From: David Lechner <david@lechnology.com>
To: Sekhar Nori <nsekhar@ti.com>,
Michael Turquette <mturquette@baylibre.com>,
Stephen Boyd <sboyd@kernel.org>
Cc: linux-clk@vger.kernel.org, linux-arm-kernel@lists.infradead.org,
Kevin Hilman <khilman@kernel.org>,
linux-kernel@vger.kernel.org
Subject: Re: [PATCH 9/9] clk: davinci: Fix link errors when not all SoCs are enabled
Date: Mon, 28 May 2018 11:54:04 -0500 [thread overview]
Message-ID: <679a6a16-213c-c766-5d82-0c4ce2bdcf75@lechnology.com> (raw)
In-Reply-To: <a1b62396-534c-bc68-63a4-ac493690a3e7@ti.com>
On 05/28/2018 08:43 AM, Sekhar Nori wrote:
> On Friday 25 May 2018 11:41 PM, David Lechner wrote:
>> This fixes linker errors due to undefined symbols when one or more of
>> the TI DaVinci SoCs is not enabled in the kernel config.
>>
>> Signed-off-by: David Lechner <david@lechnology.com>
>> ---
>> drivers/clk/davinci/pll.c | 16 ++++++++++++++++
>> drivers/clk/davinci/pll.h | 11 ++++++++---
>> drivers/clk/davinci/psc.c | 14 ++++++++++++++
>> drivers/clk/davinci/psc.h | 12 ++++++++++++
>> include/linux/clk/davinci.h | 19 +++++++++++++++----
>> 5 files changed, 65 insertions(+), 7 deletions(-)
>>
>> diff --git a/drivers/clk/davinci/pll.c b/drivers/clk/davinci/pll.c
>> index 84a343060bc8..65abd371692d 100644
>> --- a/drivers/clk/davinci/pll.c
>> +++ b/drivers/clk/davinci/pll.c
>> @@ -860,25 +860,41 @@ static struct davinci_pll_platform_data *davinci_pll_get_pdata(struct device *de
>> }
>>
>> /* needed in early boot for clocksource/clockevent */
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> CLK_OF_DECLARE(da850_pll0, "ti,da850-pll0", of_da850_pll0_init);
>> +#endif
>>
>> static const struct of_device_id davinci_pll_of_match[] = {
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> { .compatible = "ti,da850-pll1", .data = of_da850_pll1_init },
>> +#endif
>> { }
>> };
>>
>> static const struct platform_device_id davinci_pll_id_table[] = {
>> +#ifdef CONFIG_ARCH_DAVINCI_DA830
>> { .name = "da830-pll", .driver_data = (kernel_ulong_t)da830_pll_init },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> { .name = "da850-pll0", .driver_data = (kernel_ulong_t)da850_pll0_init },
>> { .name = "da850-pll1", .driver_data = (kernel_ulong_t)da850_pll1_init },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM355
>> { .name = "dm355-pll1", .driver_data = (kernel_ulong_t)dm355_pll1_init },
>> { .name = "dm355-pll2", .driver_data = (kernel_ulong_t)dm355_pll2_init },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM365
>> { .name = "dm365-pll1", .driver_data = (kernel_ulong_t)dm365_pll1_init },
>> { .name = "dm365-pll2", .driver_data = (kernel_ulong_t)dm365_pll2_init },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM644x
>> { .name = "dm644x-pll1", .driver_data = (kernel_ulong_t)dm644x_pll1_init },
>> { .name = "dm644x-pll2", .driver_data = (kernel_ulong_t)dm644x_pll2_init },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM646x
>> { .name = "dm646x-pll1", .driver_data = (kernel_ulong_t)dm646x_pll1_init },
>> { .name = "dm646x-pll2", .driver_data = (kernel_ulong_t)dm646x_pll2_init },
>> +#endif
>> { }
>> };
>>
>> diff --git a/drivers/clk/davinci/pll.h b/drivers/clk/davinci/pll.h
>> index b2e5c4496645..7cc354dd29e2 100644
>> --- a/drivers/clk/davinci/pll.h
>> +++ b/drivers/clk/davinci/pll.h
>> @@ -122,14 +122,19 @@ int of_davinci_pll_init(struct device *dev, struct device_node *node,
>>
>> /* Platform-specific callbacks */
>>
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> int da850_pll1_init(struct device *dev, void __iomem *base, struct regmap *cfgchip);
>> void of_da850_pll0_init(struct device_node *node);
>> int of_da850_pll1_init(struct device *dev, void __iomem *base, struct regmap *cfgchip);
>> -
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM355
>> int dm355_pll2_init(struct device *dev, void __iomem *base, struct regmap *cfgchip);
>> -
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM644x
>> int dm644x_pll2_init(struct device *dev, void __iomem *base, struct regmap *cfgchip);
>> -
>> +#endif
>
> The traditional way of dealing with this for functions is to provide a
> static inline stub when CONFIG_ARCH_DAVINCI_DM644x is not defined. Can
> we use that? That helps in avoiding ifdefs elsewhere.
I suppose we could. But then we would be taking function pointers to
static inline functions, which seems kind of weird to me.
>
>> +#ifdef CONFIG_ARCH_DAVINCI_DM646x
>> int dm646x_pll2_init(struct device *dev, void __iomem *base, struct regmap *cfgchip);
>> +#endif
>>
>> #endif /* __CLK_DAVINCI_PLL_H___ */
>> diff --git a/drivers/clk/davinci/psc.c b/drivers/clk/davinci/psc.c
>> index 6326ba1fe3cc..fffbed5e263b 100644
>> --- a/drivers/clk/davinci/psc.c
>> +++ b/drivers/clk/davinci/psc.c
>> @@ -513,20 +513,34 @@ int of_davinci_psc_clk_init(struct device *dev,
>> }
>>
>> static const struct of_device_id davinci_psc_of_match[] = {
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> { .compatible = "ti,da850-psc0", .data = &of_da850_psc0_init_data },
>> { .compatible = "ti,da850-psc1", .data = &of_da850_psc1_init_data },
>> +#endif
>> { }
>> };
>>
>> static const struct platform_device_id davinci_psc_id_table[] = {
>> +#ifdef CONFIG_ARCH_DAVINCI_DA830
>> { .name = "da830-psc0", .driver_data = (kernel_ulong_t)&da830_psc0_init_data },
>> { .name = "da830-psc1", .driver_data = (kernel_ulong_t)&da830_psc1_init_data },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> { .name = "da850-psc0", .driver_data = (kernel_ulong_t)&da850_psc0_init_data },
>> { .name = "da850-psc1", .driver_data = (kernel_ulong_t)&da850_psc1_init_data },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM355
>> { .name = "dm355-psc", .driver_data = (kernel_ulong_t)&dm355_psc_init_data },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM365
>> { .name = "dm365-psc", .driver_data = (kernel_ulong_t)&dm365_psc_init_data },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM644x
>> { .name = "dm644x-psc", .driver_data = (kernel_ulong_t)&dm644x_psc_init_data },
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM646x
>> { .name = "dm646x-psc", .driver_data = (kernel_ulong_t)&dm646x_psc_init_data },
>> +#endif
>> { }
>> };
>>
>> diff --git a/drivers/clk/davinci/psc.h b/drivers/clk/davinci/psc.h
>> index c2a7df6413fe..6a42529d31a9 100644
>> --- a/drivers/clk/davinci/psc.h
>> +++ b/drivers/clk/davinci/psc.h
>> @@ -94,15 +94,27 @@ struct davinci_psc_init_data {
>> int (*psc_init)(struct device *dev, void __iomem *base);
>> };
>>
>> +#ifdef CONFIG_ARCH_DAVINCI_DA830
>> extern const struct davinci_psc_init_data da830_psc0_init_data;
>> extern const struct davinci_psc_init_data da830_psc1_init_data;
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DA850
>> extern const struct davinci_psc_init_data da850_psc0_init_data;
>> extern const struct davinci_psc_init_data da850_psc1_init_data;
>> extern const struct davinci_psc_init_data of_da850_psc0_init_data;
>> extern const struct davinci_psc_init_data of_da850_psc1_init_data;
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM355
>> extern const struct davinci_psc_init_data dm355_psc_init_data;
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM356
>> extern const struct davinci_psc_init_data dm365_psc_init_data;
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM644x
>> extern const struct davinci_psc_init_data dm644x_psc_init_data;
>> +#endif
>> +#ifdef CONFIG_ARCH_DAVINCI_DM646x
>> extern const struct davinci_psc_init_data dm646x_psc_init_data;
>> +#endif
>
> ifdefs around these extern declarations are not needed.
True. I guess I just prefer a compiler error to a linker error.
next prev parent reply other threads:[~2018-05-28 16:54 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-05-25 18:11 [PATCH 0/9] clk: davinci: outstanding fixes David Lechner
2018-05-25 18:11 ` [PATCH 1/9] clk: davinci: pll-dm355: drop pll2_sysclk2 David Lechner
2018-05-25 18:11 ` [PATCH 2/9] clk: davinci: pll-dm355: fix SYSCLKn parent names David Lechner
2018-05-25 18:11 ` [PATCH 3/9] clk: davinci: psc-dm355: fix ASP0/1 clkdev lookups David Lechner
2018-05-25 18:11 ` [PATCH 4/9] clk: davinci: pll-dm646x: keep PLL2 SYSCLK1 always enabled David Lechner
2018-05-30 17:22 ` Michael Turquette
2018-05-30 19:20 ` David Lechner
2018-05-25 18:11 ` [PATCH 5/9] clk: davinci: psc-dm365: fix few clocks David Lechner
2018-05-25 18:11 ` [PATCH 6/9] clk: davinci: pll: allow dev == NULL David Lechner
2018-05-30 19:46 ` Michael Turquette
2018-05-30 19:59 ` David Lechner
2018-05-25 18:11 ` [PATCH 7/9] clk: davinci: da850-pll: change PLL0 to CLK_OF_DECLARE David Lechner
2018-05-25 18:11 ` [PATCH 8/9] clk: davinci: psc: allow for dev == NULL David Lechner
2018-05-25 18:11 ` [PATCH 9/9] clk: davinci: Fix link errors when not all SoCs are enabled David Lechner
2018-05-28 13:43 ` Sekhar Nori
2018-05-28 16:54 ` David Lechner [this message]
2018-05-30 20:09 ` [PATCH 0/9] clk: davinci: outstanding fixes Michael Turquette
2018-05-31 4:43 ` Sekhar Nori
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=679a6a16-213c-c766-5d82-0c4ce2bdcf75@lechnology.com \
--to=david@lechnology.com \
--cc=khilman@kernel.org \
--cc=linux-arm-kernel@lists.infradead.org \
--cc=linux-clk@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mturquette@baylibre.com \
--cc=nsekhar@ti.com \
--cc=sboyd@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
all inboxes | Powered by JetHome®