litmus-rt.git/drivers/video/matrox, branch tracing-devel

matroxfb: get rid of CONFIG_FB_MATROX_32MB

2009-09-23T14:39:57+00:00

CONFIG_FB_MATROX_32MB is always enabled, so there is no point in having
ifdefs all around.  And it is bad practice to use CONFIG_* as a name for
something which is not a Kconfig option.

Signed-off-by: Jean Delvare 
Acked-by: Petr Vandrovec 
Cc: Krzysztof Helt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

matroxfb: get rid of unneeded macro MINFO_FROM

2009-09-23T14:39:57+00:00

With multihead support always enabled, macros MINFO_FROM and
MINFO_FROM_INFO are no longer needed and make the code harder to read.

Signed-off-by: Jean Delvare 
Acked-by: Petr Vandrovec 
Cc: Krzysztof Helt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

matroxfb: get rid of unneeded macros WPMINFO and friends

2009-09-23T14:39:57+00:00

With multihead support always enabled, these macros are no longer needed
and make the code harder to read.

Signed-off-by: Jean Delvare 
Acked-by: Petr Vandrovec 
Cc: Krzysztof Helt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

matroxfb: get rid of unneeded macros ACCESS_FBINFO and MINFO

2009-09-23T14:39:57+00:00

With multihead support always enabled, these macros are no longer needed
and make the code harder to read.

Signed-off-by: Jean Delvare 
Acked-by: Petr Vandrovec 
Cc: Krzysztof Helt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

matroxfb: make CONFIG_FB_MATROX_MULTIHEAD=y mandatory

2009-09-23T14:39:56+00:00

I would like to get rid of option CONFIG_FB_MATROX_MULTIHEAD and just
always enable it.  There are many reasons for doing this:

* CONFIG_FB_MATROX_MULTIHEAD=y is what all x86 distributions do, so it
  definitely works or we would know by now.

* Building the matroxfb driver with CONFIG_FB_MATROX_MULTIHEAD not set
  results in the following build warning:

drivers/video/matrox/matroxfb_crtc2.c: In function 'matroxfb_dh_open':
drivers/video/matrox/matroxfb_crtc2.c:265: warning: the address of 'matroxfb_global_mxinfo' will always evaluate as 'true'
drivers/video/matrox/matroxfb_crtc2.c: In function 'matroxfb_dh_release':
drivers/video/matrox/matroxfb_crtc2.c:285: warning: the address of 'matroxfb_global_mxinfo' will always evaluate as 'true'

This is nothing to be worried about, the driver will work fine, but build
warnings are still annoying.

* The trick to get multihead support without CONFIG_FB_MATROX_MULTIHEAD,
  which is described in the config help text, no longer works: you can't
  load the same kernel module more than once.

* I fail to see how CONFIG_FB_MATROX_MULTIHEAD=y would make the code
  significantly slower, contrary to what the help text says.  A few extra
  parameters on the stack here and there can't really slow things down in
  comaprison to the rest of the code, and register access.

* The driver built without CONFIG_FB_MATROX_MULTIHEAD is larger than the
  driver build with CONFIG_FB_MATROX_MULTIHEAD=y by 8%.

* One less configuration option makes things simpler.  We add options
  all the time, being able to remove one for once is nice.  It improves
  testing coverage.  And I don't think the Matrox adapters are still
  popular enough to warrant overdetailed configuration settings.

* We should be able to unobfuscate the driver code quite a bit after
  this change (patches follow.)

Signed-off-by: Jean Delvare 
Acked-by: Petr Vandrovec 
Cc: Krzysztof Helt 
Signed-off-by: Andrew Morton 
Signed-off-by: Linus Torvalds

Remove multiple KERN_ prefixes from printk formats

2009-07-08T17:30:03+00:00

Commit 5fd29d6ccbc98884569d6f3105aeca70858b3e0f ("printk: clean up
handling of log-levels and newlines") changed printk semantics.  printk
lines with multiple KERN_ prefixes are no longer emitted as
before the patch.

 is now included in the output on each additional use.

Remove all uses of multiple KERN_s in formats.

Signed-off-by: Joe Perches 
Signed-off-by: Linus Torvalds

Revert "fb: Initialize fb_info mutexes in framebuffer_alloc()"

2009-07-08T16:20:11+00:00

This reverts commit 4148df9b0f38bdd362dd91d52076926c11cbe5a9.

Let's hope that the mm_lock initialization is now correct with all
drivers, following Krzysztof's patches.

Requested-by: Krzysztof Helt 
Signed-off-by: Linus Torvalds

matroxfb: fix regression with uninitalized fb_info->mm_lock mutex (second head)

2009-07-08T16:19:50+00:00

Remove redundant locking by the mm_lock mutex before a second head of
matrox framebuffer is registered.

This fixes a problem with uninitialized the fb_info->mm_lock mutex
introduced by the commit 537a1bf059f " fbdev: add mutex for fb_mmap
locking"

Signed-off-by: Krzysztof Helt 
Signed-off-by: Linus Torvalds

matroxfb: fix regression with uninitalized fb_info->mm_lock mutex

2009-07-08T16:19:49+00:00

Remove redundant call to the matroxfb_update_fix() before matrox
frambuffer is registered.

This fixes a problem with uninitialized the fb_info->mm_lock mutex
introduced by the commit 537a1bf059f " fbdev: add mutex for fb_mmap
locking"

Signed-off-by: Krzysztof Helt 
Signed-off-by: Linus Torvalds

fb: Initialize fb_info mutexes in framebuffer_alloc()

2009-07-06T21:02:50+00:00

This way they'll be properly initialized early enough for users that may
touch them before the framebuffer has been registered.

Drivers that allocate their fb_info structure some other way (like
matrocfb's broken static allocation) need to be fixed up appropriately.

Signed-off-by: Paul Mundt 
Signed-off-by: Linus Torvalds