As explained in more detail in the second patch, I have observed a soft lockup under some loads. These lockups were in flush_tlb_kernel_range() which was looping through a very large address range. While looking into this, I also noticed the flush routines in tlb.S were not properly handling pages larger than 4k. This is corrected in the first patch. The second patch limits the loop size for the flush_tlb_[kernel_]range functions. It uses an arbitrary constant to limit the loop, but it would be better if it were based on actual tlb size or some other heuristic.
Mark Salter (2): arm64: fix unnecessary tlb flushes arm64: fix soft lockup due to large tlb flush range arch/arm64/include/asm/tlbflush.h | 28 +++++++++++++++++++++++++--- arch/arm64/mm/tlb.S | 4 ++-- 2 files changed, 27 insertions(+), 5 deletions(-) -- 1.9.0 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
