[PATCH] V4L1: make PMS not autoprobe when builtin.

[PATCH] V4L1: make PMS not autoprobe when builtin.

[Rene Herman]

[-- Attachment #1: Type: text/plain, Size: 339 bytes --]

Hi Andrew.

Alternate version of the PMS patch sent yesterday. This one makes it 
need explicit enabling when builtin and doesn't change anything when 
modular as per Alan Cox's comments.

This is a deprecated, unused driver meaning it doesn't matter. It still 
fixes that (randconfig testing breakage) which it is supposed to fix.

Rene.

[-- Attachment #2: 0001-V4L1-make-PMS-not-autoprobe-when-builtin.patch --]
[-- Type: text/plain, Size: 1325 bytes --]

>From e089b11a8eee4876c2c65988a9f423a1ae111d7a Mon Sep 17 00:00:00 2001
From: Rene Herman <rene.herman@gmail.com>
Date: Sat, 9 Aug 2008 20:39:19 +0200
Subject: [PATCH] V4L1: make PMS not autoprobe when builtin.

The old Mediavision Pro Movie Studio legacy ISA V4L1 driver was
found to hang the boot during Ingo Molnar's testing of randconfig
kernels. Have it require a "pms.enable=1" kernel parameter to
enable the driver when builtin which avoids such problems.

This is a deprecated and, very likely, unused driver. Nothing
changes modular moreover.

Signed-off-by: Rene Herman <rene.herman@gmail.com>
---
 drivers/media/video/pms.c |   12 ++++++++++++
 1 files changed, 12 insertions(+), 0 deletions(-)

diff --git a/drivers/media/video/pms.c b/drivers/media/video/pms.c
index 00425d7..ede522a 100644
--- a/drivers/media/video/pms.c
+++ b/drivers/media/video/pms.c
@@ -1019,10 +1019,22 @@ static int init_mediavision(void)
  *	Initialization and module stuff
  */
 
+#ifndef MODULE
+static int enable;
+module_param(enable, int, 0);
+#endif
+
 static int __init init_pms_cards(void)
 {
 	printk(KERN_INFO "Mediavision Pro Movie Studio driver 0.02\n");
 
+#ifndef MODULE
+	if (!enable) {
+		printk(KERN_INFO "Not enabled\n");
+		return -ENODEV;
+	}
+#endif
+
 	data_port = io_port +1;
 
 	if(init_mediavision())
-- 
1.5.5

Re: [PATCH] V4L1: make PMS not autoprobe when builtin.

[Alan Cox]

On Sun, 10 Aug 2008 23:37:29 +0200
Rene Herman <rene.herman@keyaccess.nl> wrote:

> Hi Andrew.
> 
> Alternate version of the PMS patch sent yesterday. This one makes it 
> need explicit enabling when builtin and doesn't change anything when 
> modular as per Alan Cox's comments.
> 
> This is a deprecated, unused driver meaning it doesn't matter. It still 
> fixes that (randconfig testing breakage) which it is supposed to fix.


Acked-by: Alan Cox <alan@redhat.com>


Would probably make the printk "pms: not enabled, use pms.enable=1 to
probe"

So you know
a) What is wittering about not being probed
b) How to undo it.

But thats trivia really.

Re: [PATCH] V4L1: make PMS not autoprobe when builtin.

[Rene Herman]

[-- Attachment #1: Type: text/plain, Size: 830 bytes --]

On 10-08-08 23:37, Alan Cox wrote:

> On Sun, 10 Aug 2008 23:37:29 +0200
> Rene Herman <rene.herman@keyaccess.nl> wrote:

>> Alternate version of the PMS patch sent yesterday. This one makes it 
>> need explicit enabling when builtin and doesn't change anything when 
>> modular as per Alan Cox's comments.
>>
>> This is a deprecated, unused driver meaning it doesn't matter. It still 
>> fixes that (randconfig testing breakage) which it is supposed to fix.
> 
> 
> Acked-by: Alan Cox <alan@redhat.com>
> 
> 
> Would probably make the printk "pms: not enabled, use pms.enable=1 to
> probe"
> 
> So you know
> a) What is wittering about not being probed
> b) How to undo it.
> 
> But thats trivia really.

I suppose I wouldn't be messing with these drivers in the first place if 
I weren't deeply into trivia, so <bow> ...

Rene.

[-- Attachment #2: 0001-V4L1-make-PMS-not-autoprobe-when-builtin.patch --]
[-- Type: text/plain, Size: 1367 bytes --]

>From 87f37f786d4c73275feaa36ea51f9272f0b0a709 Mon Sep 17 00:00:00 2001
From: Rene Herman <rene.herman@gmail.com>
Date: Sat, 9 Aug 2008 20:39:19 +0200
Subject: [PATCH] V4L1: make PMS not autoprobe when builtin.

The old Mediavision Pro Movie Studio legacy ISA V4L1 driver was
found to hang the boot during Ingo Molnar's testing of randconfig
kernels. Have it require a "pms.enable=1" kernel parameter to
enable the driver when builtin which avoids such problems.

This is a deprecated and, very likely, unused driver. Nothing
changes modular moreover.

Signed-off-by: Rene Herman <rene.herman@gmail.com>
---
 drivers/media/video/pms.c |   13 +++++++++++++
 1 files changed, 13 insertions(+), 0 deletions(-)

diff --git a/drivers/media/video/pms.c b/drivers/media/video/pms.c
index 00425d7..7c84f79 100644
--- a/drivers/media/video/pms.c
+++ b/drivers/media/video/pms.c
@@ -1019,10 +1019,23 @@ static int init_mediavision(void)
  *	Initialization and module stuff
  */
 
+#ifndef MODULE
+static int enable;
+module_param(enable, int, 0);
+#endif
+
 static int __init init_pms_cards(void)
 {
 	printk(KERN_INFO "Mediavision Pro Movie Studio driver 0.02\n");
 
+#ifndef MODULE
+	if (!enable) {
+		printk(KERN_INFO "PMS: not enabled, use pms.enable=1 to "
+				 "probe\n");
+		return -ENODEV;
+	}
+#endif
+
 	data_port = io_port +1;
 
 	if(init_mediavision())
-- 
1.5.5