linux-stable/arch/riscv/mm/tlbflush.c
Alexandre Ghiti 54d7431af7
riscv: Add support for BATCHED_UNMAP_TLB_FLUSH
Allow to defer the flushing of the TLB when unmapping pages, which allows
to reduce the numbers of IPI and the number of sfence.vma.

The ubenchmarch used in commit 43b3dfdd0455 ("arm64: support
batched/deferred tlb shootdown during page reclamation/migration") that
was multithreaded to force the usage of IPI shows good performance
improvement on all platforms:

* Unmatched: ~34%
* TH1520   : ~78%
* Qemu     : ~81%

In addition, perf on qemu reports an important decrease in time spent
dealing with IPIs:

Before:  68.17%  main     [kernel.kallsyms]            [k] __sbi_rfence_v02_call
After :   8.64%  main     [kernel.kallsyms]            [k] __sbi_rfence_v02_call

* Benchmark:

int stick_this_thread_to_core(int core_id) {
        int num_cores = sysconf(_SC_NPROCESSORS_ONLN);
        if (core_id < 0 || core_id >= num_cores)
           return EINVAL;

        cpu_set_t cpuset;
        CPU_ZERO(&cpuset);
        CPU_SET(core_id, &cpuset);

        pthread_t current_thread = pthread_self();
        return pthread_setaffinity_np(current_thread,
sizeof(cpu_set_t), &cpuset);
}

static void *fn_thread (void *p_data)
{
        int ret;
        pthread_t thread;

        stick_this_thread_to_core((int)p_data);

        while (1) {
                sleep(1);
        }

        return NULL;
}

int main()
{
        volatile unsigned char *p = mmap(NULL, SIZE, PROT_READ | PROT_WRITE,
                                         MAP_SHARED | MAP_ANONYMOUS, -1, 0);
        pthread_t threads[4];
        int ret;

        for (int i = 0; i < 4; ++i) {
                ret = pthread_create(&threads[i], NULL, fn_thread, (void *)i);
                if (ret)
                {
                        printf("%s", strerror (ret));
                }
        }

        memset(p, 0x88, SIZE);

        for (int k = 0; k < 10000; k++) {
                /* swap in */
                for (int i = 0; i < SIZE; i += 4096) {
                        (void)p[i];
                }

                /* swap out */
                madvise(p, SIZE, MADV_PAGEOUT);
        }

        for (int i = 0; i < 4; i++)
        {
                pthread_cancel(threads[i]);
        }

        for (int i = 0; i < 4; i++)
        {
                pthread_join(threads[i], NULL);
        }

        return 0;
}

Signed-off-by: Alexandre Ghiti <alexghiti@rivosinc.com>
Reviewed-by: Jisheng Zhang <jszhang@kernel.org>
Tested-by: Jisheng Zhang <jszhang@kernel.org> # Tested on TH1520
Tested-by: Nam Cao <namcao@linutronix.de>
Link: https://lore.kernel.org/r/20240108193640.344929-1-alexghiti@rivosinc.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
2024-01-11 08:01:53 -08:00

232 lines
5.4 KiB
C

// SPDX-License-Identifier: GPL-2.0
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/sched.h>
#include <linux/hugetlb.h>
#include <asm/sbi.h>
#include <asm/mmu_context.h>
static inline void local_flush_tlb_all_asid(unsigned long asid)
{
if (asid != FLUSH_TLB_NO_ASID)
__asm__ __volatile__ ("sfence.vma x0, %0"
:
: "r" (asid)
: "memory");
else
local_flush_tlb_all();
}
static inline void local_flush_tlb_page_asid(unsigned long addr,
unsigned long asid)
{
if (asid != FLUSH_TLB_NO_ASID)
__asm__ __volatile__ ("sfence.vma %0, %1"
:
: "r" (addr), "r" (asid)
: "memory");
else
local_flush_tlb_page(addr);
}
/*
* Flush entire TLB if number of entries to be flushed is greater
* than the threshold below.
*/
static unsigned long tlb_flush_all_threshold __read_mostly = 64;
static void local_flush_tlb_range_threshold_asid(unsigned long start,
unsigned long size,
unsigned long stride,
unsigned long asid)
{
unsigned long nr_ptes_in_range = DIV_ROUND_UP(size, stride);
int i;
if (nr_ptes_in_range > tlb_flush_all_threshold) {
local_flush_tlb_all_asid(asid);
return;
}
for (i = 0; i < nr_ptes_in_range; ++i) {
local_flush_tlb_page_asid(start, asid);
start += stride;
}
}
static inline void local_flush_tlb_range_asid(unsigned long start,
unsigned long size, unsigned long stride, unsigned long asid)
{
if (size <= stride)
local_flush_tlb_page_asid(start, asid);
else if (size == FLUSH_TLB_MAX_SIZE)
local_flush_tlb_all_asid(asid);
else
local_flush_tlb_range_threshold_asid(start, size, stride, asid);
}
static void __ipi_flush_tlb_all(void *info)
{
local_flush_tlb_all();
}
void flush_tlb_all(void)
{
if (riscv_use_ipi_for_rfence())
on_each_cpu(__ipi_flush_tlb_all, NULL, 1);
else
sbi_remote_sfence_vma_asid(NULL, 0, FLUSH_TLB_MAX_SIZE, FLUSH_TLB_NO_ASID);
}
struct flush_tlb_range_data {
unsigned long asid;
unsigned long start;
unsigned long size;
unsigned long stride;
};
static void __ipi_flush_tlb_range_asid(void *info)
{
struct flush_tlb_range_data *d = info;
local_flush_tlb_range_asid(d->start, d->size, d->stride, d->asid);
}
static void __flush_tlb_range(struct cpumask *cmask, unsigned long asid,
unsigned long start, unsigned long size,
unsigned long stride)
{
struct flush_tlb_range_data ftd;
bool broadcast;
if (cpumask_empty(cmask))
return;
if (cmask != cpu_online_mask) {
unsigned int cpuid;
cpuid = get_cpu();
/* check if the tlbflush needs to be sent to other CPUs */
broadcast = cpumask_any_but(cmask, cpuid) < nr_cpu_ids;
} else {
broadcast = true;
}
if (broadcast) {
if (riscv_use_ipi_for_rfence()) {
ftd.asid = asid;
ftd.start = start;
ftd.size = size;
ftd.stride = stride;
on_each_cpu_mask(cmask,
__ipi_flush_tlb_range_asid,
&ftd, 1);
} else
sbi_remote_sfence_vma_asid(cmask,
start, size, asid);
} else {
local_flush_tlb_range_asid(start, size, stride, asid);
}
if (cmask != cpu_online_mask)
put_cpu();
}
static inline unsigned long get_mm_asid(struct mm_struct *mm)
{
return static_branch_unlikely(&use_asid_allocator) ?
atomic_long_read(&mm->context.id) & asid_mask : FLUSH_TLB_NO_ASID;
}
void flush_tlb_mm(struct mm_struct *mm)
{
__flush_tlb_range(mm_cpumask(mm), get_mm_asid(mm),
0, FLUSH_TLB_MAX_SIZE, PAGE_SIZE);
}
void flush_tlb_mm_range(struct mm_struct *mm,
unsigned long start, unsigned long end,
unsigned int page_size)
{
__flush_tlb_range(mm_cpumask(mm), get_mm_asid(mm),
start, end - start, page_size);
}
void flush_tlb_page(struct vm_area_struct *vma, unsigned long addr)
{
__flush_tlb_range(mm_cpumask(vma->vm_mm), get_mm_asid(vma->vm_mm),
addr, PAGE_SIZE, PAGE_SIZE);
}
void flush_tlb_range(struct vm_area_struct *vma, unsigned long start,
unsigned long end)
{
unsigned long stride_size;
if (!is_vm_hugetlb_page(vma)) {
stride_size = PAGE_SIZE;
} else {
stride_size = huge_page_size(hstate_vma(vma));
/*
* As stated in the privileged specification, every PTE in a
* NAPOT region must be invalidated, so reset the stride in that
* case.
*/
if (has_svnapot()) {
if (stride_size >= PGDIR_SIZE)
stride_size = PGDIR_SIZE;
else if (stride_size >= P4D_SIZE)
stride_size = P4D_SIZE;
else if (stride_size >= PUD_SIZE)
stride_size = PUD_SIZE;
else if (stride_size >= PMD_SIZE)
stride_size = PMD_SIZE;
else
stride_size = PAGE_SIZE;
}
}
__flush_tlb_range(mm_cpumask(vma->vm_mm), get_mm_asid(vma->vm_mm),
start, end - start, stride_size);
}
void flush_tlb_kernel_range(unsigned long start, unsigned long end)
{
__flush_tlb_range((struct cpumask *)cpu_online_mask, FLUSH_TLB_NO_ASID,
start, end - start, PAGE_SIZE);
}
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
void flush_pmd_tlb_range(struct vm_area_struct *vma, unsigned long start,
unsigned long end)
{
__flush_tlb_range(mm_cpumask(vma->vm_mm), get_mm_asid(vma->vm_mm),
start, end - start, PMD_SIZE);
}
#endif
bool arch_tlbbatch_should_defer(struct mm_struct *mm)
{
return true;
}
void arch_tlbbatch_add_pending(struct arch_tlbflush_unmap_batch *batch,
struct mm_struct *mm,
unsigned long uaddr)
{
cpumask_or(&batch->cpumask, &batch->cpumask, mm_cpumask(mm));
}
void arch_flush_tlb_batched_pending(struct mm_struct *mm)
{
flush_tlb_mm(mm);
}
void arch_tlbbatch_flush(struct arch_tlbflush_unmap_batch *batch)
{
__flush_tlb_range(&batch->cpumask, FLUSH_TLB_NO_ASID, 0,
FLUSH_TLB_MAX_SIZE, PAGE_SIZE);
}