh@...uxfoundation.org
Cc:     devel@...verdev.osuosl.org, greybus-dev@...ts.linaro.org,
        linux-kernel@...r.kernel.org
Subject: Re: [greybus-dev] [PATCH] staging: greybus: Use gpio_is_valid()

On 04/27/2018 07:50 AM, Arvind Yadav wrote:
> 
> 
> On Friday 27 April 2018 05:47 PM, Alex Elder wrote:
>> On 04/27/2018 05:52 AM, Arvind Yadav wrote:
>>> Replace the manual validity checks for the GPIO with the
>>> gpio_is_valid().
>> I haven't looked through the code paths very closely, but I
>> think that get_named_gpio() might return -EPROBE_DEFER, which
>> would be something we want to pass to the caller.
> Yes of_get_name_gpio() can return other error value apart from
> -EPROBE_DEFER.
>> So rather than returning -ENODEV and hiding the reason the
>> call to of_get_named_gpio() failed, you should continue
>> returning the errno it supplies (if not a valid gpio number).
>>
>>                     -Alex
> I have return -ENODEV because invalid gpio pin can be positive.
> static inline bool gpio_is_valid(int number)
> {
>     return number >= 0 && number < ARCH_NR_GPIOS;
> }
> Here if number > ARCH_NR_GPIOS then it's invalid but return value
> will be positive.

Your reasoning is good.  However in all three of these cases,
the GPIO number you're checking is the value returned from
of_get_named_gpio().  The return value is a "GPIO number to
use with Linux generic GPIO API, or one of the errno value."

So unless the API of of_get_named_gpio() changes, you can be
sure that if the value returned is invalid, it is a negative
errno.  (And if the API did change, the person making that
change would be responsible for fixing all callers to ensure
the change didn't break them.)

This distinction may be why the code you're changing was only
testing for negative, rather than using gpio_is_valid() (you'll
see it's used elsewhere in the Greybus code--even in the same
source files.)

Anyway, changing the code to use gpio_is_valid() is fine.  But
you should avoid obscuring the reason for the error that the
return value from of_get_named_gpio() provides.

					-Alex

> We can return like this
>     " return (gpio > 0) ? -ENODEV: gpio;"
> 
> But not sure this is worth to handle this.
> 
> ~arvind
>>
>>> Signed-off-by: Arvind Yadav <arvind.yadav.cs@...il.com>
>>> ---
>>>   drivers/staging/greybus/arche-platform.c | 12 ++++++------
>>>   1 file changed, 6 insertions(+), 6 deletions(-)
>>>
>>> diff --git a/drivers/staging/greybus/arche-platform.c b/drivers/staging/greybus/arche-platform.c
>>> index 83254a7..fc6bf60 100644
>>> --- a/drivers/staging/greybus/arche-platform.c
>>> +++ b/drivers/staging/greybus/arche-platform.c
>>> @@ -448,9 +448,9 @@ static int arche_platform_probe(struct platform_device *pdev)
>>>       arche_pdata->svc_reset_gpio = of_get_named_gpio(np,
>>>                               "svc,reset-gpio",
>>>                               0);
>>> -    if (arche_pdata->svc_reset_gpio < 0) {
>>> +    if (!gpio_is_valid(arche_pdata->svc_reset_gpio)) {
>>>           dev_err(dev, "failed to get reset-gpio\n");
>>> -        return arche_pdata->svc_reset_gpio;
>>> +        return -ENODEV;
>>>       }
>>>       ret = devm_gpio_request(dev, arche_pdata->svc_reset_gpio, "svc-reset");
>>>       if (ret) {
>>> @@ -468,9 +468,9 @@ static int arche_platform_probe(struct platform_device *pdev)
>>>       arche_pdata->svc_sysboot_gpio = of_get_named_gpio(np,
>>>                                 "svc,sysboot-gpio",
>>>                                 0);
>>> -    if (arche_pdata->svc_sysboot_gpio < 0) {
>>> +    if (!gpio_is_valid(arche_pdata->svc_sysboot_gpio)) {
>>>           dev_err(dev, "failed to get sysboot gpio\n");
>>> -        return arche_pdata->svc_sysboot_gpio;
>>> +        return -ENODEV;
>>>       }
>>>       ret = devm_gpio_request(dev, arche_pdata->svc_sysboot_gpio, "sysboot0");
>>>       if (ret) {
>>> @@ -487,9 +487,9 @@ static int arche_platform_probe(struct platform_device *pdev)
>>>       arche_pdata->svc_refclk_req = of_get_named_gpio(np,
>>>                               "svc,refclk-req-gpio",
>>>                               0);
>>> -    if (arche_pdata->svc_refclk_req < 0) {
>>> +    if (!gpio_is_valid(arche_pdata->svc_refclk_req)) {
>>>           dev_err(dev, "failed to get svc clock-req gpio\n");
>>> -        return arche_pdata->svc_refclk_req;
>>> +        return -ENODEV;
>>>       }
>>>       ret = devm_gpio_request(dev, arche_pdata->svc_refclk_req,
>>>                   "svc-clk-req");
>>>
> 

Powered by blists - more mailing lists