[lkml]   [2010]   [Jan]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
    Subject[PATCH -v4 0/37] x86: not use bootmem for x86
    please check the patches regarding with early_res and bootmem

    and at last it will use early_res instead of bootmem with x86 64bits

    -v2: allocate vmemmap on one node together, and also seperate early_res
    -v3: make x86 32 bit support early_res to use bootmem too
    move related early_res to kernel/
    sparse vmemmap together: address Ingo.
    -v4: some patches could go with tip with acked-by Jesse
    radix and logical flat etc
    Ingo said:
    I think we could remove the bootmem allocator middle man altogether.

    This can be done by initializing the page allocator sooner and by
    extending (already existing) 'reserve memory early on' mechanisms in
    architecture code. (the reserve_early*() APIs in x86 for example)

    Right now we have 5 memory allocation models on x86, initialized

    - allocator (buddy) [generic]
    - early allocator (bootmem) [generic]
    - very early allocator (reserve_early*()) [x86]
    - very very early allocator (early brk model) [x86]
    - very very very early allocator (build time .data/.bss) [generic]

    Seems excessive.

    The reserve_early() method is list/range based and can handle vast
    amounts of not very fragmented memory - perfect for basically all the
    real bootmem purposes (which is to bootstrap the buddy).

    reserve_early() allocated memory could be freed into the buddy later on
    as well. The main reason why bootmem is 'destroyed' during free-to-buddy
    is because it has excessive internal bitmaps we want to free. With a
    list/range based reserve_early() mechanism there's no such problem -
    they can linger indefinitely and there's near zero allocation management

    reserve_early() might need some small amount of extra work before it can
    be used as a generic early allocator - like adding a node field to it
    (so that the buddy can then pick those ranges up in a NUMA aware
    fashion) - but nothing very complex.

    early_res related:
    6f632c9: x86: move range related operation to one file
    b62d592: x86: check range in update range
    2a49dba: x86/pci: use u64 instead of size_t in amd_bus.c
    b9406c1: x86/pci: add cap_resource
    84dd3b2: x86/pci: enable pci root res read out for 32bit too
    e9a7f12: x86: call early_res_to_bootmem one time
    c3685a1: x86: introduce max_early_res and early_res_count
    4b951ed: x86: dynamic increase early_res array size
    5a84eb6: x86: print bootmem free before pci_iommu_alloc and free_all_bootmem -v2
    f00c3bb: x86: make early_node_mem get mem > 4g if possible
    e3e7efe: x86: only call dma32_reserve_bootmem 64bit !CONFIG_NUMA
    9bee1a1: x86: make 64 bit use early_res instead of bootmem before slab
    453b0be: sparsemem: put usemap for one node together
    5f01a21: sparsemem: put mem map for one node together.
    ef6e006: x86: change range end to start+size
    480085b: x86: move bios page reserve early to head32/64.c
    95630e9: x86: seperate early_res related code from e820.c
    a91ebdc: x86: add find_early_area_size
    8e98a3a: x86: move back find_e820_area to e820.c
    98f958f: early_res: enhance check_and_double_early_res
    0e5b16a: x86: make 32bit support NO_BOOTMEM
    59cca3b: move round_up/down to kernel.h
    27911e4: x86: add find_fw_memmap_area
    7809f98: core: move early_res
    46afde8: ram_buffer_extend_print
    1f32abd: x86: remove bios data range from e820
    d1bbc62: irq: remove not need bootmem code

    radix_tree for spare irq:
    3ac5b09: radix: move radix init early
    3b862ce: sparseirq: change irq_desc_ptrs to static
    9d32b70: sparseirq: use radix_tree instead of ptrs array
    ff0c3de: x86: remove arch_probe_nr_irqs

    logical flat cleanup:
    1a8b97c: x86, apic: Use logical flat on intel with <= 8 logical cpus
    eaf5895: use nr_cpus= to set nr_cpu_ids early
    6efca5a: x86: using logical flat for amd cpu too.
    4607ec4: x86: according to nr_cpu_ids to decide if need to leave logical flat
    d2e2375: x86: make 32bit apic flat to physflat switch like 64bit
    99ec7c7: x86: use num_processors for possible cpus



     \ /
      Last update: 2010-01-16 04:15    [W:0.025 / U:0.140 seconds]
    ©2003-2016 Jasper Spaans. hosted at Digital OceanAdvertise on this site