Set-associative caches on all v7 implementations map the index bits
to physical addresses LSBs and tag bits to MSBs. As the last level
of cache on current and upcoming ARM systems grows in size,
this means that under normal DRAM controller configurations, the
current v7 cache flush routine using set/way operations triggers a
DRAM memory controller precharge/activate for every cache line
writeback since the cache routine cleans lines by first fixing the
index and then looping through ways (index bits are mapped to lower
physical addresses on all v7 cache implementations; this means that,
with last level cache sizes in the order of MBytes, lines belonging
to the same set but different ways map to different DRAM pages).
Given the random content of cache tags, swapping the order between
indexes and ways loops do not prevent DRAM pages precharge and
activate cycles but at least, on average, improves the chances that
either multiple lines hit the same page or multiple lines belong to
different DRAM banks, improving throughput significantly.
This patch swaps the inner loops in the v7 cache flushing routine
to carry out the clean operations first on all sets belonging to
a given way (looping through sets) and then decrementing the way.
Benchmarks showed that by swapping the ordering in which sets and
ways are decremented in the v7 cache flushing routine, that uses
set/way operations, time required to flush caches is reduced
significantly, owing to improved writebacks throughput to the DRAM
controller.
Benchmarks results vary and depend heavily on the last level of
cache tag RAM content when cache is cleaned and invalidated, ranging
from 2x throughput when all tag RAM entries contain dirty lines
mapping to sequential pages of RAM to 1x (ie no improvement) when
all tag RAM accesses trigger a DRAM precharge/activate cycle, as the
current code implies on most DRAM controller configurations.
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: Dave Martin <Dave.Martin@arm.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
System-wide barriers aren't required for situations where we only need
to make visibility and ordering guarantees in the inner-shareable domain
(i.e. we are not dealing with devices or potentially incoherent CPUs).
This patch changes the v7 TLB operations, coherent_user_range and
dcache_clean_area functions to user inner-shareable barriers. For cache
maintenance, only the store access type is required to ensure completion.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
On Cortex-A9 before version r1p0, the LoUIS bit field of the CLIDR
register returns zero when it should return one. This leads to cache
maintenance operations which rely on this value to not function as
intended, causing data corruption.
The workaround for this errata is to detect affected CPUs and correct
the LoUIS value read.
Acked-by: Will Deacon <will.deacon@arm.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Cc: stable@vger.kernel.org
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
mach-socfpga is another platform that needs to use
v7_invalidate_l1 to bringup additional cores. There was a comment that
the ideal place for v7_invalidate_l1 should be in arm/mm/cache-v7.S
Signed-off-by: Dinh Nguyen <dinguyen@altera.com>
Acked-by: Simon Horman <horms+renesas@verge.net.au>
Acked-by: Stephen Warren <swarren@nvidia.com>
Reviewed-by: Pavel Machek <pavel@denx.de>
Reviewed-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Tested-by: Pavel Machek <pavel@denx.de>
Tested-by: Stephen Warren <swarren@nvidia.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Olof Johansson <olof@lixom.net>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Rob Herring <rob.herring@calxeda.com>
Cc: Sascha Hauer <kernel@pengutronix.de>
Cc: Magnus Damm <magnus.damm@gmail.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
flush_cache_louis flushes the D-side caches to the point of unification
inner-shareable. On uniprocessor CPUs, this is defined as zero and
therefore no flushing will take place. Rather than invent a new interface
for UP systems, instead use our SMP_ON_UP patching code to read the
LoUU from the CLIDR instead.
Cc: <stable@vger.kernel.org>
Cc: Lorenzo Pieralisi <Lorenzo.Pieralisi@arm.com>
Tested-by: Guennadi Liakhovetski <g.liakhovetski@gmx.de>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
arm: Add ARM ERRATA 775420 workaround
Workaround for the 775420 Cortex-A9 (r2p2, r2p6,r2p8,r2p10,r3p0) erratum.
In case a date cache maintenance operation aborts with MMU exception, it
might cause the processor to deadlock. This workaround puts DSB before
executing ISB if an abort may occur on cache maintenance.
Based on work by Kouei Abe and feedback from Catalin Marinas.
Signed-off-by: Kouei Abe <kouei.abe.cp@rms.renesas.com>
[ horms@verge.net.au: Changed to implementation
suggested by catalin.marinas@arm.com ]
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch renames jump labels in v7_flush_dcache_all in order to define
a specific flush cache levels entry point.
Acked-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
ARM v7 architecture introduced the concept of cache levels and related
control registers. New processors like A7 and A15 embed an L2 unified cache
controller that becomes part of the cache level hierarchy. Some operations in
the kernel like cpu_suspend and __cpu_disable do not require a flush of the
entire cache hierarchy to DRAM but just the cache levels belonging to the
Level of Unification Inner Shareable (LoUIS), which in most of ARM v7 systems
correspond to L1.
The current cache flushing API used in cpu_suspend and __cpu_disable,
flush_cache_all(), ends up flushing the whole cache hierarchy since for
v7 it cleans and invalidates all cache levels up to Level of Coherency
(LoC) which cripples system performance when used in hot paths like hotplug
and cpuidle.
Therefore a new kernel cache maintenance API must be added to cope with
latest ARM system requirements.
This patch adds flush_cache_louis() to the ARM kernel cache maintenance API.
This function cleans and invalidates all data cache levels up to the
Level of Unification Inner Shareable (LoUIS) and invalidates the instruction
cache for processors that support it (> v7).
This patch also creates an alias of the cache LoUIS function to flush_kern_all
for all processor versions prior to v7, so that the current cache flushing
behaviour is unchanged for those processors.
v7 cache maintenance code implements a cache LoUIS function that cleans and
invalidates the D-cache up to LoUIS and invalidates the I-cache, according
to the new API.
Reviewed-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Tested-by: Shawn Guo <shawn.guo@linaro.org>
The cacheflush syscall can fail for two reasons:
(1) The arguments are invalid (nonsensical address range or no VMA)
(2) The region generates a translation fault on a VIPT or PIPT cache
This patch allows do_cache_op to return an error code to userspace in
the case of the above. The various coherent_user_range implementations
are modified to return 0 in the case of VIVT caches or -EFAULT in the
case of an abort on v6/v7 cores.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Bootup with lockdep enabled has been broken on v7 since b46c0f7465
("ARM: 7321/1: cache-v7: Disable preemption when reading CCSIDR").
This is because v7_setup (which is called very early during boot) calls
v7_flush_dcache_all, and the save_and_disable_irqs added by that patch
ends up attempting to call into lockdep C code (trace_hardirqs_off())
when we are in no position to execute it (no stack, MMU off).
Fix this by using a notrace variant of save_and_disable_irqs. The code
already uses the notrace variant of restore_irqs.
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Acked-by: Stephen Boyd <sboyd@codeaurora.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Rabin Vincent <rabin@rab.in>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
armv7's flush_cache_all() flushes caches via set/way. To
determine the cache attributes (line size, number of sets,
etc.) the assembly first writes the CSSELR register to select a
cache level and then reads the CCSIDR register. The CSSELR register
is banked per-cpu and is used to determine which cache level CCSIDR
reads. If the task is migrated between when the CSSELR is written and
the CCSIDR is read the CCSIDR value may be for an unexpected cache
level (for example L1 instead of L2) and incorrect cache flushing
could occur.
Disable interrupts across the write and read so that the correct
cache attributes are read and used for the cache flushing
routine. We disable interrupts instead of disabling preemption
because the critical section is only 3 instructions and we want
to call v7_dcache_flush_all from __v7_setup which doesn't have a
full kernel stack with a struct thread_info.
This fixes a problem we see in scm_call() when flush_cache_all()
is called from preemptible context and sometimes the L2 cache is
not properly flushed out.
Signed-off-by: Stephen Boyd <sboyd@codeaurora.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Nicolas Pitre <nico@linaro.org>
Cc: stable@vger.kernel.org
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch implements a workaround for erratum 764369 affecting
Cortex-A9 MPCore with two or more processors (all current revisions).
Under certain timing circumstances, a data cache line maintenance
operation by MVA targeting an Inner Shareable memory region may fail to
proceed up to either the Point of Coherency or to the Point of
Unification of the system. This workaround adds a DSB instruction before
the relevant cache maintenance functions and sets a specific bit in the
diagnostic control register of the SCU.
Cc: <stable@kernel.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The v6 and v7 implementations of flush_kern_dcache_area do not align
the passed MVA to the size of a cacheline in the data cache. If a
misaligned address is used, only a subset of the requested area will
be flushed. This has been observed to cause failures in SMP boot where
the secondary_data initialised by the primary CPU is not cacheline
aligned, causing the secondary CPUs to read incorrect values for their
pgd and stack pointers.
This patch ensures that the base address is cacheline aligned before
flushing the d-cache.
Cc: <stable@kernel.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The current implementation of the v7_coherent_*_range function assumes
that the D and I cache lines have the same size, which is incorrect
architecturally. This patch adds the icache_line_size macro which reads
the CTR register. The main loop in v7_coherent_*_range is split in two
independent loops or the D and I caches. This also has the performance
advantage that the DSB is moved outside the main loop.
Reported-by: Kevin Sapp <ksapp@quicinc.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Do this by adding flush_icache_all to cache_fns for ARMv6 and 7.
As flush_icache_all may neeed to be called from flush_kern_cache_all,
add it as the first entry in the cache_fns.
Note that now we can remove the ARM_ERRATA_411920 dependency
to !SMP so it can be selected on UP ARMv6 processors, such
as omap2.
Signed-off-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Anand Gadiyar <gadiyar@ti.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
UP systems do not implement all the instructions that SMP systems have,
so in order to boot a SMP kernel on a UP system, we need to rewrite
parts of the kernel.
Do this using an 'alternatives' scheme, where the kernel code and data
is modified prior to initialization to replace the SMP instructions,
thereby rendering the problematical code ineffectual. We use the linker
to generate a list of 32-bit word locations and their replacement values,
and run through these replacements when we detect a UP system.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch fixes the flush_cache_all for ARMv7 SMP.It was
missing from commit b8349b569a
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: <stable@kernel.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The standard I-cache Invalidate All (ICIALLU) and Branch Predication
Invalidate All (BPIALL) operations are not automatically broadcast to
the other CPUs in an ARMv7 MP system. The patch adds the Inner Shareable
variants, ICIALLUIS and BPIALLIS, if ARMv7 and SMP.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
ARMv6 and ARMv7 CPUs can perform speculative prefetching, which makes
DMA cache coherency handling slightly more interesting. Rather than
being able to rely upon the CPU not accessing the DMA buffer until DMA
has completed, we now must expect that the cache could be loaded with
possibly stale data from the DMA buffer.
Where DMA involves data being transferred to the device, we clean the
cache before handing it over for DMA, otherwise we invalidate the buffer
to get rid of potential writebacks. On DMA Completion, if data was
transferred from the device, we invalidate the buffer to get rid of
any stale speculative prefetches.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Tested-By: Santosh Shilimkar <santosh.shilimkar@ti.com>
These are now unused, and so can be removed.
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Tested-By: Santosh Shilimkar <santosh.shilimkar@ti.com>
This is needed because applications using the sys_cacheflush system call
can pass a memory range which isn't mapped yet even though the
corresponding vma is valid. The patch also adds unwinding annotations
for correct backtraces from the coherent_user_range() functions.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
The flush_cache_all function on ARMv7 is implemented as a series of
cache operations by set/way. These are not guaranteed to be ordered with
previous memory accesses, requiring a DMB. This patch also adds barriers
for the TLB operations in compressed/head.S
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
This declaration specifies the "function" type and size for various
assembly functions, mainly needed for generating the correct branch
instructions in Thumb-2.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
This patch adds support for the ARMv7 cores.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Lorenzo Pieralisi