A recent commit (inadvertently?) changed how failed probe of a gpmc
child node was handled. Instead of proceeding with setting up any other
children as before, a single error now aborts the whole process.

This change broke networking on some Overo boards due to probe failing
for an unrelated nand node. This second issue should obviously be
fixed, but let's restore the old behaviour of allowing child-node
probe to fail to avoid further similar breakage on other systems.

Fixes: d2d00862 ("memory: omap-gpmc: Support general purpose input
for WAITPINs")

Cc: stable@vger.kernel.org  # v4.7+
Signed-off-by: NJohan Hovold <johan@kernel.org>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

Use helper of_platform_default_populate() in linux/of_platform
when possible, instead of calling of_platform_populate() with
the default match table.
Acked-by: NRoger Quadros <rogerq@ti.com>
Cc: Roger Quadros <rogerq@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
Signed-off-by: NKefeng Wang <wangkefeng.wang@huawei.com>
Signed-off-by: NRob Herring <robh@kernel.org>

The gpio related functions were mistakenly built only if CONFIG_OF
is defined. They are needed even otherwise and will cause build
failures if CONFIG_OF is not defined. Move the gpio functions
outside #ifdef CONFIG_OF.

Fixes: d2d00862 ("memory: omap-gpmc: Support general purpose input for WAITPINs")
Signed-off-by: NRoger Quadros <rogerq@ti.com>

The Kconfig currently controlling compilation of this code is:

drivers/memory/Kconfig:config OMAP_GPMC
drivers/memory/Kconfig:  bool

...meaning that it currently is not being built as a module by anyone.

Lets remove the modular code that is essentially orphaned, so that
when reading the driver there is no doubt it is builtin-only.

Since module_init was not in use by this code, the init ordering
remains unchanged with this commit.

Also note that MODULE_DEVICE_TABLE is a no-op for non-modular code.

We don't replace module.h with init.h since the file already has that.

Cc: Roger Quadros <rogerq@ti.com>
Cc: Tony Lindgren <tony@atomide.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: NPaul Gortmaker <paul.gortmaker@windriver.com>

In the omap gpmc driver it can be noticed that GPMC_CONFIG4_OEEXTRADELAY
is overwritten by the WEEXTRADELAY value from the device tree and
GPMC_CONFIG4_WEEXTRADELAY is not updated by the value from the device
tree.

As a consequence, the memory accesses cannot be configured properly when
the extra delay are needed for OE and WE.

Fix the update of GPMC_CONFIG4_WEEXTRADELAY with the value from the
device tree file and prevents GPMC_CONFIG4_OEXTRADELAY
being overwritten by the WEXTRADELAY value from the device tree.

Cc: stable@vger.kernel.org
Signed-off-by: NOcquidant, Sebastien <sebastienocquidant@eaton.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

The core now takes care of parsing generic DT properties in
nand_scan_ident() when nand_set_flash_node() has been called.
Rely on this initialization instead of calling of_get_nand_xxx()
manually.
Signed-off-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Acked-by: NRoger Quadros <rogerq@ti.com>
Tested-by: NFranklin S Cooper Jr. <fcooper@ti.com>

GPMC_STATUS register is private to the GPMC module and must not be
accessed directly by NAND driver through the gpmc_regs.

They must use gpmc_omap_get_nand_ops() instead.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>

OMAPs can have 2 to 4 WAITPINs that can be used as edge triggered
interrupts if not used for memory wait state insertion.

Support these interrupts via the gpmc IRQ domain.

The gpmc IRQ domain interrupt map is:

0 - NAND_fifoevent
1 - NAND_termcount
2 - GPMC_WAIT0 edge
3 - GPMC_WAIT1 edge, and so on
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NRob Herring <robh@kernel.org>
Acked-by: NTony Lindgren <tony@atomide.com>

If the device attached to GPMC wants to use the WAIT pin
for WAIT monitoring then we reserve it internally for
exclusive use.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>

OMAPs can have 2 to 4 WAITPINs that can be used as general purpose
input if not used for memory wait state insertion.

The first user will be the OMAP NAND chip to get the NAND
read/busy status using gpiolib.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NRob Herring <robh@kernel.org>
Acked-by: NTony Lindgren <tony@atomide.com>

We have been preventing mapping GPMC children in the
first 1MB but really it has to be the first 16MB as
the minimum GPMC partition size is 16MB.

Also print an error message if CS mapping fails
due to DT requesting address outside the GPMC
map.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>

Move NAND specific device tree parsing to NAND driver.

The NAND controller node must have a compatible id, register space
resource and interrupt resource.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NBrian Norris <computersforpeace@gmail.com>
Acked-by: NTony Lindgren <tony@atomide.com>

GPMC provides 2 interrupts for NAND use. i.e. fifoevent and termcount.
Use IRQ domain for this. NAND device tree node can then
get the necessary interrupts by using gpmc as the interrupt parent.

Legacy boot uses gpmc_get_client_irq to get the
NAND interrupts from the GPMC IRQ domain.
Get rid of custom bitmasks and use IRQ domain for that
as well.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NRob Herring <robh@kernel.org>
Acked-by: NTony Lindgren <tony@atomide.com>

This is needed by OMAP NAND driver to poll the empty status
of the writebuffer.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>

The OMAP GPMC module has certain registers dedicated for NAND
access and some NAND bits mixed with other GPMC functionality.

For the NAND dedicated registers we have the struct gpmc_nand_regs.

The NAND driver needs to access NAND specific bits from the
following non-dedicated registers
- EMPTYWRITEBUFFERSTATUS from GPMC_STATUS

For accessing these bits we introduce the struct gpmc_nand_ops.

Add gpmc_omap_get_nand_ops() that returns the gpmc_nand_ops along
with updating the gpmc_nand_regs. This API will be called by the
OMAP NAND driver to access the necessary bits in GPMC register space.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>

In order to support extended timings parameters on hardware supporting the
"AAD" mode like the AM335x or DM816x, add these entries into the GPMC driver
if the hardware is capable.

Tested on DM816x and AM335x.
Signed-off-by: NNeil Armstrong <narmstrong@baylibre.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

When gpmc_cs_show_timings is called in gpmc_cs_set_timings()
gpmc_cs_program_settings() was already run which modifies the CONFIG1
register. So to be more useful do the "before" dump earlier.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

set_irq_flags is ARM specific with custom flags which have genirq
equivalents. Convert drivers to use the genirq interfaces directly, so we
can kill off set_irq_flags. The translation of flags is as follows:

IRQF_VALID -> !IRQ_NOREQUEST
IRQF_PROBE -> !IRQ_NOPROBE
IRQF_NOAUTOEN -> IRQ_NOAUTOEN

For IRQs managed by an irqdomain, the irqdomain core code handles clearing
and setting IRQ_NOREQUEST already, so there is no need to do this in
.map() functions and we can simply remove the set_irq_flags calls. Some
users also set IRQ_NOPROBE and this has been maintained although it is not
clear that is really needed. There appears to be a great deal of blind
copy and paste of this code.
Signed-off-by: NRob Herring <robh@kernel.org>
Acked-by: NRoger Quadros <rogerq@ti.com>
Cc: linux-omap@vger.kernel.org
Signed-off-by: NTony Lindgren <tony@atomide.com>

If for some reason the GPMC device hasn't been probed yet, gpmc_base is
going to be NULL. Because there's no context yet to be saved, just turn
these functions into no-ops until that device gets probed.

Unable to handle kernel NULL pointer dereference at virtual address 00000010
pgd = c0204000
[00000010] *pgd=00000000
Internal error: Oops: 5 [#1] SMP ARM
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.2.0-rc5-next-20150804-05947-g23f38fe8eda9 #1
Hardware name: Generic OMAP3-GP (Flattened Device Tree)
task: c0e623e8 ti: c0e5c000 task.ti: c0e5c000
PC is at omap3_gpmc_save_context+0x8/0xc4
LR is at omap_sram_idle+0x154/0x23c
pc : [<c087c7ac>]    lr : [<c023262c>]    psr: 60000193
sp : c0e5df40  ip : c0f92a80  fp : c0999eb0
r10: c0e57364  r9 : c0e66f14  r8 : 00000003
r7 : 00000000  r6 : 00000003  r5 : 00000000  r4 : c0f5f174
r3 : c0fa4fe8  r2 : 00000000  r1 : 00000000  r0 : fa200280
Flags: nZCv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment kernel
Control: 10c5387d  Table: 80204019  DAC: 00000015
Process swapper/0 (pid: 0, stack limit = 0xc0e5c220)
Stack: (0xc0e5df40 to 0xc0e5e000)
df40: 00000000 c0e66ef8 c0f5f1a4 00000000 00000003 c02333a4 c3813822 00000000
df60: 00000000 c0e5a5c8 cfb8a5d0 c07f0c44 0e4f1d7e 00000000 00000000 00000000
df80: c3813822 00000000 cfb8a5d0 c0e5e4e4 cfb8a5d0 c0e66f14 c0e5a5c8 c0e5e54c
dfa0: c0e5e544 c0e57364 c0999eb0 c0277758 000000fa c0f5d000 00000000 c0d61c18
dfc0: ffffffff ffffffff 00000000 c0d61674 00000000 c0df7a48 00000000 c0f5d5d4
dfe0: c0e5e4c0 c0df7a44 c0e634f8 80204059 00000000 8020807c 00000000 00000000
[<c087c7ac>] (omap3_gpmc_save_context) from [<c023262c>] (omap_sram_idle+0x154/0x23c)
[<c023262c>] (omap_sram_idle) from [<c02333a4>] (omap3_enter_idle_bm+0xec/0x1a8)
[<c02333a4>] (omap3_enter_idle_bm) from [<c07f0c44>] (cpuidle_enter_state+0xbc/0x284)
[<c07f0c44>] (cpuidle_enter_state) from [<c0277758>] (cpu_startup_entry+0x174/0x24c)
[<c0277758>] (cpu_startup_entry) from [<c0d61c18>] (start_kernel+0x358/0x3c0)
[<c0d61c18>] (start_kernel) from [<8020807c>] (0x8020807c)
Code: c0ccace8 c0ccacc0 e59f30b4 e5932000 (e5921010)
Signed-off-by: NTomeu Vizoso <tomeu.vizoso@collabora.com>
Suggested-by: NJavier Martinez Canillas <javier@dowhile0.org>
Reviewed-by: NJavier Martinez Canillas <javier@osg.samsung.com>
Acked-by: NRoger Quadros <rogerq@ti.com>
[tony@atomide.com: updated description as suggested by Javier]
Signed-off-by: NTony Lindgren <tony@atomide.com>

We support decoding the bootloader values if DEBUG is defined.
But we also need to change the struct omap_hwmod flags to have
HWMOD_INIT_NO_RESET to avoid the GPMC being reset during the
boot. Otherwise just the default timings will be displayed
instead of the bootloader configured timings.

This also allows us to clean up the various GPMC related
hwmod flags. For debugging, we only need HWMOD_INIT_NO_RESET,
and HWMOD_INIT_NO_IDLE is not needed.

Cc: Brian Hutchinson <b.hutchman@gmail.com>
Cc: Paul Walmsley <paul@pwsan.com>
Cc: Roger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

We currently artificially limit the parsing of GPMC connected
devices based on the device name. Let's stop doing that, it's
confusing as adding devices to .dts files with using normal
names like fpga and usb will currently cause them to not probe.

Cc: Roger Quadros <rogerq@ti.com>
Reported-by: NBrian Hutchinson <b.hutchman@gmail.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

GPMC_CONFIG1_i parameters CLKACTIVATIONTIME and WAITMONITORINGTIME
have reserved values.
Raise an error if calculated timings try to program reserved values.

GPMC_CONFIG1_i ATTACHEDDEVICEPAGELENGTH and DEVICESIZE were already checked
when parsing the DT.

Explicitly comment invalid values on gpmc_cs_show_timings for
-CLKACTIVATIONTIME
-WAITMONITORINGTIME
-DEVICESIZE
-ATTACHEDDEVICEPAGELENGTH
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

The WAITMONITORINGTIME is expressed as a number of GPMC_CLK clock cycles,
even though the access is defined as asynchronous, and no GPMC_CLK clock
is provided to the external device. Still, GPMCFCLKDIVIDER is used as a divider
for the GPMC clock, so it must be programmed to define the
correct WAITMONITORINGTIME delay.

This patch correctly computes WAITMONITORINGTIME in GPMC_CLK cycles instead of GPMC_FCLK cycles,
both during programming (gpmc_cs_set_timings) and during retrieval (gpmc_cs_show_timings).
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

The WAITMONITORINGTIME is expressed as a number of GPMC_CLK clock cycles,
even though the access is defined as asynchronous, and no GPMC_CLK clock
is provided to the external device. Still, GPMCFCLKDIVIDER is used as a divider
for the GPMC clock, so it must be programmed to define the
correct WAITMONITORINGTIME delay.

Calculate GPMCFCLKDIVIDER independent of gpmc,sync-clk-ps in DT for
pure asynchronous accesses, i.e. both read and write asynchronous.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

The WAITMONITORINGTIME is expressed as a number of GPMC_CLK clock cycles,
even though the access is defined as asynchronous, and no GPMC_CLK clock
is provided to the external device. Still, GPMCFCLKDIVIDER is used as a divider
for the GPMC clock, so it must be programmed to define the
correct WAITMONITORINGTIME delay.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

DTS output was formatted to require additional work when copy-pasting into DTS.
Nano-second timings were replaced with interval of values that produce the same
number of clock ticks.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

GPMC debug output is aligned to 10 characters for field names.
However, some fields have bigger names, screwing up the alignment.
Consequently, alignment was changed to longest field name (17 chars) for now.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

This patch adds support for spawning buses as children of the GPMC.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

OMAP2+ GPMC driver undefines DEBUG, which makes it unnecessarily
hard to turn DEBUG on. Remove the offending lines.
Signed-off-by: NRobert ABEL <rabel@cit-ec.uni-bielefeld.de>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

Fix sparse warning:
  warning: symbol 'gpmc_cs_get_name' was not declared. Should it be static?
Signed-off-by: NSemen Protsenko <semen.protsenko@globallogic.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

Some GPMC_CONFIG7 register bits marked as "RESERVED", means they
shouldn't be overwritten. A typical approach to handle such bits called
"Read-Modify-Write". Writing procedure used in gpmc_cs_set_memconf()
utilizes RMW technique, but implemented incorrectly. Due to obvious typo
in code read register value is being rewritten by another value, which
leads to loss of read RESERVED bits. This patch fixes this.

While at it, replace magic numbers with named constants to improve code
readability.
Signed-off-by: NSemen Protsenko <semen.protsenko@globallogic.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>

Just move to drivers as further clean-up can now happen there
finally.

Let's also add Roger and me to the MAINTAINERS so we get
notified for any patches related to GPMC.

Cc: Arnd Bergmann <arnd@arndb.de>
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

This will us allow to just move gpmc.c to live under drivers
in the next patch.

Note that we now also remove the omap specific check for the
initcall. That's OK as gpmc_probe() checks for the pdata
and bails out for other platforms compiled in.

Also the postcore_initcall() maybe possible to change to
just regular module_init(), but let's do that in separate
patch after the move to drivers is done.

Cc: Arnd Bergmann <arnd@arndb.de>
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

We still need to support platform data for omap3 until it's booting
in device tree only mode. So let's add platform_data/omap-gpmc.h for
that, and a minimal linux/omap-gpmc.h for the save and restore used
by the PM code.

Let's also keep a minimal mach-omap2/gpmc.h still around to avoid
churn on the board-*.c files. Once omap3 boots in device tree only
mode, we can drop mach-omap2/gpmc.h and we can make the data
structures in platform_data/omap-gpmc.h private to the GPMC driver.

Note that we can now also remove gpmc-nand.h and gpmc-onenand.h.

Cc: Arnd Bergmann <arnd@arndb.de>
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

OMAP3 and lower SoCs don't have the ELM module so this warning
is annoying. Get rid of it.
Reported-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

The GPMC driver includes arch/arm/mach-omap2/common.h but
does not use anything on that header so it can be removed.
Signed-off-by: NJavier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Now that we have timings in the .dts files for smc91x
and 8250, we can remove the device specific checks and
just print out the bootloader timings for devices that
may not have timings in the .dts files.
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

As we still have some devices with GPMC timings missing from the
.dts files, let's make it a bit easier to use the bootloader
values and print them out.

Note that we now need to move the parsing of the device tree provided
configuration a bit earlier so we can use that for checking if anything
was configured.
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

There are cases where we have multiple device instances
connected to a single GPMC chip select. For example, there
are four UARTs on the Zoom debug boards that all share a
single chip select and a GPIO interrupt.

We do have support for this already in theory, but it's broken
because we're bailing out if the chip select is already taken.

To be able to provide checks on the chip select usage, let's
add new struct gpmc_cs_data so we can start using already
registered device names for checks.

Later on we probably want to start using struct gpmc_cs_data
as a wrapper for all the GPMC chipselect related data.
Acked-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

This prevents potential division by zero errors
if GPMC fck turns out to be zero due to faulty clock
data.

Use resource managed clk_get() API.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NSekhar Nori <nsekhar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>