Skip to contentSkip to conversation

drivers/pci: avoid module_init in non-modular host/pci*

PATCH In review Paul Gortmaker 11 participants 33 messages Dec 13, 2015 – Jan 8, 2016

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

cover drivers/pci: avoid module_init in non-modular host/pci*

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

I have more "avoid module usage in non-modular code" cleanups for the PCI subsystem, but these all have a common theme and it makes for a more maintainer friendly series size to just ask to digest these 1st.

Testing was done on linux-next, using an ARCH=arm allmodconfig and then explicitly building the files changed in this series. If desired, I can provide a v4.4-rc4 based branch for merging vs e-mail processing, since I don't think the underlying baseline is overly important for this (largely trivial) series of patches.

Paul.
---

Cc Alexandre Courbot <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Jason Cooper <[email protected]>
Cc Kishon Vijay Abraham I <[email protected]>
Cc Ley Foon Tan <[email protected]>
Cc Lucas Stach <[email protected]>
Cc Michal Simek <[email protected]>
Cc Murali Karicheri <[email protected]>
Cc Pratyush Anand <[email protected]>
Cc Richard Zhu <[email protected]>
Cc Simon Horman <[email protected]>
Cc "S?ren Brinkmann" <[email protected]>
Cc Stephen Warren <[email protected]>
Cc Thierry Reding <[email protected]>
Cc Thomas Petazzoni <[email protected]>
Cc linux-arm-kernel at lists.infradead.org
Cc linux-omap at vger.kernel.org
Cc linux-pci at vger.kernel.org
Cc linux-sh at vger.kernel.org
Cc linux-tegra at vger.kernel.org
Cc rfi at lists.rocketboards.org

Paul Gortmaker (10):
drivers/pci: make host/pci-imx6.c driver explicitly non-modular drivers/pci: make host/pcie-spear13xx.c driver explicitly non-modular drivers/pci: make host/pci-mvebu.c explicitly non-modular drivers/pci: make host/pci-dra7xx.c explicitly non-modular drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular drivers/pci: make host/pci-tegra.c explicitly non-modular drivers/pci: make host/pcie-rcar.c explicitly non-modular drivers/pci: make host/pcie-xilinx.c explicitly non-modular drivers/pci: make host/pci-keystone.c explicitly non-modular drivers/pci: make host/pcie-altera.c explicitly non-modular

 drivers/pci/host/pci-dra7xx.c     | 31 +++--------------------
 drivers/pci/host/pci-imx6.c       | 12 +++------
 drivers/pci/host/pci-keystone.c   | 21 +++-------------
 drivers/pci/host/pci-mvebu.c      | 11 +++-----
 drivers/pci/host/pci-rcar-gen2.c  | 12 +++------
 drivers/pci/host/pci-tegra.c      | 11 +++-----
 drivers/pci/host/pcie-altera.c    | 12 ++++-----
 drivers/pci/host/pcie-rcar.c      | 11 +++-----
 drivers/pci/host/pcie-spear13xx.c | 10 +++-----
 drivers/pci/host/pcie-xilinx.c    | 53 ++-------------------------------------
 10 files changed, 35 insertions(+), 149 deletions(-)
Signature · 1 line
2.6.1
#1 lore raw

PATCH 1/10 drivers/pci: make host/pci-imx6.c driver explicitly non-modular

The Kconfig for this option is currently:

config PCI_IMX6
bool "Freescale i.MX6 PCIe controller"

...meaning that it currently is not being built as a module by anyone. Lets remove the couple traces of modularity, so that when reading the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular case, the init ordering remains unchanged with this commit. An alternate init level might be worth considering at a later date.

We explicitly disallow a driver unbind, since that doesn't have a sensible use case anyway, and there wasn't a supplied ".remove" function to be called either.

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

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

Cc Richard Zhu <[email protected]>
Cc Lucas Stach <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc linux-pci at vger.kernel.org
Cc linux-arm-kernel at lists.infradead.org
---
 drivers/pci/host/pci-imx6.c | 12 ++++--------
 1 file changed, 4 insertions(+), 8 deletions(-)
drivers/pci/host/pci-imx6.c +4 -8
index 22e8224126fd..ddd6a752b554 100644
@@ -17,7 +17,7 @@
 #include <linux/kernel.h>
 #include <linux/mfd/syscon.h>
 #include <linux/mfd/syscon/imx6q-iomuxc-gpr.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of_gpio.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
@@ -641,24 +641,20 @@ static const struct of_device_id imx6_pcie_of_match[] = {
 	{ .compatible = "fsl,imx6q-pcie", },
 	{},
 };
-MODULE_DEVICE_TABLE(of, imx6_pcie_of_match);
 
 static struct platform_driver imx6_pcie_driver = {
 	.driver = {
 		.name	= "imx6q-pcie",
 		.of_match_table = imx6_pcie_of_match,
+		.suppress_bind_attrs = true,
 	},
 	.shutdown = imx6_pcie_shutdown,
 };
 
-/* Freescale PCIe driver does not allow module unload */
+/* Freescale PCIe driver does not support module configuration */
 
 static int __init imx6_pcie_init(void)
 {
 	return platform_driver_probe(&imx6_pcie_driver, imx6_pcie_probe);
 }
-module_init(imx6_pcie_init);
-
-MODULE_AUTHOR("Sean Cross <[email protected]>");
-MODULE_DESCRIPTION("Freescale i.MX6 PCIe host controller driver");
-MODULE_LICENSE("GPL v2");
+device_initcall(imx6_pcie_init);
Signature · 1 line
2.6.1
#2 lore raw

Re: [PATCH 01/10] drivers/pci: make host/pci-imx6.c driver explicitly non-modular

Quoted 8 lines · On Saturday 12 December 2015 20:41:48 Paul Gortmaker wrote:

The Kconfig for this option is currently:

config PCI_IMX6
bool "Freescale i.MX6 PCIe controller"

...meaning that it currently is not being built as a module by anyone. Lets remove the couple traces of modularity, so that when reading the driver there is no doubt it is builtin-only.

Can you fix Kconfig instead?

Arnd

#3 lore raw

PATCH 2/10 drivers/pci: make host/pcie-spear13xx.c driver explicitly non-modular

In reply to Paul Gortmaker

The Kconfig for this option is currently:

config PCIE_SPEAR13XX
bool "STMicroelectronics SPEAr PCIe controller"

...meaning that it currently is not being built as a module by anyone. Lets remove the couple traces of modularity, so that when reading the driver there is no doubt it is builtin-only.

Since module_init translates to device_initcall in the non-modular case, the init ordering remains unchanged with this commit. An alternate init level might be worth considering at a later date.

We explicitly disallow a driver unbind, since that doesn't have a sensible use case anyway, and this driver had not populated ".remove" with any function to be called either.

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

---
 drivers/pci/host/pcie-spear13xx.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-spear13xx.c +3 -7
index b95b7563c052..8248faa45f6f 100644
@@ -16,7 +16,7 @@
 #include <linux/delay.h>
 #include <linux/interrupt.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/pci.h>
 #include <linux/phy/phy.h>
@@ -366,13 +366,13 @@ static const struct of_device_id spear13xx_pcie_of_match[] = {
 	{ .compatible = "st,spear1340-pcie", },
 	{},
 };
-MODULE_DEVICE_TABLE(of, spear13xx_pcie_of_match);
 
 static struct platform_driver spear13xx_pcie_driver = {
 	.probe		= spear13xx_pcie_probe,
 	.driver = {
 		.name	= "spear-pcie",
 		.of_match_table = of_match_ptr(spear13xx_pcie_of_match),
+		.suppress_bind_attrs = true,
 	},
 };
 
@@ -382,8 +382,4 @@ static int __init spear13xx_pcie_init(void)
 {
 	return platform_driver_register(&spear13xx_pcie_driver);
 }
-module_init(spear13xx_pcie_init);
-
-MODULE_DESCRIPTION("ST Microelectronics SPEAr13xx PCIe host controller driver");
-MODULE_AUTHOR("Pratyush Anand <[email protected]>");
-MODULE_LICENSE("GPL v2");
+device_initcall(spear13xx_pcie_init);
Signature · 1 line
2.6.1
#4 lore raw

PATCH 3/10 drivers/pci: make host/pci-mvebu.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_MVEBU
drivers/pci/host/Kconfig: bool "Marvell EBU PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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 Thomas Petazzoni <[email protected]>
Cc Jason Cooper <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc linux-pci at vger.kernel.org
Cc linux-arm-kernel at lists.infradead.org
---
 drivers/pci/host/pci-mvebu.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pci-mvebu.c +4 -7
index 53b79c5f0559..17222f1a3ef3 100644
@@ -1,6 +1,8 @@
 /*
  * PCIe driver for Marvell Armada 370 and Armada XP SoCs
  *
+ * Module Author: Thomas Petazzoni <[email protected]>
+ *
  * This file is licensed under the terms of the GNU General Public
  * License version 2.  This program is licensed "as is" without any
  * warranty of any kind, whether express or implied.
@@ -11,7 +13,7 @@
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/gpio.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/mbus.h>
 #include <linux/msi.h>
 #include <linux/slab.h>
@@ -1296,7 +1298,6 @@ static const struct of_device_id mvebu_pcie_of_match_table[] = {
 	{ .compatible = "marvell,kirkwood-pcie", },
 	{},
 };
-MODULE_DEVICE_TABLE(of, mvebu_pcie_of_match_table);
 
 static struct dev_pm_ops mvebu_pcie_pm_ops = {
 	.suspend_noirq = mvebu_pcie_suspend,
@@ -1313,8 +1314,4 @@ static struct platform_driver mvebu_pcie_driver = {
 	},
 	.probe = mvebu_pcie_probe,
 };
-module_platform_driver(mvebu_pcie_driver);
-
-MODULE_AUTHOR("Thomas Petazzoni <[email protected]>");
-MODULE_DESCRIPTION("Marvell EBU PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(mvebu_pcie_driver);
Signature · 1 line
2.6.1
Reviewed-by on v1 3/10
#5 lore raw

Re: [PATCH 03/10] drivers/pci: make host/pci-mvebu.c explicitly non-modular

Paul,

Quoted 28 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_MVEBU
drivers/pci/host/Kconfig: bool "Marvell EBU PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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 Thomas Petazzoni <[email protected]>
Cc Jason Cooper <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc linux-pci at vger.kernel.org
Cc linux-arm-kernel at lists.infradead.org
Quoted 3 lines
---
 drivers/pci/host/pci-mvebu.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

I think the general direction should rather be to change the PCI subsystem to make it possible for those drivers to be built as modules. However, since this is quite certainly a much larger effort, there is no reason to not clean things up as they are today, so:

Thanks,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#6 lore raw

Re: [PATCH 03/10] drivers/pci: make host/pci-mvebu.c explicitly non-modular

In reply to Paul Gortmaker
Quoted 6 lines · On Saturday 12 December 2015 20:41:50 Paul Gortmaker wrote:

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_MVEBU
drivers/pci/host/Kconfig: bool "Marvell EBU PCIe controller"

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

Here too. I think we should really try to let all PCI host drivers be loadable modules if possible. The module unload path is currently a bit iffy because of the way it interacts with the arch/arm PCI infrastructure, but we are in the process of fixing it.

For now, I think allowing all PCIe host drivers for ARM to be loadable but non-removably modules would be ideal.

Arnd

#7 lore raw

PATCH 4/10 drivers/pci: make host/pci-dra7xx.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_DRA7XX
drivers/pci/host/Kconfig: bool "TI DRA7xx PCIe controller"

...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.

We don't replace module.h with init.h since the file already has that.

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.

---
 drivers/pci/host/pci-dra7xx.c | 31 +++----------------------------
 1 file changed, 3 insertions(+), 28 deletions(-)
drivers/pci/host/pci-dra7xx.c +3 -28
index 8c3688046c02..3769b7654c5d 100644
@@ -16,7 +16,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of_gpio.h>
 #include <linux/pci.h>
 #include <linux/phy/phy.h>
@@ -451,25 +451,6 @@ err_phy:
 	return ret;
 }
 
-static int __exit dra7xx_pcie_remove(struct platform_device *pdev)
-{
-	struct dra7xx_pcie *dra7xx = platform_get_drvdata(pdev);
-	struct pcie_port *pp = &dra7xx->pp;
-	struct device *dev = &pdev->dev;
-	int count = dra7xx->phy_count;
-
-	if (pp->irq_domain)
-		irq_domain_remove(pp->irq_domain);
-	pm_runtime_put(dev);
-	pm_runtime_disable(dev);
-	while (count--) {
-		phy_power_off(dra7xx->phy[count]);
-		phy_exit(dra7xx->phy[count]);
-	}
-
-	return 0;
-}
-
 #ifdef CONFIG_PM_SLEEP
 static int dra7xx_pcie_suspend(struct device *dev)
 {
@@ -553,19 +534,13 @@ static const struct of_device_id of_dra7xx_pcie_match[] = {
 	{ .compatible = "ti,dra7-pcie", },
 	{},
 };
-MODULE_DEVICE_TABLE(of, of_dra7xx_pcie_match);
 
 static struct platform_driver dra7xx_pcie_driver = {
-	.remove		= __exit_p(dra7xx_pcie_remove),
 	.driver = {
 		.name	= "dra7-pcie",
 		.of_match_table = of_dra7xx_pcie_match,
+		.suppress_bind_attrs = true,
 		.pm	= &dra7xx_pcie_pm_ops,
 	},
 };
-
-module_platform_driver_probe(dra7xx_pcie_driver, dra7xx_pcie_probe);
-
-MODULE_AUTHOR("Kishon Vijay Abraham I <[email protected]>");
-MODULE_DESCRIPTION("TI PCIe controller driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver_probe(dra7xx_pcie_driver, dra7xx_pcie_probe);
Signature · 1 line
2.6.1
#8 lore raw

PATCH 5/10 drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2
drivers/pci/host/Kconfig: bool "Renesas R-Car Gen2 Internal PCI controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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

We don't replace module.h with init.h since the file already has that.

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.

---
 drivers/pci/host/pci-rcar-gen2.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)
drivers/pci/host/pci-rcar-gen2.c +3 -9
index c4f64bfee551..81dda40c7a9f 100644
@@ -4,6 +4,8 @@
  * Copyright (C) 2013 Renesas Solutions Corp.
  * Copyright (C) 2013 Cogent Embedded, Inc.
  *
+ * Module Author: Valentine Barshak <[email protected]>
+ *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
@@ -14,7 +16,6 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
 #include <linux/of_pci.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
@@ -366,8 +367,6 @@ static struct of_device_id rcar_pci_of_match[] = {
 	{ },
 };
 
-MODULE_DEVICE_TABLE(of, rcar_pci_of_match);
-
 static struct platform_driver rcar_pci_driver = {
 	.driver = {
 		.name = "pci-rcar-gen2",
@@ -376,9 +375,4 @@ static struct platform_driver rcar_pci_driver = {
 	},
 	.probe = rcar_pci_probe,
 };
-
-module_platform_driver(rcar_pci_driver);
-
-MODULE_LICENSE("GPL v2");
-MODULE_DESCRIPTION("Renesas R-Car Gen2 internal PCI");
-MODULE_AUTHOR("Valentine Barshak <[email protected]>");
+builtin_platform_driver(rcar_pci_driver);
Signature · 1 line
2.6.1
#9 lore raw

Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular

CC MODULE_AUTHOR

Quoted 64 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2
drivers/pci/host/Kconfig: bool "Renesas R-Car Gen2 Internal PCI controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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

We don't replace module.h with init.h since the file already has that.

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.

---
 drivers/pci/host/pci-rcar-gen2.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)
drivers/pci/host/pci-rcar-gen2.c +2 -3
index c4f64bfee551..81dda40c7a9f 100644
@@ -4,6 +4,8 @@
  * Copyright (C) 2013 Renesas Solutions Corp.
  * Copyright (C) 2013 Cogent Embedded, Inc.
  *
+ * Module Author: Valentine Barshak <[email protected]>
+ *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
@@ -14,7 +16,6 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
 #include <linux/of_pci.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
@@ -366,8 +367,6 @@ static struct of_device_id rcar_pci_of_match[] = {
        { },
 };
 
-MODULE_DEVICE_TABLE(of, rcar_pci_of_match);
-
 static struct platform_driver rcar_pci_driver = {
        .driver = {
                .name = "pci-rcar-gen2",

drivers/pci/host/pci-rcar-gen2.c static struct platform_driver rcar_pci_driver = { · line 378

Quoted 10 lines
drivers/pci/host/pci-rcar-gen2.c +1 -6
@@ -376,9 +375,4 @@ static struct platform_driver rcar_pci_driver = {
        },
        .probe = rcar_pci_probe,
 };
-
-module_platform_driver(rcar_pci_driver);
-
-MODULE_LICENSE("GPL v2");
-MODULE_DESCRIPTION("Renesas R-Car Gen2 internal PCI");
-MODULE_AUTHOR("Valentine Barshak <[email protected]>");
+builtin_platform_driver(rcar_pci_driver);

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#10 lore raw

Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular

Quoted 1 line · [Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly…

CC MODULE_AUTHOR

Thanks, I just assumed get-maintainer.pl would have automatically collected that up with the other names it emits. Apparently not.

Joe: is there a reason it doesn't use the module author field?

Paul.
--

Quoted 88 lines
Quoted 64 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2
drivers/pci/host/Kconfig: bool "Renesas R-Car Gen2 Internal PCI controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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

We don't replace module.h with init.h since the file already has that.

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.

---
 drivers/pci/host/pci-rcar-gen2.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)
drivers/pci/host/pci-rcar-gen2.c +2 -3
index c4f64bfee551..81dda40c7a9f 100644
@@ -4,6 +4,8 @@
  * Copyright (C) 2013 Renesas Solutions Corp.
  * Copyright (C) 2013 Cogent Embedded, Inc.
  *
+ * Module Author: Valentine Barshak <[email protected]>
+ *
  * This program is free software; you can redistribute it and/or modify
  * it under the terms of the GNU General Public License version 2 as
  * published by the Free Software Foundation.
@@ -14,7 +16,6 @@
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
 #include <linux/of_pci.h>
 #include <linux/pci.h>
 #include <linux/platform_device.h>
@@ -366,8 +367,6 @@ static struct of_device_id rcar_pci_of_match[] = {
        { },
 };
 
-MODULE_DEVICE_TABLE(of, rcar_pci_of_match);
-
 static struct platform_driver rcar_pci_driver = {
        .driver = {
                .name = "pci-rcar-gen2",

drivers/pci/host/pci-rcar-gen2.c static struct platform_driver rcar_pci_driver = { · line 378

Quoted 10 lines
drivers/pci/host/pci-rcar-gen2.c +1 -6
@@ -376,9 +375,4 @@ static struct platform_driver rcar_pci_driver = {
        },
        .probe = rcar_pci_probe,
 };
-
-module_platform_driver(rcar_pci_driver);
-
-MODULE_LICENSE("GPL v2");
-MODULE_DESCRIPTION("Renesas R-Car Gen2 internal PCI");
-MODULE_AUTHOR("Valentine Barshak <[email protected]>");
+builtin_platform_driver(rcar_pci_driver);

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#11 lore raw

Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular

Quoted 9 lines · Paul Gortmaker

[Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly

Quoted 1 line · non-modular] On 13/12/2015 (Sun 11:59) Geert Uytterhoeven wrote:

CC MODULE_AUTHOR

Thanks, I just assumed get-maintainer.pl would have automatically collected that up with the other names it emits.  Apparently not.

Joe: is there a reason it doesn't use the module author field?

Because it's frequently out-of-date and the
MAINTAINERS entry is generally enough.

If you want to add it, there's an option for it
--file-emails

Quoted 1 line

Paul

cheers, Joe

Acked-by on v1 5/10
#12 lore raw

Re: [PATCH 05/10] drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular

In reply to Paul Gortmaker
Quoted 21 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2
drivers/pci/host/Kconfig: bool "Renesas R-Car Gen2 Internal PCI controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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

We don't replace module.h with init.h since the file already has that.

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

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.

#13 lore raw

PATCH 6/10 drivers/pci: make host/pci-tegra.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_TEGRA
drivers/pci/host/Kconfig: bool "NVIDIA Tegra PCIe controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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 Thierry Reding <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Stephen Warren <[email protected]>
Cc Alexandre Courbot <[email protected]>
---
 drivers/pci/host/pci-tegra.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pci-tegra.c +4 -7
index 3018ae52e092..b49ea58787c1 100644
@@ -9,6 +9,8 @@
  *
  * Bits taken from arch/arm/mach-dove/pcie.c
  *
+ * Module Author: Thierry Reding <[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; either version 2 of the License, or
@@ -32,7 +34,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_pci.h>
@@ -1864,7 +1866,6 @@ static const struct of_device_id tegra_pcie_of_match[] = {
 	{ .compatible = "nvidia,tegra20-pcie", .data = &tegra20_pcie_data },
 	{ },
 };
-MODULE_DEVICE_TABLE(of, tegra_pcie_of_match);
 
 static void *tegra_pcie_ports_seq_start(struct seq_file *s, loff_t *pos)
 {
@@ -2055,8 +2056,4 @@ static struct platform_driver tegra_pcie_driver = {
 	},
 	.probe = tegra_pcie_probe,
 };
-module_platform_driver(tegra_pcie_driver);
-
-MODULE_AUTHOR("Thierry Reding <[email protected]>");
-MODULE_DESCRIPTION("NVIDIA Tegra PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(tegra_pcie_driver);
Signature · 1 line
2.6.1
#14 lore raw

Re: [PATCH 06/10] drivers/pci: make host/pci-tegra.c explicitly non-modular

Quoted 29 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_TEGRA
drivers/pci/host/Kconfig: bool "NVIDIA Tegra PCIe controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

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 Thierry Reding <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc Stephen Warren <[email protected]>
Cc Alexandre Courbot <[email protected]>
Quoted 3 lines
---
 drivers/pci/host/pci-tegra.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

I'd rather not do this. There is work underway to make it possible to turn PCI host drivers into modules. I have a local tree with the necessary changes and, though I suspect it might take a few releases to get everything merged, I think we'll get there eventually.

Thierry

#15 lore raw

PATCH 7/10 drivers/pci: make host/pcie-rcar.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2_PCIE
drivers/pci/host/Kconfig: bool "Renesas R-Car PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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.

---
 drivers/pci/host/pcie-rcar.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-rcar.c +4 -7
index f4fa6c537448..2c619277d265 100644
@@ -7,6 +7,8 @@
  *  arch/sh/drivers/pci/ops-sh7786.c
  *  Copyright (C) 2009 - 2011  Paul Mundt
  *
+ * Module Author: Phil Edworthy <[email protected]>
+ *
  * This file is licensed under the terms of the GNU General Public
  * License version 2.  This program is licensed "as is" without any
  * warranty of any kind, whether express or implied.
@@ -18,7 +20,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
@@ -921,7 +923,6 @@ static const struct of_device_id rcar_pcie_of_match[] = {
 	{ .compatible = "renesas,pcie-r8a7791", .data = rcar_pcie_hw_init },
 	{},
 };
-MODULE_DEVICE_TABLE(of, rcar_pcie_of_match);
 
 static int rcar_pcie_probe(struct platform_device *pdev)
 {
@@ -1007,8 +1008,4 @@ static struct platform_driver rcar_pcie_driver = {
 	},
 	.probe = rcar_pcie_probe,
 };
-module_platform_driver(rcar_pcie_driver);
-
-MODULE_AUTHOR("Phil Edworthy <[email protected]>");
-MODULE_DESCRIPTION("Renesas R-Car PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(rcar_pcie_driver);
Signature · 1 line
2.6.1
#16 lore raw

Re: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular

CC MODULE_AUTHOR

Quoted 62 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2_PCIE
drivers/pci/host/Kconfig: bool "Renesas R-Car PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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.

---
 drivers/pci/host/pcie-rcar.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-rcar.c +3 -2
index f4fa6c537448..2c619277d265 100644
@@ -7,6 +7,8 @@
  *  arch/sh/drivers/pci/ops-sh7786.c
  *  Copyright (C) 2009 - 2011  Paul Mundt
  *
+ * Module Author: Phil Edworthy <[email protected]>
+ *
  * This file is licensed under the terms of the GNU General Public
  * License version 2.  This program is licensed "as is" without any
  * warranty of any kind, whether express or implied.
@@ -18,7 +20,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
@@ -921,7 +923,6 @@ static const struct of_device_id rcar_pcie_of_match[] = {
        { .compatible = "renesas,pcie-r8a7791", .data = rcar_pcie_hw_init },
        {},
 };
-MODULE_DEVICE_TABLE(of, rcar_pcie_of_match);
 
 static int rcar_pcie_probe(struct platform_device *pdev)
 {

drivers/pci/host/pcie-rcar.c static struct platform_driver rcar_pcie_driver = { · line 1011

Quoted 9 lines
drivers/pci/host/pcie-rcar.c +1 -5
@@ -1007,8 +1008,4 @@ static struct platform_driver rcar_pcie_driver = {
        },
        .probe = rcar_pcie_probe,
 };
-module_platform_driver(rcar_pcie_driver);
-
-MODULE_AUTHOR("Phil Edworthy <[email protected]>");
-MODULE_DESCRIPTION("Renesas R-Car PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(rcar_pcie_driver);

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#17 lore raw

Re: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular

Quoted 1 line · [Re: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular]…

CC MODULE_AUTHOR

Yep, thanks -- I'll check for future sends to ensure get_maintainer.pl hasn't omitted the module author. Looks like it may emit the module author if they have been actively committing, but that won't work in the authored and then largely left it alone case.

P.
--

Quoted 85 lines
Quoted 62 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2_PCIE
drivers/pci/host/Kconfig: bool "Renesas R-Car PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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.

---
 drivers/pci/host/pcie-rcar.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-rcar.c +3 -2
index f4fa6c537448..2c619277d265 100644
@@ -7,6 +7,8 @@
  *  arch/sh/drivers/pci/ops-sh7786.c
  *  Copyright (C) 2009 - 2011  Paul Mundt
  *
+ * Module Author: Phil Edworthy <[email protected]>
+ *
  * This file is licensed under the terms of the GNU General Public
  * License version 2.  This program is licensed "as is" without any
  * warranty of any kind, whether express or implied.
@@ -18,7 +20,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
@@ -921,7 +923,6 @@ static const struct of_device_id rcar_pcie_of_match[] = {
        { .compatible = "renesas,pcie-r8a7791", .data = rcar_pcie_hw_init },
        {},
 };
-MODULE_DEVICE_TABLE(of, rcar_pcie_of_match);
 
 static int rcar_pcie_probe(struct platform_device *pdev)
 {

drivers/pci/host/pcie-rcar.c static struct platform_driver rcar_pcie_driver = { · line 1011

Quoted 9 lines
drivers/pci/host/pcie-rcar.c +1 -5
@@ -1007,8 +1008,4 @@ static struct platform_driver rcar_pcie_driver = {
        },
        .probe = rcar_pcie_probe,
 };
-module_platform_driver(rcar_pcie_driver);
-
-MODULE_AUTHOR("Phil Edworthy <[email protected]>");
-MODULE_DESCRIPTION("Renesas R-Car PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(rcar_pcie_driver);

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Acked-by
#18 lore raw

RE: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular

In reply to Geert Uytterhoeven

SGkgUGF1bCwNCg0KT24gMTMgRGVjZW1iZXIgMjAxNSAxMDo1OSwgR2VlcnQgVXl0dGVyaG9ldmVu IHdyb3RlOg0KPiANCj4gQ0MgTU9EVUxFX0FVVEhPUg0KPiANCj4gT24gU3VuLCBEZWMgMTMsIDIw MTUgYXQgMjo0MSBBTSwgUGF1bCBHb3J0bWFrZXINCj4gPHBhdWwuZ29ydG1ha2VyQHdpbmRyaXZl ci5jb20+IHdyb3RlOg0KPiA+IFRoZSBLY29uZmlnIGN1cnJlbnRseSBjb250cm9sbGluZyBjb21w aWxhdGlvbiBvZiB0aGlzIGNvZGUgaXM6DQo+ID4NCj4gPiBkcml2ZXJzL3BjaS9ob3N0L0tjb25m aWc6Y29uZmlnIFBDSV9SQ0FSX0dFTjJfUENJRQ0KPiA+IGRyaXZlcnMvcGNpL2hvc3QvS2NvbmZp ZzogICAgICAgYm9vbCAiUmVuZXNhcyBSLUNhciBQQ0llIGNvbnRyb2xsZXIiDQo+ID4NCj4gPiAu Li5tZWFuaW5nIHRoYXQgaXQgY3VycmVudGx5IGlzIG5vdCBiZWluZyBidWlsdCBhcyBhIG1vZHVs ZSBieSBhbnlvbmUuDQo+ID4NCj4gPiBMZXRzIHJlbW92ZSB0aGUgbW9kdWxhciBjb2RlIHRoYXQg aXMgZXNzZW50aWFsbHkgb3JwaGFuZWQsIHNvIHRoYXQNCj4gPiB3aGVuIHJlYWRpbmcgdGhlIGRy aXZlciB0aGVyZSBpcyBubyBkb3VidCBpdCBpcyBidWlsdGluLW9ubHkuDQo+ID4NCj4gPiBTaW5j ZSBtb2R1bGVfcGxhdGZvcm1fZHJpdmVyKCkgdXNlcyB0aGUgc2FtZSBpbml0IGxldmVsIHByaW9y aXR5IGFzDQo+ID4gYnVpbHRpbl9wbGF0Zm9ybV9kcml2ZXIoKSB0aGUgaW5pdCBvcmRlcmluZyBy ZW1haW5zIHVuY2hhbmdlZCB3aXRoDQo+ID4gdGhpcyBjb21taXQuDQo+ID4NCj4gPiBXZSBkb24n dCBoYXZlIHRvIGRpc2FsbG93IGEgZHJpdmVyIHVuYmluZCwgc2luY2UgdGhhdCBpcyBhbHJlYWR5 DQo+ID4gZG9uZSBmb3IgdXMgaW4gdGhpcyBkcml2ZXIuDQo+ID4NCj4gPiBBbHNvIG5vdGUgdGhh dCBNT0RVTEVfREVWSUNFX1RBQkxFIGlzIGEgbm8tb3AgZm9yIG5vbi1tb2R1bGFyIGNvZGUuDQo+ ID4NCj4gPiBXZSBhbHNvIGRlbGV0ZSB0aGUgTU9EVUxFX0xJQ0VOU0UgdGFnIGV0Yy4gc2luY2Ug YWxsIHRoYXQgaW5mb3JtYXRpb24NCj4gPiB3YXMgKG9yIGlzIG5vdykgY29udGFpbmVkIGF0IHRo ZSB0b3Agb2YgdGhlIGZpbGUgaW4gdGhlIGNvbW1lbnRzLg0KPiA+DQo+ID4gQ2M6IFNpbW9uIEhv cm1hbiA8aG9ybXNAdmVyZ2UubmV0LmF1Pg0KPiA+IENjOiBCam9ybiBIZWxnYWFzIDxiaGVsZ2Fh c0Bnb29nbGUuY29tPg0KPiA+IENjOiBsaW51eC1wY2lAdmdlci5rZXJuZWwub3JnDQo+ID4gQ2M6 IGxpbnV4LXNoQHZnZXIua2VybmVsLm9yZw0KPiA+IFNpZ25lZC1vZmYtYnk6IFBhdWwgR29ydG1h a2VyIDxwYXVsLmdvcnRtYWtlckB3aW5kcml2ZXIuY29tPg0KDQpJdCBkb2Vzbid0IGFwcGx5LCB3 b3VsZCB5b3UgbWluZCByZWJhc2luZyBvbnRvOg0KaHR0cHM6Ly9naXQua2VybmVsLm9yZy9jZ2l0 L2xpbnV4L2tlcm5lbC9naXQvaGVsZ2Fhcy9wY2kuZ2l0L2xvZy8/aD1uZXh0DQoNCkFzIGZvciB0 aGUgY2hhbmdlcyB0aGVtc2VsdmVzLA0KQWNrZWQtYnk6IFBoaWwgRWR3b3J0aHkgPHBoaWwuZWR3 b3J0aHlAcmVuZXNhcy5jb20+DQoNClRoYW5rcw0KUGhpbA0KDQo+ID4gLS0tDQo+ID4gIGRyaXZl cnMvcGNpL2hvc3QvcGNpZS1yY2FyLmMgfCAxMSArKysrLS0tLS0tLQ0KPiA+ICAxIGZpbGUgY2hh bmdlZCwgNCBpbnNlcnRpb25zKCspLCA3IGRlbGV0aW9ucygtKQ0KPiA+DQo+ID4gZGlmZiAtLWdp dCBhL2RyaXZlcnMvcGNpL2hvc3QvcGNpZS1yY2FyLmMgYi9kcml2ZXJzL3BjaS9ob3N0L3BjaWUt cmNhci5jDQo+ID4gaW5kZXggZjRmYTZjNTM3NDQ4Li4yYzYxOTI3N2QyNjUgMTAwNjQ0DQo+ID4g LS0tIGEvZHJpdmVycy9wY2kvaG9zdC9wY2llLXJjYXIuYw0KPiA+ICsrKyBiL2RyaXZlcnMvcGNp L2hvc3QvcGNpZS1yY2FyLmMNCj4gPiBAQCAtNyw2ICs3LDggQEANCj4gPiAgICogIGFyY2gvc2gv ZHJpdmVycy9wY2kvb3BzLXNoNzc4Ni5jDQo+ID4gICAqICBDb3B5cmlnaHQgKEMpIDIwMDkgLSAy MDExICBQYXVsIE11bmR0DQo+ID4gICAqDQo+ID4gKyAqIE1vZHVsZSBBdXRob3I6IFBoaWwgRWR3 b3J0aHkgPHBoaWwuZWR3b3J0aHlAcmVuZXNhcy5jb20+DQo+ID4gKyAqDQo+ID4gICAqIFRoaXMg ZmlsZSBpcyBsaWNlbnNlZCB1bmRlciB0aGUgdGVybXMgb2YgdGhlIEdOVSBHZW5lcmFsIFB1Ymxp Yw0KPiA+ICAgKiBMaWNlbnNlIHZlcnNpb24gMi4gIFRoaXMgcHJvZ3JhbSBpcyBsaWNlbnNlZCAi YXMgaXMiIHdpdGhvdXQgYW55DQo+ID4gICAqIHdhcnJhbnR5IG9mIGFueSBraW5kLCB3aGV0aGVy IGV4cHJlc3Mgb3IgaW1wbGllZC4NCj4gPiBAQCAtMTgsNyArMjAsNyBAQA0KPiA+ICAjaW5jbHVk ZSA8bGludXgvaXJxLmg+DQo+ID4gICNpbmNsdWRlIDxsaW51eC9pcnFkb21haW4uaD4NCj4gPiAg I2luY2x1ZGUgPGxpbnV4L2tlcm5lbC5oPg0KPiA+IC0jaW5jbHVkZSA8bGludXgvbW9kdWxlLmg+ DQo+ID4gKyNpbmNsdWRlIDxsaW51eC9pbml0Lmg+DQo+ID4gICNpbmNsdWRlIDxsaW51eC9tc2ku aD4NCj4gPiAgI2luY2x1ZGUgPGxpbnV4L29mX2FkZHJlc3MuaD4NCj4gPiAgI2luY2x1ZGUgPGxp bnV4L29mX2lycS5oPg0KPiA+IEBAIC05MjEsNyArOTIzLDYgQEAgc3RhdGljIGNvbnN0IHN0cnVj dCBvZl9kZXZpY2VfaWQgcmNhcl9wY2llX29mX21hdGNoW10gPSB7DQo+ID4gICAgICAgICB7IC5j b21wYXRpYmxlID0gInJlbmVzYXMscGNpZS1yOGE3NzkxIiwgLmRhdGEgPSByY2FyX3BjaWVfaHdf aW5pdCB9LA0KPiA+ICAgICAgICAge30sDQo+ID4gIH07DQo+ID4gLU1PRFVMRV9ERVZJQ0VfVEFC TEUob2YsIHJjYXJfcGNpZV9vZl9tYXRjaCk7DQo+ID4NCj4gPiAgc3RhdGljIGludCByY2FyX3Bj aWVfcHJvYmUoc3RydWN0IHBsYXRmb3JtX2RldmljZSAqcGRldikNCj4gPiAgew0KPiA+IEBAIC0x MDA3LDggKzEwMDgsNCBAQCBzdGF0aWMgc3RydWN0IHBsYXRmb3JtX2RyaXZlciByY2FyX3BjaWVf ZHJpdmVyID0gew0KPiA+ICAgICAgICAgfSwNCj4gPiAgICAgICAgIC5wcm9iZSA9IHJjYXJfcGNp ZV9wcm9iZSwNCj4gPiAgfTsNCj4gPiAtbW9kdWxlX3BsYXRmb3JtX2RyaXZlcihyY2FyX3BjaWVf ZHJpdmVyKTsNCj4gPiAtDQo+ID4gLU1PRFVMRV9BVVRIT1IoIlBoaWwgRWR3b3J0aHkgPHBoaWwu ZWR3b3J0aHlAcmVuZXNhcy5jb20+Iik7DQo+ID4gLU1PRFVMRV9ERVNDUklQVElPTigiUmVuZXNh cyBSLUNhciBQQ0llIGRyaXZlciIpOw0KPiA+IC1NT0RVTEVfTElDRU5TRSgiR1BMIHYyIik7DQo+ ID4gK2J1aWx0aW5fcGxhdGZvcm1fZHJpdmVyKHJjYXJfcGNpZV9kcml2ZXIpOw0KPiANCj4gR3J7 b2V0amUsZWV0aW5nfXMsDQo+IA0KPiAgICAgICAgICAgICAgICAgICAgICAgICBHZWVydA0KPiAN Cj4gLS0NCj4gR2VlcnQgVXl0dGVyaG9ldmVuIC0tIFRoZXJlJ3MgbG90cyBvZiBMaW51eCBiZXlv bmQgaWEzMiAtLSBnZWVydEBsaW51eC1tNjhrLm9yZw0KPiANCj4gSW4gcGVyc29uYWwgY29udmVy c2F0aW9ucyB3aXRoIHRlY2huaWNhbCBwZW9wbGUsIEkgY2FsbCBteXNlbGYgYSBoYWNrZXIuIEJ1 dA0KPiB3aGVuIEknbSB0YWxraW5nIHRvIGpvdXJuYWxpc3RzIEkganVzdCBzYXkgInByb2dyYW1t ZXIiIG9yIHNvbWV0aGluZyBsaWtlIHRoYXQuDQo+ICAgICAgICAgICAgICAgICAgICAgICAgICAg ICAgICAgLS0gTGludXMgVG9ydmFsZHMNCg==

#19 lore raw

Re: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular

Quoted 30 lines · [RE: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular]…

Hi Paul,

Quoted 27 lines · Geert Uytterhoeven

CC MODULE_AUTHOR

Quoted 22 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2_PCIE
drivers/pci/host/Kconfig: bool "Renesas R-Car PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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.

Quoted 9 lines
Quoted 5 lines
Quoted 5 lines

It doesn't apply, would you mind rebasing onto:
https://git.kernel.org/cgit/linux/kernel/git/helgaas/pci.git/log/?h=next

Sure I can do that and just resend the 2 renesas commits if there is no interest in tristate conversions for them. As for the others, I'm still awaiting feedback on what approach people want to take; I listed what I figured our possible options are.

Paul.
--

Quoted 61 lines

As for the changes themselves,

Thanks
Phil

Quoted 55 lines
Quoted 35 lines
---
 drivers/pci/host/pcie-rcar.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-rcar.c +3 -2
index f4fa6c537448..2c619277d265 100644
@@ -7,6 +7,8 @@
  *  arch/sh/drivers/pci/ops-sh7786.c
  *  Copyright (C) 2009 - 2011  Paul Mundt
  *
+ * Module Author: Phil Edworthy <[email protected]>
+ *
  * This file is licensed under the terms of the GNU General Public
  * License version 2.  This program is licensed "as is" without any
  * warranty of any kind, whether express or implied.
@@ -18,7 +20,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
@@ -921,7 +923,6 @@ static const struct of_device_id rcar_pcie_of_match[] = {
        { .compatible = "renesas,pcie-r8a7791", .data = rcar_pcie_hw_init },
        {},
 };
-MODULE_DEVICE_TABLE(of, rcar_pcie_of_match);
 
 static int rcar_pcie_probe(struct platform_device *pdev)
 {

drivers/pci/host/pcie-rcar.c static struct platform_driver rcar_pcie_driver = { · line 1011

Quoted 9 lines
drivers/pci/host/pcie-rcar.c +1 -5
@@ -1007,8 +1008,4 @@ static struct platform_driver rcar_pcie_driver = {
        },
        .probe = rcar_pcie_probe,
 };
-module_platform_driver(rcar_pcie_driver);
-
-MODULE_AUTHOR("Phil Edworthy <[email protected]>");
-MODULE_DESCRIPTION("Renesas R-Car PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(rcar_pcie_driver);

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- [email protected]

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

Acked-by on v1 7/10
#20 lore raw

Re: [PATCH 07/10] drivers/pci: make host/pcie-rcar.c explicitly non-modular

In reply to Paul Gortmaker
Quoted 19 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_RCAR_GEN2_PCIE
drivers/pci/host/Kconfig: bool "Renesas R-Car PCIe controller"

...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.

We don't have to disallow a driver unbind, since that is already done for us in this driver.

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

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.

#21 lore raw

PATCH 8/10 drivers/pci: make host/pcie-xilinx.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCIE_XILINX
drivers/pci/host/Kconfig: bool "Xilinx AXI PCIe host bridge support"

...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. This makes xilinx_pcie_free_irq_domain orphaned so we remove it too.

We don't have to worry about disallowing a driver unbind, since this driver already does that.

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 is already contained at the top of the file in the comments.

Cc Bjorn Helgaas <[email protected]>
Cc Michal Simek <[email protected]>
Cc "S?ren Brinkmann" <[email protected]>
Cc linux-pci at vger.kernel.org
Cc linux-arm-kernel at lists.infradead.org
---
 drivers/pci/host/pcie-xilinx.c | 53 ++----------------------------------------
 1 file changed, 2 insertions(+), 51 deletions(-)
drivers/pci/host/pcie-xilinx.c +2 -51
index 3c7a0d580b1e..73b5e856f9c2 100644
@@ -18,7 +18,7 @@
 #include <linux/irq.h>
 #include <linux/irqdomain.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_address.h>
 #include <linux/of_pci.h>
@@ -516,35 +516,6 @@ static irqreturn_t xilinx_pcie_intr_handler(int irq, void *data)
 }
 
 /**
- * xilinx_pcie_free_irq_domain - Free IRQ domain
- * @port: PCIe port information
- */
-static void xilinx_pcie_free_irq_domain(struct xilinx_pcie_port *port)
-{
-	int i;
-	u32 irq, num_irqs;
-
-	/* Free IRQ Domain */
-	if (IS_ENABLED(CONFIG_PCI_MSI)) {
-
-		free_pages(port->msi_pages, 0);
-
-		num_irqs = XILINX_NUM_MSI_IRQS;
-	} else {
-		/* INTx */
-		num_irqs = 4;
-	}
-
-	for (i = 0; i < num_irqs; i++) {
-		irq = irq_find_mapping(port->irq_domain, i);
-		if (irq > 0)
-			irq_dispose_mapping(irq);
-	}
-
-	irq_domain_remove(port->irq_domain);
-}
-
-/**
  * xilinx_pcie_init_irq_domain - Initialize IRQ domain
  * @port: PCIe port information
  *
@@ -858,21 +829,6 @@ static int xilinx_pcie_probe(struct platform_device *pdev)
 	return 0;
 }
 
-/**
- * xilinx_pcie_remove - Remove function
- * @pdev: Platform device pointer
- *
- * Return: '0' always
- */
-static int xilinx_pcie_remove(struct platform_device *pdev)
-{
-	struct xilinx_pcie_port *port = platform_get_drvdata(pdev);
-
-	xilinx_pcie_free_irq_domain(port);
-
-	return 0;
-}
-
 static struct of_device_id xilinx_pcie_of_match[] = {
 	{ .compatible = "xlnx,axi-pcie-host-1.00.a", },
 	{}
@@ -885,10 +841,5 @@ static struct platform_driver xilinx_pcie_driver = {
 		.suppress_bind_attrs = true,
 	},
 	.probe = xilinx_pcie_probe,
-	.remove = xilinx_pcie_remove,
 };
-module_platform_driver(xilinx_pcie_driver);
-
-MODULE_AUTHOR("Xilinx Inc");
-MODULE_DESCRIPTION("Xilinx AXI PCIe driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(xilinx_pcie_driver);
Signature · 1 line
2.6.1
Acked-by on v1 8/10
#22 lore raw

Re: [PATCH 08/10] drivers/pci: make host/pcie-xilinx.c explicitly non-modular

Quoted 10 lines · Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCIE_XILINX
drivers/pci/host/Kconfig: bool "Xilinx AXI PCIe host bridge support"

...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. This makes xilinx_pcie_free_irq_domain orphaned so we remove it too.

Right.

Thanks,
Michal

#23 lore raw

PATCH 9/10 drivers/pci: make host/pci-keystone.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCI_KEYSTONE
drivers/pci/host/Kconfig: bool "TI Keystone PCIe controller"

...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 Murali Karicheri <[email protected]>
Cc Bjorn Helgaas <[email protected]>
Cc linux-pci at vger.kernel.org
Cc linux-arm-kernel at lists.infradead.org
---
 drivers/pci/host/pci-keystone.c | 21 +++------------------
 1 file changed, 3 insertions(+), 18 deletions(-)
drivers/pci/host/pci-keystone.c +3 -18
index 0aa81bd3de12..0493d4257bde 100644
@@ -16,7 +16,7 @@
 #include <linux/clk.h>
 #include <linux/delay.h>
 #include <linux/irqdomain.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/msi.h>
 #include <linux/of_irq.h>
 #include <linux/of.h>
@@ -329,16 +329,6 @@ static const struct of_device_id ks_pcie_of_match[] = {
 	},
 	{ },
 };
-MODULE_DEVICE_TABLE(of, ks_pcie_of_match);
-
-static int __exit ks_pcie_remove(struct platform_device *pdev)
-{
-	struct keystone_pcie *ks_pcie = platform_get_drvdata(pdev);
-
-	clk_disable_unprepare(ks_pcie->clk);
-
-	return 0;
-}
 
 static int __init ks_pcie_probe(struct platform_device *pdev)
 {
@@ -398,15 +388,10 @@ fail_clk:
 
 static struct platform_driver ks_pcie_driver __refdata = {
 	.probe  = ks_pcie_probe,
-	.remove = __exit_p(ks_pcie_remove),
 	.driver = {
 		.name	= "keystone-pcie",
 		.of_match_table = of_match_ptr(ks_pcie_of_match),
+		.suppress_bind_attrs = true,
 	},
 };
-
-module_platform_driver(ks_pcie_driver);
-
-MODULE_AUTHOR("Murali Karicheri <[email protected]>");
-MODULE_DESCRIPTION("Keystone PCIe host controller driver");
-MODULE_LICENSE("GPL v2");
+builtin_platform_driver(ks_pcie_driver);
Signature · 1 line
2.6.1
#24 lore raw

PATCH 10/10 drivers/pci: make host/pcie-altera.c explicitly non-modular

In reply to Paul Gortmaker

The Kconfig currently controlling compilation of this code is:

drivers/pci/host/Kconfig:config PCIE_ALTERA
drivers/pci/host/Kconfig: bool "Altera PCIe controller"

...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 don't have to disallow a driver unbind, since that is already done for us in this driver.

Since module_init translates to device_initcall in the non-modular case, 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.

---
 drivers/pci/host/pcie-altera.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)
drivers/pci/host/pcie-altera.c +5 -7
index 99da549d5d06..ba55885d508f 100644
@@ -1,6 +1,9 @@
 /*
  * Copyright Altera Corporation (C) 2013-2015. All rights reserved
  *
+ * Author: Ley Foon Tan <[email protected]>
+ * Description: Altera PCIe host controller driver
+ *
  * This program is free software; you can redistribute it and/or modify it
  * under the terms and conditions of the GNU General Public License,
  * version 2, as published by the Free Software Foundation.
@@ -17,7 +20,7 @@
 #include <linux/delay.h>
 #include <linux/interrupt.h>
 #include <linux/irqchip/chained_irq.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of_address.h>
 #include <linux/of_irq.h>
 #include <linux/of_pci.h>
@@ -566,7 +569,6 @@ static const struct of_device_id altera_pcie_of_match[] = {
 	{ .compatible = "altr,pcie-root-port-1.0", },
 	{},
 };
-MODULE_DEVICE_TABLE(of, altera_pcie_of_match);
 
 static struct platform_driver altera_pcie_driver = {
 	.probe		= altera_pcie_probe,
@@ -581,8 +583,4 @@ static int altera_pcie_init(void)
 {
 	return platform_driver_register(&altera_pcie_driver);
 }
-module_init(altera_pcie_init);
-
-MODULE_AUTHOR("Ley Foon Tan <[email protected]>");
-MODULE_DESCRIPTION("Altera PCIe host controller driver");
-MODULE_LICENSE("GPL v2");
+device_initcall(altera_pcie_init);
Signature · 1 line
2.6.1
#25 lore raw
In reply to Paul Gortmaker

Hi Paul,

Quoted 26 lines · Paul Gortmaker

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Quoted 4 lines

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

Personally, I think all of these should become tristate, so distro kernels don't have to build in PCI(e) support for all SoCs. multi_v7_defconfig kernels are becoming too big.

That does not preclude making these modules un-unloadable, though.

Quoted 11 lines

Paul Gortmaker (10):
drivers/pci: make host/pci-imx6.c driver explicitly non-modular drivers/pci: make host/pcie-spear13xx.c driver explicitly non-modular drivers/pci: make host/pci-mvebu.c explicitly non-modular drivers/pci: make host/pci-dra7xx.c explicitly non-modular drivers/pci: make host/pci-rcar-gen2.c explicitly non-modular drivers/pci: make host/pci-tegra.c explicitly non-modular drivers/pci: make host/pcie-rcar.c explicitly non-modular drivers/pci: make host/pcie-xilinx.c explicitly non-modular drivers/pci: make host/pci-keystone.c explicitly non-modular drivers/pci: make host/pcie-altera.c explicitly non-modular

Gr{oetje,eeting}s,

Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert at linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds

#26 lore raw
Quoted 35 lines · Geert Uytterhoeven

Hi Paul,

Quoted 26 lines · Paul Gortmaker

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Quoted 4 lines

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

Quoted 5 lines

Personally, I think all of these should become tristate, so distro kernels don't have to build in PCI(e) support for all SoCs. multi_v7_defconfig kernels are becoming too big.

That does not preclude making these modules un-unloadable, though.

Most of these can't be made tristate as-is, because they use symbols that aren't exported. Many of those symbols can easily be exported, so its just a matter of getting the respective patches merged. I disagree with making the modules non-unloadable, though. I have a local branch with changes necessary to unload the host controller driver and it works just fine.

Thierry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20151214/21036b5d/attachment.sig>

#27 lore raw
Quoted 40 lines · Thierry Reding
Quoted 39 lines · Geert Uytterhoeven

Hi Paul,

Quoted 26 lines · Paul Gortmaker

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Quoted 4 lines

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

Personally, I think all of these should become tristate, so distro kernels don't have to build in PCI(e) support for all SoCs. multi_v7_defconfig kernels are becoming too big.

Quoted 8 lines
Quoted 1 line

That does not preclude making these modules un-unloadable, though.

Most of these can't be made tristate as-is, because they use symbols that aren't exported. Many of those symbols can easily be exported, so its just a matter of getting the respective patches merged. I disagree with making the modules non-unloadable, though. I have a local branch with changes necessary to unload the host controller driver and it works just fine.

Great.

Send them out.

Thanks,
Michal

#28 lore raw
In reply to Thierry Reding
Quoted 40 lines · Thierry Reding
Quoted 39 lines · Geert Uytterhoeven

Hi Paul,

Quoted 26 lines · Paul Gortmaker

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Quoted 4 lines

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

Personally, I think all of these should become tristate, so distro kernels don't have to build in PCI(e) support for all SoCs. multi_v7_defconfig kernels are becoming too big.

Quoted 9 lines
Quoted 1 line

That does not preclude making these modules un-unloadable, though.

Most of these can't be made tristate as-is, because they use symbols that aren't exported. Many of those symbols can easily be exported, so its just a matter of getting the respective patches merged. I disagree with making the modules non-unloadable, though. I have a local branch with changes necessary to unload the host controller driver and it works just fine.

PCIe host driver that use fixup (DECLARE_PCI_FIXUP_*) can't use tristate. Fixup region is in kernel region and this region if not updated when loading a module.

Regards
Ley Foon

#29 lore raw
Quoted 51 lines · Ley Foon Tan
Quoted 40 lines · Thierry Reding
Quoted 39 lines · Geert Uytterhoeven

Hi Paul,

Quoted 26 lines · Paul Gortmaker

This series of commits is a slice of a larger project to ensure people don't have dead code for module removal in non-modular drivers. Overall there was roughly 5k lines of dead code in the kernel due to this. So far we've fixed several areas, like tty, x86, net, etc. and we continue to work on other areas.

There are several reasons to not use module_init for code that can never be built as a module, but the big ones are:

(1) it is easy to accidentally code up 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.

Here we convert some module_init() calls into device_initcall() and delete any module_exit and remove code that gets orphaned in the process, for an overall net code reduction, which is always welcome.

The use of device_initcall ensures that the init function ordering remains unchanged, but one could argue that PCI host code might be more appropriate to be handled under subsys_initcall. Fortunately we can revisit making this extra change at a later date if desired; it does not need to happen now, and we reduce the risk of introducing regressions at this point in time by separating the two changes.

Quoted 4 lines

Over half of the drivers changed here already explicitly disallowed any unbind operations. For the rest we make them the same, since there is not really any sensible use case to unbind any built-in bus support that I can think of.

Personally, I think all of these should become tristate, so distro kernels don't have to build in PCI(e) support for all SoCs. multi_v7_defconfig kernels are becoming too big.

Quoted 9 lines
Quoted 1 line

That does not preclude making these modules un-unloadable, though.

Most of these can't be made tristate as-is, because they use symbols that aren't exported. Many of those symbols can easily be exported, so its just a matter of getting the respective patches merged. I disagree with making the modules non-unloadable, though. I have a local branch with changes necessary to unload the host controller driver and it works just fine.

Quoted 3 lines

PCIe host driver that use fixup (DECLARE_PCI_FIXUP_*) can't use tristate. Fixup region is in kernel region and this region if not updated when loading a module.

Interesting, I hadn't thought about that. I suppose this means that the module will end up containing an unused section with the fixup code. It might be useful to add a way for that to trigger a warning at build time.

Perhaps to fix this a mechanism could be introduced to add a table of fixups to a host controller driver and that will get applied to all children of the bridge. It could be problematic to cover all of the different fixup stages, though.

Thierry
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: not available
URL: <http://lists.infradead.org/pipermail/linux-arm-kernel/attachments/20151214/46147b8a/attachment-0001.sig>

#30 lore raw
Quoted 9 lines · On Monday 14 December 2015 10:19:40 Thierry Reding wrote:
Quoted 3 lines

PCIe host driver that use fixup (DECLARE_PCI_FIXUP_*) can't use tristate. Fixup region is in kernel region and this region if not updated when loading a module.

Interesting, I hadn't thought about that. I suppose this means that the module will end up containing an unused section with the fixup code. It might be useful to add a way for that to trigger a warning at build time.

Quoted 5 lines

Perhaps to fix this a mechanism could be introduced to add a table of fixups to a host controller driver and that will get applied to all children of the bridge. It could be problematic to cover all of the different fixup stages, though.

I think a lot of the fixups shouldn't really be there in the first place, they are about stuff that we can fix up in the probe function, or that should be fixed up in the probe function with some appropriate core support added.

Arnd

#31 lore raw
Quoted 17 lines · [Re: [PATCH 00/10] drivers/pci: avoid module_init in non-modular host/pci*] On…
Quoted 9 lines · On Monday 14 December 2015 10:19:40 Thierry Reding wrote:
Quoted 3 lines

PCIe host driver that use fixup (DECLARE_PCI_FIXUP_*) can't use tristate. Fixup region is in kernel region and this region if not updated when loading a module.

Interesting, I hadn't thought about that. I suppose this means that the module will end up containing an unused section with the fixup code. It might be useful to add a way for that to trigger a warning at build time.

Quoted 5 lines

Perhaps to fix this a mechanism could be introduced to add a table of fixups to a host controller driver and that will get applied to all children of the bridge. It could be problematic to cover all of the different fixup stages, though.

Quoted 3 lines

I think a lot of the fixups shouldn't really be there in the first place, they are about stuff that we can fix up in the probe function, or that should be fixed up in the probe function with some appropriate core support added.

So, the feedback on this is a bit all over the map, leaving me unsure what to do next. And is the choice we make on a per board/bsp basis or ideally across all platforms? I see the choices as:

1) do nothing; which IMHO is least desirable as it leaves the code misrepresenting itself as modular; one of the key issues I wanted to fix

2) use the patches I've sent ; then as they are genuinely made modular, the person doing so essentially "patch -R" or reverts the change as step one. This has the advantage of solving the "we'll get to it someday" issue if someday never comes.

3) make them all tristate; beat it with a stick until it compiles [M] and modposts -- leaving the fixups and functional testing to people with the boards and low level knowledge to make it _work_ as a module. The downside here is the code is still kind of misrepresenting itself as modularly functional -- a ban of unloading might mitigate that some.

Paul.
--

Quoted 2 lines

Arnd

#32 lore raw
Quoted 33 lines · Paul Gortmaker
Quoted 10 lines · [Re: [PATCH 00/10] drivers/pci: avoid module_init in non-modular host/pci*] On…
Quoted 9 lines · On Monday 14 December 2015 10:19:40 Thierry Reding wrote:
Quoted 3 lines

PCIe host driver that use fixup (DECLARE_PCI_FIXUP_*) can't use tristate. Fixup region is in kernel region and this region if not updated when loading a module.

Interesting, I hadn't thought about that. I suppose this means that the module will end up containing an unused section with the fixup code. It might be useful to add a way for that to trigger a warning at build time.

Quoted 8 lines
Quoted 4 lines

Perhaps to fix this a mechanism could be introduced to add a table of fixups to a host controller driver and that will get applied to all children of the bridge. It could be problematic to cover all of the different fixup stages, though.

I think a lot of the fixups shouldn't really be there in the first place, they are about stuff that we can fix up in the probe function, or that should be fixed up in the probe function with some appropriate core support added.

So, the feedback on this is a bit all over the map, leaving me unsure what to do next. And is the choice we make on a per board/bsp basis or ideally across all platforms? I see the choices as:

1) do nothing; which IMHO is least desirable as it leaves the code misrepresenting itself as modular; one of the key issues I wanted to fix

2) use the patches I've sent ; then as they are genuinely made modular, the person doing so essentially "patch -R" or reverts the change as step one. This has the advantage of solving the "we'll get to it someday" issue if someday never comes.

Quoted 5 lines

3) make them all tristate; beat it with a stick until it compiles [M] and modposts -- leaving the fixups and functional testing to people with the boards and low level knowledge to make it _work_ as a module. The downside here is the code is still kind of misrepresenting itself as modularly functional -- a ban of unloading might mitigate that some.

I'd like to preserve the mind-set that host controller drivers are *expected* to be modular, even if we aren't there yet. I guess that means I'm in favor of option 3, at least for drivers that don't use fixups.

Bjorn