Since page cache pages are now considered 'dirty' by default, the cache
flushing is handled via __flush_dcache_page() when a page gets mapped to
user space. Highmem pages on VIVT systems are flushed during kunmap()
and flush_kernel_dcache_page() was already a no-op in this case.
ARCH_HAS_FLUSH_KERNEL_DCACHE_PAGE is still defined since ARM needs
specific implementations for flush_kernel_vmap_range() and
invalidate_kernel_vmap_range().
Cc: Nicolas Pitre <nicolas.pitre@linaro.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
#define ARCH_HAS_FLUSH_KERNEL_DCACHE_PAGE
static inline void flush_kernel_dcache_page(struct page *page)
{
- /* highmem pages are always flushed upon kunmap already */
- if ((cache_is_vivt() || cache_is_vipt_aliasing()) && !PageHighMem(page))
- __cpuc_flush_dcache_area(page_address(page), PAGE_SIZE);
}
#define flush_dcache_mmap_lock(mapping) \