59bd00c850
The patch added a special get_unmapped_area for framebuffer which was hooked to the file ops in drivers/video/fbmem.c. This is needed since v2.6.29-rc1 where nommu vma management was updated, and mmap of framebuffer caused kernel BUG panic. You may turn on "Debug the global anon/private NOMMU mapping region tree" config to such message. As Documentation/nommu-mmap.txt said, "To provide shareable character device support, a driver must provide a file->f_op->get_unmapped_area() operation. The mmap() routines will call this to get a proposed address for the mapping." With this change, user space should call mmap for framebuffer using shared map. Or it can try shared map first, then private map if failed. This shared map usage is now consistent between mmu and nommu. The sys_ file may not be a good place for this patch. But there is a similar one for sparc. I tested a similar patch on nios2nommu, though I don't have a blackfin board to test. Signed-off-by: Thomas Chou <thomas@wytron.com.tw> Signed-off-by: Graf Yang <graf.yang@analog.com> Signed-off-by: Michael Hennerich <michael.hennerich@analog.com> Signed-off-by: Mike Frysinger <vapier@gentoo.org> |
||
---|---|---|
.. | ||
cplb-mpu | ||
cplb-nompu | ||
.gitignore | ||
asm-offsets.c | ||
bfin_dma_5xx.c | ||
bfin_gpio.c | ||
bfin_ksyms.c | ||
cplbinfo.c | ||
dma-mapping.c | ||
early_printk.c | ||
entry.S | ||
fixed_code.S | ||
flat.c | ||
ftrace-entry.S | ||
ftrace.c | ||
gptimers.c | ||
init_task.c | ||
ipipe.c | ||
irqchip.c | ||
kgdb_test.c | ||
kgdb.c | ||
Makefile | ||
module.c | ||
process.c | ||
ptrace.c | ||
reboot.c | ||
setup.c | ||
shadow_console.c | ||
signal.c | ||
stacktrace.c | ||
sys_bfin.c | ||
time-ts.c | ||
time.c | ||
traps.c | ||
vmlinux.lds.S |