/hugepage

Primary LanguageRust

hugepage

关闭透明巨页

echo "never" > /sys/kernel/mm/transparent_hugepage/enabled

cat /sys/kernel/mm/transparent_hugepage/enabled
# 如果返回 如下输出则说明透明巨页的特性被关闭了
# always madvise [never]

也可以通过修改 /boot/grub2/grub.cfg 里面的内核启动参数来禁止（加上内核启动参数 transparent_hugepage=never），这里不再赘述。

相关阅读: Disable Transparent Hugepages

动态调整巨页数量

# 预留巨页数量
echo 70 > /proc/sys/vm/nr_hugepages

# 查看巨页相关的内存信息
cat /proc/meminfo | grep Huge

需要注意的是，如果系统运行很久之后，可能会导致无法保留足够多的巨页数量的情况，所以，最好在开机的时候执行该操作。

修改内核的启动参数

打开 /boot/grub/grub.cfg 文件，找到 menuentry 区块里面的 linux 行，并在末尾增加 default_hugepagesz=1G hugepagesz=1G 参数。

menuentry 'Debian ....' {
    linux   /boot/vmlinuz-4.15.0-135-generic root=UUID=01a8 ro  default_hugepagesz=1G hugepagesz=1G transparent_hugepage=never
}

相关的 内核启动参数 :

hugepages=      [HW,X86-32,IA-64] HugeTLB pages to allocate at boot.
hugepagesz=     [HW,IA-64,PPC,X86-64] The size of the HugeTLB pages.
                On x86-64 and powerpc, this option can be specified
                multiple times interleaved with hugepages= to reserve
                huge pages of different sizes. Valid pages sizes on
                x86-64 are 2M (when the CPU supports "pse") and 1G
                (when the CPU supports the "pdpe1gb" cpuinfo flag).

default_hugepagesz=
                [same as hugepagesz=] The size of the default
                HugeTLB page size. This is the size represented by
                the legacy /proc/ hugepages APIs, used for SHM, and
                default size when mounting hugetlbfs filesystems.
                Defaults to the default architecture's huge page size
                if not specified.
transparent_hugepage=
                [KNL]
                Format: [always|madvise|never]
                Can be used to control the default behavior of the system
                with respect to transparent hugepages.
                See Documentation/vm/transhuge.txt for more details.
nohugeiomap     [KNL,x86] Disable kernel huge I/O mappings.