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

Re: [Xen-devel] Re: how to transfer virtual address into pyhsical address



Related to this, can you tell me whether the memory allocated to a domain is
contiguous in machine RAM (given the calls to xen_create_contiguous, seems 
like it is not)? 

Thanks,
Himanshu

On Wed, Feb 15, 2006 at 02:13:52PM +0100, Mathieu Ropert wrote:
> Yong LIU wrote:
> >I write a very simple module to test memory address translation.
> >When I install the module under xen 3.0, kernel panic. But when
> >I install it under xen3.0 & kernel 2.6.16-rc2, it can work well.
> >
> >My question is: Is there a different way to translate virtual address
> >into physical address under Xen 3.0?
> >thanks for help. 
> >
> >dmesg information.
> >--------------------------------------------------------
> >hello: module license 'unspecified' taints kernel.
> >Unable to handle kernel paging request at virtual address 3e0c5b0c
> > printing eip:
> >f483c023
> >*pde = ma 00000000 pa 55555000
> >Oops: 0000 [#1]
> >SMP
> >Modules linked in: hello iptable_filter ip_tables video thermal
> >processor fan button battery ac
> >CPU:    1
> >EIP:    0061:[<f483c023>]    Tainted: P      VLI
> >EFLAGS: 00010206   (2.6.12.6-xen0-smp)
> >EIP is at kvirt_to_pa+0x23/0x3c [hello]
> >eax: 7e0c5000   ebx: f22c3000   ecx: 00000b0c   edx: 000003c8
> >esi: f2026000   edi: c0000000   ebp: f2026000   esp: f2027f90
> >ds: 007b   es: 007b   ss: 0069
> >Process insmod (pid: 4710, threadinfo=f2026000 task=f27c4520)
> >Stack: f483e01a f22c3000 f483c380 c013a1f8 c05a10a8 00000001 f483c380
> >0804a060
> >       b7fb3ff4 b7fb5538 c0109021 0804a060 00000dda 0804a050 b7fb3ff4
> >b7fb5538
> >       bffd6558 00000080 0000007b c010007b 00000080 b7f5667e 00000073
> >00010246
> >Call Trace:
> > [<f483e01a>] hello_init_module+0x1a/0x2d [hello]
> > [<c013a1f8>] sys_init_module+0x145/0x1e1
> > [<c0109021>] syscall_call+0x7/0xb
> >Code:  Bad EIP value.
> >-------------------------------------------------------
> >here is the code.
> >#include <linux/init.h>
> >#include <linux/kernel.h>
> >#include <linux/module.h>
> >
> >#include <asm/pgtable.h>
> >#include <asm/page.h>
> >
> >static void kvirt_to_pa(void *vaddr)
> >{
> >        unsigned long addr=(unsigned long) vaddr;
> >        unsigned long pte_value;
> >        pgd_t *pgd=NULL;
> >        pmd_t *pmd=NULL;
> >        pte_t *pte=NULL;
> >
> >        pgd=pgd_offset_k(addr);
> >        if(pgd_none(*pgd))
> >                goto hello_failed;
> >        pmd=pmd_offset(pgd, addr);
> >        if(pmd_none(*pmd))
> >                goto hello_failed;
> >        pte=pte_offset_kernel(pmd, addr);
> >        if(!pte_present(*pte))
> >                goto hello_failed;
> >
> >        pte_value= pte_val(*pte) & PAGE_MASK | (addr & (PAGE_SIZE - 1));
> >        return;
> >
> >hello_failed:
> >        printk(" failed\n");
> >        return;
> >}
> >static int __init hello_init_module(void)
> >{
> >        void * page=NULL;
> >        page=__get_free_pages(GFP_KERNEL, 0);
> >        kvirt_to_pa(page);
> >        free_page(page);
> >        return 1;
> >}
> >
> >
> >static void __exit hello_exit_module(void)
> >{
> >        printk(" say bye\n");
> >        return;
> >}
> >module_init(hello_init_module);
> >module_exit(hello_exit_module);
> 
> __alloc_pages() returns a pointer to a page struct, not the virtual 
> address of the allocated page. You can't use it directly like you do.
> If you want the virtual address, either use page_address(page) or try 
> some get_*_page function directly instead of alloc_pages().
> 
> Regards,
> Mathieu
> 
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@xxxxxxxxxxxxxxxxxxx
> http://lists.xensource.com/xen-devel

-- 
-------------------------------------------------------------------------
Himanshu Raj
PhD Student, GaTech (www.cc.gatech.edu/~rhim)
I prefer to receive attachments in an open, non-proprietary format.
-------------------------------------------------------------------------

_______________________________________________
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®.