The bitfield allocation function returns error condition
as a negative value, but in two cases its result
was assigned to an unsigned member of the hw_perf_event
structure, thus the error would not be ever detected.

Fixed by using an intermediate, signed variable.
Reported-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NPawel Moll <pawel.moll@arm.com>
Signed-off-by: NOlof Johansson <olof@lixom.net>

In commit ae91d60b, a bug was fixed that
involved converting !x & y to !(x & y).  The code below shows the same
pattern, and thus should perhaps be fixed in the same way.

The Coccinelle semantic patch that makes this change is as follows:

// <smpl>
@@ expression E1,E2; @@
(
  !E1 & !E2
|
- !E1 & E2
+ !(E1 & E2)
)
// </smpl>
Signed-off-by: NHimangi Saraogi <himangi774@gmail.com>
Acked-by: NJulia Lawall <julia.lawall@lip6.fr>
Acked-by: NPunit Agrawal <punit.agrawal@arm.com>
Signed-off-by: NOlof Johansson <olof@lixom.net>

The CCN driver makes no sense without PERF_EVENTS, and trying to
build it when that option is disabled results in compile errors,
so it's best to just add a strong Kconfig dependency.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

Driver providing perf backend for ARM Cache Coherent Network
interconnect. Supports counting all hardware events and crosspoint
watchpoints.

Currently works with CCN-504 only, although there should be
no changes required for CCN-508 (just impossible to test it now).
Signed-off-by: NPawel Moll <pawel.moll@arm.com>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

There could be some memory map devices located in
a certain chip select region of the i.MX WEIM.
The devices could be attached to a simple bus(for
example, a AXI bus) whose root node is one child
device tree node of the i.MX WEIM device tree node.
There should be a bridge(very likely, software
transparent) bewteen the i.MX WEIM and the simple bus.
This patch makes the i.MX WEIM driver possible to
populate devices on a simple bus.  In this way, people
may try various IPs(in a FPGA, maybe) outside of i.MX
chips with the i.MX WEIM embedded.
Signed-off-by: NLiu Ying <Ying.Liu@freescale.com>
Signed-off-by: NShawn Guo <shawn.guo@freescale.com>

devm_request_and_ioremap() was obsoleted by the commit 75096579
("lib: devres: Introduce devm_ioremap_resource()") and has been
deprecated for a long time. So, let's remove this function.
In addition, all usages of devm_request_and_ioremap() are also
removed.
Signed-off-by: NJingoo Han <jg1.han@samsung.com>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

The arm-cci code uses device tree helpers for initialization
that don't work on kernels built without CONFIG_OF. Further,
it contains an inline assembly in cci_enable_port_for_self()
that uses ARMv7 instructions and fails to build when targetting
other ARM instruction set versions.

This works around both issues by limiting the scope of the
Kconfig symbol to platforms that can actually build this driver
cleanly.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NLorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Shawn Guo <shawn.guo@linaro.org>

The versatile express changes for 3.16 introduced a number of
build regressions for randconfig kernels by not tracking dependencies
between the components right.

This patch tries to rectify that:

* the mach-vexpress code cannot link without the syscfg driver,
  which in turn needs MFD_VEXPRESS_SYSREG
* various drivers call devm_regmap_init_vexpress_config(), which
  has to be exported so it can be used by loadable modules
* the configuration bus uses OF DT helper functions that are not
  available to platforms disable CONFIG_OF
* The sysreg driver exports GPIOs through gpiolib, which can
  be disabled on some platforms.
* The clocksource code cannot be built on platforms that don't
  use modern timekeeping but rely on gettimeoffset.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

This patch adds support for the Broadcom GISB arbiter bus timeout/error
handler. GISB is a proprietary bus used by Broadcom Set Top Box
System-on-a-chip devices (BCM7xxx) which allows multiple masters and
clients to be interfaced with each other.

The bus arbiter offers support for generating two interrupts towards the
host CPU, thus allowing us to "catch" clock gated masters, or masters
being volontarily blocked for powersaving purposes, or do general system
troubleshooting.

We also register a hook with the ARM fault exception handling to allow
printing a more informative message than "imprecise external abort at
0x00000000" for instance.
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

Components of the Versatile Express platform (configuration
microcontrollers on motherboard and daughterboards in particular)
talk to each other over a custom configuration bus. They
provide miscellaneous functions (from clock generator control
to energy sensors) which are represented as platform devices
(and Device Tree nodes). The transactions on the bus can
be generated by different "bridges" in the system, some
of which are universal for the whole platform (for the price
of high transfer latencies), others restricted to a subsystem
(but much faster).

Until now drivers for such functions were using custom "func"
API, which is being replaced in this patch by regmap calls.
This required:

* a rework (and move to drivers/bus directory, as suggested
  by Samuel and Arnd) of the config bus core, which is much
  simpler now and uses device model infrastructure (class)
  to keep track of the bridges; non-DT case (soon to be
  retired anyway) is simply covered by a special device
  registration function

* the new config-bus driver also takes over device population,
  so there is no need for special matching table for
  of_platform_populate nor "simple-bus" hack in the arm64
  model dtsi file (relevant bindings documentation has
  been updated); this allows all the vexpress devices
  fit into normal device model, making it possible
  to remove plenty of early inits and other hacks in
  the near future

* adaptation of the syscfg bridge implementation in the
  sysreg driver, again making it much simpler; there is
  a special case of the "energy" function spanning two
  registers, where they should be both defined in the tree
  now, but backward compatibility is maintained in the code

* modification of the relevant drivers:

  * hwmon - just a straight-forward API change
  * power/reset driver - API change
  * regulator - API change plus error handling
    simplification
  * osc clock driver - this one required larger rework
    in order to turn in into a standard platform driver
Signed-off-by: NPawel Moll <pawel.moll@arm.com>
Acked-by: NMark Brown <broonie@linaro.org>
Acked-by: NLee Jones <lee.jones@linaro.org>
Acked-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NMike Turquette <mturquette@linaro.org>

Add AM4372 information to handle L3 error.

AM4372 has two clk domains 100f and 200s. Provide flagmux and data
associated with it.

NOTE: Timeout doesn't have STDERRLOG_MAIN register. And per hardware
team, L3 timeout error cannot be cleared the normal way (by setting
bit 31 in STDERRLOG_MAIN), instead it may be required to do system
reset. L3 error handler can't help in such scenarios.

Hence indicate timeout target offset as L3_TARGET_NOT_SUPPORTED as
done for undocumented bits.
Signed-off-by: NDave Gerlach <d-gerlach@ti.com>
Signed-off-by: NAfzal Mohammed <afzal@ti.com>
Signed-off-by: NSekhar Nori <nsekhar@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

DRA7 is distinctly different from OMAP4 in terms of masters and clock
domain organization. There two main clock domains which is divided as
follows:
     <0x44000000 0x1000000> is clk1 and clk2 is the sub clock domain
     <0x45000000 0x1000> is clk3

Add all the data needed to handle L3 error handling on DRA7 devices
and mark clk2 as subdomain and provide a compatible flag for
functionality. Other than the data difference the hardware blocks
involved are essentially the same.
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
[nm@ti.com: bugfixes and generic improvements, documentation]
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

While OMAP4 and OMAP5 had 3 separate clock domains, DRA7 has only 2
and the first one then is internally divided into 2 sub clock domains.

To better represent this in the driver, we use the concept of submodule.

The address defintions in the devicetree is as per the high level
clock domain(module) base, the sub clockdomain/subdomain which shares
the same register space of a clockdomain is marked in the SoC data as
L3_BASE_IS_SUBMODULE.

L3_BASE_IS_SUBMODULE is used as an indication that it's base address is
the same as the parent module and offsets are considered from the same
base address as they are usually intermingled.

Other than the base address, the submodule is same as a module as it is
functionally so.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

L3 error may be triggered using Debug interface (example JTAG) or
due to other errors, for example an opcode fetch (due to function
pointer or stack corruption) or a data access (due to some other
failure). NOC registers contain additional information to help aid
debug information.

With this, we can enhance the error information to more detailed form:
"
L3 Custom Error: MASTER MPU TARGET L4PER2 (Read): Data Access in User mode
during Functional access
"
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Today we get error such as
L3 Custom Error: MASTER MPU TARGET L4PER2

But since the actual instruction triggerring the error Vs the point
at which we report error may not be aligned, it makes sense to try
and provide additional information - example the type of operation
that was attempted to being performed can help narrow the debug down
further.

This helps provide log such as:
L3 Custom Error: MASTER MPU TARGET L4PER2 (Read)
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Errors that cannot be cleared (determined by reading REGERR register)
are currently handled by masking it. Documentation states that REGERR
"Checks which application/debug error sources are active" - it does not
indicate that this is "interrupt status" - masked out status represented
eventually in the irq line to MPU.
For example:

Lets say module 0 bit 8(0x100) was unclearable, we do the mask it from
generating further errors. However in the following cases:
a) bit 9 of Module 0
OR
b) any bit of Module 1+
occur, the interrupt handler wrongly assumes that the raw interrupt
status of module 0 bit 8 is the root cause of the interrupt, and
returns. This causes unhandled interrupt and resultant infinite
interrupts.

Fix this scenario by storing the events we masked out and masking raw
status with masked ones before identifying and handling the error.
Reported-by: NVaibhav Hiremath <hvaibhav@ti.com>
Signed-off-by: NAfzal Mohammed <afzal@ti.com>
Tested-by: NVaibhav Hiremath <hvaibhav@gmail.com>
Signed-off-by: NSekhar Nori <nsekhar@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Current interrupt handler does the first level parse to identify the
slave and then handles the slave even identification, reporting and
clearing of event as well. It is hence logical to split the handler
into two where the primary handler just parses the flagmux till it
identifies a slave and the slave handling, reporting and clearing is
done in a helper function.

While at it update the documentation in kerneldoc style.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

The logic between handling CUSTOM_ERROR and STANDARD_ERROR is just the
reporting style.

So make it generic, simplify and standardize the reporting with both
master and target information printed to log.

Handle the register address difference for master code for standard
error and custom error as well.

While at it, fix a minor indentation error.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

As per Documentation (OMAP4+), then masterid is infact encoded as
follows:
"L3_TARG_STDERRLOG_MSTADDR[7:0] STDERRLOG_MSTADDR stores the NTTP
master address. The master address is the concatenation of Prefix &
Initiator ConnID. It is defined on 8 bits. The 6 MSBs are used to
distinguish the different initiators."

So, when we matchup currently with the master ID list, we never get a
proper match other than when MPU is the master (thanks to 0).

Now, on other platforms such as AM437x, this tends to be bits[5:0].

Fix this by using the relevant 6MSBits to identify the master ID for
standard and custom errors.
Reported-by: NDarren Etheridge <detheridge@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

This allows us to encompass target information and flag mux offset that
points to the target information into a singular structure. This saves
us the need to look up two different arrays indexed by module ID for
information.

This allows us to reduce the static target information allocation to
just the ones that are documented.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

DRA7xx SoC has the same l3-noc interconnect ip (as OMAP4 and OMAP5), but
AM437x SoC has just 2 modules instead of 3 which other SoCs have.

So, stop using direct access of array indices and use of->match data and
simplify implementation to benefit future usage.

While at it, rename a few very generic variables to make them omap
specific. This helps us differentiate from DRA7 and AM43xx data in the
future.

NOTE: None of the platforms that use omap_l3_noc are non-device tree
anymore. So, it is safe to assume OF match here.
Signed-off-by: NSricharan R <r.sricharan@ti.com>
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
[nm@ti.com: split, refactor and optimize logic]
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

On DRA7, unlike on OMAP4 and OMAP5, the flag mux input numbers used
to indicate the source of errors are not continous. Have a way in the
driver to catch these and WARN the user of the flag mux input thats
either undocumented or wrong.

In the similar vein, Timeout errors in AM43x can't be cleared per h/w
team, neither does it have a STDERRLOG_MAIN to clear the error.

Further, the mux bit offset might not even be indexed into our array
of known mux input description, in which case we'd have a abort.

So, define a static range check for bit description and any definition
which has target_name set to NULL (the ones that are not populated or
ones that are specifically marked in the case of discontinous input
numbers), can handle the same gracefully. Upon occurance of error from
such sources, mask it. Otherwise, we'd have an infinite interrupt
source without any means to clear it.

NOTE: follow on patch ensures that these masked bits are ignored.

[nm@ti.com: rebase, squash and improve]
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
Signed-off-by: NAfzal Mohammed <afzal@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Currently the target instance information is organized indexed by bit
field offset into multiple arrays.

1. We currently have offsets specific to each target associated with each
clock domains are in seperate arrays:

l3_targ_inst_clk1
l3_targ_inst_clk2
l3_targ_inst_clk3

2. Then they are organized per master index in l3_targ.

3. We have names in l3_targ_inst_name as an array to array of strings
corresponding to the above with offsets.

Simplify the same by defining a structure for information containing
both target offset and name. this is then stored in arrays per domain
and organized into an array indexed off domain.

The array is still indexed based on bit field offset.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Move the L3 master structure out of the static definition to enable
reuse for other SoCs.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

just simplify derefencing that is equivalent.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Currently we use __raw_readl and writel in this driver. Considering
there is no specific need for a memory barrier, replacing writel
with endian-neutral writel_relaxed and replacing __raw_readls with
the corresponding endian-neutral readl_relaxed allows us to have a
standard set of register operations for the driver.

While at it, simplify address computation using variables for
register.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NFelipe Balbi <balbi@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

l3->dev is not populated, so populate it and use it to print information
relevant to the device instead of using a generic pr_*.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

we do not use iclk directly anymore. And, even if we had to, we
should be using pm_runtime APIs to do the same to be completely SoC
independent.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Since omap_l3_noc driver is now being used for OMAP5 and reusable with
DRA7 and AM437x, using omap4 specific naming is misleading.
Signed-off-by: NSricharan R <r.sricharan@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

This is an embarrassing patch :(.

Texas Corporation does not make OMAP. Texas Instruments Inc does.

For that matter I dont seem to be able to find a Texas Corporation on
the internet either.

While at it, update coverage to the current year and update the template
to remove redundant information and use the standard boiler plate
licensing.
Signed-off-by: NNishanth Menon <nm@ti.com>
Acked-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: NDarren Etheridge <detheridge@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Use dev_err() which will going to print the driver's name as well and the
KERN_ERR level is sufficient in this case (we also print via dev_err when
there is an error with the mem resources)
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

It is NOP after the devm_* conversion.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

With this we can remove the free_irq() calls from probe and remove.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

We can then remove the iounmap() calls from probe and remove.
Since the driver requests the resources via index we can do the mem resource
request within a for loop.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

We can remove the kfree() calls from probe and remove.
Signed-off-by: NPeter Ujfalusi <peter.ujfalusi@ti.com>
Reviewed-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Tested-by: NSekhar Nori <nsekhar@ti.com>

Until now, the mvebu-mbus was guessing by itself whether hardware I/O
coherency was available or not by poking into the Device Tree to see
if the coherency fabric Device Tree node was present or not.

However, on some upcoming SoCs, the presence or absence of the
coherency fabric DT node isn't sufficient: in CONFIG_SMP, the
coherency can be enabled, but not in !CONFIG_SMP.

In order to clean this up, the mvebu_mbus_dt_init() function is
extended to get a boolean argument telling whether coherency is
enabled or not. Therefore, the logic to decide whether coherency is
available or not now belongs to the core SoC code instead of the
mvebu-mbus driver itself, which is much better.
Signed-off-by: NThomas Petazzoni <thomas.petazzoni@free-electrons.com>
Link: https://lkml.kernel.org/r/1397483228-25625-4-git-send-email-thomas.petazzoni@free-electrons.comSigned-off-by: NJason Cooper <jason@lakedaemon.net>

Having multiple windows with the same target and attribute is actually
legal, and can be useful for PCIe windows, when PCIe BARs have a size
that isn't a power of two, and we therefore need to create several
MBus windows to cover the PCIe BAR for a given PCIe interface.

Fixes: fddddb52 ('bus: introduce an Marvell EBU MBus driver')
Cc: <stable@vger.kernel.org> # v3.10+
Signed-off-by: NThomas Petazzoni <thomas.petazzoni@free-electrons.com>
Link: https://lkml.kernel.org/r/1397823593-1932-7-git-send-email-thomas.petazzoni@free-electrons.comTested-by: NNeil Greatorex <neil@fatboyfat.co.uk>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

The mbus hardware requires a power of two size, and size aligned base.
Currently, if a non-power of two is passed in to the low level routines
they configure the register in a way that results in undefined behaviour.

Call WARN and return EINVAL instead.

Also, update the debugfs routines to show a message if there is an
invalid register setting.

All together this makes the recent problems with silent failure
of PCI very obvious, noisy and debuggable.
Signed-off-by: NJason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: NThomas Petazzoni <thomas.petazzoni@free-electrons.com>
Link: https://lkml.kernel.org/r/1397823593-1932-6-git-send-email-thomas.petazzoni@free-electrons.comSigned-off-by: NJason Cooper <jason@lakedaemon.net>

For imx50-weim and imx6q-weim type of devices, there might a WEIM CS
space configuration register in General Purpose Register controller,
e.g. IOMUXC_GPR1 on i.MX6Q.

Depending on which configuration of the following 4 is chosen for given
system, IOMUXC_GPR1[11:0] should be set up as 05, 033, 0113 or 01111
correspondingly.

	CS0(128M) CS1(0M)  CS2(0M)  CS3(0M)
	CS0(64M)  CS1(64M) CS2(0M)  CS3(0M)
	CS0(64M)  CS1(32M) CS2(32M) CS3(0M)
	CS0(32M)  CS1(32M) CS2(32M) CS3(32M)

The patch creates a function for such type of devices, which scans
'ranges' property of WEIM node and build the GPR value incrementally.
Thus the WEIM CS GPR can be set up automatically at boot time.
Signed-off-by: NShawn Guo <shawn.guo@linaro.org>
Reviewed-by: NPhilippe De Muyter <phdm@macqel.be>
Tested-by: NPhilippe De Muyter <phdm@macqel.be>

The event numbering changed between revision r0 and r1 of the CCI
PMU. Expose this to userspace to allow tooling to handle the
differences in event numbers.
Suggested-by: NDrew Richardson <Drew.Richardson@arm.com>
Signed-off-by: NPunit Agrawal <punit.agrawal@arm.com>
Reviewed-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>