Re: [PATCH 02/39] xen: document Xen is using an unused bit for thepagetables

From: Konrad Rzeszutek Wilk
Date: Fri Mar 30 2012 - 17:45:29 EST


On Mon, Mar 26, 2012 at 07:45:49PM +0200, Andrea Arcangeli wrote:
> Xen has taken over the last reserved bit available for the pagetables
> which is set through ioremap, this documents it and makes the code
> more readable.

About a year ago we redid the P2M code to ditch the major use case for this.
But there were two left over cases that I hadn't found a good solution
for that would allow us to completly eliminate the use of this bit:

1). When setting a PTE of a PFN which overlaps an E820 hole or any of the
non-E820-RAM entries, we lookup in the P2M and find out that
the PFN is a 1:1 and return a pte.pte | pfn << PAGE_SIZE.

But we also stick the _PAGE_IOMAP on it so that when the call to
xen_pte_val is done we don't end up doing the lookup in the P2M tree
once more and just set the pte as is.

So this is the dance between xen_pte_val and xen_make_pte.

2). When the userspace tries to mmap a guest memory for save/migrate
or to setup something in the guest, it would use the xen_remap_domain_mfn_range
to setup PTE's with the guest's PFN (gpfn). The _PAGE_IOMAP
is used again to tell xen_pte_val to not bother looking it up in the
P2M tree and use it as is.

So.. any thoughts on how to eliminate the usage of this?
>
> Signed-off-by: Andrea Arcangeli <aarcange@xxxxxxxxxx>
> ---
> arch/x86/include/asm/pgtable_types.h | 11 +++++++++--
> 1 files changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/include/asm/pgtable_types.h b/arch/x86/include/asm/pgtable_types.h
> index 013286a..b74cac9 100644
> --- a/arch/x86/include/asm/pgtable_types.h
> +++ b/arch/x86/include/asm/pgtable_types.h
> @@ -17,7 +17,7 @@
> #define _PAGE_BIT_PAT 7 /* on 4KB pages */
> #define _PAGE_BIT_GLOBAL 8 /* Global TLB entry PPro+ */
> #define _PAGE_BIT_UNUSED1 9 /* available for programmer */
> -#define _PAGE_BIT_IOMAP 10 /* flag used to indicate IO mapping */
> +#define _PAGE_BIT_UNUSED2 10
> #define _PAGE_BIT_HIDDEN 11 /* hidden by kmemcheck */
> #define _PAGE_BIT_PAT_LARGE 12 /* On 2MB or 1GB pages */
> #define _PAGE_BIT_SPECIAL _PAGE_BIT_UNUSED1
> @@ -41,7 +41,7 @@
> #define _PAGE_PSE (_AT(pteval_t, 1) << _PAGE_BIT_PSE)
> #define _PAGE_GLOBAL (_AT(pteval_t, 1) << _PAGE_BIT_GLOBAL)
> #define _PAGE_UNUSED1 (_AT(pteval_t, 1) << _PAGE_BIT_UNUSED1)
> -#define _PAGE_IOMAP (_AT(pteval_t, 1) << _PAGE_BIT_IOMAP)
> +#define _PAGE_UNUSED2 (_AT(pteval_t, 1) << _PAGE_BIT_UNUSED2)
> #define _PAGE_PAT (_AT(pteval_t, 1) << _PAGE_BIT_PAT)
> #define _PAGE_PAT_LARGE (_AT(pteval_t, 1) << _PAGE_BIT_PAT_LARGE)
> #define _PAGE_SPECIAL (_AT(pteval_t, 1) << _PAGE_BIT_SPECIAL)
> @@ -49,6 +49,13 @@
> #define _PAGE_SPLITTING (_AT(pteval_t, 1) << _PAGE_BIT_SPLITTING)
> #define __HAVE_ARCH_PTE_SPECIAL
>
> +/* flag used to indicate IO mapping */
> +#ifdef CONFIG_XEN
> +#define _PAGE_IOMAP (_AT(pteval_t, 1) << _PAGE_BIT_UNUSED2)
> +#else
> +#define _PAGE_IOMAP (_AT(pteval_t, 0))
> +#endif
> +
> #ifdef CONFIG_KMEMCHECK
> #define _PAGE_HIDDEN (_AT(pteval_t, 1) << _PAGE_BIT_HIDDEN)
> #else
> --
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@xxxxxxxxxxxxxxx
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/