Xen devel - Feb 2008 - [Xen-ia64-devel] [PATCH][QEMU][IA64] Fix qemu memory access on IA64

Hi Keir,

This patch is for IA64:
- correct a parameter to xc_map_foreign_batch(). skipping vga hole.
- sync_icache() might overrun a page boundary. qemu-dm will die.

Thanks,
Kouya

Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>



_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@lists.xensource.com
http://lists.xensource.com/xen-ia64-devel

Please apply it to the IA64 tree.

 -- Keir

On 21/2/08 10:35, "Kouya Shimura" <kouya@jp.fujitsu.com> wrote:
> Hi Keir,
> 
> This patch is for IA64:
> - correct a parameter to xc_map_foreign_batch(). skipping vga hole.
> - sync_icache() might overrun a page boundary. qemu-dm will die.
> 
> Thanks,
> Kouya
> 
> Signed-off-by: Kouya Shimura <kouya@jp.fujitsu.com>
> 
> diff -r f853c0497095 tools/ioemu/hw/xen_machine_fv.c
> --- a/tools/ioemu/hw/xen_machine_fv.c Tue Feb 19 11:14:40 2008 -0700
> +++ b/tools/ioemu/hw/xen_machine_fv.c Thu Feb 21 17:52:55 2008 +0900
> @@ -246,12 +246,18 @@ static void xen_init_fv(uint64_t ram_siz
>  
>      for (i = 0; i < nr_pages; i++)
>          page_array[i] = i;
> - 
> +
>      /* VTI will not use memory between 3G~4G, so we just pass a legal pfn
>         to make QEMU map continuous virtual memory space */
> -    if (ram_size > MMIO_START) { 
> +    if (ram_size > MMIO_START) {
>          for (i = 0 ; i < (MEM_G >> XC_PAGE_SHIFT); i++)
>              page_array[(MMIO_START >> XC_PAGE_SHIFT) + i] > +    
(STORE_PAGE_START >> XC_PAGE_SHIFT);
> +    }
> +    /* skipping VGA hole, same as above */
> +    if (ram_size > VGA_IO_START) {
> +        for (i = 0 ; i < (VGA_IO_SIZE >> XC_PAGE_SHIFT); i++)
> +            page_array[(VGA_IO_START >> XC_PAGE_SHIFT) + i] >    
(STORE_PAGE_START >> XC_PAGE_SHIFT);
>      }
>  
> diff -r f853c0497095 tools/ioemu/target-i386-dm/exec-dm.c
> --- a/tools/ioemu/target-i386-dm/exec-dm.c Tue Feb 19 11:14:40 2008 -0700
> +++ b/tools/ioemu/target-i386-dm/exec-dm.c Thu Feb 21 19:36:20 2008 +0900
> @@ -350,12 +350,13 @@ CPUReadMemoryFunc **cpu_get_io_memory_re
>   * So to emulate right behavior that guest OS is assumed, we need to flush
>   * I/D cache here.
>   */
> -static void sync_icache(unsigned long address, int len)
> -{
> -    int l;
> -
> -    for(l = 0; l < (len + 32); l += 32)
> -        __ia64_fc(address + l);
> +static void sync_icache(uint8_t *address, int len)
> +{
> +    unsigned long addr = (unsigned long)address;
> +    unsigned long end = addr + len;
> +
> +    for (addr &= ~(32UL-1); addr < end; addr += 32UL)
> +        __ia64_fc(addr);
>  
>      ia64_sync_i();
>      ia64_srlz_i();


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

On Thu, 2008-02-21 at 19:35 +0900, Kouya Shimura wrote:
> Hi Keir,
> 
> This patch is for IA64:
> - correct a parameter to xc_map_foreign_batch(). skipping vga hole.
> - sync_icache() might overrun a page boundary. qemu-dm will die.
   Applied.  Thanks,

	Alex

-- 
Alex Williamson                             HP Open Source & Linux Org.


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