Skip to contentSkip to conversation

Allow core fb to be built as a module

PATCH In review Jon Smirl 4 participants 14 messages Feb 24, 2005 – Mar 2, 2005 Framebuffer layer

lore
Share & tools
lore

Jon Smirl

author
#0 lore raw

PATCH 1/1 Allow core fb to be built as a module

I'm starting to get too many patches hanging around. Is there any reason this one can't be pushed up stream? My next target would be the sysfs patch.

Signature · 71 lines
Jon Smirl
[email protected]


diff -Nru a/drivers/video/Kconfig b/drivers/video/Kconfig
--- a/drivers/video/Kconfig	2005-02-17 19:11:37 -05:00
+++ b/drivers/video/Kconfig	2005-02-17 19:11:37 -05:00
@@ -5,7 +5,7 @@
 menu "Graphics support"
 
 config FB
-	bool "Support for frame buffer devices"
+	tristate "Support for frame buffer devices"
 	---help---
 	  The frame buffer device provides an abstraction for the graphics
 	  hardware. It represents the frame buffer of some video hardware and
diff -Nru a/drivers/video/Makefile b/drivers/video/Makefile
--- a/drivers/video/Makefile	2005-02-17 19:11:37 -05:00
+++ b/drivers/video/Makefile	2005-02-17 19:11:37 -05:00
@@ -8,11 +8,13 @@
 obj-$(CONFIG_LOGO)		  += logo/
 obj-$(CONFIG_SYSFS)		  += backlight/
 
-obj-$(CONFIG_FB)                  += fbmem.o fbmon.o fbcmap.o
fbsysfs.o modedb.o softcursor.o
+obj-$(CONFIG_FB)                  += fb.o
+fb-y                              := fbmem.o fbmon.o fbcmap.o
fbsysfs.o modedb.o softcursor.o
 # Only include macmodes.o if we have FB support and are PPC
-ifeq ($(CONFIG_FB),y)
-obj-$(CONFIG_PPC)                 += macmodes.o
+ifneq ($(CONFIG_FB),n)
+fb-$(CONFIG_PPC)                  += macmodes.o
 endif
+fb-objs                           := $(fb-y)
 
 # Hardware specific drivers go first
 obj-$(CONFIG_FB_RETINAZ3)         += retz3fb.o
diff -Nru a/drivers/video/fbmem.c b/drivers/video/fbmem.c
--- a/drivers/video/fbmem.c	2005-02-17 19:11:37 -05:00
+++ b/drivers/video/fbmem.c	2005-02-17 19:11:37 -05:00
@@ -1204,7 +1204,22 @@
 	}
 	return 0;
 }
+
+void __exit
+fbmem_exit(void)
+{
+	class_simple_destroy(fb_class);
+}
+
+#ifdef MODULE
+module_init(fbmem_init);
+module_exit(fbmem_exit);
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Framebuffer base");
+#else
 subsys_initcall(fbmem_init);
+subsys_exitcall(fbmem_exit);
+#endif
 
 static char *video_options[FB_MAX];
 static int ofonly;


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
#1 lore raw
Quoted 3 lines · Jon Smirl

I'm starting to get too many patches hanging around. Is there any reason this one can't be pushed up stream? My next target would be the sysfs patch.

Wait until 2.6.11 comes out. Linus will not accept, and Andrew will not push patches unless they're bug fixes.

Tony

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

author
#2 lore raw

Can I assume then that it is acceptable and all we are waiting on is 2.6.11?

Quoted 11 lines · Antonino A. Daplas
Quoted 3 lines · Jon Smirl

I'm starting to get too many patches hanging around. Is there any reason this one can't be pushed up stream? My next target would be the sysfs patch.

Wait until 2.6.11 comes out. Linus will not accept, and Andrew will not push patches unless they're bug fixes.

Tony

Signature · 9 lines
Jon Smirl
[email protected]


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
#3 lore raw
Quoted 2 lines · Jon Smirl

Can I assume then that it is acceptable and all we are waiting on is 2.6.11?

You can assume that it will get to the mm tree for a trial run, and if there are no problems, will get pushed to Linus' tree.

Tony

PS: I will also do some testing myself this weekend.

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

#4 lore raw
Quoted 10 lines
Quoted 3 lines

Can I assume then that it is acceptable and all we are waiting on is 2.6.11?

You can assume that it will get to the mm tree for a trial run, and if there are no problems, will get pushed to Linus' tree.

Tony

PS: I will also do some testing myself this weekend.

Yeap. The process is first make patch and send to list for testing. People comment and some changes happen. This can happen several times. Once the patch is acceptable then it going into Andrew's tree for wider range testing. Once Andrew determines that the patch is safe he sends it to linus. It is a sure way to make sure the patches are of the highest quality.

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

author
#5 lore raw
Quoted 6 lines · James Simmons

Yeap. The process is first make patch and send to list for testing. People comment and some changes happen. This can happen several times. Once the patch is acceptable then it going into Andrew's tree for wider range testing. Once Andrew determines that the patch is safe he sends it to linus. It is a sure way to make sure the patches are of the highest quality.

There is already code of mine in the kernel so I have been through the process. Let's get the process moving on these patches so that they will be in place by the time OLS comes around. This patch is small and non-controversial. Let's get it started moving forward so that we can spend time on the complex ones. The complex ones will also need to get hammered out on LKML since they involve complicated hotplug code.

The next patch will be the sysfs one but it is not ready yet. After that is the reset one.
Next will be secondary head support in the radeon driver.

These three will require a lot more code review.

Signature · 9 lines
Jon Smirl
[email protected]


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
#6 lore raw
Quoted 8 lines

There is already code of mine in the kernel so I have been through the process. Let's get the process moving on these patches so that they will be in place by the time OLS comes around. This patch is small and non-controversial. Let's get it started moving forward so that we can spend time on the complex ones. The complex ones will also need to get hammered out on LKML since they involve complicated hotplug code.

Quoted 5 lines

The next patch will be the sysfs one but it is not ready yet. After that is the reset one.
Next will be secondary head support in the radeon driver.

These three will require a lot more code review.

Have patience. Also if your patch doesn't go in time you still can demo a patched kernel. When I first started out maintaning the fbdev layer I made radical changes all the time. This just got people upset. Now I take my time and make sure everything works and its clean.

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

#7 lore raw
In reply to Jon Smirl

Tony are you okay with this? I have no problem myself. I haven't tested with various driver with various conditions yet.

Quoted 72 lines · Jon Smirl

I'm starting to get too many patches hanging around. Is there any reason this one can't be pushed up stream? My next target would be the sysfs patch.

Signature · 5 lines
Jon Smirl
[email protected]


diff -Nru a/drivers/video/Kconfig b/drivers/video/Kconfig
drivers/video/Kconfig +1 -1
@@ -5,7 +5,7 @@
 menu "Graphics support"
 
 config FB
-	bool "Support for frame buffer devices"
+	tristate "Support for frame buffer devices"
 	---help---
 	  The frame buffer device provides an abstraction for the graphics
 	  hardware. It represents the frame buffer of some video hardware and

diff -Nru a/drivers/video/Makefile b/drivers/video/Makefile

drivers/video/Makefile +0 -1
@@ -8,11 +8,13 @@
 obj-$(CONFIG_LOGO)		  += logo/
 obj-$(CONFIG_SYSFS)		  += backlight/
 
-obj-$(CONFIG_FB)                  += fbmem.o fbmon.o fbcmap.o

fbsysfs.o modedb.o softcursor.o

drivers/video/Makefile +2 -0
+obj-$(CONFIG_FB)                  += fb.o
+fb-y                              := fbmem.o fbmon.o fbcmap.o

fbsysfs.o modedb.o softcursor.o
# Only include macmodes.o if we have FB support and are PPC

drivers/video/Makefile +3 -2
-ifeq ($(CONFIG_FB),y)
-obj-$(CONFIG_PPC)                 += macmodes.o
+ifneq ($(CONFIG_FB),n)
+fb-$(CONFIG_PPC)                  += macmodes.o
 endif
+fb-objs                           := $(fb-y)
 
 # Hardware specific drivers go first
 obj-$(CONFIG_FB_RETINAZ3)         += retz3fb.o

diff -Nru a/drivers/video/fbmem.c b/drivers/video/fbmem.c

drivers/video/fbmem.c +15 -1
@@ -1204,7 +1204,22 @@
 	}
 	return 0;
 }
+
+void __exit
+fbmem_exit(void)
+{
+	class_simple_destroy(fb_class);
+}
+
+#ifdef MODULE
+module_init(fbmem_init);
+module_exit(fbmem_exit);
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Framebuffer base");
+#else
 subsys_initcall(fbmem_init);
+subsys_exitcall(fbmem_exit);
+#endif
 
 static char *video_options[FB_MAX];
 static int ofonly;
 
 
-------------------------------------------------------
Quoted 9 lines

SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
_______________________________________________
Linux-fbdev-devel mailing list [email protected]
https://lists.sourceforge.net/lists/listinfo/linux-fbdev-devel

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

#8 lore raw
Quoted 3 lines · James Simmons

Tony are you okay with this? I have no problem myself. I haven't tested with various driver with various conditions yet.

It does have a few problems. Drivers that are bool, instead of tristate such as vesafb, can still be selected even if fbcore is selected as module.

Tony

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

author
#9 lore raw
Quoted 7 lines · Antonino A. Daplas
Quoted 3 lines · James Simmons

Tony are you okay with this? I have no problem myself. I haven't tested with various driver with various conditions yet.

It does have a few problems. Drivers that are bool, instead of tristate such as vesafb, can still be selected even if fbcore is selected as module.

Then it is needs (FB = y) like this...
If you leave them as checks they will be built and linked into the FB module.

[jonsmirl@jonsmirl video]$ bk diffs -u
===== Kconfig 1.76 vs edited =====

Kconfig 2005-03-02 12:25:28 -05:00 renamed from drivers/video/Kconfig 2005-02-17 19:09:04 -05:00 +3 -3
@@ -269,11 +269,11 @@
 
 config FB_ASILIANT
        bool "Chips 69000 display support"
-       depends on FB && PCI
+       depends on (FB = y) && PCI
 
 config FB_IMSTT
        bool "IMS Twin Turbo display support"
-       depends on FB && PCI
+       depends on (FB = y) && PCI
        help
          The IMS Twin Turbo is a PCI-based frame buffer card bundled with
          many Macintosh and compatible computers.
@@ -330,7 +330,7 @@
 
 config FB_VESA
        bool "VESA VGA graphics support"
-       depends on FB && (X86 || X86_64)
+       depends on (FB = y) && (X86 || X86_64)
        help
          This is the frame buffer device driver for generic VESA 2.0
          compliant graphic cards. The older VESA 1.2 cards are not supported.

[jonsmirl@jonsmirl video]$

Signature · 9 lines
Jon Smirl
[email protected]


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
#10 lore raw
Quoted 6 lines
Quoted 3 lines

It does have a few problems. Drivers that are bool, instead of tristate such as vesafb, can still be selected even if fbcore is selected as module.

Then it is needs (FB = y) like this...
If you leave them as checks they will be built and linked into the FB module.

Send another patch with the below and we will include it.

Quoted 27 lines

[jonsmirl@jonsmirl video]$ bk diffs -u
===== Kconfig 1.76 vs edited =====

Kconfig 2005-03-02 12:25:28 -05:00 renamed from drivers/video/Kconfig 2005-02-17 19:09:04 -05:00 +3 -3
@@ -269,11 +269,11 @@
 
 config FB_ASILIANT
        bool "Chips 69000 display support"
-       depends on FB && PCI
+       depends on (FB = y) && PCI
 
 config FB_IMSTT
        bool "IMS Twin Turbo display support"
-       depends on FB && PCI
+       depends on (FB = y) && PCI
        help
          The IMS Twin Turbo is a PCI-based frame buffer card bundled with
          many Macintosh and compatible computers.
@@ -330,7 +330,7 @@
 
 config FB_VESA
        bool "VESA VGA graphics support"
-       depends on FB && (X86 || X86_64)
+       depends on (FB = y) && (X86 || X86_64)
        help
          This is the frame buffer device driver for generic VESA 2.0
          compliant graphic cards. The older VESA 1.2 cards are not supported.
Quoted 7 lines

[jonsmirl@jonsmirl video]$

Signature · 2 lines
Jon Smirl
[email protected]

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

author
#11 lore raw

Patch to allow building of the fbdev core as a module.

Signature · 93 lines
Jon Smirl
[email protected]

diff -Nru a/drivers/video/Kconfig b/drivers/video/Kconfig
--- a/drivers/video/Kconfig	2005-03-02 12:51:46 -05:00
+++ b/drivers/video/Kconfig	2005-03-02 12:51:46 -05:00
@@ -5,7 +5,7 @@
 menu "Graphics support"
 
 config FB
-	bool "Support for frame buffer devices"
+	tristate "Support for frame buffer devices"
 	---help---
 	  The frame buffer device provides an abstraction for the graphics
 	  hardware. It represents the frame buffer of some video hardware and
@@ -269,11 +269,11 @@
 
 config FB_ASILIANT
 	bool "Chips 69000 display support"
-	depends on FB && PCI
+	depends on (FB = y) && PCI
 
 config FB_IMSTT
 	bool "IMS Twin Turbo display support"
-	depends on FB && PCI
+	depends on (FB = y) && PCI
 	help
 	  The IMS Twin Turbo is a PCI-based frame buffer card bundled with
 	  many Macintosh and compatible computers.
@@ -330,7 +330,7 @@
 
 config FB_VESA
 	bool "VESA VGA graphics support"
-	depends on FB && (X86 || X86_64)
+	depends on (FB = y) && (X86 || X86_64)
 	help
 	  This is the frame buffer device driver for generic VESA 2.0
 	  compliant graphic cards. The older VESA 1.2 cards are not supported.
diff -Nru a/drivers/video/Makefile b/drivers/video/Makefile
--- a/drivers/video/Makefile	2005-03-02 12:51:46 -05:00
+++ b/drivers/video/Makefile	2005-03-02 12:51:46 -05:00
@@ -8,11 +8,13 @@
 obj-$(CONFIG_LOGO)		  += logo/
 obj-$(CONFIG_SYSFS)		  += backlight/
 
-obj-$(CONFIG_FB)                  += fbmem.o fbmon.o fbcmap.o
fbsysfs.o modedb.o softcursor.o
+obj-$(CONFIG_FB)                  += fb.o
+fb-y                              := fbmem.o fbmon.o fbcmap.o
fbsysfs.o modedb.o softcursor.o
 # Only include macmodes.o if we have FB support and are PPC
-ifeq ($(CONFIG_FB),y)
-obj-$(CONFIG_PPC)                 += macmodes.o
+ifneq ($(CONFIG_FB),n)
+fb-$(CONFIG_PPC)                  += macmodes.o
 endif
+fb-objs                           := $(fb-y)
 
 # Hardware specific drivers go first
 obj-$(CONFIG_FB_RETINAZ3)         += retz3fb.o
diff -Nru a/drivers/video/fbmem.c b/drivers/video/fbmem.c
--- a/drivers/video/fbmem.c	2005-03-02 12:51:46 -05:00
+++ b/drivers/video/fbmem.c	2005-03-02 12:51:46 -05:00
@@ -1204,7 +1204,22 @@
 	}
 	return 0;
 }
+
+void __exit
+fbmem_exit(void)
+{
+	class_simple_destroy(fb_class);
+}
+
+#ifdef MODULE
+module_init(fbmem_init);
+module_exit(fbmem_exit);
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Framebuffer base");
+#else
 subsys_initcall(fbmem_init);
+subsys_exitcall(fbmem_exit);
+#endif
 
 static char *video_options[FB_MAX];
 static int ofonly;


-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users.
Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
author
#12 lore raw

Patch to allow building of the fbdev core as a module.

Forgot to fix up non-x86. The algorithm is that any FB driver that is not tristate needs (FB = y) instead of FB. It is really minor to convert these non-module drivers to module based ones It seems to have taken Paul about 5 minutes to convert his.

--
Jon Smirl
[email protected]

drivers/video/Kconfig +25 -26
diff -Nru a/drivers/video/Kconfig b/drivers/video/Kconfig
@@ -5,7 +5,7 @@
 menu "Graphics support"
 
 config FB
-	bool "Support for frame buffer devices"
+	tristate "Support for frame buffer devices"
 	---help---
 	  The frame buffer device provides an abstraction for the graphics
 	  hardware. It represents the frame buffer of some video hardware and
@@ -112,7 +112,7 @@
 
 config FB_ACORN
 	bool "Acorn VIDC support"
-	depends on FB && ARM && ARCH_ACORN
+	depends on (FB = y) && ARM && ARCH_ACORN
 	help
 	  This is the frame buffer device driver for the Acorn VIDC graphics
 	  hardware found in Acorn RISC PCs and other ARM-based machines.  If
@@ -120,11 +120,11 @@
 
 config FB_CLPS711X
 	bool "CLPS711X LCD support"
-	depends on FB && ARM && ARCH_CLPS711X
+	depends on (FB = y) && ARM && ARCH_CLPS711X
 
 config FB_SA1100
 	bool "SA-1100 LCD support"
-	depends on FB && ARM && ARCH_SA1100
+	depends on (FB = y) && ARM && ARCH_SA1100
 	help
 	  This is a framebuffer device for the SA-1100 LCD Controller.
 	  See <http://www.linux-fbdev.org/> for information on framebuffer
@@ -144,12 +144,12 @@
 
 config FB_APOLLO
 	bool
-	depends on FB && APOLLO
+	depends on (FB = y) && APOLLO
 	default y
 
 config FB_Q40
 	bool
-	depends on FB && Q40
+	depends on (FB = y) && Q40
 	default y
 
 config FB_AMIGA
@@ -201,7 +201,7 @@
 
 config FB_VIRGE
 	bool "Amiga CyberVision 64/3D support "
-	depends on FB && ZORRO && BROKEN
+	depends on (FB = y) && ZORRO && BROKEN
 	help
 	  This enables support for the Cybervision 64/3D graphics card from
 	  Phase5. Please note that its use is not all that intuitive (i.e. if
@@ -212,7 +212,7 @@
 
 config FB_RETINAZ3
 	tristate "Amiga Retina Z3 support"
-	depends on FB && ZORRO && BROKEN
+	depends on (FB = y) && ZORRO && BROKEN
 	help
 	  This enables support for the Retina Z3 graphics card. Say N unless
 	  you have a Retina Z3 or plan to get one before you next recompile
@@ -220,67 +220,67 @@
 
 config FB_FM2
 	bool "Amiga FrameMaster II/Rainbow II support"
-	depends on FB && ZORRO
+	depends on (FB = y) && ZORRO
 	help
 	  This is the frame buffer device driver for the Amiga FrameMaster
 	  card from BSC (exhibited 1992 but not shipped as a CBM product).
 
 config FB_ATARI
 	bool "Atari native chipset support"
-	depends on FB && ATARI && BROKEN
+	depends on (FB = y) && ATARI && BROKEN
 	help
 	  This is the frame buffer device driver for the builtin graphics
 	  chipset found in Ataris.
 
 config FB_OF
 	bool "Open Firmware frame buffer device support"
-	depends on FB && (PPC64 || PPC_OF)
+	depends on (FB = y) && (PPC64 || PPC_OF)
 	help
 	  Say Y if you want support with Open Firmware for your graphics
 	  board.
 
 config FB_CONTROL
 	bool "Apple \"control\" display support"
-	depends on FB && PPC_PMAC
+	depends on (FB = y) && PPC_PMAC
 	help
 	  This driver supports a frame buffer for the graphics adapter in the
 	  Power Macintosh 7300 and others.
 
 config FB_PLATINUM
 	bool "Apple \"platinum\" display support"
-	depends on FB && PPC_PMAC
+	depends on (FB = y) && PPC_PMAC
 	help
 	  This driver supports a frame buffer for the "platinum" graphics
 	  adapter in some Power Macintoshes.
 
 config FB_VALKYRIE
 	bool "Apple \"valkyrie\" display support"
-	depends on FB && (MAC || PPC_PMAC)
+	depends on (FB = y) && (MAC || PPC_PMAC)
 	help
 	  This driver supports a frame buffer for the "valkyrie" graphics
 	  adapter in some Power Macintoshes.
 
 config FB_CT65550
 	bool "Chips 65550 display support"
-	depends on FB && PPC
+	depends on (FB = y) && PPC
 	help
 	  This is the frame buffer device driver for the Chips & Technologies
 	  65550 graphics chip in PowerBooks.
 
 config FB_ASILIANT
 	bool "Chips 69000 display support"
-	depends on FB && PCI
+	depends on (FB = y) && PCI
 
 config FB_IMSTT
 	bool "IMS Twin Turbo display support"
-	depends on FB && PCI
+	depends on (FB = y) && PCI
 	help
 	  The IMS Twin Turbo is a PCI-based frame buffer card bundled with
 	  many Macintosh and compatible computers.
 
 config FB_S3TRIO
 	bool "S3 Trio display support"
-	depends on FB && PPC && BROKEN
+	depends on (FB = y) && PPC && BROKEN
 	help
 	  If you have a S3 Trio say Y. Say N for S3 Virge.
 
@@ -313,12 +313,12 @@
 
 config FB_MAC
 	bool "Generic Macintosh display support"
-	depends on FB && MAC
+	depends on (FB = y) && MAC
 
 #      bool '  Apple DAFB display support' CONFIG_FB_DAFB
 config FB_HP300
 	bool
-	depends on FB && HP300
+	depends on (FB = y) && HP300
 	default y
 
 config FB_TGA
@@ -330,7 +330,7 @@
 
 config FB_VESA
 	bool "VESA VGA graphics support"
-	depends on FB && (X86 || X86_64)
+	depends on (FB = y) && (X86 || X86_64)
 	help
 	  This is the frame buffer device driver for generic VESA 2.0
 	  compliant graphic cards. The older VESA 1.2 cards are not supported.
@@ -364,7 +364,7 @@
 
 config VIDEO_SELECT
 	bool
-	depends on FB && X86
+	depends on (FB = y) && X86
 	default y
 
 config FB_SGIVW
@@ -375,7 +375,7 @@
 
 config FB_GBE
 	bool "SGI Graphics Backend frame buffer support"
-	depends on FB && (SGI_IP32 || X86_VISWS)
+	depends on (FB = y) && (SGI_IP32 || X86_VISWS)
  	help
 	  This is the frame buffer device driver for SGI Graphics Backend.
 	  This chip is used in SGI O2 and Visual Workstation 320/540.
@@ -390,28 +390,28 @@
 
 config BUS_I2C
 	bool
-	depends on FB && VISWS
+	depends on (FB = y) && VISWS
 	default y
 
 config FB_SUN3
 	bool "Sun3 framebuffer support"
-	depends on FB && (SUN3 || SUN3X) && BROKEN
+	depends on (FB = y) && (SUN3 || SUN3X) && BROKEN
 
 config FB_BW2
 	bool "BWtwo support"
-	depends on FB && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 || SUN3X)

&& FB_SUN3)
+ depends on (FB = y) && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 ||
SUN3X) && FB_SUN3)
help
This is the frame buffer device driver for the BWtwo frame buffer.

config FB_CG3
bool "CGthree support"
- depends on FB && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 || SUN3X)
&& FB_SUN3)
+ depends on (FB = y) && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 ||
SUN3X) && FB_SUN3)
help
This is the frame buffer device driver for the CGthree frame buffer.

config FB_CG6
bool "CGsix (GX,TurboGX) support"
- depends on FB && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 || SUN3X)
&& FB_SUN3)
+ depends on (FB = y) && ((SPARC32 || SPARC64) && FB_SBUS || (SUN3 ||
SUN3X) && FB_SUN3)
help
This is the frame buffer device driver for the CGsix (GX, TurboGX) frame buffer.
@@ -436,7 +436,7 @@

config FB_EPSON1355
bool "Epson 1355 framebuffer support"
- depends on FB && (SUPERH || ARCH_CEIVA)
+ depends on (FB = y) && (SUPERH || ARCH_CEIVA)
help
Build in support for the SED1355 Epson Research Embedded RAMDAC LCD/CRT Controller (since redesignated as the S1D13505) as a
@@ -953,11 +953,11 @@

config FB_AU1100
bool "Au1100 LCD Driver"
- depends on FB && EXPERIMENTAL && PCI && MIPS && MIPS_PB1100=y
+ depends on (FB = y) && EXPERIMENTAL && PCI && MIPS && MIPS_PB1100=y

config FB_SBUS
bool "SBUS and UPA framebuffers"
- depends on FB && (SPARC32 || SPARC64)
+ depends on (FB = y) && (SPARC32 || SPARC64)
help
Say Y if you want support for SBUS or UPA based frame buffer device.

@@ -998,7 +998,7 @@

config FB_PCI
bool "PCI framebuffers"
- depends on FB && PCI && (SPARC64 || SPARC32)
+ depends on (FB = y) && PCI && (SPARC64 || SPARC32)

config FB_IGA
bool "IGA 168x display support"
@@ -1016,21 +1016,21 @@

config FB_PMAG_AA
bool "PMAG-AA TURBOchannel framebuffer support"
- depends on FB && MACH_DECSTATION && TC
+ depends on (FB = y) && MACH_DECSTATION && TC
help
Support for the PMAG-AA TURBOchannel framebuffer card (1280x1024x1) used mainly in the MIPS-based DECstation series.

config FB_PMAG_BA
bool "PMAG-BA TURBOchannel framebuffer support"
- depends on FB && MACH_DECSTATION && TC
+ depends on (FB = y) && MACH_DECSTATION && TC
help
Support for the PMAG-BA TURBOchannel framebuffer card (1024x864x8) used mainly in the MIPS-based DECstation series.

config FB_PMAGB_B
bool "PMAGB-B TURBOchannel framebuffer support"
- depends on FB && MACH_DECSTATION && TC
+ depends on (FB = y) && MACH_DECSTATION && TC
help
Support for the PMAGB-B TURBOchannel framebuffer card used mainly in the MIPS-based DECstation series. The card is currently only
@@ -1038,7 +1038,7 @@

config FB_MAXINE
bool "Maxine (Personal DECstation) onboard framebuffer support"
- depends on FB && MACH_DECSTATION && TC
+ depends on (FB = y) && MACH_DECSTATION && TC
help
Support for the onboard framebuffer (1024x768x8) in the Personal DECstation series (Personal DECstation 5000/20, /25, /33, /50,
@@ -1046,7 +1046,7 @@

config FB_TX3912
bool "TMPTX3912/PR31700 frame buffer support"
- depends on FB && NINO
+ depends on (FB = y) && NINO
help
The TX3912 is a Toshiba RISC processor based on the MIPS 3900 core see <http://www.toshiba.com/taec/components/Generic/risc/tx3912.htm>.

drivers/video/Makefile +0 -1
diff -Nru a/drivers/video/Makefile b/drivers/video/Makefile
@@ -8,11 +8,13 @@
 obj-$(CONFIG_LOGO)		  += logo/
 obj-$(CONFIG_SYSFS)		  += backlight/
 
-obj-$(CONFIG_FB)                  += fbmem.o fbmon.o fbcmap.o

fbsysfs.o modedb.o softcursor.o
+obj-$(CONFIG_FB) += fb.o
+fb-y := fbmem.o fbmon.o fbcmap.o
fbsysfs.o modedb.o softcursor.o
# Only include macmodes.o if we have FB support and are PPC
-ifeq ($(CONFIG_FB),y)
-obj-$(CONFIG_PPC) += macmodes.o
+ifneq ($(CONFIG_FB),n)
+fb-$(CONFIG_PPC) += macmodes.o
endif
+fb-objs := $(fb-y)

# Hardware specific drivers go first
obj-$(CONFIG_FB_RETINAZ3) += retz3fb.o

drivers/video/fbmem.c +15 -0
diff -Nru a/drivers/video/fbmem.c b/drivers/video/fbmem.c
@@ -1204,7 +1204,22 @@
 	}
 	return 0;
 }
+
+void __exit
+fbmem_exit(void)
+{
+	class_simple_destroy(fb_class);
+}
+
+#ifdef MODULE
+module_init(fbmem_init);
+module_exit(fbmem_exit);
+MODULE_LICENSE("GPL");
+MODULE_DESCRIPTION("Framebuffer base");
+#else
 subsys_initcall(fbmem_init);
+subsys_exitcall(fbmem_exit);
+#endif
 
 static char *video_options[FB_MAX];
 static int ofonly;

-------------------------------------------------------
SF email is sponsored by - The IT Product Guide
Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now.
http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click

#13 lore raw
In reply to Jon Smirl
Quoted 8 lines · Jon Smirl

Then it is needs (FB = y) like this...
If you leave them as checks they will be built and linked into the FB module.

[jonsmirl@jonsmirl video]$ bk diffs -u
===== Kconfig 1.76 vs edited =====

Kconfig 2005-03-02 12:25:28 -05:00 line 273

Quoted 6 lines
Kconfig 2005-03-02 12:25:28 -05:00 renamed from drivers/video/Kconfig 2005-02-17 19:09:04 -05:00 +1 -1
@@ -269,11 +269,11 @@
 
 config FB_ASILIANT
        bool "Chips 69000 display support"
-       depends on FB && PCI
+       depends on (FB = y) && PCI
 

There's no reason asiliantfb can't be a module, lets get that cleaned up instead..

It would also be nice to get this fb_get_options() stuff done in a clean way, having all these MODULE ifdefs/ifndefs here is just as ugly as when the module_init()'s had them.

 drivers/video/Kconfig      |    2 +-
 drivers/video/asiliantfb.c |    9 ++++++---
 2 files changed, 7 insertions(+), 4 deletions(-)

===== drivers/video/Kconfig 1.74 vs edited =====

drivers/video/Kconfig +1 -1
@@ -268,7 +268,7 @@
 	  65550 graphics chip in PowerBooks.
 
 config FB_ASILIANT
-	bool "Chips 69000 display support"
+	tristate "Chips 69000 display support"
 	depends on FB && PCI
 
 config FB_IMSTT

===== drivers/video/asiliantfb.c 1.6 vs edited =====

drivers/video/asiliantfb.c +6 -3
@@ -606,17 +606,20 @@
 
 int __init asiliantfb_init(void)
 {
+#ifndef MODULE
 	if (fb_get_options("asiliantfb", NULL))
 		return -ENODEV;
+#endif
 
-	return pci_module_init(&asiliantfb_driver);
+	return pci_register_driver(&asiliantfb_driver);
 }
 
-module_init(asiliantfb_init);
-
 static void __exit asiliantfb_exit(void)
 {
 	pci_unregister_driver(&asiliantfb_driver);
 }
+
+module_init(asiliantfb_init);
+module_exit(asiliantfb_exit);
 
 MODULE_LICENSE("GPL");