This reverts commit 1f85b42a69.
The internal dma-direct.h API has changed in -next, which collides with
us trying to use it to manage non-coherent DMA devices on systems with
unreasonably large cache writeback granules.
This isn't at all trivial to resolve, so revert our changes for now and
we can revisit this after the merge window. Effectively, this just
restores our behaviour back to that of 4.16.
Signed-off-by: Will Deacon <will.deacon@arm.com>
The DCache clean & ICache invalidation requirements for instructions
to be data coherence are discoverable through new fields in CTR_EL0.
The following two control bits DIC and IDC were defined for this
purpose. No need to perform point of unification cache maintenance
operations from software on systems where CPU caches are transparent.
This patch optimize the three functions __flush_cache_user_range(),
clean_dcache_area_pou() and invalidate_icache_range() if the hardware
reports CTR_EL0.IDC and/or CTR_EL0.IDC. Basically it skips the two
instructions 'DC CVAU' and 'IC IVAU', and the associated loop logic
in order to avoid the unnecessary overhead.
CTR_EL0.DIC: Instruction cache invalidation requirements for
instruction to data coherence. The meaning of this bit[29].
0: Instruction cache invalidation to the point of unification
is required for instruction to data coherence.
1: Instruction cache cleaning to the point of unification is
not required for instruction to data coherence.
CTR_EL0.IDC: Data cache clean requirements for instruction to data
coherence. The meaning of this bit[28].
0: Data cache clean to the point of unification is required for
instruction to data coherence, unless CLIDR_EL1.LoC == 0b000
or (CLIDR_EL1.LoUIS == 0b000 && CLIDR_EL1.LoUU == 0b000).
1: Data cache clean to the point of unification is not required
for instruction to data coherence.
Co-authored-by: Philip Elcan <pelcan@codeaurora.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Shanker Donthineni <shankerd@codeaurora.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Commit 9730348075 ("arm64: Increase the max granular size") increased
the cache line size to 128 to match Cavium ThunderX, apparently for some
performance benefit which could not be confirmed. This change, however,
has an impact on the network packets allocation in certain
circumstances, requiring slightly over a 4K page with a significant
performance degradation.
This patch reverts L1_CACHE_SHIFT back to 6 (64-byte cache line) while
keeping ARCH_DMA_MINALIGN at 128. The cache_line_size() function was
changed to default to ARCH_DMA_MINALIGN in the absence of a meaningful
CTR_EL0.CWG bit field.
In addition, if a system with ARCH_DMA_MINALIGN < CTR_EL0.CWG is
detected, the kernel will force swiotlb bounce buffering for all
non-coherent devices since DMA cache maintenance on sub-CWG ranges is
not safe, leading to data corruption.
Cc: Tirumalesh Chalamarla <tchalamarla@cavium.com>
Cc: Timur Tabi <timur@codeaurora.org>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Add support for detecting VPIPT I-caches, as introduced by ARMv8.2.
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
cachetype.h and cache.h are small and both obviously related to caches.
Merge them together to reduce clutter.
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Increase the standard cacheline size to avoid having locks in the same
cacheline.
Cavium's ThunderX core implements cache lines of 128 byte size. With
current granulare size of 64 bytes (L1_CACHE_SHIFT=6) two locks could
share the same cache line leading a performance degradation.
Increasing the size fixes that.
Increasing the size has no negative impact to cache invalidation on
systems with a smaller cache line. There is an impact on memory usage,
but that's not too important for arm64 use cases.
Signed-off-by: Tirumalesh Chalamarla <tchalamarla@cavium.com>
Signed-off-by: Robert Richter <rrichter@cavium.com>
Acked-by: Timur Tabi <timur@codeaurora.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
As putting data which is read mostly together, we can avoid
unnecessary cache line bouncing.
Other architectures, such as ARM and x86, adopted the same idea.
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Jungseok Lee <jungseoklee85@gmail.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
The hardware provides the maximum cache line size in the system via the
CTR_EL0.CWG bits. This patch implements the cache_line_size() function
to read such information, together with a sanity check if the statically
defined L1_CACHE_BYTES is smaller than the hardware value.
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
The patch adds functionality required for cache maintenance. The AArch64
architecture mandates non-aliasing VIPT or PIPT D-cache and VIPT (may
have aliases) or ASID-tagged VIVT I-cache. Cache maintenance operations
are automatically broadcast in hardware between CPUs.
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Nicolas Pitre <nico@linaro.org>
Acked-by: Olof Johansson <olof@lixom.net>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
