reset: make non-modular drivers really non modular

PATCH In review Paul Gortmaker 4 participants 11 messages Jun 13, 2016 – Oct 24, 2016

Subsystems (4)

Arm/Zynq architecture Arm/Allwinner sunXi SoC support Reset controller framework Arm/LPC18XX architecture

lore

Share & tools

lore

Paul Gortmaker

author

2016-06-13 18:04 UTC #0 lore raw

cover reset: make non-modular drivers really non modular

For anyone new to the underlying goal of this cleanup, we are trying to not use module support for code that can never be built as a module since:

(1) it is easy to accidentally write unused module_exit and remove code
(2) it can be misleading when reading the source, thinking it can be modular when the Makefile and/or Kconfig prohibit it
(3) it requires the include of the module.h header file which in turn includes nearly everything else, thus adding to CPP overhead.
(4) it gets copied/replicated into other code and spreads like weeds.

Fortunately the reset dir is in pretty good shape, and there are only five instances of non-modular code using modular funcionality/macros.

Changes seen here cover the following categories:

-just replacement of modular macros with their non-modular
equivalents that CPP would have inserted anyway

-the removal of including module.h ; replaced with init.h
as required based on whether the file already had it.

-the removal of any ".remove" functions that were hooked into
the driver struct. This ".remove" function would of
course not be called from the __exit function since that was never run. However in theory, someone could have triggered it via sysfs unbind, even though there isn't a sensible use case for doing so. So to cover that possibility, we've also disabled sysfs unbind in the driver.

Recently several of these drivers got removal of the ".remove" fcns via other changes, so only one remains for which that happens here.

There are no initcall level changes here; everything was at the level of device_initcall and remains so, by using the builtin equivalents.

Build tested for several different key arch on a recent linux-next tree to ensure no silly typos crept in.

If there is a desire for any of these to be modular, we can definitely consider that, but by default the changes here keep the code consistent with existing behaviour and do not expand functionality into the modular realm that I can't run time test.

Paul.

---

Cc Antoine Tenart <[email protected]>

Cc Joachim Eastwood <[email protected]>

Cc Masahiro Yamada <[email protected]>

Cc Maxime Ripard <[email protected]>

Cc Michal Simek <[email protected]>

Cc Moritz Fischer <[email protected]>

Cc Philipp Zabel <[email protected]>

Cc Sebastian Hesselbarth <[email protected]>

Cc "Sören Brinkmann" <[email protected]>

Cc Steffen Trumtrar <[email protected]>

Paul Gortmaker (5):
reset: berlin: make it explicitly non-modular
reset: socfpga: make it explicitly non-modular
reset: sunxi: make it explicitly non-modular
reset: zynq: make it explicitly non-modular
reset: lpc18xx: make it explicitly non-modular

 drivers/reset/reset-berlin.c  | 12 ++++--------
 drivers/reset/reset-lpc18xx.c | 32 +++++---------------------------
 drivers/reset/reset-socfpga.c | 10 ++++------
 drivers/reset/reset-sunxi.c   |  9 ++-------
 drivers/reset/reset-zynq.c    | 10 ++++------
 5 files changed, 19 insertions(+), 54 deletions(-)

Signature · 1 line

2.8.4

Paul Gortmaker

author

2016-06-13 18:05 UTC #1 lore raw

PATCH 1/5 reset: berlin: make it explicitly non-modular

The Kconfig currently controlling compilation of this code is:

arch/arm/mach:berlin/Kconfig:menuconfig ARCH_BERLIN
arch/arm/mach:berlin/Kconfig: bool "Marvell Berlin SoCs" if ARCH_MULTI_V7

arch/arm64/Kconfig.platforms:config ARCH_BERLIN
arch/arm64/Kconfig.platforms: bool "Marvell Berlin SoC Family"

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that when reading the driver there is no doubt it is builtin-only.

Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information was (or is now) contained at the top of the file in the comments.

Cc Philipp Zabel <[email protected]>

Cc Antoine Tenart <[email protected]>

Cc Sebastian Hesselbarth <[email protected]>

Signed-off-by Paul Gortmaker <[email protected]>

---
 drivers/reset/reset-berlin.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)

drivers/reset/reset-berlin.c +4 -8

index 369f3917fd8e..371197bbd055 100644
@@ -1,6 +1,8 @@
 /*
  * Copyright (C) 2014 Marvell Technology Group Ltd.
  *
+ * Marvell Berlin reset driver
+ *
  * Antoine Tenart <[email protected]>
  * Sebastian Hesselbarth <[email protected]>
  *

@@ -12,7 +14,7 @@
 #include <linux/delay.h>
 #include <linux/io.h>
 #include <linux/mfd/syscon.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
 #include <linux/platform_device.h>

@@ -91,7 +93,6 @@ static const struct of_device_id berlin_reset_dt_match[] = {
 	{ .compatible = "marvell,berlin2-reset" },
 	{ },
 };
-MODULE_DEVICE_TABLE(of, berlin_reset_dt_match);
 
 static struct platform_driver berlin_reset_driver = {
 	.probe	= berlin2_reset_probe,

@@ -100,9 +101,4 @@ static struct platform_driver berlin_reset_driver = {
 		.of_match_table = berlin_reset_dt_match,
 	},
 };
-module_platform_driver(berlin_reset_driver);
-
-MODULE_AUTHOR("Antoine Tenart <[email protected]>");
-MODULE_AUTHOR("Sebastian Hesselbarth <[email protected]>");
-MODULE_DESCRIPTION("Marvell Berlin reset driver");
-MODULE_LICENSE("GPL");
+builtin_platform_driver(berlin_reset_driver);

Signature · 1 line

2.8.4

Paul Gortmaker

author

2016-06-13 18:04 UTC #2 lore raw

PATCH 2/5 reset: socfpga: make it explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

arch/arm/mach:socfpga/Kconfig:menuconfig ARCH_SOCFPGA arch/arm/mach:socfpga/Kconfig: bool "Altera SOCFPGA family" if ARCH_MULTI_V7

...meaning that it currently is not being built as a module by anyone.

Lets remove the small amount of modular evidence that remains, so that when reading the driver there is no doubt it is builtin-only.

Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

We also delete the MODULE_LICENSE tag etc. since all that information was (or is now) contained at the top of the file in the comments.

Cc Steffen Trumtrar <[email protected]>

Cc Masahiro Yamada <[email protected]>

Cc Philipp Zabel <[email protected]>

Signed-off-by Paul Gortmaker <[email protected]>

---
 drivers/reset/reset-socfpga.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

drivers/reset/reset-socfpga.c +4 -6

index 12add9b0fa49..153a5a02abf8 100644
@@ -1,4 +1,6 @@
 /*
+ * Socfpga Reset Controller Driver
+ *
  * Copyright 2014 Steffen Trumtrar <[email protected]>
  *
  * based on

@@ -16,7 +18,7 @@
 
 #include <linux/err.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/reset-controller.h>

@@ -149,8 +151,4 @@ static struct platform_driver socfpga_reset_driver = {
 		.of_match_table	= socfpga_reset_dt_ids,
 	},
 };
-module_platform_driver(socfpga_reset_driver);
-
-MODULE_AUTHOR("Steffen Trumtrar <[email protected]");
-MODULE_DESCRIPTION("Socfpga Reset Controller Driver");
-MODULE_LICENSE("GPL");
+builtin_platform_driver(socfpga_reset_driver);

Signature · 1 line

2.8.4

Paul Gortmaker

author

2016-06-13 18:04 UTC #3 lore raw

PATCH 3/5 reset: sunxi: make it explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

arch/arm/mach:sunxi/Kconfig:menuconfig ARCH_SUNXI
arch/arm/mach:sunxi/Kconfig: bool "Allwinner SoCs" if ARCH_MULTI_V7

...meaning that it currently is not being built as a module by anyone.

Lets remove the few remaining traces of modular macro usage, so that when reading the driver there is no doubt it is builtin-only.

Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information is already contained at the top of the file in the comments.

Cc Philipp Zabel <[email protected]>

Cc Maxime Ripard <[email protected]>

Signed-off-by Paul Gortmaker <[email protected]>

---
 drivers/reset/reset-sunxi.c | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

drivers/reset/reset-sunxi.c +2 -7

index 3080190b3f90..b44f6b5f87b6 100644
@@ -13,7 +13,7 @@
 
 #include <linux/err.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/of_address.h>
 #include <linux/platform_device.h>

@@ -142,7 +142,6 @@ static const struct of_device_id sunxi_reset_dt_ids[] = {
 	 { .compatible = "allwinner,sun6i-a31-clock-reset", },
 	 { /* sentinel */ },
 };
-MODULE_DEVICE_TABLE(of, sunxi_reset_dt_ids);
 
 static int sunxi_reset_probe(struct platform_device *pdev)
 {

@@ -175,8 +174,4 @@ static struct platform_driver sunxi_reset_driver = {
 		.of_match_table	= sunxi_reset_dt_ids,
 	},
 };
-module_platform_driver(sunxi_reset_driver);
-
-MODULE_AUTHOR("Maxime Ripard <[email protected]");
-MODULE_DESCRIPTION("Allwinner SoCs Reset Controller Driver");
-MODULE_LICENSE("GPL");
+builtin_platform_driver(sunxi_reset_driver);

Signature · 1 line

2.8.4

Paul Gortmaker

author

2016-06-13 18:04 UTC #4 lore raw

PATCH 4/5 reset: zynq: make it explicitly non-modular

In reply to Paul Gortmaker

The Makefile/Kconfig currently controlling compilation of this code is:

drivers/reset/Makefile:obj-$(CONFIG_ARCH_ZYNQ) += reset-zynq.o arch/arm/mach:zynq/Kconfig:config ARCH_ZYNQ
arch/arm/mach:zynq/Kconfig: bool "Xilinx Zynq ARM Cortex A9 Platform" if ARCH_MULTI_V7

...meaning that it currently is not being built as a module by anyone.

Lets remove the few remaining traces of modular macro usage, so that when reading the driver there is no doubt it is builtin-only.

Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

We also delete the MODULE_LICENSE tag etc. since all that information was (or is now) contained at the top of the file in the comments.

Cc Philipp Zabel <[email protected]>

Cc Michal Simek <[email protected]>

Cc "Sören Brinkmann" <[email protected]>

Cc Moritz Fischer <[email protected]>

Signed-off-by Paul Gortmaker <[email protected]>

---
 drivers/reset/reset-zynq.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

drivers/reset/reset-zynq.c +4 -6

index 138f2f205662..87a4e355578f 100644
@@ -3,6 +3,8 @@
  *
  * Xilinx Zynq Reset controller driver
  *
+ * Author: Moritz Fischer <[email protected]>
+ *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
  * the Free Software Foundation; version 2 of the License.

@@ -15,7 +17,7 @@
 
 #include <linux/err.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/mfd/syscon.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>

@@ -137,8 +139,4 @@ static struct platform_driver zynq_reset_driver = {
 		.of_match_table	= zynq_reset_dt_ids,
 	},
 };
-module_platform_driver(zynq_reset_driver);
-
-MODULE_LICENSE("GPL v2");
-MODULE_AUTHOR("Moritz Fischer <[email protected]>");
-MODULE_DESCRIPTION("Zynq Reset Controller Driver");
+builtin_platform_driver(zynq_reset_driver);

Signature · 1 line

2.8.4

Moritz Fischer

Acked-by on v1 4/5

2016-06-13 19:15 UTC #5 lore raw

Re: [PATCH 4/5] reset: zynq: make it explicitly non-modular

Hi Paul,

Quoted 15 lines · Paul Gortmaker

The Makefile/Kconfig currently controlling compilation of this code is:
drivers/reset/Makefile:obj-$(CONFIG_ARCH_ZYNQ) += reset-zynq.o arch/arm/mach:zynq/Kconfig:config ARCH_ZYNQ
arch/arm/mach:zynq/Kconfig: bool "Xilinx Zynq ARM Cortex A9 Platform" if ARCH_MULTI_V7
...meaning that it currently is not being built as a module by anyone.
Lets remove the few remaining traces of modular macro usage, so that when reading the driver there is no doubt it is builtin-only.
Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

Quoted 8 lines

We also delete the MODULE_LICENSE tag etc. since all that information was (or is now) contained at the top of the file in the comments.
Cc Philipp Zabel <[email protected]>
Cc Michal Simek <[email protected]>
Cc "Sören Brinkmann" <[email protected]>
Cc Moritz Fischer <[email protected]>
Signed-off-by Paul Gortmaker <[email protected]>

Acked-by Moritz Fischer <[email protected]>

Thanks,

Moritz

Michal Simek

Acked-by on v1 4/5

2016-06-14 06:12 UTC #6 lore raw

Re: [PATCH 4/5] reset: zynq: make it explicitly non-modular

In reply to Paul Gortmaker

Quoted 51 lines · Paul Gortmaker

The Makefile/Kconfig currently controlling compilation of this code is:
drivers/reset/Makefile:obj-$(CONFIG_ARCH_ZYNQ) += reset-zynq.o arch/arm/mach:zynq/Kconfig:config ARCH_ZYNQ
arch/arm/mach:zynq/Kconfig: bool "Xilinx Zynq ARM Cortex A9 Platform" if ARCH_MULTI_V7
...meaning that it currently is not being built as a module by anyone.
Lets remove the few remaining traces of modular macro usage, so that when reading the driver there is no doubt it is builtin-only.
Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.
We also delete the MODULE_LICENSE tag etc. since all that information was (or is now) contained at the top of the file in the comments.
Cc Philipp Zabel <[email protected]>
Cc Michal Simek <[email protected]>
Cc "Sören Brinkmann" <[email protected]>
Cc Moritz Fischer <[email protected]>
Signed-off-by Paul Gortmaker <[email protected]>
---
 drivers/reset/reset-zynq.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)
drivers/reset/reset-zynq.c +3 -1
index 138f2f205662..87a4e355578f 100644
@@ -3,6 +3,8 @@
  *
  * Xilinx Zynq Reset controller driver
  *
+ * Author: Moritz Fischer <[email protected]>
+ *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License as published by
  * the Free Software Foundation; version 2 of the License.
@@ -15,7 +17,7 @@
 
 #include <linux/err.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/mfd/syscon.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>

drivers/reset/reset-zynq.c static struct platform_driver zynq_reset_driver = { · line 142

Quoted 9 lines

drivers/reset/reset-zynq.c +1 -5

@@ -137,8 +139,4 @@ static struct platform_driver zynq_reset_driver = {
 		.of_match_table	= zynq_reset_dt_ids,
 	},
 };
-module_platform_driver(zynq_reset_driver);
-
-MODULE_LICENSE("GPL v2");
-MODULE_AUTHOR("Moritz Fischer <[email protected]>");
-MODULE_DESCRIPTION("Zynq Reset Controller Driver");
+builtin_platform_driver(zynq_reset_driver);

Acked-by Michal Simek <[email protected]>

Thanks,
Michal

Paul Gortmaker

author

2016-06-13 18:04 UTC #7 lore raw

PATCH 5/5 reset: lpc18xx: make it explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

arch/arm/Kconfig:config ARCH_LPC18XX
arch/arm/Kconfig: bool "NXP LPC18xx/LPC43xx"

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that when reading the driver there is no doubt it is builtin-only.

We explicitly disallow a driver unbind, since that doesn't have a sensible use case anyway, and it allows us to drop the ".remove" code for non-modular drivers.

Since module_platform_driver() uses the same init level priority as builtin_platform_driver() the init ordering remains unchanged with this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We also delete the MODULE_LICENSE tag etc. since all that information is already contained at the top of the file in the comments.

Cc Philipp Zabel <[email protected]>

Cc Joachim Eastwood <[email protected]>

Signed-off-by Paul Gortmaker <[email protected]>

---
 drivers/reset/reset-lpc18xx.c | 32 +++++---------------------------
 1 file changed, 5 insertions(+), 27 deletions(-)

drivers/reset/reset-lpc18xx.c +5 ~~-27~~

index 54cca0055171..a62ad52e262b 100644
@@ -13,7 +13,7 @@
 #include <linux/delay.h>
 #include <linux/err.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
 #include <linux/reboot.h>

@@ -218,39 +218,17 @@ dis_clk_reg:
 	return ret;
 }
 
-static int lpc18xx_rgu_remove(struct platform_device *pdev)
-{
-	struct lpc18xx_rgu_data *rc = platform_get_drvdata(pdev);
-	int ret;
-
-	ret = unregister_restart_handler(&rc->restart_nb);
-	if (ret)
-		dev_warn(&pdev->dev, "failed to unregister restart handler\n");
-
-	reset_controller_unregister(&rc->rcdev);
-
-	clk_disable_unprepare(rc->clk_delay);
-	clk_disable_unprepare(rc->clk_reg);
-
-	return 0;
-}
-
 static const struct of_device_id lpc18xx_rgu_match[] = {
 	{ .compatible = "nxp,lpc1850-rgu" },
 	{ }
 };
-MODULE_DEVICE_TABLE(of, lpc18xx_rgu_match);
 
 static struct platform_driver lpc18xx_rgu_driver = {
 	.probe	= lpc18xx_rgu_probe,
-	.remove	= lpc18xx_rgu_remove,
 	.driver	= {
-		.name		= "lpc18xx-reset",
-		.of_match_table	= lpc18xx_rgu_match,
+		.name			= "lpc18xx-reset",
+		.of_match_table		= lpc18xx_rgu_match,
+		.suppress_bind_attrs	= true,
 	},
 };
-module_platform_driver(lpc18xx_rgu_driver);
-
-MODULE_AUTHOR("Joachim Eastwood <[email protected]>");
-MODULE_DESCRIPTION("Reset driver for LPC18xx/43xx RGU");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(lpc18xx_rgu_driver);

Signature · 1 line

2.8.4

Paul Gortmaker

author

2016-09-05 22:25 UTC #8 lore raw

In reply to Paul Gortmaker

Quoted 2 lines · [[PATCH 0/5] reset: make non-modular drivers really non modular] On 13/06/2016…

For anyone new to the underlying goal of this cleanup, we are trying to not use module support for code that can never be built as a module since:

Hi Phillip,

Wondering what the status of these patches is, since I didn't see them in this recent merge in linux-next now that it has restarted again:

----------------------
commit 7d3ef43f1f99d4250ff120c8965b3e5cd81b3c71
Merge: 2b97f10b3896 cddb4800162e
Author: Arnd Bergmann <[email protected]>
Date: Fri Sep 2 18:29:04 2016 +0200

Merge tag 'reset-for-4.9' of git://git.pengutronix.de/git/pza/linux into next/drivers

Merge "Reset controller changes for v4.9" from Philipp Zabel:
----------------------

I believe there are at least two new instances in the reset drivers since these original five were sent to you back in June.

The V1 didn't get any change requests, just some acks. Do you want a resend with the additional commits, and the Ack'd by from V1 added?

Thanks,
Paul.
--

Quoted 74 lines

(1) it is easy to accidentally write unused module_exit and remove code
(2) it can be misleading when reading the source, thinking it can be modular when the Makefile and/or Kconfig prohibit it
(3) it requires the include of the module.h header file which in turn includes nearly everything else, thus adding to CPP overhead.
(4) it gets copied/replicated into other code and spreads like weeds.
Fortunately the reset dir is in pretty good shape, and there are only five instances of non-modular code using modular funcionality/macros.
Changes seen here cover the following categories:
-just replacement of modular macros with their non-modular
equivalents that CPP would have inserted anyway
-the removal of including module.h ; replaced with init.h
as required based on whether the file already had it.
-the removal of any ".remove" functions that were hooked into
the driver struct. This ".remove" function would of
course not be called from the __exit function since that was never run. However in theory, someone could have triggered it via sysfs unbind, even though there isn't a sensible use case for doing so. So to cover that possibility, we've also disabled sysfs unbind in the driver.
Recently several of these drivers got removal of the ".remove" fcns via other changes, so only one remains for which that happens here.
There are no initcall level changes here; everything was at the level of device_initcall and remains so, by using the builtin equivalents.
Build tested for several different key arch on a recent linux-next tree to ensure no silly typos crept in.
If there is a desire for any of these to be modular, we can definitely consider that, but by default the changes here keep the code consistent with existing behaviour and do not expand functionality into the modular realm that I can't run time test.
Paul.
---
Cc Antoine Tenart <[email protected]>
Cc Joachim Eastwood <[email protected]>
Cc Masahiro Yamada <[email protected]>
Cc Maxime Ripard <[email protected]>
Cc Michal Simek <[email protected]>
Cc Moritz Fischer <[email protected]>
Cc Philipp Zabel <[email protected]>
Cc Sebastian Hesselbarth <[email protected]>
Cc "Sören Brinkmann" <[email protected]>
Cc Steffen Trumtrar <[email protected]>
Paul Gortmaker (5):
reset: berlin: make it explicitly non-modular
reset: socfpga: make it explicitly non-modular
reset: sunxi: make it explicitly non-modular
reset: zynq: make it explicitly non-modular
reset: lpc18xx: make it explicitly non-modular
 drivers/reset/reset-berlin.c  | 12 ++++--------
 drivers/reset/reset-lpc18xx.c | 32 +++++---------------------------
 drivers/reset/reset-socfpga.c | 10 ++++------
 drivers/reset/reset-sunxi.c   |  9 ++-------
 drivers/reset/reset-zynq.c    | 10 ++++------
 5 files changed, 19 insertions(+), 54 deletions(-)
Signature · 1 line
2.8.4

Paul Gortmaker

author

2016-10-22 21:49 UTC #9 lore raw

In reply to Paul Gortmaker

Quoted 2 lines · [[PATCH 0/5] reset: make non-modular drivers really non modular] On 13/06/2016…

For anyone new to the underlying goal of this cleanup, we are trying to not use module support for code that can never be built as a module since:

Hi Philipp,

I sent this several months ago, and aside from a couple Ack'd by, there was no negative feedback or similar.

Since then there have been two more drivers added with the same issue.

reset-make-ath79-explicitly-non-modular.patch reset-pistachio-make-it-explicitly-non-modular.patch

Should I assume that the original send has fallen through the cracks and resend with the two new patches, or do you still have the original five in a queue somewhere?

I'd like to merge as many of these as I can via their respective maintainers vs. trying to send them to Linus directly.

Thanks,
Paul.
--

Quoted 74 lines

(1) it is easy to accidentally write unused module_exit and remove code
(2) it can be misleading when reading the source, thinking it can be modular when the Makefile and/or Kconfig prohibit it
(3) it requires the include of the module.h header file which in turn includes nearly everything else, thus adding to CPP overhead.
(4) it gets copied/replicated into other code and spreads like weeds.
Fortunately the reset dir is in pretty good shape, and there are only five instances of non-modular code using modular funcionality/macros.
Changes seen here cover the following categories:
-just replacement of modular macros with their non-modular
equivalents that CPP would have inserted anyway
-the removal of including module.h ; replaced with init.h
as required based on whether the file already had it.
-the removal of any ".remove" functions that were hooked into
the driver struct. This ".remove" function would of
course not be called from the __exit function since that was never run. However in theory, someone could have triggered it via sysfs unbind, even though there isn't a sensible use case for doing so. So to cover that possibility, we've also disabled sysfs unbind in the driver.
Recently several of these drivers got removal of the ".remove" fcns via other changes, so only one remains for which that happens here.
There are no initcall level changes here; everything was at the level of device_initcall and remains so, by using the builtin equivalents.
Build tested for several different key arch on a recent linux-next tree to ensure no silly typos crept in.
If there is a desire for any of these to be modular, we can definitely consider that, but by default the changes here keep the code consistent with existing behaviour and do not expand functionality into the modular realm that I can't run time test.
Paul.
---
Cc Antoine Tenart <[email protected]>
Cc Joachim Eastwood <[email protected]>
Cc Masahiro Yamada <[email protected]>
Cc Maxime Ripard <[email protected]>
Cc Michal Simek <[email protected]>
Cc Moritz Fischer <[email protected]>
Cc Philipp Zabel <[email protected]>
Cc Sebastian Hesselbarth <[email protected]>
Cc "Sören Brinkmann" <[email protected]>
Cc Steffen Trumtrar <[email protected]>
Paul Gortmaker (5):
reset: berlin: make it explicitly non-modular
reset: socfpga: make it explicitly non-modular
reset: sunxi: make it explicitly non-modular
reset: zynq: make it explicitly non-modular
reset: lpc18xx: make it explicitly non-modular
 drivers/reset/reset-berlin.c  | 12 ++++--------
 drivers/reset/reset-lpc18xx.c | 32 +++++---------------------------
 drivers/reset/reset-socfpga.c | 10 ++++------
 drivers/reset/reset-sunxi.c   |  9 ++-------
 drivers/reset/reset-zynq.c    | 10 ++++------
 5 files changed, 19 insertions(+), 54 deletions(-)
Signature · 1 line
2.8.4

Philipp Zabel

maintainer

2016-10-24 08:56 UTC #10 lore raw

Hi Paul,

Am Samstag, den 22.10.2016, 17:48 -0400 schrieb Paul Gortmaker:

Quoted 10 lines

Quoted 2 lines · [[PATCH 0/5] reset: make non-modular drivers really non modular] On 13/06/2016…
For anyone new to the underlying goal of this cleanup, we are trying to not use module support for code that can never be built as a module since:
Hi Philipp,
I sent this several months ago, and aside from a couple Ack'd by, there was no negative feedback or similar.

Quoted 8 lines

Since then there have been two more drivers added with the same issue.
reset-make-ath79-explicitly-non-modular.patch reset-pistachio-make-it-explicitly-non-modular.patch
Should I assume that the original send has fallen through the cracks and resend with the two new patches, or do you still have the original five in a queue somewhere?

thanks for the reminder. I had them stashed away because I wanted to see whether the introduction of per-driver Kconfig options would change some of them to become modular. That wasn't the case, so I have pushed your patches to the reset/next branch now, with fixed up commit messages.

Quoted 5 lines

I'd like to merge as many of these as I can via their respective maintainers vs. trying to send them to Linus directly.
Thanks,
Paul.

regards
Philipp