When booting with device-tree, retrieve GPMC settings for ONENAND from
the device-tree blob. This will allow us to remove all static settings
stored in the gpmc-nand.c in the future once the migration to
device-tree is complete.

The user must now specify the ONENAND device width in the device-tree
binding so that the GPMC can be programmed correctly. Therefore, update
the device-tree binding documentation for ONENAND devices connected to
the GPMC to reflect this.

Please note that this does not include GPMC timings for ONENAND. The
timings are being calculated at runtime.

There is some legacy code that only enables read wait monitoring for
non-OMAP3 devices. There are no known OMAP3 device issues that prevent
this feature being enabled and so when booting with device-tree use the
wait-monitoring settings described in the device-tree blob.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

When booting with device-tree, retrieve GPMC settings for NAND from
the device-tree blob. This will allow us to remove all static settings
stored in the gpmc-nand.c in the future once the migration to
device-tree is complete.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

NOR flash is not currently supported when booting with device-tree
on OMAP2+ devices. Add support to detect and configure NOR devices
when booting with device-tree.

Add documentation for the TI GPMC NOR binding.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Some of the GPMC timings parameters are currently missing from the GPMC
device-tree binding. Add these parameters to the binding documentation
as well as code to read them. Also add either "-ps" or "-ns" suffix to
the GPMC timing properties to indicate whether the timing is in
picoseconds or nanoseconds.

The existing code in gpmc_read_timings_dt() is checking the value of
of_property_read_u32() and only is successful storing the value read
in the gpmc_timings structure. Checking the return value in this case
is not necessary and we can simply read the value, if present, and
store directly in the gpmc_timings structure. Therefore, simplify the
code by removing these checks.

The comment in the gpmc_read_timings_dt() function, "only for OMAP3430"
is also incorrect as it is applicable to all OMAP3+ devices. So correct
this too.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Adds a function to read the various GPMC chip-select settings from
device-tree and store them in the gpmc_settings structure.

Update the GPMC device-tree binding documentation to describe these
options.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

With the addition of the gpmc_cs_program_settings(), we no longer need
or use gpmc_cs_configure() to configure some of the GPMC chip-select
options. So rename the function to gpmc_configure() and remove code that
modifies options in the CONFIG1 register.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Convert the OMAP2+ TUSB code to use the gpmc_cs_program_settings()
function for configuring the various GPMC options instead of directly
programming the CONFIG1 register.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Convert the OMAP2+ SMC91x code to use the gpmc_cs_program_settings()
function for configuring the various GPMC options instead of directly
programming the CONFIG1 register.

Move configuration of the GPMC settings outside retime function and
this does not need to be done if the timings are changed dynamically
at runtime.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Convert the OMAP2+ NAND code to use the gpmc_cs_program_settings()
function for configuring the various GPMC options instead of directly
programming the CONFIG1 register.

This moves the configuration of some GPMC options outside the
nand_gpmc_retime() because these options should only need to be set once
regardless of whether the gpmc timing is changing dynamically at runtime.
The programming of where the wait-pin is also moved slightly, but this
will not have any impact to existing devices as no boards are currently
setting the dev_ready variable.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

Convert the OMAP2+ ONENAND code to use the gpmc_cs_program_settings()
function for configuring the various GPMC options instead of directly
programming the CONFIG1 register.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

The GPMC has various different configuration options such as bus-width,
synchronous or asychronous mode selection, burst mode options etc.
Currently, there is no common function for configuring these options and
various devices set these options by either programming the GPMC CONFIG1
register directly or by calling gpmc_cs_configure() to set some of the
options.

Add a new function for configuring all of the GPMC options. Having a common
function for configuring this options will simplify code and ease the
migration to device-tree.

Also add a new capability flag to detect devices that support the
address-address-data multiplexing mode.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

The GPMC has various different configuration options such as bus-width,
synchronous or asychronous mode selection, burst mode options etc.
Currently, there is no central structure for storing all these options
when configuring the GPMC for a given device. Some of the options are
stored in the GPMC timing structure and some are directly programmed
into the GPMC configuration register. Add a new structure to store
these options and convert code to use this structure. Adding this
structure will allow us to create a common function for configuring
these options.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

The GPMC has wait-pin signals that can be assigned to a chip-select
to monitor the ready signal of an external device. Add a variable to
indicate the total number of wait-pins for a given device. This will
allow us to detect if the wait-pin being selected is valid or not.

When booting with device-tree read the number of wait-pins from the
device-tree blob. When device-tree is not used set the number of
wait-pins to 4 which is valid for OMAP2-5 devices. Newer devices
that have less wait-pins (such as AM335x) only support booting with
device-tree and so hard-coding the wait-pin number when not using
device-tree is fine.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

The OMAP2+ code that configures the GPMC for ONENAND devices is copying
structures between functions unnecessarily. Avoid this by passing
pointers instead and simplify the code.

A pointer to structure "omap_onenand_platform_data" is passed to the
function omap2_onenand_calc_sync_timings(), but only the flags member
of the structure is used. Simplify the code by only passing the flags
member and not the entire structure.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>

This check is done before the call to gpmc_cs_reserved() and
gpmc_cs_set_reserved() and it's redundant to do it again in each
function. This simplifies the code a bit.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

Since the condition is not an error but a warning, replace
printk KERN_ERR with dev_warn.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

Do this becasue dev_err() is preferred over pr_err() and because
it will match gpmc-nand, thus the code shows looks more consistent.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

If CS request fails the current error message is rather unhelpful.
Fix it by printing the failing chip select and the error code.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

If CS request fails the current error message is rather unhelpful.
Fix it by printing the failing chip select and the error code.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

Currently gpmc_cs_reserved() return value is somewhat inconsistent,
returning a negative value on an error condition, a positive value
if the chip select is reserved and zero if it's available.

Fix this by returning a boolean value as the function name suggests:
  * true if the chip select is reserved,
  * false if it's available
Suggested-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

This function is not used anywhere, so it's safe to remove it.
This means less code to maintain.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

This patch marks a bunch of functions that are local
to gpmc.c file only as static.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Reviewed-by: NJon Hunter <jon-hunter@ti.com>
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

When setting the GPMC device type, make sure any previous
bits are cleared down, before applying the new setting.

For OMAP4+ devices MUXADDDATA is a 2-bit field (bits 9:8)
where as for OMAP2/3 devices it was only a one bit field
(bit 9). For OMAP2/3 devices bit 8 is reserved and the
OMAP documentation says to write a 0 to this bit. So
clearing bit 8 on OMAP2/3 devices should not be a problem.
Hence update the code to handle both bits 8 and 9 for all
devices.
Signed-off-by: NMark Jackson <mpfj@newflow.co.uk>
[jon-hunter@ti.com: updated changelog]
Signed-off-by: NJon Hunter <jon-hunter@ti.com>

Consistently check errors using the usual method used in the kernel
for much of its history.  For instance:

int gpmc_cs_set_timings(int cs, const struct gpmc_timings *t)
{
	int div;
	div = gpmc_calc_divider(t->sync_clk);
	if (div < 0)
		return div;
static int gpmc_set_async_mode(int cs, struct gpmc_timings *t)
{
...
	return gpmc_cs_set_timings(cs, t);

.....
	ret = gpmc_set_async_mode(gpmc_onenand_data->cs, &t);
	if (IS_ERR_VALUE(ret))
		return ret;

So, gpmc_cs_set_timings() thinks any negative return value is an error,
but where we check that in higher levels, only a limited range are
errors...

There is only _one_ use of IS_ERR_VALUE() in arch/arm which is really
appropriate, and that is in arch/arm/include/asm/syscall.h:

static inline long syscall_get_error(struct task_struct *task,
				     struct pt_regs *regs)
{
	unsigned long error = regs->ARM_r0;
	return IS_ERR_VALUE(error) ? error : 0;
}

because this function really does have to differentiate between error
return values and addresses which look like negative numbers (eg, from
mmap()).

So, here's a patch to remove them from OMAP, except for the above.
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Commit 6797b4fe (ARM: OMAP2+: Prevent potential crash if GPMC probe fails)
added code to ensure that GPMC chip-selects could not be requested until the
device probe was successful. The chip-selects should have been
unreserved at the end of the probe function, but the code to unreserve
them appears to have ended up in the gpmc_calc_timings() function and
hence, this is causing problems requesting chip-selects. Fix this merge
error by unreserving the chip-selects at the end of the probe, but
before we call the gpmc child probe functions (for device-tree) which
request a chip-select.
Signed-off-by: NJon Hunter <jon-hunter@ti.com>
Tested-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: NPhilip Avinash <avinashphilip@ti.com>
Tested-by: NGrazvydas Ignotas <notasas@gmail.com>
[tony@atomide.com: updated description to add breaking commit id]
Signed-off-by: NTony Lindgren <tony@atomide.com>

Commit 16559ae4 (kgdb: remove #include <linux/serial_8250.h> from kgdb.h)
had a side effect of breaking omap1_defconfig build as some headers
were included indirectly:

arch/arm/mach-omap1/board-h2.c:249: error: ‘INT_KEYBOARD’ undeclared here (not in a function)
...

This worked earlier as linux/serial_8250.h included linux/serial_core.h,
via linux/serial_8250.h from linux/kgdb.h. Fix this by including the
necessary headers directly.
Signed-off-by: NTony Lindgren <tony@atomide.com>

metag/allmodconfig:

drivers/media/v4l2-core/videobuf2-dma-contig.c: In function 'vb2_dc_get_base_sgt':
drivers/media/v4l2-core/videobuf2-dma-contig.c:387: error: implicit declaration of function 'dma_get_sgtable'

For architectures using dma_map_ops, dma_get_sgtable() is provided in
<asm-generic/dma-mapping-common.h>.

Metag does not use dma_map_ops yet, hence it should implement it as an
inline stub using dma_common_get_sgtable().
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Acked-by: NMarek Szyprowski <m.szyprowski@samsung.com>

Metag doesn't have a metag_dt_memblock_reserve() function so remove the
declaration from asm/prom.h.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

Make a copy of the device tree blob in non-init memory. It is required
when using built-in device tree files that the platform code copies the
blob to non-init memory prior to calling unflatten_device_tree(),
otherwise the strings that the device tree refer to will get poisoned
and potentially reused, breaking later reading of the device tree
post-init (such as compatible matching in modules, debugfs, and the
procfs interface).
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>
Reviewed-by: NVineet Gupta <vgupta@synopsys.com>

Minimise metag_ksyms.c includes to directly include the <asm/*.h> files
that declare a particular symbol, and not include any unnecessary ones.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

It's less error prone to have function symbols exported immediately
after the function rather than in metag_ksyms.c. Move each EXPORT_SYMBOL
in metag_ksyms.c for symbols defined in mm/init.c into mm/init.c.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

It's less error prone to have function symbols exported immediately
after the function rather than in metag_ksyms.c. Move each EXPORT_SYMBOL
in metag_ksyms.c for symbols defined in usercopy.c into usercopy.c
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

It's less error prone to have function symbols exported immediately
after the function rather than in metag_ksyms.c. Move each EXPORT_SYMBOL
in metag_ksyms.c for symbols defined in setup.c into setup.c
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

It's less error prone to have function symbols exported immediately
after the function rather than in metag_ksyms.c. Move each EXPORT_SYMBOL
in metag_ksyms.c for symbols defined in kick.c into kick.c
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

It's less error prone to have function symbols exported immediately
after the function rather than in metag_ksyms.c. Move each EXPORT_SYMBOL
in metag_ksyms.c for symbols defined in traps.c into traps.c
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

The SMP version of arch_local_irq_enable() uses preempt_disable(), but
<asm/irqflags.h> doesn't include <linux/preempt.h> causing the following
errors on SMP when pstore/ftrace is enabled (caught by buildbot smp
allyesconfig):

In file included from include/linux/irqflags.h:15,
                 from fs/pstore/ftrace.c:16:
arch/metag/include/asm/irqflags.h: In function 'arch_local_irq_enable':
arch/metag/include/asm/irqflags.h:84: error: implicit declaration of function 'preempt_disable'
arch/metag/include/asm/irqflags.h:86: error: implicit declaration of function 'preempt_enable_no_resched'

However <linux/preempt.h> cannot be easily included from
<asm/irqflags.h> as it can cause circular include dependencies in the
!SMP case, and potentially in the SMP case in the future. Therefore move
the SMP implementation of arch_local_irq_enable() into traps.c and use
an inline version of get_trigger_mask() which is also defined in traps.c
for SMP.

This adds an extra layer of function call / stack push when
preempt_disable needs to call other functions, however in the
non-preemptive SMP case it should be about as fast, as it was already
calling the get_trigger_mask() function which is now used inline.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

Convert hugetlb_get_unmapped_area_new_pmd() to use vm_unmapped_area()
rather than searching the virtual address space itself. This fixes the
following errors in linux-next due to the specified members being
removed after other architectures have already been converted:

arch/metag/mm/hugetlbpage.c: In function 'hugetlb_get_unmapped_area_new_pmd':
arch/metag/mm/hugetlbpage.c:199: error: 'struct mm_struct' has no member named 'cached_hole_size'
arch/metag/mm/hugetlbpage.c:200: error: 'struct mm_struct' has no member named 'free_area_cache'
arch/metag/mm/hugetlbpage.c:215: error: 'struct mm_struct' has no member named 'cached_hole_size'
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>
Acked-by: NMichel Lespinasse <walken@google.com>

Various file systems use clear_page() and copy_page(), so when they're
built as modules we get build errors like the following:

ERROR: "clear_page" [fs/ntfs/ntfs.ko] undefined!
ERROR: "copy_page" [fs/nilfs2/nilfs2.ko] undefined!

Therefore export these functions to modules from metag_ksyms.c to fix
the errors. This was hit by a randconfig build.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

Various file systems indirectly use metag_code_cache_flush_all(), so
when they're built as modules we get build errors like the following:

ERROR: "metag_code_cache_flush_all" [fs/xfs/xfs.ko] undefined!

Therefore export this function to modules to fix the errors. This was
hit by a randconfig build.
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>

Rename setup_txprivext() to setup_priv() and add initialisation of some
more per-thread privilege protection registers:

 - TxPRIVSYSR: 0x04400000-0x047fffff
               0x05000000-0x07ffffff
               0x84000000-0x87ffffff
 - TxPIOREG:   0x02000000-0x02ffffff
               0x04800000-0x048fffff
 - TxSYREG:    0x04000000-0x04000fff (except write fetch system event)
Signed-off-by: NJames Hogan <james.hogan@imgtec.com>