On OMAP4 SOC, intecronnects has many write buffers in the async bridges
and they need to be drained before CPU enters into standby state.

Patch 'OMAP4: PM: Add CPUX OFF mode support' added CPU PM support
but OMAP errata i688 (Async Bridge Corruption) needs to be taken
care to avoid issues like system freeze, CPU deadlocks, random
crashes with register accesses, synchronisation loss on initiators
operating on both interconnect port simultaneously.

As per the errata, if a data is stalled inside asynchronous bridge
because of back pressure, it may be accepted multiple times, creating
pointer misalignment that will corrupt next transfers on that data
path until next reset of the system (No recovery procedure once
the issue is hit, the path remains consistently broken).
Async bridge can be found on path between MPU to EMIF and
MPU to L3 interconnect. This situation can happen only when the
idle is initiated by a Master Request Disconnection (which is
trigged by software when executing WFI on CPU).

The work-around for this errata needs all the initiators
connected through async bridge must ensure that data path
is properly drained before issuing WFI. This condition will be
met if one Strongly ordered access is performed to the
target right before executing the WFI. In MPU case, L3 T2ASYNC
FIFO and DDR T2ASYNC FIFO needs to be drained. IO barrier ensure
that there is no synchronisation loss on initiators operating
on both interconnect port simultaneously.

Thanks to Russell for a tip to conver assembly function to
C fuction there by reducing 40 odd lines of code from the patch.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: NRichard Woodruff <r-woodruff2@ti.com>
Acked-by: NJean Pihet <j-pihet@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Tested-by: NVishwanath BS <vishwanath.bs@ti.com>
Signed-off-by: NKevin Hilman <khilman@ti.com>

OMAP WakeupGen is the interrupt controller extension used along
with ARM GIC to wake the CPU out from low power states on
external interrupts.

The WakeupGen unit is responsible for generating the wakeup event
from the incoming interrupts and enable bits. It is implemented
in the MPU always ON power domain. During normal operation,
WakeupGen delivers the external interrupts directly to the GIC.

WakeupGen specification has one restriction as per Veyron version 1.6.
It is SW responsibility to program interrupt enabling/disabling
coherently in the GIC and in the WakeupGen enable registers. That is, a
given interrupt for a given CPU is either enable at both GIC and WakeupGen,
or disable at both, but no mix. That's the reason the WakeupGen is
implemented as an extension of GIC.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NJean Pihet <j-pihet@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Tested-by: NVishwanath BS <vishwanath.bs@ti.com>
Signed-off-by: NKevin Hilman <khilman@ti.com>

This patch adds SAR RAM support on OMAP4430. SAR RAM used to save
and restore the HW context in low power modes.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NJean Pihet <j-pihet@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Tested-by: NVishwanath BS <vishwanath.bs@ti.com>
Signed-off-by: NKevin Hilman <khilman@ti.com>

This patch exports APIs to get base address for GIC
distributor, CPU interface, SCU and PL310 L2 Cache which
are used in OMAP4 PM code.

This was suggested by Kevin Hilman <khilman@ti.com> during
OMAP4 PM code review.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NJean Pihet <j-pihet@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Tested-by: NVishwanath BS <vishwanath.bs@ti.com>
Signed-off-by: NKevin Hilman <khilman@ti.com>

OMAP4 L2X0 initialisation code uses BUG_ON() for the ioremap()
failure scenarios.

Use WARN_ON() instead and allow graceful function exits.

This was suggsted by Kevin Hilman <khilman@ti.com> during
OMAP4 PM code review.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: NJean Pihet <j-pihet@ti.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>
Tested-by: NVishwanath BS <vishwanath.bs@ti.com>
Signed-off-by: NKevin Hilman <khilman@ti.com>

As suggested by Russell King - ARM Linux <linux@arm.linux.org.uk>,
there's no need to keep local prototypes in non-local headers.

Add mach-omap1/common.h and mach-omap2/common.h and move the
local prototypes there from plat/common.h and mach/omap4-common.h.
Signed-off-by: NTony Lindgren <tony@atomide.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

After the MULTI_IRQ_HANDLER conversion, a couple of global
variables can be removed.
Signed-off-by: NMarc Zyngier <marc.zyngier@arm.com>

This allows us to remove cpu_is_omap calls from init_irq functions.
There should not be any need for cpu_is_omap calls as at this point.
During the timer init we only care about SoC generation, and not about
subrevisions.

The main reason for the patch is that we want to initialize only
minimal omap specific code from the init_early call.
Signed-off-by: NTony Lindgren <tony@atomide.com>
Reviewed-by: NKevin Hilman <khilman@ti.com>

Populate the l2x0 set_debug function pointer with OMAP secure call
and enable the PL310 Errata 727915

This patch has dependency on the earlier patch
ARM: l2x0: Errata fix for flush by Way operation can cause data
corruption
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NCatalin Marinas <catalin.marinas@arm.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

The AXI protocol specifies that the write response can only
be sent back to an AXI master when the last write data has been
accepted. This optimization enables the PL310 to send the write
response of certain write transactions as soon as the store buffer
accepts the write address. This behavior is not compatible with
the AXI protocol and is disabled by default. You enable this
optimization by setting the Early BRESP Enable bit in the
Auxiliary Control Register (bit [30]).
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: NMans Rullgard <mans@mansr.com>
Tested-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Clearing bit 22 in the PL310 Auxiliary Control register (shared
attribute override enable) has the side effect of transforming Normal
Shared Non-cacheable reads into Cacheable no-allocate reads.

Coherent DMA buffers in Linux always have a Cacheable alias via the
kernel linear mapping and the processor can speculatively load cache
lines into the PL310 controller. With bit 22 cleared, Non-cacheable
reads would unexpectedly hit such cache lines leading to buffer
corruption
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Tested-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Enabling L2 prefetching improves performance as shown on Panda
ES2.1 board with mem test, and it has measurable impact on
performances. I think we should consider it, even though it damages
"writes" a bit. (rebased to k.org)
Usually the prefetch is used at both levels together L1 + L2, however,
to enable the CP15 prefetch engines, these are under security, and on
GP devices, we cannot enable it(e.g. on PandaBoard). However, just
enabling PL310 prefetch seems to provide performance improvement,
as shown in the data below (from Ubuntu) and would be a great thing
to pull in.

What prefetch does is enable automatic next line prefetching. With this
enabled, whenever the PL310 receives a cachable read request, it
automatically prefetches the following cache line as well.

Measurement Data:
==
STOCK 10.10 WITHOUT PATCH

========================
~# ./memspeed
size    8388608 8192k 8M
offset  8388608, 0
buffers 0x2aaad000 0x2b2ad000
copy  libc          133 MB/s
copy  Android v5    273 MB/s
copy  Android NEON  235 MB/s
copy  INT32         116 MB/s
copy  ASM ARM       187 MB/s
copy  ASM VLDM 64   204 MB/s
copy  ASM VLDM 128  173 MB/s
copy  ASM VLD1      216 MB/s
read  ASM ARM       286 MB/s
read  ASM VLDM      242 MB/s
read  ASM VLD1      286 MB/s
write libc         1947 MB/s
write ASM ARM      1943 MB/s
write ASM VSTM     1942 MB/s
write ASM VST1     1935 MB/s

10.10 + PATCH
=============
~# ./memspeed
size    8388608 8192k 8M
offset  8388608, 0
buffers 0x2ab17000 0x2b317000
copy  libc          129 MB/s
copy  Android v5    256 MB/s
copy  Android NEON  356 MB/s
copy  INT32         127 MB/s
copy  ASM ARM       321 MB/s
copy  ASM VLDM 64   337 MB/s
copy  ASM VLDM 128  321 MB/s
copy  ASM VLD1      350 MB/s
read  ASM ARM       496 MB/s
read  ASM VLDM      470 MB/s
read  ASM VLD1      488 MB/s
write libc         1701 MB/s
write ASM ARM      1682 MB/s
write ASM VSTM     1693 MB/s
write ASM VST1     1681 MB/s
Signed-off-by: NMans Rullgard <mans@mansr.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Tested-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

This patch removes the hardcoded value of auxctrl value and
construct it using bitfields

Bit 25 is reserved and is always set to 1. Same value
of this bit is retained in this patch
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Tested-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Every architecture using the GIC has a gic_cpu_base_addr pointer for
GIC 0 for their entry assembly code to use to decode the cause of the
current interrupt.  Move this into the common GIC code.
Reviewed-by: NCatalin Marinas <catalin.marinas@arm.com>
Tested-by: NAbhijeet Dharmapurikar <adharmap@codeaurora.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Provide gic_init() which initializes the GIC distributor and current
CPU's GIC interface for the boot (or single) CPU.
Reviewed-by: NCatalin Marinas <catalin.marinas@arm.com>
Tested-by: NAbhijeet Dharmapurikar <adharmap@codeaurora.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

The machine_kexec() calls outer_disable which can crash on OMAP4
becasue of trustzone restrictions.

This patch overrides the default l2x0_disable with a OMAP4
specific implementation taking care of trustzone
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NTony Lindgren <tony@atomide.com>
Acked-by: NLinus Walleij <linus.walleij@stericsson.com>

On ES2.0 the L2 cache init parameter ineeds to be changed to take
care of cache size. The cache size is 1MB on ES2.0 vs 512KB on ES1.0

This patch fixes the init parameter to update the same using
dynamic cpu version check
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>

This patch moves OMAP4 soc specific code from 4430sdp board file.
The change is necessary so that newer board support can be added
with minimal changes. This will be also problematic for
multi-board, multi-omap builds.
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>