| Message ID | 20240627082426.394937-2-pengfei.li_1@nxp.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | clk: imx93: Move IMX93_CLK_END macro to clk driver | expand |
On 27/06/2024 10:24, Pengfei Li wrote: > IMX93_CLK_END was previously defined in imx93-clock.h to indicate > the number of clocks, but it is not part of the ABI, so it should > be moved to clk driver. > > Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > --- > > Notes: > Change for v2: > - Use pre-processor define to simplify code. > > drivers/clk/imx/clk-imx93.c | 2 ++ > 1 file changed, 2 insertions(+) Are you sure this builds fine? If you need to define it here, it means some part of that unit file already uses the define. If so, you include the header. If you include the header, you should see fat warning. Really, people, just build your patches... Best regards, Krzysztof
On Fri, Jun 28, 2024 at 09:38:54AM +0200, Krzysztof Kozlowski wrote: > On 27/06/2024 10:24, Pengfei Li wrote: > > IMX93_CLK_END was previously defined in imx93-clock.h to indicate > > the number of clocks, but it is not part of the ABI, so it should > > be moved to clk driver. > > > > Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > > --- > > > > Notes: > > Change for v2: > > - Use pre-processor define to simplify code. > > > > drivers/clk/imx/clk-imx93.c | 2 ++ > > 1 file changed, 2 insertions(+) > > Are you sure this builds fine? If you need to define it here, it means > some part of that unit file already uses the define. If so, you include > the header. If you include the header, you should see fat warning. > > Really, people, just build your patches... > > Best regards, > Krzysztof > > Hi Krzysztof, Yes, you're right, this macro definition has indeed been used. But this clk-imx93.c driver file is the only place where this macro is used. So maybe it would be more appropriate to define this macro in C rather than in the header file. Meanwhile, I also built my patch, but no warnings were found. BR, Pengfei Li
On Sat, Jun 29, 2024 at 10:49:49AM +0800, Pengfei Li wrote: > On Fri, Jun 28, 2024 at 09:38:54AM +0200, Krzysztof Kozlowski wrote: > > On 27/06/2024 10:24, Pengfei Li wrote: > > > IMX93_CLK_END was previously defined in imx93-clock.h to indicate > > > the number of clocks, but it is not part of the ABI, so it should > > > be moved to clk driver. > > > > > > Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > > > --- > > > > > > Notes: > > > Change for v2: > > > - Use pre-processor define to simplify code. > > > > > > drivers/clk/imx/clk-imx93.c | 2 ++ > > > 1 file changed, 2 insertions(+) > > > > Are you sure this builds fine? If you need to define it here, it means > > some part of that unit file already uses the define. If so, you include > > the header. If you include the header, you should see fat warning. > > > > Really, people, just build your patches... > > > > Best regards, > > Krzysztof > > > > > > Hi Krzysztof, > > Yes, you're right, this macro definition has indeed been used. But > this clk-imx93.c driver file is the only place where this macro is > used. So maybe it would be more appropriate to define this macro in > C rather than in the header file. Meanwhile, I also built my patch, > but no warnings were found. > > BR, > Pengfei Li > Hi Krzysztof, I've built my patches, but no warnings were found. This C file is the only place where this macro definition is used, so why would I see fat warning? BR, Pengfei
On 19/07/2024 08:08, Pengfei Li wrote: > On Sat, Jun 29, 2024 at 10:49:49AM +0800, Pengfei Li wrote: >> On Fri, Jun 28, 2024 at 09:38:54AM +0200, Krzysztof Kozlowski wrote: >>> On 27/06/2024 10:24, Pengfei Li wrote: >>>> IMX93_CLK_END was previously defined in imx93-clock.h to indicate >>>> the number of clocks, but it is not part of the ABI, so it should >>>> be moved to clk driver. >>>> >>>> Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> >>>> --- >>>> >>>> Notes: >>>> Change for v2: >>>> - Use pre-processor define to simplify code. >>>> >>>> drivers/clk/imx/clk-imx93.c | 2 ++ >>>> 1 file changed, 2 insertions(+) >>> >>> Are you sure this builds fine? If you need to define it here, it means >>> some part of that unit file already uses the define. If so, you include >>> the header. If you include the header, you should see fat warning. >>> >>> Really, people, just build your patches... >>> >>> Best regards, >>> Krzysztof >>> >>> >> >> Hi Krzysztof, >> >> Yes, you're right, this macro definition has indeed been used. But >> this clk-imx93.c driver file is the only place where this macro is >> used. So maybe it would be more appropriate to define this macro in >> C rather than in the header file. Meanwhile, I also built my patch, >> but no warnings were found. >> >> BR, >> Pengfei Li >> > > Hi Krzysztof, > > I've built my patches, but no warnings were found. This C file is the only place where this macro definition is used, so why would I see fat warning? I could be wrong, I really thought there should be a warning of duplicated define. But if there is none, sure, sounds good. Best regards, Krzysztof
On Sat, Jul 20, 2024 at 08:28:14PM +0200, Krzysztof Kozlowski wrote: > On 19/07/2024 08:08, Pengfei Li wrote: > > On Sat, Jun 29, 2024 at 10:49:49AM +0800, Pengfei Li wrote: > >> On Fri, Jun 28, 2024 at 09:38:54AM +0200, Krzysztof Kozlowski wrote: > >>> On 27/06/2024 10:24, Pengfei Li wrote: > >>>> IMX93_CLK_END was previously defined in imx93-clock.h to indicate > >>>> the number of clocks, but it is not part of the ABI, so it should > >>>> be moved to clk driver. > >>>> > >>>> Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > >>>> --- > >>>> > >>>> Notes: > >>>> Change for v2: > >>>> - Use pre-processor define to simplify code. > >>>> > >>>> drivers/clk/imx/clk-imx93.c | 2 ++ > >>>> 1 file changed, 2 insertions(+) > >>> > >>> Are you sure this builds fine? If you need to define it here, it means > >>> some part of that unit file already uses the define. If so, you include > >>> the header. If you include the header, you should see fat warning. > >>> > >>> Really, people, just build your patches... > >>> > >>> Best regards, > >>> Krzysztof > >>> > >>> > >> > >> Hi Krzysztof, > >> > >> Yes, you're right, this macro definition has indeed been used. But > >> this clk-imx93.c driver file is the only place where this macro is > >> used. So maybe it would be more appropriate to define this macro in > >> C rather than in the header file. Meanwhile, I also built my patch, > >> but no warnings were found. > >> > >> BR, > >> Pengfei Li > >> > > > > Hi Krzysztof, > > > > I've built my patches, but no warnings were found. This C file is the only place where this macro definition is used, so why would I see fat warning? > > I could be wrong, I really thought there should be a warning of > duplicated define. But if there is none, sure, sounds good. > > Best regards, > Krzysztof > > Hi Krzysztof, If this patchset is ok, could you help merge it? Otherwise I won't be able to send subsequent patches. BR, Pengfei
On 06/08/2024 10:36, Pengfei Li wrote: > Hi Krzysztof, > > If this patchset is ok, could you help merge it? Otherwise I won't be able to send subsequent patches. Sure, let me apply it to my clk tree. Oh wait... Best regards, Krzysztof
On 25/09/2024 00:54, Pengfei Li wrote: > On Tue, Aug 06, 2024 at 11:39:18AM +0200, Krzysztof Kozlowski wrote: >> On 06/08/2024 10:36, Pengfei Li wrote: >>> Hi Krzysztof, >>> >>> If this patchset is ok, could you help merge it? Otherwise I won't be able to send subsequent patches. >> >> Sure, let me apply it to my clk tree. Oh wait... >> >> Best regards, >> Krzysztof >> >> > Hi Krzysztof, > > These patches are still not merged into linux-next, can you help with that? Then I'll be able to send the subsequent patches! What do you need from me? Anyway, why are you pinging during merge window? Best regards, Krzysztof
On Tue, Aug 06, 2024 at 11:39:18AM +0200, Krzysztof Kozlowski wrote: > On 06/08/2024 10:36, Pengfei Li wrote: > > Hi Krzysztof, > > > > If this patchset is ok, could you help merge it? Otherwise I won't be able to send subsequent patches. > > Sure, let me apply it to my clk tree. Oh wait... > > Best regards, > Krzysztof > > Hi Krzysztof, These patches are still not merged into linux-next, can you help with that? Then I'll be able to send the subsequent patches! BR, Pengfei Li
On 24-06-27 16:24:25, Pengfei Li wrote: > IMX93_CLK_END was previously defined in imx93-clock.h to indicate > the number of clocks, but it is not part of the ABI, so it should > be moved to clk driver. > > Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > --- > > Notes: > Change for v2: > - Use pre-processor define to simplify code. > > drivers/clk/imx/clk-imx93.c | 2 ++ > 1 file changed, 2 insertions(+) > > diff --git a/drivers/clk/imx/clk-imx93.c b/drivers/clk/imx/clk-imx93.c > index c6a9bc8ecc1f..c8b65146e76e 100644 > --- a/drivers/clk/imx/clk-imx93.c > +++ b/drivers/clk/imx/clk-imx93.c > @@ -15,6 +15,8 @@ > > #include "clk.h" > > +#define IMX93_CLK_END 202 > + TBH, I don't like this. And I still don't understand the reason. Where are the clocks you want to add ? And why wouldn't it be OK to just keep this in the bindings header even if you need to add more clocks ? For the moment, I don't see the value of this change. Sorry, but NACK. > enum clk_sel { > LOW_SPEED_IO_SEL, > NON_IO_SEL, > -- > 2.34.1 >
On Wed, Oct 09, 2024 at 02:51:41PM +0300, Abel Vesa wrote: > On 24-06-27 16:24:25, Pengfei Li wrote: > > IMX93_CLK_END was previously defined in imx93-clock.h to indicate > > the number of clocks, but it is not part of the ABI, so it should > > be moved to clk driver. > > > > Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> > > --- > > > > Notes: > > Change for v2: > > - Use pre-processor define to simplify code. > > > > drivers/clk/imx/clk-imx93.c | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/drivers/clk/imx/clk-imx93.c b/drivers/clk/imx/clk-imx93.c > > index c6a9bc8ecc1f..c8b65146e76e 100644 > > --- a/drivers/clk/imx/clk-imx93.c > > +++ b/drivers/clk/imx/clk-imx93.c > > @@ -15,6 +15,8 @@ > > > > #include "clk.h" > > > > +#define IMX93_CLK_END 202 > > + > > TBH, I don't like this. And I still don't understand the reason. > > Where are the clocks you want to add ? Hi Abel, I re-sent a new version of the patch set which includes changes to add the new clocks: https://lore.kernel.org/all/20241011045736.308338-1-pengfei.li_1@nxp.com/ > > And why wouldn't it be OK to just keep this in the bindings header even > if you need to add more clocks ? Krzysztof has help explained this: https://lore.kernel.org/all/a07828a4-8040-42cb-8c62-8939cac4d9de@kernel.org/. > > For the moment, I don't see the value of this change. > > Sorry, but NACK. > > > enum clk_sel { > > LOW_SPEED_IO_SEL, > > NON_IO_SEL, > > -- > > 2.34.1 > > >
diff --git a/drivers/clk/imx/clk-imx93.c b/drivers/clk/imx/clk-imx93.c index c6a9bc8ecc1f..c8b65146e76e 100644 --- a/drivers/clk/imx/clk-imx93.c +++ b/drivers/clk/imx/clk-imx93.c @@ -15,6 +15,8 @@ #include "clk.h" +#define IMX93_CLK_END 202 + enum clk_sel { LOW_SPEED_IO_SEL, NON_IO_SEL,
IMX93_CLK_END was previously defined in imx93-clock.h to indicate the number of clocks, but it is not part of the ABI, so it should be moved to clk driver. Signed-off-by: Pengfei Li <pengfei.li_1@nxp.com> --- Notes: Change for v2: - Use pre-processor define to simplify code. drivers/clk/imx/clk-imx93.c | 2 ++ 1 file changed, 2 insertions(+)