Re: [GIT PULL] x86: add brk allocator for very early allocations

[Yinghai Lu <yinghai@kernel.org>]

Jeremy Fitzhardinge wrote:
> H. Peter Anvin wrote:
>> Those references are just the actual reservation of memory.  If all
>> the users of that memory are converted to either brk or bss,
>> INIT_MAP_BEYOND_END should be removed.  If all the users of that
>> memory aren't converted to brk to bss, we should do so.
> 
> I just added an adapted Yinghai's patch and added it to push/x86/brk.
> 
>    J
> 
> 
> The following changes since commit
> 8131667360004a0b74e4dcadfee8a18d4e2b074f:
>  Jeremy Fitzhardinge (1):
>        x86: allow extend_brk users to reserve brk space
> 
> are available in the git repository at:
> 
>  git://git.kernel.org/pub/scm/linux/kernel/git/jeremy/xen.git push/x86/brk
> 
> Yinghai Lu (1):
>      x86: put initial_pg_tables into .bss -v4
> 
> arch/x86/kernel/head_32.S        |   47
> ++++++++++++++-----------------------
> arch/x86/kernel/vmlinux_32.lds.S |    6 +++++
> 2 files changed, 24 insertions(+), 29 deletions(-)
> 
> diff --git a/arch/x86/kernel/head_32.S b/arch/x86/kernel/head_32.S
> index 80dc05e..ad7dbbb 100644
> --- a/arch/x86/kernel/head_32.S
> +++ b/arch/x86/kernel/head_32.S
> @@ -38,42 +38,30 @@
> #define X86_VENDOR_ID    new_cpu_data+CPUINFO_x86_vendor_id
> 
> /*
> - * This is how much memory *in addition to the memory covered up to
> - * and including _end* we need mapped initially.
> - * We need:
> - *  - one bit for each possible page, but only in low memory, which means
> - *     2^32/4096/8 = 128K worst case (4G/4G split.)
> + * This is how much memory for page table to and including _end
> + * we need mapped initially.
>  *  - enough space to map all low memory, which means
> - *     (2^32/4096) / 1024 pages (worst case, non PAE)
> - *     (2^32/4096) / 512 + 4 pages (worst case for PAE)
> - *  - a few pages for allocator use before the kernel pagetable has
> - *     been set up
> + *     (KERNEL_IMAGE_SIZE/4096) / 1024 pages (worst case, non PAE)
> + *     (KERNEL_IMAGE_SIZE/4096) / 512 + 4 pages (worst case for PAE)
>  *
>  * Modulo rounding, each megabyte assigned here requires a kilobyte of
>  * memory, which is currently unreclaimed.
>  *
>  * This should be a multiple of a page.
> + *
> + * KERNEL_IMAGE_SIZE should be greater than pa(_end)
> + * and small than max_low_pfn, otherwise will waste some page table
> entries
>  */
> LOW_PAGES = (KERNEL_IMAGE_SIZE + PAGE_SIZE_asm - 1)>>PAGE_SHIFT
> 
> -/*
> - * To preserve the DMA pool in PAGEALLOC kernels, we'll allocate
> - * pagetables from above the 16MB DMA limit, so we'll have to set
> - * up pagetables 16MB more (worst-case):
> - */
> -#if defined(CONFIG_DEBUG_PAGEALLOC) || defined(CONFIG_KMEMCHECK)
> -LOW_PAGES = LOW_PAGES + 0x1000000
> -#endif
> -
> #if PTRS_PER_PMD > 1
> PAGE_TABLE_SIZE = (LOW_PAGES / PTRS_PER_PMD) + PTRS_PER_PGD
> #else
> PAGE_TABLE_SIZE = (LOW_PAGES / PTRS_PER_PGD)
> #endif
> -BOOTBITMAP_SIZE = LOW_PAGES / 8
> ALLOCATOR_SLOP = 4
> 
> -INIT_MAP_BEYOND_END = BOOTBITMAP_SIZE + (PAGE_TABLE_SIZE +
> ALLOCATOR_SLOP)*PAGE_SIZE_asm
> +INIT_MAP_SIZE = (PAGE_TABLE_SIZE + ALLOCATOR_SLOP) * PAGE_SIZE_asm
> 
> RESERVE_BRK(pagetables, PAGE_TABLE_SIZE * PAGE_SIZE)

use INIT_MAP_SIZE in RESERVE_BRK directly?

> 
> @@ -168,10 +156,10 @@ num_subarch_entries = (. - subarch_entries) / 4
> 
> /*
>  * Initialize page tables.  This creates a PDE and a set of page
> - * tables, which are located immediately beyond _end.  The variable
> + * tables, which are located immediately beyond __brk_base.  The variable
>  * _brk_end is set up to point to the first "safe" location.
>  * Mappings are created both at virtual address 0 (identity mapping)
> - * and PAGE_OFFSET for up to _end+sizeof(page tables)+INIT_MAP_BEYOND_END.
> + * and PAGE_OFFSET for up to _end.
>  *
>  * Note that the stack is not yet set up!
>  */
> @@ -210,10 +198,9 @@ default_entry:
>     loop 11b
> 
>     /*
> -     * End condition: we must map up to and including INIT_MAP_BEYOND_END
> -     * bytes beyond the end of our own page tables.
> +     * End condition: we must map up to the end.
>      */
> -    leal (INIT_MAP_BEYOND_END+PTE_IDENT_ATTR)(%edi),%ebp
> +    movl $pa(_end) + PTE_IDENT_ATTR, %ebp
>     cmpl %ebp,%eax
>     jb 10b
> 1:
> @@ -243,11 +230,9 @@ page_pde_offset = (__PAGE_OFFSET >> 20);
>     addl $0x1000,%eax
>     loop 11b
>     /*
> -     * End condition: we must map up to and including INIT_MAP_BEYOND_END
> -     * bytes beyond the end of our own page tables; the +0x007 is
> -     * the attribute bits
> +     * End condition: we must map up to end
>      */
> -    leal (INIT_MAP_BEYOND_END+PTE_IDENT_ATTR)(%edi),%ebp
> +    movl $pa(_end) + PTE_IDENT_ATTR, %ebp
>     cmpl %ebp,%eax
>     jb 10b
>     addl $__PAGE_OFFSET, %edi
> @@ -638,6 +623,10 @@ swapper_pg_fixmap:
>     .fill 1024,4,0
> ENTRY(empty_zero_page)
>     .fill 4096,1,0
> +
> +.section ".bss.extra_page_aligned","wa"
> +    .align PAGE_SIZE_asm
> +    .fill INIT_MAP_SIZE,1,0

this 4 lines are not needed. you already had RESERVE_BRK for it.

YH

> /*
>  * This starts the data section.
>  */
> diff --git a/arch/x86/kernel/vmlinux_32.lds.S
> b/arch/x86/kernel/vmlinux_32.lds.S
> index 4005279..c318dee 100644
> --- a/arch/x86/kernel/vmlinux_32.lds.S
> +++ b/arch/x86/kernel/vmlinux_32.lds.S
> @@ -210,6 +210,12 @@ SECTIONS
>   DWARF_DEBUG
> }
> 
> +/*
> + * Build-time check on the image size:
> + */
> +ASSERT((_end - LOAD_OFFSET <= KERNEL_IMAGE_SIZE),
> +    "kernel image bigger than KERNEL_IMAGE_SIZE")
> +
> #ifdef CONFIG_KEXEC
> /* Link time checks */
> #include <asm/kexec.h>