In 537a1bf05, a per-fbinfo mutex was added and framebuffer drivers were
adopted accordingly. This new lock is currently initialized in
register_framebuffer().
The mx3fb driver takes this lock in mx3fb_map_video_memory() _before_
register_framebuffer() is called and hence crashes the system very
early.
Moving this mutex_init() to framebuffer_alloc() solves this and also
seems a lot more straight forward.
Signed-off-by: Daniel Mack <[email protected]>
Cc:: Krzysztof Helt <[email protected]>
Cc: Peter Zijlstra <[email protected]>
Cc: Rafael J. Wysocki <[email protected]>
Cc: <[email protected]>
Cc: Andrew Morton <[email protected]>
---
drivers/video/fbmem.c | 3 ---
drivers/video/fbsysfs.c | 3 +++
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
index 53ea056..1d8d098 100644
--- a/drivers/video/fbmem.c
+++ b/drivers/video/fbmem.c
@@ -1513,9 +1513,6 @@ register_framebuffer(struct fb_info *fb_info)
if (!registered_fb[i])
break;
fb_info->node = i;
- mutex_init(&fb_info->lock);
- mutex_init(&fb_info->mm_lock);
-
fb_info->dev = device_create(fb_class, fb_info->device,
MKDEV(FB_MAJOR, i), NULL, "fb%d", i);
if (IS_ERR(fb_info->dev)) {
diff --git a/drivers/video/fbsysfs.c b/drivers/video/fbsysfs.c
index d4a2c11..60ddd1d 100644
--- a/drivers/video/fbsysfs.c
+++ b/drivers/video/fbsysfs.c
@@ -58,6 +58,9 @@ struct fb_info *framebuffer_alloc(size_t size, struct device *dev)
info->device = dev;
+ mutex_init(&info->lock);
+ mutex_init(&info->mm_lock);
+
#ifdef CONFIG_FB_BACKLIGHT
mutex_init(&info->bl_curve_mutex);
#endif
--
1.6.3.1
On Wed, 2009-07-01 at 18:26 +0200, Daniel Mack wrote:
> In 537a1bf05, a per-fbinfo mutex was added and framebuffer drivers were
> adopted accordingly. This new lock is currently initialized in
> register_framebuffer().
>
> The mx3fb driver takes this lock in mx3fb_map_video_memory() _before_
> register_framebuffer() is called and hence crashes the system very
> early.
>
> Moving this mutex_init() to framebuffer_alloc() solves this and also
> seems a lot more straight forward.
>
but this may introduce deadlock to some _old_ drivers which have not
used framebuffer_alloc to allocate fb_info structure.
to avoid deadlock, these drivers need to change kmalloc to
framebuffer_alloc, otherwise, they need to add their own mutexes
initialization source code:
mutex_init(&info->lock); mutex_init(&info->mm_lock);
here is only a record for the _old_ drivers' maintainers. hope they can
get this info from google or find this commit:
commit 4148df9b0f38bdd362dd91d52076926c11cbe5a9
Author: Paul Mundt <[email protected]>
Date: Mon Jul 6 00:25:57 2009 +0900
fb: Initialize fb_info mutexes in framebuffer_alloc()
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 <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
Regards,
Wu Zhangjin
> Signed-off-by: Daniel Mack <[email protected]>
> Cc:: Krzysztof Helt <[email protected]>
> Cc: Peter Zijlstra <[email protected]>
> Cc: Rafael J. Wysocki <[email protected]>
> Cc: <[email protected]>
> Cc: Andrew Morton <[email protected]>
> ---
> drivers/video/fbmem.c | 3 ---
> drivers/video/fbsysfs.c | 3 +++
> 2 files changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/video/fbmem.c b/drivers/video/fbmem.c
> index 53ea056..1d8d098 100644
> --- a/drivers/video/fbmem.c
> +++ b/drivers/video/fbmem.c
> @@ -1513,9 +1513,6 @@ register_framebuffer(struct fb_info *fb_info)
> if (!registered_fb[i])
> break;
> fb_info->node = i;
> - mutex_init(&fb_info->lock);
> - mutex_init(&fb_info->mm_lock);
> -
> fb_info->dev = device_create(fb_class, fb_info->device,
> MKDEV(FB_MAJOR, i), NULL, "fb%d", i);
> if (IS_ERR(fb_info->dev)) {
> diff --git a/drivers/video/fbsysfs.c b/drivers/video/fbsysfs.c
> index d4a2c11..60ddd1d 100644
> --- a/drivers/video/fbsysfs.c
> +++ b/drivers/video/fbsysfs.c
> @@ -58,6 +58,9 @@ struct fb_info *framebuffer_alloc(size_t size, struct device *dev)
>
> info->device = dev;
>
> + mutex_init(&info->lock);
> + mutex_init(&info->mm_lock);
> +
> #ifdef CONFIG_FB_BACKLIGHT
> mutex_init(&info->bl_curve_mutex);
> #endif
On Wed, Jul 08, 2009 at 05:42:25PM +0800, Wu Zhangjin wrote:
> On Wed, 2009-07-01 at 18:26 +0200, Daniel Mack wrote:
> > In 537a1bf05, a per-fbinfo mutex was added and framebuffer drivers were
> > adopted accordingly. This new lock is currently initialized in
> > register_framebuffer().
> >
> > The mx3fb driver takes this lock in mx3fb_map_video_memory() _before_
> > register_framebuffer() is called and hence crashes the system very
> > early.
> >
> > Moving this mutex_init() to framebuffer_alloc() solves this and also
> > seems a lot more straight forward.
> >
>
> but this may introduce deadlock to some _old_ drivers which have not
> used framebuffer_alloc to allocate fb_info structure.
>
> to avoid deadlock, these drivers need to change kmalloc to
> framebuffer_alloc, otherwise, they need to add their own mutexes
> initialization source code:
> mutex_init(&info->lock); mutex_init(&info->mm_lock);
>
> here is only a record for the _old_ drivers' maintainers. hope they can
> get this info from google or find this commit:
>
Yes, I'm going through and gradually converting all of the outstanding
drivers (both old and new) over to framebuffer_alloc(). The old ones are
a bit more tedious given all of the casting games they play. I should
have the outstanding ones finished tomorrow.