Re: backlight: oops in __mutex_lock_slowpath during head /sys/class/graphics/fb0/* in 2.6.18
From: Michael Hanselmann <hidden>
Date: 2006-09-21 21:56:23
Also in:
lkml
On Thu, Sep 21, 2006 at 12:37:42PM -0700, Andrew Morton wrote:
On Thu, 21 Sep 2006 14:19:52 +0200 Olaf Hering [off-list ref] wrote:quoted
The bl_curve code has some room for improvement.
Agreed.
quoted
inst-sys:~ # dmesg valkyriefb: vmode 13 does not support cmode 1. Unable to handle kernel paging request for data at address 0x00000000 Faulting instruction address: 0xc02c7bd8
Seems like not all drivers use the framebuffer_alloc() function and won't have an initialized mutex. But those don't have a backlight, anyway. What about the patch below? Does it work for you? Signed-off-by: Michael Hanselmann <redacted> --- diff -Nrup --exclude-from linux-exclude-from linux-2.6.18.orig/drivers/video/fbsysfs.c linux-2.6.18/drivers/video/fbsysfs.c
--- linux-2.6.18.orig/drivers/video/fbsysfs.c 2006-09-21 23:40:06.000000000 +0200
+++ linux-2.6.18/drivers/video/fbsysfs.c 2006-09-21 23:50:00.000000000 +0200@@ -397,6 +397,12 @@ static ssize_t store_bl_curve(struct cla u8 tmp_curve[FB_BACKLIGHT_LEVELS]; unsigned int i; + /* Some drivers don't use framebuffer_alloc(), but those also + * don't have backlights. + */ + if (!fb_info || !fb_info->bl_dev) + return -ENODEV; + if (count != (FB_BACKLIGHT_LEVELS / 8 * 24)) return -EINVAL;
@@ -430,6 +436,12 @@ static ssize_t show_bl_curve(struct clas ssize_t len = 0; unsigned int i; + /* Some drivers don't use framebuffer_alloc(), but those also + * don't have backlights. + */ + if (!fb_info || !fb_info->bl_dev) + return -ENODEV; + mutex_lock(&fb_info->bl_mutex); for (i = 0; i < FB_BACKLIGHT_LEVELS; i += 8) len += snprintf(&buf[len], PAGE_SIZE,