Re: [PATCH v2] minios: don't rely on specific page table allocation scheme

[Wei Liu <wei.liu2@citrix.com>]

On Fri, Nov 20, 2015 at 02:52:57PM +0100, Juergen Gross wrote:
> Today mini-os is making assumptions how the page tables it is started
> with are being allocated. Especially it is using the number of page
> table frames to calculate which is the first unmapped pfn.
> 
> Instead of relying on page table number assumptions just look into the
> page tables to find the first pfn not already mapped.
> 
> Signed-off-by: Juergen Gross <jgross@suse.com>
> ---
> Changes in V2:
> - remove need_pt_frame() as it simplifies code (suggested by Wei Liu)
> 
> ---
>  arch/x86/mm.c         | 81 ++++++++++-----------------------------------------
>  include/x86/arch_mm.h |  7 -----
>  2 files changed, 15 insertions(+), 73 deletions(-)
> 
> diff --git a/arch/x86/mm.c b/arch/x86/mm.c
> index 9c6d1b8..b828efc 100644
> --- a/arch/x86/mm.c
> +++ b/arch/x86/mm.c
> @@ -132,61 +132,6 @@ static void new_pt_frame(unsigned long *pt_pfn, unsigned long prev_l_mfn,
>  }
>  
>  /*
> - * Checks if a pagetable frame is needed at 'level' to map a given
> - * address. Note, this function is specific to the initial page table
> - * building.
> - */
> -static int need_pt_frame(unsigned long va, int level)
> -{
> -    unsigned long hyp_virt_start = HYPERVISOR_VIRT_START;
> -#if defined(__x86_64__)
> -    unsigned long hyp_virt_end = HYPERVISOR_VIRT_END;
> -#else
> -    unsigned long hyp_virt_end = 0xffffffff;
> -#endif
> -
> -    /* In general frames will _not_ be needed if they were already
> -       allocated to map the hypervisor into our VA space */
> -#if defined(__x86_64__)
> -    if ( level == L3_FRAME )
> -    {
> -        if ( l4_table_offset(va) >= 
> -             l4_table_offset(hyp_virt_start) &&
> -             l4_table_offset(va) <= 
> -             l4_table_offset(hyp_virt_end))
> -            return 0;
> -        return 1;
> -    } 
> -    else
> -#endif
> -
> -    if ( level == L2_FRAME )
> -    {
> -#if defined(__x86_64__)
> -        if ( l4_table_offset(va) >= 
> -             l4_table_offset(hyp_virt_start) &&
> -             l4_table_offset(va) <= 
> -             l4_table_offset(hyp_virt_end))
> -#endif
> -            if ( l3_table_offset(va) >= 
> -                 l3_table_offset(hyp_virt_start) &&
> -                 l3_table_offset(va) <= 
> -                 l3_table_offset(hyp_virt_end))
> -                return 0;
> -
> -        return 1;
> -    } 
> -    else 
> -        /* Always need l1 frames */
> -        if ( level == L1_FRAME )
> -            return 1;
> -
> -    printk("ERROR: Unknown frame level %d, hypervisor %llx,%llx\n", 
> -           level, hyp_virt_start, hyp_virt_end);
> -    return -1;
> -}
> -
> -/*

Strangely git am rejects the above hunk. I manually fix that up in my
tree and test again.

All my tests in previous email passed.

Reviewed-by: Wei Liu <wei.liu2@citrix.com>
Tested-by: Wei Liu <wei.liu2@citrix.com>