[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] query on physical memory map for Xen?


  • To: "Peter Teoh" <htmldeveloper@xxxxxxxxx>
  • From: "pradeep singh rautela" <rautelap@xxxxxxxxx>
  • Date: Mon, 14 Apr 2008 11:21:52 +0530
  • Cc: xen-devel <xen-devel@xxxxxxxxxxxxxxxxxxx>, Keir Fraser <keir.fraser@xxxxxxxxxxxxx>
  • Delivery-date: Sun, 13 Apr 2008 22:52:18 -0700
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:cc:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=wzP6pZsmvAb3A+zA53QQCy5lJRdA2ZxHjsUEqwxYt45Em5Z5dijPjKeNnjfuoIeoopnJV3KwHiRZ8E0RRu2EPPJkPeoIPSVdueex9ZsdYHpxKR8K1ljR/8SO5/qP57dBwdF0WDf4PG/KdrZ4u98Jd+mQW9CdGiW5FTEdDNw2t+w=
  • List-id: Xen developer discussion <xen-devel.lists.xensource.com>

Nada, this is definetly the wrong piece of code you are looking at.:-)
Let me check once again for you ...

Ahh...got something.

see xen/include/asm-x86/config.h

296 #ifdef CONFIG_X86_PAE
297 # define LINEARPT_MBYTES         8
298 # define MACHPHYS_MBYTES        16 /* 1 MB needed per 1 GB memory */
299 # define FRAMETABLE_MBYTES (MACHPHYS_MBYTES * 6)
300 #else
301 # define LINEARPT_MBYTES         4
302 # define MACHPHYS_MBYTES         4
303 # define FRAMETABLE_MBYTES      24

a 32 bit kernel/hypervisor cannot access more than 4GB of physical
memory.So if PAE is enabled while compiling the kernel/hypervisor you
get the first part of the above conditional macro.

for each GB you have 1MB of memory for a m2p table.isn't it?
16MB means 16GB of physical mem max you can address in 32 bit xen hypervisor.

Thanks,

On Mon, Apr 14, 2008 at 11:13 AM, Peter Teoh <htmldeveloper@xxxxxxxxx> wrote:
> Just curious, can anyone give a hint on where this 16GB max physical
>  memory is defined?
>
>  Doing some random search by strings I found these:
>
>  #define MAXMEM           0x3fffffffffffUL
>  #define MAXMEM                  
> (__FIXADDR_TOP-__PAGE_OFFSET-__VMALLOC_RESERVE)
>  #define MAXMEM_PFN      PFN_DOWN(MAXMEM)
>  #define ACPI_MEM_LIST_MAX               6
>  #define NR_NODE_MEMBLKS (MAX_NUMNODES*2)
>  #define EFI_MAX_MEMORY_TYPE             14
>  #define  PCI_X_CMD_MAX_READ     0x000c  /* Max Memory Read Byte Count */
>  #define  PCI_X_STATUS_MAX_READ  0x00600000      /* Designed Max Memory
>  Read Count */
>  #define GEO_TYPE_MAX            (GEO_TYPE_MEM+1)
>  #define NR_NODE_MEMBLKS (MAX_NUMNODES*2)
>     *max_mfn = xc_memory_op(xc_handle, XENMEM_maximum_ram_page, NULL);
>  #define EFI_MAX_MEMORY_TYPE             14
>  #define MAXMEM           0x3fffffffffffUL
>   * KEXEC_SOURCE_MEMORY_LIMIT maximum page get_free_page can return.
>  #define MAXMEM                  
> (__FIXADDR_TOP-__PAGE_OFFSET-__VMALLOC_RESERVE)
>  #define MAXMEM_PFN      PFN_DOWN(MAXMEM)
>   * KEXEC_SOURCE_MEMORY_LIMIT maximum page get_free_page can return.
>  #define ACPI_MEM_LIST_MAX               6
>  #define XENMEM_maximum_ram_page     2
>  #define XENMEM_maximum_reservation  4
>  #define XENMEM_maximum_gpfn         14
>  #define NR_NODE_MEMBLKS (MAX_NUMNODES*2)
>  #define EFI_MAX_MEMORY_TYPE             14
>  #define GEO_TYPE_MAX            (GEO_TYPE_MEM+1)
>  #define NR_NODE_MEMBLKS (MAX_NUMNODES*2)
>
>  Not sure which one maps to this 16GB thing?   Thanks for the sharing :-).
>
>
>
>  On Fri, Apr 11, 2008 at 2:54 PM, Keir Fraser <keir.fraser@xxxxxxxxxxxxx> 
> wrote:
>  > On 11/4/08 07:14, "pradeep singh rautela" <rautelap@xxxxxxxxx> wrote:
>  >
>  >  > Based on my understanding of the Xen code, I have a small query for you 
> all.
>  >  >
>  >  > Xen hypervisor can access only a max of 16GB of physical memory
>  >  > directly on x86 architecture.
>  >  > Is this correct inference?
>  >  >
>  >  > So does Xen hypervisor cannot touch memory about 16G directly? Does it
>  >  > employ some kind of temporary mapping like linux kernel to access
>  >  > memory above 16G then?
>  >
>  >  32-bit Xen does not use memory above 16GB at all. 64-bit Xen can use any
>  >  amount of memory, and is able to keep a permanent mapping of all of it.
>  >
>  >   -- Keir
>  >
>  >
>  >
>  >
>  >
>  >  _______________________________________________
>  >  Xen-devel mailing list
>  >  Xen-devel@xxxxxxxxxxxxxxxxxxx
>  >  http://lists.xensource.com/xen-devel
>  >
>
>
>
>  --
>  Regards,
>  Peter Teoh
>



-- 
Pradeep Singh Rautela
http://eagain.wordpress.com
http://emptydomain.googlepages.com

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.