Re: [PATCH 10/10] xen: arm: Use a direct mapping of RAM on arm64

[Julien Grall <julien.grall@citrix.com>]

On 06/18/2013 02:26 PM, Ian Campbell wrote:

> We have plenty of virtual address space so we can avoid needing to map and
> unmap pages all the time.
> 
> A totally arbitrarily chosen 32GB frame table leads to support for 5TB of RAM.
> I haven't tested with anything near that amount of RAM though. There is plenty
> of room to expand further when that becomes necessary.
> 
> Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
> ---
>  xen/arch/arm/mm.c            |  155 ++++++++++++++++++++++++++++++++++--------
>  xen/arch/arm/setup.c         |   92 +++++++++++++++++++++++++
>  xen/include/asm-arm/config.h |   96 +++++++++++++++++++-------
>  xen/include/asm-arm/mm.h     |   16 ++++
>  4 files changed, 307 insertions(+), 52 deletions(-)
> 
> diff --git a/xen/arch/arm/mm.c b/xen/arch/arm/mm.c
> index 5ad687f..35ef56d 100644
> --- a/xen/arch/arm/mm.c
> +++ b/xen/arch/arm/mm.c
> @@ -37,6 +37,7 @@
>  #include <public/memory.h>
>  #include <xen/sched.h>
>  #include <xen/vmap.h>
> +#include <asm/early_printk.h>
>  #include <xsm/xsm.h>
>  #include <xen/pfn.h>
>  
> @@ -48,6 +49,14 @@ struct domain *dom_xen, *dom_io, *dom_cow;
>  lpae_t boot_pgtable[LPAE_ENTRIES] __attribute__((__aligned__(4096)));
>  #ifdef CONFIG_ARM_64
>  lpae_t boot_first[LPAE_ENTRIES] __attribute__((__aligned__(4096)));
> +/* The first page of the first level mapping of the xenheap. The
> + * subsequent xenheap first level pages are dynamically allocated, but
> + * we need this one to bootstrap ourselves. */
> +lpae_t xenheap_first_first[LPAE_ENTRIES] __attribute__((__aligned__(4096)));
> +/* The zeroeth level slot which uses xenheap_first_first. Used because
> + * setup_xenheap_mappings otherwise relies on mfn_to_virt which isn't
> + * valid for a non-xenheap mapping. */
> +static __initdata int xenheap_first_first_slot = -1;
>  #endif
>  
>  /*
> @@ -57,6 +66,9 @@ lpae_t boot_first[LPAE_ENTRIES] __attribute__((__aligned__(4096)));
>   * xen_second, xen_fixmap and xen_xenmap are shared between all PCPUs.
>   */
>  
> +#ifdef CONFIG_ARM_64
> +#define THIS_CPU_PGTABLE boot_pgtable
> +#else
>  /* Per-CPU pagetable pages */
>  /* xen_pgtable == root of the trie (zeroeth level on 64-bit, first on 32-bit) */
>  static DEFINE_PER_CPU(lpae_t *, xen_pgtable);
> @@ -65,6 +77,7 @@ static DEFINE_PER_CPU(lpae_t *, xen_pgtable);
>   * the second level pagetables which mapp the domheap region
>   * DOMHEAP_VIRT_START...DOMHEAP_VIRT_END in 2MB chunks. */
>  static DEFINE_PER_CPU(lpae_t *, xen_dommap);
> +#endif
>  
>  /* Common pagetable leaves */
>  /* Second level page tables.
> @@ -73,10 +86,16 @@ static DEFINE_PER_CPU(lpae_t *, xen_dommap);
>   * addresses from 0 to 0x7fffffff. Offsets into it are calculated
>   * with second_linear_offset(), not second_table_offset().
>   *
> - * Addresses 0x80000000 to 0xffffffff are covered by the per-cpu
> - * xen_domheap mappings described above. However we allocate 4 pages
> + * On 32bit addresses 0x80000000 to 0xffffffff are covered by the
> + * per-cpu xen_domheap mappings described above. We allocate 4 pages
>   * here for use in the boot page tables and the second two pages
>   * become the boot CPUs xen_dommap pages.
> + *
> + * On 64bit addresses 0x80000000 to 0xffffffff are unused. However we
> + * allocate 4 pages here for use while relocating Xen, which currently
> + * expects a second level page to exist for all addresses in the first
> + * 4GB. We need to keep these extra mappings in place for seconary CPU
> + * bring up too. For now we just leave them forever.
>   */
>  lpae_t xen_second[LPAE_ENTRIES*4] __attribute__((__aligned__(4096*4)));
>  /* First level page table used for fixmap */
> @@ -92,7 +111,7 @@ uint64_t boot_ttbr;
>  static paddr_t phys_offset;
>  
>  /* Limits of the Xen heap */
> -unsigned long xenheap_mfn_start __read_mostly;
> +unsigned long xenheap_mfn_start __read_mostly = ~0UL;
>  unsigned long xenheap_mfn_end __read_mostly;
>  unsigned long xenheap_virt_end __read_mostly;
>  
> @@ -112,7 +131,9 @@ static inline void check_memory_layout_alignment_constraints(void) {
>      BUILD_BUG_ON(BOOT_MISC_VIRT_START & ~SECOND_MASK);
>      /* 1GB aligned regions */
>      BUILD_BUG_ON(XENHEAP_VIRT_START & ~FIRST_MASK);
> +#ifdef CONFIG_DOMAIN_PAGE
>      BUILD_BUG_ON(DOMHEAP_VIRT_START & ~FIRST_MASK);
> +#endif
>  }
>  
>  void dump_pt_walk(lpae_t *first, paddr_t addr)
> @@ -180,6 +201,7 @@ void clear_fixmap(unsigned map)
>      flush_xen_data_tlb_range_va(FIXMAP_ADDR(map), PAGE_SIZE);
>  }
>  
> +#ifdef CONFIG_DOMAIN_PAGE
>  void *map_domain_page_global(unsigned long mfn)
>  {
>      return vmap(&mfn, 1);
> @@ -284,6 +306,7 @@ unsigned long domain_page_map_to_mfn(const void *va)
>  
>      return map[slot].pt.base + offset;
>  }
> +#endif
>  
>  void __init arch_init_memory(void)
>  {
> @@ -431,6 +454,7 @@ void __init setup_pagetables(unsigned long boot_phys_offset, paddr_t xen_paddr)
>      /* Flush everything after setting WXN bit. */
>      flush_xen_text_tlb();
>  
> +#ifdef CONFIG_ARM_32
>      per_cpu(xen_pgtable, 0) = boot_pgtable;
>      per_cpu(xen_dommap, 0) = xen_second +
>          second_linear_offset(DOMHEAP_VIRT_START);
> @@ -440,43 +464,31 @@ void __init setup_pagetables(unsigned long boot_phys_offset, paddr_t xen_paddr)
>      memset(this_cpu(xen_dommap), 0, DOMHEAP_SECOND_PAGES*PAGE_SIZE);
>      flush_xen_dcache_va_range(this_cpu(xen_dommap),
>                                DOMHEAP_SECOND_PAGES*PAGE_SIZE);
> +#endif
>  }
>  
>  int init_secondary_pagetables(int cpu)
>  {
> -    lpae_t *root, *first, *domheap, pte;
> -    int i;
> -
> -    root = alloc_xenheap_page();
>  #ifdef CONFIG_ARM_64
> -    first = alloc_xenheap_page();
> +    /* All CPUs share a single page table on 64 bit */
> +    return 0;
>  #else
> -    first = root; /* root == first level on 32-bit 3-level trie */
> -#endif
> +    lpae_t *first, *domheap, pte;
> +    int i;
> +
> +    first = alloc_xenheap_page(); /* root == first level on 32-bit 3-level trie */
>      domheap = alloc_xenheap_pages(get_order_from_pages(DOMHEAP_SECOND_PAGES), 0);
>  
> -    if ( root == NULL || domheap == NULL || first == NULL )
> +    if ( domheap == NULL || first == NULL )
>      {
>          printk("Not enough free memory for secondary CPU%d pagetables\n", cpu);
>          free_xenheap_pages(domheap, get_order_from_pages(DOMHEAP_SECOND_PAGES));
> -#ifdef CONFIG_ARM_64
>          free_xenheap_page(first);
> -#endif
> -        free_xenheap_page(root);
>          return -ENOMEM;
>      }
>  
>      /* Initialise root pagetable from root of boot tables */
> -    memcpy(root, boot_pgtable, PAGE_SIZE);
> -
> -#ifdef CONFIG_ARM_64
> -    /* Initialise first pagetable from first level of boot tables, and
> -     * hook into the new root. */
> -    memcpy(first, boot_first, PAGE_SIZE);
> -    pte = mfn_to_xen_entry(virt_to_mfn(first));
> -    pte.pt.table = 1;
> -    write_pte(root, pte);
> -#endif
> +    memcpy(first, boot_pgtable, PAGE_SIZE);
>  
>      /* Ensure the domheap has no stray mappings */
>      memset(domheap, 0, DOMHEAP_SECOND_PAGES*PAGE_SIZE);
> @@ -490,16 +502,14 @@ int init_secondary_pagetables(int cpu)
>          write_pte(&first[first_table_offset(DOMHEAP_VIRT_START+i*FIRST_SIZE)], pte);
>      }
>  
> -    flush_xen_dcache_va_range(root, PAGE_SIZE);
> -#ifdef CONFIG_ARM_64
>      flush_xen_dcache_va_range(first, PAGE_SIZE);
> -#endif
>      flush_xen_dcache_va_range(domheap, DOMHEAP_SECOND_PAGES*PAGE_SIZE);
>  
> -    per_cpu(xen_pgtable, cpu) = root;
> +    per_cpu(xen_pgtable, cpu) = first;
>      per_cpu(xen_dommap, cpu) = domheap;
>  
>      return 0;
> +#endif
>  }
>  
>  /* MMU setup for secondary CPUS (which already have paging enabled) */
> @@ -544,6 +554,7 @@ static void __init create_32mb_mappings(lpae_t *second,
>      flush_xen_data_tlb();
>  }
>  
> +#ifdef CONFIG_ARM_32
>  /* Set up the xenheap: up to 1GB of contiguous, always-mapped memory. */
>  void __init setup_xenheap_mappings(unsigned long base_mfn,
>                                     unsigned long nr_mfns)
> @@ -556,19 +567,107 @@ void __init setup_xenheap_mappings(unsigned long base_mfn,
>      xenheap_mfn_end = base_mfn + nr_mfns;
>  }
>  
> +#else /* CONFIG_ARM_64 */
> +
> +void __init setup_xenheap_mappings(unsigned long base_mfn,
> +                                   unsigned long nr_mfns)
> +{
> +    lpae_t *first, pte;
> +    unsigned long offset, end_mfn;
> +    vaddr_t vaddr;
> +
> +    /* First call sets the xenheap physical offset. */
> +    if ( xenheap_mfn_start == ~0UL )
> +        xenheap_mfn_start = base_mfn;
> +
> +    if ( base_mfn < xenheap_mfn_start )
> +        early_panic("cannot add xenheap mapping at %lx below heap start %lx\n",
> +                    base_mfn, xenheap_mfn_start);
> +
> +    end_mfn = base_mfn + nr_mfns;
> +
> +    /* Align to previous 1GB boundary */
> +    base_mfn &= ~FIRST_MASK;
> +
> +    offset = base_mfn - xenheap_mfn_start;
> +    vaddr = DIRECTMAP_VIRT_START + offset*PAGE_SIZE;
> +
> +    while ( base_mfn < end_mfn )
> +    {
> +        int slot = zeroeth_table_offset(vaddr);
> +        lpae_t *p = &boot_pgtable[slot];
> +
> +        if ( p->pt.valid )
> +        {
> +            /* mfn_to_virt is not valid on the 1st 1st mfn, since it
> +             * is not within the xenheap. */
> +            first = slot == xenheap_first_first_slot ?
> +                xenheap_first_first : mfn_to_virt(p->pt.base);
> +        }
> +        else if ( xenheap_first_first_slot == -1)
> +        {
> +            /* Use xenheap_first_first to bootstrap the mappings */
> +            first = xenheap_first_first;
> +
> +            pte = pte_of_xenaddr((vaddr_t)xenheap_first_first);
> +            pte.pt.table = 1;
> +            write_pte(p, pte);
> +
> +            xenheap_first_first_slot = slot;
> +        }
> +        else
> +        {
> +            unsigned long first_mfn = alloc_boot_pages(1, 1);
> +            pte = mfn_to_xen_entry(first_mfn);
> +            pte.pt.table = 1;
> +            write_pte(p, pte);
> +            first = mfn_to_virt(first_mfn);
> +        }
> +
> +        pte = mfn_to_xen_entry(base_mfn);
> +        //pte.pt.contig = /* XXX need to think about the logic */
> +        write_pte(&first[first_table_offset(vaddr)], pte);
> +
> +        base_mfn += FIRST_SIZE>>PAGE_SHIFT;
> +        vaddr += FIRST_SIZE;
> +    }
> +
> +    flush_xen_data_tlb();
> +}
> +#endif
> +
>  /* Map a frame table to cover physical addresses ps through pe */
>  void __init setup_frametable_mappings(paddr_t ps, paddr_t pe)
>  {
>      unsigned long nr_pages = (pe - ps) >> PAGE_SHIFT;
>      unsigned long frametable_size = nr_pages * sizeof(struct page_info);
>      unsigned long base_mfn;
> +#ifdef CONFIG_ARM_64
> +    lpae_t *second, pte;
> +    unsigned long nr_second, second_base;
> +    int i;
> +#endif
>  
>      frametable_base_mfn = ps >> PAGE_SHIFT;
>  
>      /* Round up to 32M boundary */
>      frametable_size = (frametable_size + 0x1ffffff) & ~0x1ffffff;
>      base_mfn = alloc_boot_pages(frametable_size >> PAGE_SHIFT, 32<<(20-12));
> +
> +#ifdef CONFIG_ARM_64
> +    nr_second = frametable_size >> SECOND_SHIFT;
> +    second_base = alloc_boot_pages(nr_second, 1);
> +    second = mfn_to_virt(second_base);
> +    for ( i = 0; i < nr_second; i++ )
> +    {
> +        pte = mfn_to_xen_entry(second_base + i);
> +        pte.pt.table = 1;
> +        write_pte(&boot_first[first_table_offset(FRAMETABLE_VIRT_START)+i], pte);
> +    }
> +    create_32mb_mappings(second, 0, base_mfn, frametable_size >> PAGE_SHIFT);
> +#else
>      create_32mb_mappings(xen_second, FRAMETABLE_VIRT_START, base_mfn, frametable_size >> PAGE_SHIFT);
> +#endif
>  
>      memset(&frame_table[0], 0, nr_pages * sizeof(struct page_info));
>      memset(&frame_table[nr_pages], -1,
> diff --git a/xen/arch/arm/setup.c b/xen/arch/arm/setup.c
> index da2a734..3efa063 100644
> --- a/xen/arch/arm/setup.c
> +++ b/xen/arch/arm/setup.c
> @@ -279,6 +279,7 @@ static paddr_t __init get_xen_paddr(void)
>      return paddr;
>  }
>  
> +#ifdef CONFIG_ARM_32
>  static void __init setup_mm(unsigned long dtb_paddr, size_t dtb_size)
>  {
>      paddr_t ram_start;
> @@ -392,6 +393,97 @@ static void __init setup_mm(unsigned long dtb_paddr, size_t dtb_size)
>  
>      end_boot_allocator();
>  }
> +#else /* CONFIG_ARM_64 */
> +static void __init setup_mm(unsigned long dtb_paddr, size_t dtb_size)
> +{
> +    paddr_t ram_start = ~0;
> +    paddr_t ram_end = 0;
> +    int bank;
> +    unsigned long  xenheap_pages = 0;
> +    unsigned long dtb_pages;
> +
> +    /*
> +     * TODO: only using the first RAM bank for now.  The heaps and the
> +     * frame table assume RAM is physically contiguous.
> +     */


I think this comment can be removed.

> +    total_pages = 0;
> +    for ( bank = 0 ; bank < early_info.mem.nr_banks; bank++ )
> +    {
> +        paddr_t bank_start = early_info.mem.bank[bank].start;
> +        paddr_t bank_size  = early_info.mem.bank[bank].size;
> +        paddr_t  bank_end = bank_start + bank_size;
> +        unsigned long bank_pages = bank_size >> PAGE_SHIFT;
> +        paddr_t s, e;
> +
> +        total_pages += bank_pages;
> +
> +        if ( bank_start < ram_start )
> +            ram_start = bank_start;
> +        if ( bank_end > ram_end )
> +            ram_end = bank_end;
> +
> +        /*
> +         * Locate the xenheap using these constraints:
> +         *
> +         *  - must be 32 MiB aligned
> +         *  - must not include Xen itself or the boot modules
> +         *  - must be at most 1 GiB
> +         *  - must be at least 128M
> +         *
> +         * We try to allocate the largest xenheap possible within these
> +         * constraints.
> +         */
> +        xenheap_pages += (bank_size >> PAGE_SHIFT);
> +
> +        /* XXX we assume that the ram regions are ordered */
> +        s = bank_start;
> +        while ( s < bank_end )
> +        {
> +            paddr_t n = bank_end;
> +
> +            e = next_module(s, &n);
> +
> +            if ( e == ~(paddr_t)0 )
> +            {
> +                e = n = bank_end;
> +            }
> +
> +            setup_xenheap_mappings(s>>PAGE_SHIFT, (e-s)>>PAGE_SHIFT);
> +
> +            xenheap_mfn_end = e;
> +
> +            init_boot_pages(s, e);
> +            s = n;
> +        }
> +    }
> +
> +    xenheap_virt_end = XENHEAP_VIRT_START + ram_end - ram_start;
> +    xenheap_mfn_start = ram_start >> PAGE_SHIFT;
> +    xenheap_mfn_end = ram_end >> PAGE_SHIFT;
> +    xenheap_max_mfn(xenheap_mfn_end);
> +
> +    /*
> +     * Need enough mapped pages for copying the DTB.
> +     *
> +     * TODO: The DTB (and other payloads) are assumed to be towards
> +     * the start of RAM.
> +     */
> +    dtb_pages = (dtb_size + PAGE_SIZE-1) >> PAGE_SHIFT;
> +
> +    /*
> +     * Copy the DTB.
> +     *
> +     * TODO: handle other payloads too.
> +     */
> +    device_tree_flattened = mfn_to_virt(alloc_boot_pages(dtb_pages, 1));
> +    copy_from_paddr(device_tree_flattened, dtb_paddr, dtb_size, BUFFERABLE);
> +
> +   setup_frametable_mappings(ram_start, ram_end);


What happen if the RAM is not contiguous? Do we have hole in the frametable?

BTW, I was looking to the mfn_valid function and I think the function
can possibly consider device memory region as valid MFN. Is it the right
behaviour?

-- 
Julien