This patch add basic support for Toradex' Colibri PXA300 module.
Ethernet is enabled conditionally, depdending on CONFIG_AX88796.
Signed-off-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Eric Miao <eric.miao@marvell.com>
Namespace cleanup: rename colibri.c to colibri-pxa270.c and change
some names in colibri.h.
Signed-off-by: Daniel Mack <daniel@caiaq.de>
Signed-off-by: Eric Miao <eric.miao@marvell.com>
The initial version of the DSA driver only supported a single switch
chip per network interface, while DSA-capable switch chips can be
interconnected to form a tree of switch chips. This patch adds support
for multiple switch chips on a network interface.
An example topology for a 16-port device with an embedded CPU is as
follows:
+-----+ +--------+ +--------+
| |eth0 10| switch |9 10| switch |
| CPU +----------+ +-------+ |
| | | chip 0 | | chip 1 |
+-----+ +---++---+ +---++---+
|| ||
|| ||
||1000baseT ||1000baseT
||ports 1-8 ||ports 9-16
This requires a couple of interdependent changes in the DSA layer:
- The dsa platform driver data needs to be extended: there is still
only one netdevice per DSA driver instance (eth0 in the example
above), but each of the switch chips in the tree needs its own
mii_bus device pointer, MII management bus address, and port name
array. (include/net/dsa.h) The existing in-tree dsa users need
some small changes to deal with this. (arch/arm)
- The DSA and Ethertype DSA tagging modules need to be extended to
use the DSA device ID field on receive and demultiplex the packet
accordingly, and fill in the DSA device ID field on transmit
according to which switch chip the packet is heading to.
(net/dsa/tag_{dsa,edsa}.c)
- The concept of "CPU port", which is the switch chip port that the
CPU is connected to (port 10 on switch chip 0 in the example), needs
to be extended with the concept of "upstream port", which is the
port on the switch chip that will bring us one hop closer to the CPU
(port 10 for both switch chips in the example above).
- The dsa platform data needs to specify which ports on which switch
chips are links to other switch chips, so that we can enable DSA
tagging mode on them. (For inter-switch links, we always use
non-EtherType DSA tagging, since it has lower overhead. The CPU
link uses dsa or edsa tagging depending on what the 'root' switch
chip supports.) This is done by specifying "dsa" for the given
port in the port array.
- The dsa platform data needs to be extended with information on via
which port to reach any given switch chip from any given switch chip.
This info is specified via the per-switch chip data struct ->rtable[]
array, which gives the nexthop ports for each of the other switches
in the tree.
For the example topology above, the dsa platform data would look
something like this:
static struct dsa_chip_data sw[2] = {
{
.mii_bus = &foo,
.sw_addr = 1,
.port_names[0] = "p1",
.port_names[1] = "p2",
.port_names[2] = "p3",
.port_names[3] = "p4",
.port_names[4] = "p5",
.port_names[5] = "p6",
.port_names[6] = "p7",
.port_names[7] = "p8",
.port_names[9] = "dsa",
.port_names[10] = "cpu",
.rtable = (s8 []){ -1, 9, },
}, {
.mii_bus = &foo,
.sw_addr = 2,
.port_names[0] = "p9",
.port_names[1] = "p10",
.port_names[2] = "p11",
.port_names[3] = "p12",
.port_names[4] = "p13",
.port_names[5] = "p14",
.port_names[6] = "p15",
.port_names[7] = "p16",
.port_names[10] = "dsa",
.rtable = (s8 []){ 10, -1, },
},
},
static struct dsa_platform_data pd = {
.netdev = &foo,
.nr_switches = 2,
.sw = sw,
};
Signed-off-by: Lennert Buytenhek <buytenh@marvell.com>
Tested-by: Gary Thomas <gary@mlbassoc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
It would seem when building kernel modules with modern binutils
(required by modern GCC) for ARM v4T targets (specifically observed
with the Samsung 24xx SoC which is an 920T) R_ARM_V4BX relocations
are emitted for function epilogues.
This manifests at module load time with an "unknown relocation: 40"
error message.
The following patch adds the R_ARM_V4BX relocation to the ARM kernel
module loader. The relocation operation is taken from that within the
binutils bfd library.
Signed-off-by: Simtec Linux Team <linux@simtec.co.uk>
Signed-off-by: Vincent Sanders <vince@simtec.co.uk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Mike Rapoport <mike@compulab.co.il>
Acked-by: Eric Miao <eric.miao@marvell.com>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
This is a device driver for the NAND flash controller found on the various
DaVinci family chips. It handles up to four SoC chipselects, and some
flavors of secondary chipselect (e.g. based on upper bits of the address
bus) as used with some multichip packages. (Including the 2 GiB chips
used on some TI devel boards.)
The 1-bit ECC hardware is supported (3 bytes ECC per 512 bytes data); but
not yet the newer 4-bit ECC (10 bytes ECC per 512 bytes data), as
available on chips like the DM355 or OMAP-L137 and needed with the more
error-prone MLC NAND chips.
This is a cleaned-up version of code that's been in use for several years
now; sanity checked with the new drivers/mtd/tests.
Signed-off-by: David Brownell <dbrownell@users.sourceforge.net>
Signed-off-by: Sudhakar Rajashekhara <sudhakar.raj@ti.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Woodhouse <David.Woodhouse@intel.com>
OMAP wishes to pass state to the boot loader upon reboot in order to
instruct it whether to wait for USB-based reflashing or not. There is
already a facility to do this via the reboot() syscall, except we ignore
the string passed to machine_restart().
This patch fixes things to pass this string to arch_reset(). This means
that we keep the reboot mode limited to telling the kernel _how_ to
perform the reboot which should be independent of what we request the
boot loader to do.
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
clk_add_alias is commonly called for platform devices that are not yet
registered in the device tree. Thus the clock alias is associated with
NULL device name. Fix this by passing the device name instead of just
device pointer.
Signed-off-by: Dmitry Eremin-Solenikov <dbaryshkov@gmail.com>
Signed-off-by: Eric Miao <eric.miao@marvell.com>
Map unused registers at the end of DMA region at 64 MB to allow PCI masters
to cross the boundary when prefetching data from SDRAM.
Signed-off-by: Krzysztof Hałasa <khc@pm.waw.pl>
Ideally, the default should be set to 0 and let the EHCI driver turn
it on as needed. This makes USB usable in the mean time.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
Here it is... HIGHMEM for the ARM architecture. :-)
If you don't have enough ram for highmem pages to be allocated and still
want to test this, then the cmdline option "vmalloc=" can be used with
a value large enough to force the highmem threshold down.
Successfully tested on a Marvell DB-78x00-BP Development Board with
2 GB of RAM.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
VIPT aliasing caches have issues of their own which are not yet handled.
Usage of discard_old_kernel_data() in copypage-v6.c is not highmem ready,
kmap/fixmap stuff doesn't take account of cache colouring, etc.
If/when those issues are handled then this could be reverted.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
On xsc3, L2 cache ops are possible only on virtual addresses. The code
is rearranged so to have a linear progression requiring the least amount
of pte setups in the highmem case. To protect the virtual mapping so
created, interrupts must be disabled currently up to a page worth of
address range.
The interrupt disabling is done in a way to minimize the overhead within
the inner loop. The alternative would consist in separate code for
the highmem and non highmem compilation which is less preferable.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
The choice is between looping over the physical range and performing
single cache line operations, or to map highmem pages somewhere, as
cache range ops are possible only on virtual addresses.
Because L2 range ops are much faster, we go with the later by factoring
the physical-to-virtual address conversion and use a fixmap entry for it
in the HIGHMEM case.
Possible future optimizations to avoid the pte setup cost:
- do the pte setup for highmem pages only
- determine a threshold for doing a line-by-line processing on physical
addresses when the range is small
Signed-off-by: Nicolas Pitre <nico@marvell.com>
If a machine class has a custom __virt_to_bus() implementation then it
must provide a __arch_page_to_dma() implementation as well which is
_not_ based on page_address() to support highmem.
This patch fixes existing __arch_page_to_dma() and provide a default
implementation otherwise. The default implementation for highmem is
based on __pfn_to_bus() which is defined only when no custom
__virt_to_bus() is provided by the machine class.
That leaves only ebsa110 and footbridge which cannot support highmem
until they provide their own __arch_page_to_dma() implementation.
But highmem support on those legacy platforms with limited memory is
certainly not a priority.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
This is a helper to be used by the DMA mapping API to handle cache
maintenance for memory identified by a page structure instead of a
virtual address. Those pages may or may not be highmem pages, and
when they're highmem pages, they may or may not be virtually mapped.
When they're not mapped then there is no L1 cache to worry about. But
even in that case the L2 cache must be processed since unmapped highmem
pages can still be L2 cached.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
The kmap virtual area borrows a 2MB range at the top of the 16MB area
below PAGE_OFFSET currently reserved for kernel modules and/or the
XIP kernel. This 2MB corresponds to the range covered by 2 consecutive
second-level page tables, or a single pmd entry as seen by the Linux
page table abstraction. Because XIP kernels are unlikely to be seen
on systems needing highmem support, there shouldn't be any shortage of
VM space for modules (14 MB for modules is still way more than twice the
typical usage).
Because the virtual mapping of highmem pages can go away at any moment
after kunmap() is called on them, we need to bypass the delayed cache
flushing provided by flush_dcache_page() in that case.
The atomic kmap versions are based on fixmaps, and
__cpuc_flush_dcache_page() is used directly in that case.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
This is the minimum fixmap interface expected to be implemented by
architectures supporting highmem.
We have a second level page table already allocated and covering
0xfff00000-0xffffffff because the exception vector page is located
at 0xffff0000, and various cache tricks already use some entries above
0xffff0000. Therefore the PTEs covering 0xfff00000-0xfffeffff are free
to be used.
However the XScale cache flushing code already uses virtual addresses
between 0xfffe0000 and 0xfffeffff.
So this reserves the 0xfff00000-0xfffdffff range for fixmap stuff.
The Documentation/arm/memory.txt information is updated accordingly,
including the information about the actual top of DMA memory mapping
region which didn't match the code.
Signed-off-by: Nicolas Pitre <nico@marvell.com>
Since now ipaq_model_ops used only for accessing h3600 EGPIOs,
drop it completely and use assign_h3600_egpio() directly.
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Replace all occurences with assign_h3600_egpio.
Also simplify code a bit by replacing couple of if-else
statements with one-line equivalents.
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Remove unused fields and associated funtions-accesors.
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Right now iPaq h3600's default MTD partitions table is a mess. It has
two #ifdefs with #else, giving total 3 variants, depending on your
kernel config. Replace all this with simple two-partitions scheme
(bootloader + rootfs), that used by both shipped WindowsCE and
most of the linux distributions (Familiar, Angstrom)
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
There's no actual code for iPAQ sleeves support in kernel that depends
on this config option.
Signed-off-by: Dmitry Artamonow <mad_soft@inbox.ru>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch adds basic support for Dave/DENX QongEVB-LITE i.MX31-based
board. It includes support for clocks initialization, UART1, NOR-flash,
FPGA-attached NAND flash and DNET ethernet controller (inside FPGA).
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch adds clkdev support for i.MX31. This is done in a
similar way done previously for i.MX27
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The UART3 had a copy-paste bug. instead of claiming rxd, txd, rts and
cts pins, cts and rts were claimed twice
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
On MX31 we can't do much without mapping the AIPS1/2 register space.
Move these mappings from individual boards to plat-mxc/mm.c
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
The i.MX35 basically features the same peripherals as the i.MX31 with
some differences:
- The i.MX35 has a FEC ethernet controller
- The NAND controller base addresses are different
- The i.MX35 has only 3 UARTs
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch adds clock support for i.MX35 SoCs. We do not support setting
of clock rates yet, but most interesting clock rates should be reported.
I couldn't test all clock rates and the datasheet contains some obvious
bugs, so expect some bugs in this code.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
We had hardcoded cpu_is_ macros for mxc architectures till now. As we
want to run the same kernel on i.MX31 and i.MX35 this patch adds cpu_is_
macros which expand to 0 or 1 if only one architecture is compiled in and
only check for the cpu type if more than one architecture is compiled
in.
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This patch moves the stuff common to i.MX31 and i.MX35 to mx3x.h and the
specifics to mx31.h/mx35.h. We can build a kernel which runs on i.MX31 and
i.MX35, so always include mx31.h and mx35.h
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
This adds the dma (ipu_dma) and fb devices for the mx31 for which drivers now are
available.
v2: merge the ipu and fb device in the same patch as suggested by Sascha
Signed-off-by: Valentin Longchamp <valentin.longchamp@epfl.ch>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Daniel Mack