Re: CVS commit: src/sys/arch

["Cherry G. Mathew" <cherry.g.mathew%gmail.com@localhost>]

On 8 November 2011 19:11, Cherry G. Mathew 
<cherry.g.mathew%gmail.com@localhost> wrote:
> On 8 November 2011 05:50, Jean-Yves Migeon 
> <jeanyves.migeon%free.fr@localhost> wrote:
>> On 06.11.2011 16:18, Cherry G. Mathew wrote:
>>> Module Name:  src
>>> Committed By: cherry
>>> Date:         Sun Nov  6 15:18:19 UTC 2011
>>>
>>> Modified Files:
>>>       src/sys/arch/amd64/include: pmap.h
>>>       src/sys/arch/i386/include: pmap.h
>>>       src/sys/arch/x86/include: cpu.h
>>>       src/sys/arch/x86/x86: pmap.c
>>>       src/sys/arch/xen/x86: cpu.c x86_xpmap.c
>>>
>>> Log Message:
>>> [merging from cherry-xenmp] make pmap_kernel() shadow PMD per-cpu and MP
>>> aware.
>>
>> Some comments.
>>
>
> ...
>>> -/* flag to be used for kernel mappings: PG_u on Xen/amd64, 0 otherwise */
>>> -#if defined(XEN)&&  defined(__x86_64__)
>>> -#define PG_k PG_u
>>> -#else
>>> -#define PG_k 0
>>> -#endif
>>> -
>>
>> Are you sure that all the mapping sites are safe (PT/PD bits), given the
>> pmap split between pmap/xen_xpmap.c?
>>
>
> Ok, I realise I've broken the build with this one - apologies. For
> some odd reason my tree built ok ( even after nuking the obj dir)
>
> The current bandaid by christos and njoly is incorrect. I propose
> re-exporting PG_k to x86/include/pmap.h until (if ?) the xen pmap is
> completely independant of the x86 one.
>
> Please let me know if there are objections to this patch below:
>
> Thanks,
>
> --
> ~Cherry
>
>
>
> Index: arch/x86/include/pmap.h
> ===================================================================
> RCS file: /cvsroot/src/sys/arch/x86/include/pmap.h,v
> retrieving revision 1.44
> diff -u -r1.44 pmap.h
> --- arch/x86/include/pmap.h     6 Nov 2011 11:40:47 -0000       1.44
> +++ arch/x86/include/pmap.h     8 Nov 2011 13:35:16 -0000
> @@ -173,6 +173,13 @@
>        ((pmap)->pm_pdirpa[0] + (index) * sizeof(pd_entry_t))
>  #endif
>
> +/* flag to be used for kernel mappings: PG_u on Xen/amd64, 0 otherwise */
> +#if defined(XEN) && defined(__x86_64__)
> +#define PG_k PG_u
> +#else
> +#define PG_k 0
> +#endif
> +
>  /*
>  * MD flags that we use for pmap_enter and pmap_kenter_pa:
>  */
> Index: arch/x86/x86/pmap.c
> ===================================================================
> RCS file: /cvsroot/src/sys/arch/x86/x86/pmap.c,v
> retrieving revision 1.140
> diff -u -r1.140 pmap.c
> --- arch/x86/x86/pmap.c 8 Nov 2011 12:44:29 -0000       1.140
> +++ arch/x86/x86/pmap.c 8 Nov 2011 13:35:20 -0000
> @@ -211,11 +211,6 @@
>  #include <xen/hypervisor.h>
>  #endif
>
> -/* If this is not needed anymore it should be GC'ed */
> -#ifndef PG_k
> -#define        PG_k    0
> -#endif
> -
>  /*
>  * general info:
>  *
> Index: arch/xen/include/xenpmap.h
> ===================================================================
> RCS file: /cvsroot/src/sys/arch/xen/include/xenpmap.h,v
> retrieving revision 1.30
> diff -u -r1.30 xenpmap.h
> --- arch/xen/include/xenpmap.h  6 Nov 2011 11:40:47 -0000       1.30
> +++ arch/xen/include/xenpmap.h  8 Nov 2011 13:35:20 -0000
> @@ -34,13 +34,6 @@
>  #include "opt_xen.h"
>  #endif
>
> -/* flag to be used for kernel mappings: PG_u on Xen/amd64, 0 otherwise */
> -#if defined(XEN) && defined(__x86_64__)
> -#define PG_k PG_u
> -#else
> -#define PG_k 0
> -#endif
> -
>  #define        INVALID_P2M_ENTRY       (~0UL)
>
>  void xpq_queue_machphys_update(paddr_t, paddr_t);
> Index: arch/xen/x86/xen_pmap.c
> ===================================================================
> RCS file: /cvsroot/src/sys/arch/xen/x86/xen_pmap.c,v
> retrieving revision 1.7
> diff -u -r1.7 xen_pmap.c
> --- arch/xen/x86/xen_pmap.c     6 Nov 2011 11:40:47 -0000       1.7
> +++ arch/xen/x86/xen_pmap.c     8 Nov 2011 13:35:21 -0000
> @@ -142,13 +142,6 @@
>  #include <xen/hypervisor.h>
>  #endif
>
> -/* flag to be used for kernel mappings: PG_u on Xen/amd64, 0 otherwise */
> -#if defined(XEN) && defined(__x86_64__)
> -#define PG_k PG_u
> -#else
> -#define PG_k 0
> -#endif
> -
>  #define COUNT(x)       /* nothing */
>
>  static pd_entry_t * const alternate_pdes[] = APDES_INITIALIZER;
>

Committed.

Thanks,

-- 
~Cherry