[PATCH] arm64: add a function to show the different types of pagetable
Mark Rutland
mark.rutland at arm.com
Fri Dec 4 09:52:33 PST 2015
On Fri, Dec 04, 2015 at 07:43:17PM +0800, zhongjiang wrote:
> The patch is mainly to show pagetable number of different level in the direct
> mapping. pagetable is created from pud to pte in arm64 , resulting in different
> resluts with x86. For instance, The kernel of two-level pages will produce three
> types of pagetable.
>
> It can also be used to detect whether there is a large page spliting and merging.
> Large page will significantly reduce the TLB miss, and improve the system
> performance.
As I mentioned previously, I still think it makes more sense to expose
this via the pagetable dumping code.
Also, the EFI runtime pagetable code uses __create_mapping, and hence
alloc_init_{pud,pmd,pte}. This code doesn't seem to account for that, so it
looks like the values will be misleading.
> Signed-off-by: zhongjiang <zhongjiang at huawei.com>
> ---
> arch/arm64/include/asm/pgtable-types.h | 19 +++++++++++++++++
> arch/arm64/mm/mmu.c | 12 +++++++++++
> arch/arm64/mm/pageattr.c | 35 ++++++++++++++++++++++++++++++++
> 3 files changed, 66 insertions(+), 0 deletions(-)
>
> diff --git a/arch/arm64/include/asm/pgtable-types.h b/arch/arm64/include/asm/pgtable-types.h
> index 2b1bd7e..a0f58d0 100644
> --- a/arch/arm64/include/asm/pgtable-types.h
> +++ b/arch/arm64/include/asm/pgtable-types.h
> @@ -86,6 +86,25 @@ typedef pteval_t pgprot_t;
>
> #endif /* STRICT_MM_TYPECHECKS */
>
> +struct seq_file;
> +extern void arch_report_meminfo(struct seq_file *m);
> +
> +enum pg_level {
> + PG_LEVEL_NONE,
This is never used. It can go.
> + PG_LEVEL_PTE,
> + PG_LEVEL_PMD,
> + PG_LEVEL_PUD,
> + PG_LEVEL_NUM
> +};
[...]
> diff --git a/arch/arm64/mm/mmu.c b/arch/arm64/mm/mmu.c
> index 0a7bee7..77aef0b 100644
> --- a/arch/arm64/mm/mmu.c
> +++ b/arch/arm64/mm/mmu.c
> @@ -30,6 +30,7 @@
> #include <linux/stop_machine.h>
> #include <linux/bootmem.h>
>
> +#include <asm/pgtable-types.h>
> #include <asm/cputype.h>
> #include <asm/fixmap.h>
> #include <asm/sections.h>
Nit: please keep includes ordered alphabetically.
[...]
> diff --git a/arch/arm64/mm/pageattr.c b/arch/arm64/mm/pageattr.c
> index 7a5ff11..a8257a2 100644
> --- a/arch/arm64/mm/pageattr.c
> +++ b/arch/arm64/mm/pageattr.c
> @@ -15,12 +15,47 @@
> #include <linux/module.h>
> #include <linux/sched.h>
>
> +#include <linux/seq_file.h>
> #include <asm/pgalloc.h>
> #include <asm/pgtable.h>
> #include <asm/tlbflush.h>
Nit: please keep includes ordered alphabetically.
[...]
> +void arch_report_meminfo(struct seq_file *m)
> +{
> +
> + seq_printf(m, "DirectMap%ldk: %8lu kB\n", PAGE_SIZE / SZ_1K,
> + direct_pages_count[PG_LEVEL_PTE] * PAGE_SIZE / SZ_1K);
> +
> +#if CONFIG_PGTABLE_LEVELS == 2
You can change this to >= 2 ...
> + seq_printf(m, "DirectMap%ldM: %8lu kB\n", PMD_SIZE / SZ_1M,
> + direct_pages_count[PG_LEVEL_PMD] * PMD_SIZE / SZ_1K);
> +
> +#endif
> +
> +#if CONFIG_PGTABLE_LEVELS > 2
> + seq_printf(m, "DirectMap%ldM: %8lu kB\n", PMD_SIZE / SZ_1M,
> + direct_pages_count[PG_LEVEL_PMD] * PMD_SIZE / SZ_1K);
.... and get rid of this line.
> + seq_printf(m, "DirectMap%ldG: %8lu kB\n", PUD_SIZE / SZ_1G,
> + direct_pages_count[PG_LEVEL_PUD] * PUD_SIZE / SZ_1K);
> +
> +#endif
> +}
Thanks,
Mark.
More information about the linux-arm-kernel
mailing list