diff mbox series

[v2] slimbus: ngd: QCOM_QMI_HELPERS has to be selected

Message ID 20181015194441.10475-1-niklas.cassel@linaro.org
State Accepted
Commit 5323ac5177c445c15f49c91fadf9539a0bc95fa9
Headers show
Series [v2] slimbus: ngd: QCOM_QMI_HELPERS has to be selected | expand

Commit Message

Niklas Cassel Oct. 15, 2018, 7:44 p.m. UTC
QCOM_QMI_HELPERS is a hidden kconfig, so the proper usage is
to select it, not depend upon it.

Because of this change, we now also need to depend on the same
Kconfigs as QCOM_QMI_HELPERS depends on.

Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>

---
Hello Greg, Srini,

I'm sorry for this.
(Although I'm a bit curious why 0-day test bot didn't catch this.)

Considering that I've just changed QCOM_QMI_HELPERS in:
ccfb464cd106 ("soc: qcom: Allow COMPILE_TEST of qcom SoC Kconfigs")
which is currently in linux-next, and that this Kconfig should
depend on the same Kconfigs as QCOM_QMI_HELPERS depends on,
I chose to have this Kconfig match the QCOM_QMI_HELPERS that is
currently in linux-next.

 drivers/slimbus/Kconfig | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

-- 
2.17.2

Comments

Srinivas Kandagatla Oct. 16, 2018, 9:35 a.m. UTC | #1
Thanks Niklas for quick v2 patch!

On 15/10/18 20:44, Niklas Cassel wrote:
> QCOM_QMI_HELPERS is a hidden kconfig, so the proper usage is

> to select it, not depend upon it.

> 

> Because of this change, we now also need to depend on the same

> Kconfigs as QCOM_QMI_HELPERS depends on.

> 

> Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>


LGTM

Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>


> ---

> Hello Greg, Srini,

> 

> I'm sorry for this.

> (Although I'm a bit curious why 0-day test bot didn't catch this.)

> 

> Considering that I've just changed QCOM_QMI_HELPERS in:

> ccfb464cd106 ("soc: qcom: Allow COMPILE_TEST of qcom SoC Kconfigs")

> which is currently in linux-next, and that this Kconfig should

> depend on the same Kconfigs as QCOM_QMI_HELPERS depends on,

> I chose to have this Kconfig match the QCOM_QMI_HELPERS that is

> currently in linux-next.

> 

>   drivers/slimbus/Kconfig | 5 +++--

>   1 file changed, 3 insertions(+), 2 deletions(-)

> 

> diff --git a/drivers/slimbus/Kconfig b/drivers/slimbus/Kconfig

> index 9d73ad806698..8cd595148d17 100644

> --- a/drivers/slimbus/Kconfig

> +++ b/drivers/slimbus/Kconfig

> @@ -22,8 +22,9 @@ config SLIM_QCOM_CTRL

>   

>   config SLIM_QCOM_NGD_CTRL

>   	tristate "Qualcomm SLIMbus Satellite Non-Generic Device Component"

> -	depends on QCOM_QMI_HELPERS

> -	depends on HAS_IOMEM && DMA_ENGINE

> +	depends on HAS_IOMEM && DMA_ENGINE && NET

> +	depends on ARCH_QCOM || COMPILE_TEST

> +	select QCOM_QMI_HELPERS

>   	help

>   	  Select driver if Qualcomm's SLIMbus Satellite Non-Generic Device

>   	  Component is programmed using Linux kernel.

>
Greg KH Nov. 7, 2018, 1:59 p.m. UTC | #2
On Mon, Oct 15, 2018 at 09:44:41PM +0200, Niklas Cassel wrote:
> QCOM_QMI_HELPERS is a hidden kconfig, so the proper usage is

> to select it, not depend upon it.

> 

> Because of this change, we now also need to depend on the same

> Kconfigs as QCOM_QMI_HELPERS depends on.

> 

> Signed-off-by: Niklas Cassel <niklas.cassel@linaro.org>

> Acked-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>

> ---

> Hello Greg, Srini,

> 

> I'm sorry for this.

> (Although I'm a bit curious why 0-day test bot didn't catch this.)

> 

> Considering that I've just changed QCOM_QMI_HELPERS in:

> ccfb464cd106 ("soc: qcom: Allow COMPILE_TEST of qcom SoC Kconfigs")

> which is currently in linux-next, and that this Kconfig should

> depend on the same Kconfigs as QCOM_QMI_HELPERS depends on,

> I chose to have this Kconfig match the QCOM_QMI_HELPERS that is

> currently in linux-next.


Really?  I do this and then I get this build error on x86:

drivers/slimbus/qcom-ngd-ctrl.c: In function ‘of_qcom_slim_ngd_register’:
drivers/slimbus/qcom-ngd-ctrl.c:1333:63: warning: dereferencing ‘void *’ pointer
  data = of_match_node(qcom_slim_ngd_dt_match, parent->of_node)->data;
                                                               ^~
drivers/slimbus/qcom-ngd-ctrl.c:1333:63: error: request for member ‘data’ in something not a structure or union


So I can't take this, something else must be wrong here...

thanks,

greg k-h
Srinivas Kandagatla Nov. 7, 2018, 2:15 p.m. UTC | #3
On 07/11/18 13:59, Greg KH wrote:
> Really?  I do this and then I get this build error on x86:

> 

> drivers/slimbus/qcom-ngd-ctrl.c: In function ‘of_qcom_slim_ngd_register’:

> drivers/slimbus/qcom-ngd-ctrl.c:1333:63: warning: dereferencing ‘void *’ pointer

>    data = of_match_node(qcom_slim_ngd_dt_match, parent->of_node)->data;

>                                                                 ^~

> drivers/slimbus/qcom-ngd-ctrl.c:1333:63: error: request for member ‘data’ in something not a structure or union

> 

> 

> So I can't take this, something else must be wrong here...

That is fine!
Yes, there seems to be one more issue here, on non DT platforms 
of_match_node is set to be NULL, which is why we are seeing this error I 
guess!

I will fix this up, do some test and send both the fixes together.

thanks,
srini
diff mbox series

Patch

diff --git a/drivers/slimbus/Kconfig b/drivers/slimbus/Kconfig
index 9d73ad806698..8cd595148d17 100644
--- a/drivers/slimbus/Kconfig
+++ b/drivers/slimbus/Kconfig
@@ -22,8 +22,9 @@  config SLIM_QCOM_CTRL
 
 config SLIM_QCOM_NGD_CTRL
 	tristate "Qualcomm SLIMbus Satellite Non-Generic Device Component"
-	depends on QCOM_QMI_HELPERS
-	depends on HAS_IOMEM && DMA_ENGINE
+	depends on HAS_IOMEM && DMA_ENGINE && NET
+	depends on ARCH_QCOM || COMPILE_TEST
+	select QCOM_QMI_HELPERS
 	help
 	  Select driver if Qualcomm's SLIMbus Satellite Non-Generic Device
 	  Component is programmed using Linux kernel.