[PATCH 0/3] memory: emif: ifdefs and platform_driver_probe()

[PATCH 0/3] memory: emif: ifdefs and platform_driver_probe()

[Uwe Kleine-König]

Hello,

while preparing the series to convert the platform drivers below
drivers/memory to use .remove_new()[1], I noticed that the emif driver
unnecessarily uses some #ifdefs and doesn't use platform_driver_probe()
correctly. Note there is a conflict between these series. Tell me if you
need support to resolve this.

The alternative to the third patch is to do s/__init_or_module/__init/
in the driver and mark emif_driver with __refdata.

Best regards
Uwe

[1] https://lore.kernel.org/r/cover.1702822744.git.u.kleine-koenig@pengutronix.de 

Uwe Kleine-König (3):
  memory: emif: Simplify code handling CONFIG_DEBUG_FS
  memory: emif: Simplify code handling CONFIG_OF
  memory: emif: Drop usage of platform_driver_probe()

 drivers/memory/emif.c | 63 ++++++++++++++++---------------------------
 1 file changed, 23 insertions(+), 40 deletions(-)

base-commit: 17cb8a20bde66a520a2ca7aad1063e1ce7382240
-- 
2.42.0

[PATCH 1/3] memory: emif: Simplify code handling CONFIG_DEBUG_FS

[Uwe Kleine-König]

Instead of using #ifdef make use of IS_ENABLED().

According to bloat-o-meter this patch doesn't change code sizes with
CONFIG_DEBUG_FS=n.

Also change emif_debugfs_init() to return void. The only caller doesn't
check the return value anyhow.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
 drivers/memory/emif.c | 32 ++++++++++++--------------------
 1 file changed, 12 insertions(+), 20 deletions(-)

diff --git a/drivers/memory/emif.c b/drivers/memory/emif.c
index f305643209f0..dd1d87f8fdc8 100644
--- a/drivers/memory/emif.c
+++ b/drivers/memory/emif.c
@@ -72,7 +72,6 @@ static DEFINE_SPINLOCK(emif_lock);
 static unsigned long	irq_state;
 static LIST_HEAD(device_list);
 
-#ifdef CONFIG_DEBUG_FS
 static void do_emif_regdump_show(struct seq_file *s, struct emif_data *emif,
 	struct emif_regs *regs)
 {
@@ -140,31 +139,24 @@ static int emif_mr4_show(struct seq_file *s, void *unused)
 
 DEFINE_SHOW_ATTRIBUTE(emif_mr4);
 
-static int __init_or_module emif_debugfs_init(struct emif_data *emif)
+static void __init_or_module emif_debugfs_init(struct emif_data *emif)
 {
-	emif->debugfs_root = debugfs_create_dir(dev_name(emif->dev), NULL);
-	debugfs_create_file("regcache_dump", S_IRUGO, emif->debugfs_root, emif,
-			    &emif_regdump_fops);
-	debugfs_create_file("mr4", S_IRUGO, emif->debugfs_root, emif,
-			    &emif_mr4_fops);
-	return 0;
+	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
+		emif->debugfs_root = debugfs_create_dir(dev_name(emif->dev), NULL);
+		debugfs_create_file("regcache_dump", S_IRUGO, emif->debugfs_root, emif,
+				    &emif_regdump_fops);
+		debugfs_create_file("mr4", S_IRUGO, emif->debugfs_root, emif,
+				    &emif_mr4_fops);
+	}
 }
 
 static void __exit emif_debugfs_exit(struct emif_data *emif)
 {
-	debugfs_remove_recursive(emif->debugfs_root);
-	emif->debugfs_root = NULL;
+	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
+		debugfs_remove_recursive(emif->debugfs_root);
+		emif->debugfs_root = NULL;
+	}
 }
-#else
-static inline int __init_or_module emif_debugfs_init(struct emif_data *emif)
-{
-	return 0;
-}
-
-static inline void __exit emif_debugfs_exit(struct emif_data *emif)
-{
-}
-#endif
 
 /*
  * Get bus width used by EMIF. Note that this may be different from the
-- 
2.42.0

Re: [PATCH 1/3] memory: emif: Simplify code handling CONFIG_DEBUG_FS

[Krzysztof Kozlowski]

On 17/12/2023 20:31, Uwe Kleine-König wrote:
> Instead of using #ifdef make use of IS_ENABLED().
> 
> According to bloat-o-meter this patch doesn't change code sizes with
> CONFIG_DEBUG_FS=n.
> 
> Also change emif_debugfs_init() to return void. The only caller doesn't
> check the return value anyhow.
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---

Applied.

Best regards,
Krzysztof

[PATCH 2/3] memory: emif: Simplify code handling CONFIG_OF

[Uwe Kleine-König]

The first thing that of_get_memory_device_details() does is calling
of_parse_phandle(). With CONFIG_OF=n this returns NULL in a static
inline function. So the compiler can determine that
of_get_memory_device_details() also returns NULL. bloat-o-meter confirms
that this patch has no effects on the size of the generated code for
CONFIG_OF=n builds.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
 drivers/memory/emif.c | 10 ----------
 1 file changed, 10 deletions(-)

diff --git a/drivers/memory/emif.c b/drivers/memory/emif.c
index dd1d87f8fdc8..5a3194b9b3bc 100644
--- a/drivers/memory/emif.c
+++ b/drivers/memory/emif.c
@@ -826,7 +826,6 @@ static int is_custom_config_valid(struct emif_custom_configs *cust_cfgs,
 	return valid;
 }
 
-#if defined(CONFIG_OF)
 static void __init_or_module of_get_custom_configs(struct device_node *np_emif,
 		struct emif_data *emif)
 {
@@ -983,15 +982,6 @@ static struct emif_data * __init_or_module of_get_memory_device_details(
 	return emif;
 }
 
-#else
-
-static struct emif_data * __init_or_module of_get_memory_device_details(
-		struct device_node *np_emif, struct device *dev)
-{
-	return NULL;
-}
-#endif
-
 static struct emif_data *__init_or_module get_device_details(
 		struct platform_device *pdev)
 {
-- 
2.42.0

Re: [PATCH 2/3] memory: emif: Simplify code handling CONFIG_OF

[Krzysztof Kozlowski]

On 17/12/2023 20:31, Uwe Kleine-König wrote:
> The first thing that of_get_memory_device_details() does is calling
> of_parse_phandle(). With CONFIG_OF=n this returns NULL in a static
> inline function. So the compiler can determine that
> of_get_memory_device_details() also returns NULL. bloat-o-meter confirms
> that this patch has no effects on the size of the generated code for
> CONFIG_OF=n builds.
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---
>  drivers/memory/emif.c | 10 ----------
>  1 file changed, 10 deletions(-)

Applied.

Best regards,
Krzysztof

[PATCH 3/3] memory: emif: Drop usage of platform_driver_probe()

[Uwe Kleine-König]

There are considerations to drop platform_driver_probe() as a concept
that isn't relevant any more today. It comes with an added complexity
that makes many users hold it wrong. (E.g. this driver should have
better used __init instead of __init_or_module to mark functions only
relevant to .probe() and mark the driver struct with __refdata.)

This fixes a W=1 build warning:

	WARNING: modpost: drivers/memory/emif: section mismatch in reference: emif_driver+0x4 (section: .data) -> emif_remove (section: .exit.text)

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
---
 drivers/memory/emif.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)

diff --git a/drivers/memory/emif.c b/drivers/memory/emif.c
index 5a3194b9b3bc..84973d7133f4 100644
--- a/drivers/memory/emif.c
+++ b/drivers/memory/emif.c
@@ -139,7 +139,7 @@ static int emif_mr4_show(struct seq_file *s, void *unused)
 
 DEFINE_SHOW_ATTRIBUTE(emif_mr4);
 
-static void __init_or_module emif_debugfs_init(struct emif_data *emif)
+static void emif_debugfs_init(struct emif_data *emif)
 {
 	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
 		emif->debugfs_root = debugfs_create_dir(dev_name(emif->dev), NULL);
@@ -150,7 +150,7 @@ static void __init_or_module emif_debugfs_init(struct emif_data *emif)
 	}
 }
 
-static void __exit emif_debugfs_exit(struct emif_data *emif)
+static void emif_debugfs_exit(struct emif_data *emif)
 {
 	if (IS_ENABLED(CONFIG_DEBUG_FS)) {
 		debugfs_remove_recursive(emif->debugfs_root);
@@ -671,7 +671,7 @@ static void disable_and_clear_all_interrupts(struct emif_data *emif)
 	clear_all_interrupts(emif);
 }
 
-static int __init_or_module setup_interrupts(struct emif_data *emif, u32 irq)
+static int setup_interrupts(struct emif_data *emif, u32 irq)
 {
 	u32		interrupts, type;
 	void __iomem	*base = emif->base;
@@ -702,7 +702,7 @@ static int __init_or_module setup_interrupts(struct emif_data *emif, u32 irq)
 
 }
 
-static void __init_or_module emif_onetime_settings(struct emif_data *emif)
+static void emif_onetime_settings(struct emif_data *emif)
 {
 	u32				pwr_mgmt_ctrl, zq, temp_alert_cfg;
 	void __iomem			*base = emif->base;
@@ -826,7 +826,7 @@ static int is_custom_config_valid(struct emif_custom_configs *cust_cfgs,
 	return valid;
 }
 
-static void __init_or_module of_get_custom_configs(struct device_node *np_emif,
+static void of_get_custom_configs(struct device_node *np_emif,
 		struct emif_data *emif)
 {
 	struct emif_custom_configs	*cust_cfgs = NULL;
@@ -875,7 +875,7 @@ static void __init_or_module of_get_custom_configs(struct device_node *np_emif,
 	emif->plat_data->custom_configs = cust_cfgs;
 }
 
-static void __init_or_module of_get_ddr_info(struct device_node *np_emif,
+static void of_get_ddr_info(struct device_node *np_emif,
 		struct device_node *np_ddr,
 		struct ddr_device_info *dev_info)
 {
@@ -909,7 +909,7 @@ static void __init_or_module of_get_ddr_info(struct device_node *np_emif,
 		dev_info->io_width = __fls(io_width) - 1;
 }
 
-static struct emif_data * __init_or_module of_get_memory_device_details(
+static struct emif_data *of_get_memory_device_details(
 		struct device_node *np_emif, struct device *dev)
 {
 	struct emif_data		*emif = NULL;
@@ -1086,7 +1086,7 @@ static struct emif_data *__init_or_module get_device_details(
 	return NULL;
 }
 
-static int __init_or_module emif_probe(struct platform_device *pdev)
+static int emif_probe(struct platform_device *pdev)
 {
 	struct emif_data	*emif;
 	int			irq, ret;
@@ -1141,7 +1141,7 @@ static int __init_or_module emif_probe(struct platform_device *pdev)
 	return -ENODEV;
 }
 
-static int __exit emif_remove(struct platform_device *pdev)
+static int emif_remove(struct platform_device *pdev)
 {
 	struct emif_data *emif = platform_get_drvdata(pdev);
 
@@ -1167,7 +1167,8 @@ MODULE_DEVICE_TABLE(of, emif_of_match);
 #endif
 
 static struct platform_driver emif_driver = {
-	.remove		= __exit_p(emif_remove),
+	.probe		= emif_probe,
+	.remove		= emif_remove,
 	.shutdown	= emif_shutdown,
 	.driver = {
 		.name = "emif",
@@ -1175,7 +1176,7 @@ static struct platform_driver emif_driver = {
 	},
 };
 
-module_platform_driver_probe(emif_driver, emif_probe);
+module_platform_driver(emif_driver);
 
 MODULE_DESCRIPTION("TI EMIF SDRAM Controller Driver");
 MODULE_LICENSE("GPL");
-- 
2.42.0

Re: [PATCH 3/3] memory: emif: Drop usage of platform_driver_probe()

[Krzysztof Kozlowski]

On 17/12/2023 20:31, Uwe Kleine-König wrote:
> There are considerations to drop platform_driver_probe() as a concept
> that isn't relevant any more today. It comes with an added complexity
> that makes many users hold it wrong. (E.g. this driver should have
> better used __init instead of __init_or_module to mark functions only
> relevant to .probe() and mark the driver struct with __refdata.)
> 
> This fixes a W=1 build warning:
> 
> 	WARNING: modpost: drivers/memory/emif: section mismatch in reference: emif_driver+0x4 (section: .data) -> emif_remove (section: .exit.text)
> 
> Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> ---
>  drivers/memory/emif.c | 23 ++++++++++++-----------

I applied #1 and #2. This does not apply. Please rebase and resend.

Best regards,
Krzysztof

Re: [PATCH 0/3] memory: emif: ifdefs and platform_driver_probe()

[Krzysztof Kozlowski]

On 17/12/2023 20:31, Uwe Kleine-König wrote:
> Hello,
> 
> while preparing the series to convert the platform drivers below
> drivers/memory to use .remove_new()[1], I noticed that the emif driver
> unnecessarily uses some #ifdefs and doesn't use platform_driver_probe()
> correctly. Note there is a conflict between these series. Tell me if you
> need support to resolve this.
> 

I was waiting here for some Reviews or Tested-by. Time passed, reviews
did not happen but it is too late for me to take it for the next merge
window. I will take it after the merge window.


Best regards,
Krzysztof

Re: [PATCH 0/3] memory: emif: ifdefs and platform_driver_probe()

[Uwe Kleine-König]

[-- Attachment #1: Type: text/plain, Size: 1046 bytes --]

Hello Krzysztof,

On Thu, Dec 21, 2023 at 09:46:29PM +0100, Krzysztof Kozlowski wrote:
> On 17/12/2023 20:31, Uwe Kleine-König wrote:
> > while preparing the series to convert the platform drivers below
> > drivers/memory to use .remove_new()[1], I noticed that the emif driver
> > unnecessarily uses some #ifdefs and doesn't use platform_driver_probe()
> > correctly. Note there is a conflict between these series. Tell me if you
> > need support to resolve this.
> 
> I was waiting here for some Reviews or Tested-by. Time passed, reviews
> did not happen but it is too late for me to take it for the next merge
> window. I will take it after the merge window.

Thanks for the status update. That's very appreciated. (And taking it
for the next development cycle is entirely fine, that's just a drive-by
series, nothing on my side depends on it.)

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | https://www.pengutronix.de/ |

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]