| Commit message (Collapse) | Author | Age | Files | Lines |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This commit adds basic mmu support, ie:
- DMA cache handling is not supported
- Remapping memory region also
The current mmu setting is:
- 4KB granularity
- 3 level lookup (skipping L0)
- 33 bits per VA
This is based on coreboot and u-boot mmu configuration.
Signed-off-by: Raphael Poggi <poggi.raph@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
remap_range is for remapping regions with different cache attributes.
It is implemented for ARM and PowerPC only, the other architectures only
provide stubs.
Currently the new cache attributes are passed in an architecture specific
way and the attributes have to be retrieved by calls to
mmu_get_pte_cached_flags() and mmu_get_pte_uncached_flags().
Make this simpler by providing architecture independent flags which can
be directly passed to remap_range()
Also provide a MAP_ARCH_DEFAULT flag and a arch_can_remap() function.
The MAP_ARCH_DEFAULT defaults to whatever caching type the architecture
has as default. the arch_can_remap() function returns true if the
architecture can change the cache attributes, false otherwise. This
allows the memtest code to better find out what it has to do.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
Otherwise entries may still be in the cache and never reach memory.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
Those should only be used internally. All users should rather
use the streaming DMA API, which does proper cache maintenance.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
| |
This better separates the DMA from the MMU functionality.
Also move all drivers that only depends on asm/mmu.h for the alloc
functions over to the common header.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
| |
That's the place where Linux has them and other architectures would
implement this.
This will help in phasing out the direct usage of the ARM asm/mmu.h
header.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
As a lot drivers currently rely on the 1:1 virt->phys mapping on ARM
we define DMA_ADDRESS_BROKEN to mark them. In order to use them on
other architectures with a different mapping they need proper fixing.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
| |
So users can pass in device memory pointers without provoking
warnings.
Signed-off-by: Lucas Stach <dev@lynxeye.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
| |
With relocatable binaries the vector addresses cannot be supplied by
the linker. This adds support for fixing them up during runtime.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
| |
Change function remap_range in arm architecture to make it
global accessable. For example command 'memtest' can change
pte flags to enable or disable cache.
Add dummy function for others architectures that doesn't
have mmu or pte support.
Signed-off-by: Alexander Aring <alex.aring@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
dma_alloc() allocates memory aligned to cache lines. We have to use
cache line aligned buffers if a driver calls dma_inv_range on the
buffer.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
| |
Add a function to remap an IO range into a virtual addresses
range. This is particulary usefull for the few devices
mapped at physical address 0, as the MTD boot devices.
Signed-off-by: Robert Jarzmik <robert.jarzmik@free.fr>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
In barebox we used 1MiB sections to map our SDRAM cachable. This
has the drawback that we have to map our sdram twice: cached for
normal sdram and uncached for DMA operations. As address space gets
sparse on newer systems we are sometines unable to find a suitably
big enough area for the dma coherent space.
This patch changes the MMU code to use second level page tables.
With it we can implement dma_alloc_coherent as normal malloc, we
just have to remap the allocated area uncached afterwards and map
it cached again after free().
This makes arm_create_section(), setup_dma_coherent() and mmu_enable()
noops.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
We'll need it later once we remap dma memory.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
So that the ifdeffery can be moved out of the board code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The idea is to panic() when there is no memory available for normal
operation. Exception: code which can consume arbitrary amount of RAM
(example: files allocated in ramfs) must report error instead of panic().
This patch also fixes code which didn't check for NULL from malloc() etc.
Usage: malloc(), memalign() return NULL when out of RAM.
xmalloc(), xmemalign() always return non-NULL or panic().
Signed-off-by: Krzysztof HaĆasa <khc@pm.waw.pl>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
|
|
| |
This is a preparation to add second level cache support.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
|
|
| |
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
|
|
Move platform independent header files to arch/arm/include/asm, leaving
those in asm/arch* and asm/proc* alone.
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
|