Re: [PATCH] x86/mm/tlb: Remove flush_tlb_info from the stack

From: Nadav Amit
Date: Tue Apr 23 2019 - 12:57:00 EST


> On Apr 23, 2019, at 9:50 AM, Andy Lutomirski <luto@xxxxxxxxxx> wrote:
>
> On Tue, Apr 23, 2019 at 12:12 AM Nadav Amit <namit@xxxxxxxxxx> wrote:
>> Remove flush_tlb_info variables from the stack. This allows to align
>> flush_tlb_info to cache-line and avoid potentially unnecessary cache
>> line movements. It also allows to have a fixed virtual-to-physical
>> translation of the variables, which reduces TLB misses.
>>
>> Use per-CPU struct for flush_tlb_mm_range() and
>> flush_tlb_kernel_range(). Add debug assertions to ensure there are
>> no nested TLB flushes that might overwrite the per-CPU data. For
>> arch_tlbbatch_flush(), use a const struct.
>>
>> Results when running a microbenchmarks that performs 10^6 MADV_DONTEED
>> operations and touching a page, in which 3 additional threads run a
>> busy-wait loop (5 runs):
>
> Can you add a memset(,,,. 0, sizeof(struct flush_tlb_info)) everywhere
> you grab it? Or, even better, perhaps do something like:
>
> static inline struct flush_tlb_info *get_flush_tlb_info(void)
> {
> /* check reentrancy, make sure that we use smp_processor_id() or
> otherwise assert that we're bound to a single CPU. */
> struct flush_tlb_info *ptr = this_cpu_ptr(...);
> memset(ptr, 0, sizeof(*ptr));
> return ptr;
> }
>
> static inline void put_flush_tlb_info(void)
> {
> /* finish checking reentrancy. */
> }

Iâll check if the compiler is smart enough to avoid redundant assignments,
and if it is not, Iâll just give all the struct arguments to
get_flush_tlb_info() instead of memset() if you donât mind.

I also want to give a try for parallelizing the remote and local
invocations, which really annoys me every time I look at the code. Please
let me know if there is any âbig reasonâ that I am missing for it not to be
done before. I just hope all the paravirt stuff will not make it too ugly.