[PATCH 0/5] drivers/tty: make more bool drivers explicitly non-modular

[PATCH 0/5] drivers/tty: make more bool drivers explicitly non-modular

[Paul Gortmaker]

This second set of patches to drivers/tty steps outside of the serial
dir, and an improved auditing finds two more serial drivers pretending
to be modular that really are not.

The reasoning for doing this is the same as the first set[1] of patches
and is largely copied below:

  In the previous merge window, we made changes to allow better
  delineation between modular and non-modular code in commit
  0fd972a7d91d6e15393c449492a04d94c0b89351 ("module: relocate module_init
  from init.h to module.h").  This allows us to now ensure module code
  looks modular and non-modular code does not accidentally look modular
  without suffering build breakage.
  
  Here we target code that is, by nature of their Kconfig settings, only
  available to be built-in, but implicitly presenting itself as being
  possibly modular by way of using modular headers, macros, and functions.
  
  The goal here is to remove that illusion of modularity from these
  drivers, but in a way that leaves the actual runtime unchanged.
  In doing so, we remove code that has never been tested and adds
  no value to the tree.  And we begin the process of expecting a
  level of consistency between the Kconfig of a driver and the code
  that the driver uses.
  
Build tested for allyesconfig on x86_64, and ARM for lpc81xx, and powerpc
for hvc_console and mpsc, layered onto tty/tty-next as a baseline.

Paul.

[1] https://lkml.kernel.org/r/1437530538-5078-1-git-send-email-paul.gortmaker@windriver.com
--

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: Joachim Eastwood <manabian@gmail.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-serial@vger.kernel.org

Paul Gortmaker (5):
  drivers/tty: make pty.c slightly more explicitly non-modular
  drivers/tty: make sysrq.c slightly more explicitly non-modular
  drivers/tty: make hvc_console.c explicitly non-modular
  drivers/tty: make serial/mpsc.c driver explicitly non-modular
  drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular

 drivers/tty/hvc/hvc_console.c          | 18 +----------------
 drivers/tty/pty.c                      |  7 +++++--
 drivers/tty/serial/8250/8250_lpc18xx.c | 19 ++++--------------
 drivers/tty/serial/mpsc.c              | 36 +++-------------------------------
 drivers/tty/sysrq.c                    |  6 +++++-
 5 files changed, 18 insertions(+), 68 deletions(-)

-- 
2.5.0

[PATCH 4/5] drivers/tty: make serial/mpsc.c driver explicitly non-modular

[Paul Gortmaker]

The Kconfig for this driver is currently:

config SERIAL_MPSC
        bool "Marvell MPSC serial port 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.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.

We leave some tags like MODULE_AUTHOR for documentation purposes.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 drivers/tty/serial/mpsc.c | 36 +++---------------------------------
 1 file changed, 3 insertions(+), 33 deletions(-)

diff --git a/drivers/tty/serial/mpsc.c b/drivers/tty/serial/mpsc.c
index 82bb6d1fe23b..edb32e3f1e84 100644
--- a/drivers/tty/serial/mpsc.c
+++ b/drivers/tty/serial/mpsc.c
@@ -55,8 +55,6 @@
 #define SUPPORT_SYSRQ
 #endif
 
-#include <linux/module.h>
-#include <linux/moduleparam.h>
 #include <linux/tty.h>
 #include <linux/tty_flip.h>
 #include <linux/ioport.h>
@@ -2108,24 +2106,8 @@ static int mpsc_drv_probe(struct platform_device *dev)
 	return rc;
 }
 
-static int mpsc_drv_remove(struct platform_device *dev)
-{
-	pr_debug("mpsc_drv_exit: Removing MPSC %d\n", dev->id);
-
-	if (dev->id < MPSC_NUM_CTLRS) {
-		uart_remove_one_port(&mpsc_reg, &mpsc_ports[dev->id].port);
-		mpsc_release_port((struct uart_port *)
-				&mpsc_ports[dev->id].port);
-		mpsc_drv_unmap_regs(&mpsc_ports[dev->id]);
-		return 0;
-	} else {
-		return -ENODEV;
-	}
-}
-
 static struct platform_driver mpsc_driver = {
 	.probe	= mpsc_drv_probe,
-	.remove	= mpsc_drv_remove,
 	.driver	= {
 		.name	= MPSC_CTLR_NAME,
 	},
@@ -2156,22 +2138,10 @@ static int __init mpsc_drv_init(void)
 
 	return rc;
 }
+device_initcall(mpsc_drv_init);
 
-static void __exit mpsc_drv_exit(void)
-{
-	platform_driver_unregister(&mpsc_driver);
-	platform_driver_unregister(&mpsc_shared_driver);
-	uart_unregister_driver(&mpsc_reg);
-	memset(mpsc_ports, 0, sizeof(mpsc_ports));
-	memset(&mpsc_shared_regs, 0, sizeof(mpsc_shared_regs));
-}
-
-module_init(mpsc_drv_init);
-module_exit(mpsc_drv_exit);
-
+/*
 MODULE_AUTHOR("Mark A. Greer <mgreer@mvista.com>");
 MODULE_DESCRIPTION("Generic Marvell MPSC serial/UART driver");
-MODULE_VERSION(MPSC_VERSION);
 MODULE_LICENSE("GPL");
-MODULE_ALIAS_CHARDEV_MAJOR(MPSC_MAJOR);
-MODULE_ALIAS("platform:" MPSC_CTLR_NAME);
+*/
-- 
2.5.0

[PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular

[Paul Gortmaker]

The Kconfig currently controlling compilation of this code is:

8250/Kconfig:config SERIAL_8250_LPC18XX
8250/Kconfig:        bool "NXP LPC18xx/43xx serial port 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.

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 leave some tags like MODULE_AUTHOR for documentation purposes.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: Joachim Eastwood <manabian@gmail.com>
Cc: linux-serial@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 drivers/tty/serial/8250/8250_lpc18xx.c | 19 ++++---------------
 1 file changed, 4 insertions(+), 15 deletions(-)

diff --git a/drivers/tty/serial/8250/8250_lpc18xx.c b/drivers/tty/serial/8250/8250_lpc18xx.c
index 99cd478851ff..f5112bee1ab6 100644
--- a/drivers/tty/serial/8250/8250_lpc18xx.c
+++ b/drivers/tty/serial/8250/8250_lpc18xx.c
@@ -15,7 +15,7 @@
 
 #include <linux/clk.h>
 #include <linux/io.h>
-#include <linux/module.h>
+#include <linux/init.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
 
@@ -198,33 +198,22 @@ dis_clk_reg:
 	return ret;
 }
 
-static int lpc18xx_serial_remove(struct platform_device *pdev)
-{
-	struct lpc18xx_uart_data *data = platform_get_drvdata(pdev);
-
-	serial8250_unregister_port(data->line);
-	clk_disable_unprepare(data->clk_uart);
-	clk_disable_unprepare(data->clk_reg);
-
-	return 0;
-}
-
 static const struct of_device_id lpc18xx_serial_match[] = {
 	{ .compatible = "nxp,lpc1850-uart" },
 	{ },
 };
-MODULE_DEVICE_TABLE(of, lpc18xx_serial_match);
 
 static struct platform_driver lpc18xx_serial_driver = {
 	.probe  = lpc18xx_serial_probe,
-	.remove = lpc18xx_serial_remove,
 	.driver = {
 		.name = "lpc18xx-uart",
 		.of_match_table = lpc18xx_serial_match,
 	},
 };
-module_platform_driver(lpc18xx_serial_driver);
+builtin_platform_driver(lpc18xx_serial_driver);
 
+/*
 MODULE_AUTHOR("Joachim Eastwood <manabian@gmail.com>");
 MODULE_DESCRIPTION("Serial port driver NXP LPC18xx/43xx devices");
 MODULE_LICENSE("GPL v2");
+*/
-- 
2.5.0

Re: [PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular

[Joachim Eastwood]

Hi Paul,

On 9 August 2015 at 02:51, Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
> The Kconfig currently controlling compilation of this code is:
>
> 8250/Kconfig:config SERIAL_8250_LPC18XX
> 8250/Kconfig:        bool "NXP LPC18xx/43xx serial port 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.

The bool in KConfig is more or less a copy-paste mistake on my side. I
really meant for this driver to be buildable as module and it builds
just fine with 'm'.

Want me send a patch changing it to tristate or will you handle it?


regards,
Joachim Eastwood

Re: [PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular

[Paul Gortmaker]

[Re: [PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular] On 10/08/2015 (Mon 20:42) Joachim Eastwood wrote:

> Hi Paul,
> 
> On 9 August 2015 at 02:51, Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
> > The Kconfig currently controlling compilation of this code is:
> >
> > 8250/Kconfig:config SERIAL_8250_LPC18XX
> > 8250/Kconfig:        bool "NXP LPC18xx/43xx serial port 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.
> 
> The bool in KConfig is more or less a copy-paste mistake on my side. I
> really meant for this driver to be buildable as module and it builds
> just fine with 'm'.
> 
> Want me send a patch changing it to tristate or will you handle it?

I will do a v2 in the next day or so with collected ACKs and this patch
replaced, unless Greg jumps in and indicates he'd like something
different.

Paul.
--

> 
> 
> regards, Joachim Eastwood

Re: [PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular

[Greg Kroah-Hartman]

On Tue, Aug 11, 2015 at 02:05:44AM -0400, Paul Gortmaker wrote:
> [Re: [PATCH 5/5] drivers/tty: make serial 8250_lpc18xx.c explicitly non-modular] On 10/08/2015 (Mon 20:42) Joachim Eastwood wrote:
> 
> > Hi Paul,
> > 
> > On 9 August 2015 at 02:51, Paul Gortmaker <paul.gortmaker@windriver.com> wrote:
> > > The Kconfig currently controlling compilation of this code is:
> > >
> > > 8250/Kconfig:config SERIAL_8250_LPC18XX
> > > 8250/Kconfig:        bool "NXP LPC18xx/43xx serial port 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.
> > 
> > The bool in KConfig is more or less a copy-paste mistake on my side. I
> > really meant for this driver to be buildable as module and it builds
> > just fine with 'm'.
> > 
> > Want me send a patch changing it to tristate or will you handle it?
> 
> I will do a v2 in the next day or so with collected ACKs and this patch
> replaced, unless Greg jumps in and indicates he'd like something
> different.

Nope, that would be fine with me, thanks.

greg k-h

[PATCH v2 0/5] drivers/tty: make more bool drivers explicitly non-modular

[Paul Gortmaker]

[v2: drop dead module code removal from 8250_lpc18xx.c ; instead convert it
 from bool to tristate ; also add ack to hvc_console commit.]

This second set of patches to drivers/tty steps outside of the serial
dir, and an improved auditing finds two more serial drivers pretending
to be modular that really are not.

The reasoning for doing this is the same as the first set[1] of patches
and is largely copied below:

  In the previous merge window, we made changes to allow better
  delineation between modular and non-modular code in commit
  0fd972a7d91d6e15393c449492a04d94c0b89351 ("module: relocate module_init
  from init.h to module.h").  This allows us to now ensure module code
  looks modular and non-modular code does not accidentally look modular
  without suffering build breakage.
  
  Here we target code that is, by nature of their Kconfig settings, only
  available to be built-in, but implicitly presenting itself as being
  possibly modular by way of using modular headers, macros, and functions.
  
  The goal here is to remove that illusion of modularity from these
  drivers, but in a way that leaves the actual runtime unchanged.
  In doing so, we remove code that has never been tested and adds
  no value to the tree.  And we begin the process of expecting a
  level of consistency between the Kconfig of a driver and the code
  that the driver uses.
  
Build tested for allyesconfig on x86_64, and ARM for lpc81xx, and powerpc
for hvc_console and mpsc, layered onto tty/tty-next as a baseline.

Paul.

[1] https://lkml.kernel.org/r/1437530538-5078-1-git-send-email-paul.gortmaker@windriver.com
--

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: Joachim Eastwood <manabian@gmail.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linuxppc-dev@lists.ozlabs.org
Cc: linux-serial@vger.kernel.org

Paul Gortmaker (5):
  drivers/tty: make pty.c slightly more explicitly non-modular
  drivers/tty: make sysrq.c slightly more explicitly non-modular
  drivers/tty: make hvc_console.c explicitly non-modular
  drivers/tty: make serial/mpsc.c driver explicitly non-modular
  drivers/tty: make serial 8250_lpc18xx.c Kconfig a tristate

 drivers/tty/hvc/hvc_console.c   | 18 +-----------------
 drivers/tty/pty.c               |  7 +++++--
 drivers/tty/serial/8250/Kconfig |  2 +-
 drivers/tty/serial/mpsc.c       | 36 +++---------------------------------
 drivers/tty/sysrq.c             |  6 +++++-
 5 files changed, 15 insertions(+), 54 deletions(-)

-- 
2.5.0

[PATCH 4/5] drivers/tty: make serial/mpsc.c driver explicitly non-modular

[Paul Gortmaker]

The Kconfig for this driver is currently:

config SERIAL_MPSC
        bool "Marvell MPSC serial port 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.

Since module_init translates to device_initcall in the non-modular
case, the init ordering remains unchanged with this commit.

We leave some tags like MODULE_AUTHOR for documentation purposes.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Jiri Slaby <jslaby@suse.com>
Cc: linux-serial@vger.kernel.org
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
 drivers/tty/serial/mpsc.c | 36 +++---------------------------------
 1 file changed, 3 insertions(+), 33 deletions(-)

diff --git a/drivers/tty/serial/mpsc.c b/drivers/tty/serial/mpsc.c
index 82bb6d1fe23b..edb32e3f1e84 100644
--- a/drivers/tty/serial/mpsc.c
+++ b/drivers/tty/serial/mpsc.c
@@ -55,8 +55,6 @@
 #define SUPPORT_SYSRQ
 #endif
 
-#include <linux/module.h>
-#include <linux/moduleparam.h>
 #include <linux/tty.h>
 #include <linux/tty_flip.h>
 #include <linux/ioport.h>
@@ -2108,24 +2106,8 @@ static int mpsc_drv_probe(struct platform_device *dev)
 	return rc;
 }
 
-static int mpsc_drv_remove(struct platform_device *dev)
-{
-	pr_debug("mpsc_drv_exit: Removing MPSC %d\n", dev->id);
-
-	if (dev->id < MPSC_NUM_CTLRS) {
-		uart_remove_one_port(&mpsc_reg, &mpsc_ports[dev->id].port);
-		mpsc_release_port((struct uart_port *)
-				&mpsc_ports[dev->id].port);
-		mpsc_drv_unmap_regs(&mpsc_ports[dev->id]);
-		return 0;
-	} else {
-		return -ENODEV;
-	}
-}
-
 static struct platform_driver mpsc_driver = {
 	.probe	= mpsc_drv_probe,
-	.remove	= mpsc_drv_remove,
 	.driver	= {
 		.name	= MPSC_CTLR_NAME,
 	},
@@ -2156,22 +2138,10 @@ static int __init mpsc_drv_init(void)
 
 	return rc;
 }
+device_initcall(mpsc_drv_init);
 
-static void __exit mpsc_drv_exit(void)
-{
-	platform_driver_unregister(&mpsc_driver);
-	platform_driver_unregister(&mpsc_shared_driver);
-	uart_unregister_driver(&mpsc_reg);
-	memset(mpsc_ports, 0, sizeof(mpsc_ports));
-	memset(&mpsc_shared_regs, 0, sizeof(mpsc_shared_regs));
-}
-
-module_init(mpsc_drv_init);
-module_exit(mpsc_drv_exit);
-
+/*
 MODULE_AUTHOR("Mark A. Greer <mgreer@mvista.com>");
 MODULE_DESCRIPTION("Generic Marvell MPSC serial/UART driver");
-MODULE_VERSION(MPSC_VERSION);
 MODULE_LICENSE("GPL");
-MODULE_ALIAS_CHARDEV_MAJOR(MPSC_MAJOR);
-MODULE_ALIAS("platform:" MPSC_CTLR_NAME);
+*/
-- 
2.5.0

Re: [PATCH 4/5] drivers/tty: make serial/mpsc.c driver explicitly non-modular

[Thierry Reding]

[-- Attachment #1: Type: text/plain, Size: 2236 bytes --]

On Wed, Aug 19, 2015 at 05:48:08PM -0400, Paul Gortmaker wrote:
> The Kconfig for this driver is currently:
> 
> config SERIAL_MPSC
>         bool "Marvell MPSC serial port 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.
> 
> Since module_init translates to device_initcall in the non-modular
> case, the init ordering remains unchanged with this commit.
> 
> We leave some tags like MODULE_AUTHOR for documentation purposes.
> 
> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
> Cc: Jiri Slaby <jslaby@suse.com>
> Cc: linux-serial@vger.kernel.org
> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
> ---
>  drivers/tty/serial/mpsc.c | 36 +++---------------------------------
>  1 file changed, 3 insertions(+), 33 deletions(-)
> 
> diff --git a/drivers/tty/serial/mpsc.c b/drivers/tty/serial/mpsc.c
> index 82bb6d1fe23b..edb32e3f1e84 100644
> --- a/drivers/tty/serial/mpsc.c
> +++ b/drivers/tty/serial/mpsc.c
> @@ -55,8 +55,6 @@
>  #define SUPPORT_SYSRQ
>  #endif
>  
> -#include <linux/module.h>
> -#include <linux/moduleparam.h>
>  #include <linux/tty.h>
>  #include <linux/tty_flip.h>
>  #include <linux/ioport.h>
> @@ -2108,24 +2106,8 @@ static int mpsc_drv_probe(struct platform_device *dev)
>  	return rc;
>  }
>  
> -static int mpsc_drv_remove(struct platform_device *dev)
> -{
> -	pr_debug("mpsc_drv_exit: Removing MPSC %d\n", dev->id);
> -
> -	if (dev->id < MPSC_NUM_CTLRS) {
> -		uart_remove_one_port(&mpsc_reg, &mpsc_ports[dev->id].port);
> -		mpsc_release_port((struct uart_port *)
> -				&mpsc_ports[dev->id].port);
> -		mpsc_drv_unmap_regs(&mpsc_ports[dev->id]);
> -		return 0;
> -	} else {
> -		return -ENODEV;
> -	}
> -}
> -
>  static struct platform_driver mpsc_driver = {
>  	.probe	= mpsc_drv_probe,
> -	.remove	= mpsc_drv_remove,
>  	.driver	= {
>  		.name	= MPSC_CTLR_NAME,
>  	},

I don't think this is right. The driver can always be unbound from the
device via sysfs, in which case it will now leak resources and leave
behind a dangling UART port.

Thierry

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]

Re: [PATCH 4/5] drivers/tty: make serial/mpsc.c driver explicitly non-modular

[Paul Gortmaker]

[Re: [PATCH 4/5] drivers/tty: make serial/mpsc.c driver explicitly non-modular] On 09/10/2015 (Fri 15:38) Thierry Reding wrote:

> On Wed, Aug 19, 2015 at 05:48:08PM -0400, Paul Gortmaker wrote:

[...]

> >  static struct platform_driver mpsc_driver = {
> >  	.probe	= mpsc_drv_probe,
> > -	.remove	= mpsc_drv_remove,
> >  	.driver	= {
> >  		.name	= MPSC_CTLR_NAME,
> >  	},
> 
> I don't think this is right. The driver can always be unbound from the
> device via sysfs, in which case it will now leak resources and leave
> behind a dangling UART port.

I suppose it does open the window for root to do something else stupid
he shouldn't be doing anyway...

I think the right thing to do here is for me to send a follow on patch
that does also set:

	.suppress_bind_attrs = true

in the above block, since this is a console device used on 2005 vintage
74xx based powerpc embedded targets, and is essentially always used in
conjunction with SERIAL_MPSC_CONSOLE=y -- there is no sane reason anyone
would ever want to unbind the builtin driver and lose the console.

I'll send a follow on patch doing the above if folks are in agreement
with the above logic.

Thanks,
Paul.
--

> 
> Thierry