RE: Switching from writeable pagetable toshadowlog-dirtyat runtime

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

RE: Switching from writeable pagetable toshadowlog-dirtyat runtime

Xin, Xiaohui

Keir,
We found that in 32-bit code, the shadow_fault() do one more thing than 64-bit code.
The following code in file shadow32.c checks the L1 page table with RW bit removed and make the shadow with RW bit, it does not inject the fault to guest.
Does this is a work-around for switch from writable pagetable to shadow mode enable?
Does this code makes sense to the writeable pagetable, and then may be one reason that 32bit log-dirty code can work?

    if ( regs->error_code & 2 )
    {
  +      int allow_writes = 0;
  +      if ( unlikely(!(l1e_get_flags(gpte) & _PAGE_RW)) )
  +      {
  +          if ( shadow_mode_page_writable(va, regs, l1e_get_pfn(gpte)) )
  +          {
  +              allow_writes = 1;
  +              l1e_add_flags(gpte, _PAGE_RW);
  +          }
  +          else
  +          {
  +              /* Write fault on a read-only mapping. */
  +              SH_VVLOG("shadow_fault - EXIT: wr fault on RO page (%" PRIpte     ")",
  +                       l1e_get_intpte(gpte));
  +              perfc_incrc(shadow_fault_bail_ro_mapping);
  +              goto fail;
  +          }
  +     }
        if ( !l1pte_write_fault(v, &gpte, &spte, va) )
        {
            SH_VVLOG("shadow_fault - EXIT: l1pte_write_fault failed");
            perfc_incrc(write_fault_bail);
            shadow_unlock(d);
            return 0;
        }
   +     if ( allow_writes )
   +         l1e_remove_flags(gpte, _PAGE_RW);
   + }

Thanks
Xiaohui
-----Original Message-----
From: [hidden email] [mailto:[hidden email]] On Behalf Of Keir Fraser
Sent: 2005年10月14日 16:44
To: Li, Chengyuan
Cc: Ian Pratt; [hidden email]; Nakajima, Jun
Subject: Re: [Xen-devel] Switching from writeable pagetable toshadowlog-dirtyat runtime


On 14 Oct 2005, at 03:46, Li, Chengyuan wrote:

>
> I don't see ptwr_flush() when shadow mode enable in both ia32 and
> x86_64 code. Why plain ia32 log dirty shadow can work?

Perhaps it can't, properly. :-)

The best fix will be to make domain_pause() call
sync_pagetable_state(d). The main problem I can see is that the
cleanup_writeable_pagetable() function is not currently safe to be
called by other than the subject domain. For example, failure paths
crash the calling domain rather than the subject domain! This needs
fixing.

> audit_adjust_pgtables() is for plain ia32 only and it doesn't work for
> both ia32 pae and x86_64, but this function only check the ref counts
> and doesn't update guest PTE entry, e.g. R/W bit.
>
> Another question is that how does guest update its guest page-table
> after shadow mode is enabled?  As writable-pagetable can't be used
> after shadow mode is enabled.

Shadow pagetable code has its own quite separate writeable pagetable
logic, I believe. Which is why the non-shadow wrpt state must be
flushed before switching to shadow pagetables.

  -- Keir


_______________________________________________
Xen-devel mailing list
[hidden email]
http://lists.xensource.com/xen-devel

_______________________________________________
Xen-devel mailing list
[hidden email]
http://lists.xensource.com/xen-devel
Reply | Threaded
Open this post in threaded view
|

RE: Switching from writeable pagetable toshadowlog-dirtyat runtime

Khoa Huynh

"Xin, Xiaohui" <[hidden email]> wrote on 10/14/2005 04:43:07 AM:

>
> Keir,
> We found that in 32-bit code, the shadow_fault() do one more thing
> than 64-bit code.
> The following code in file shadow32.c checks the L1 page table with
> RW bit removed and make the shadow with RW bit, it does not inject
> the fault to guest.
> Does this is a work-around for switch from writable pagetable to
> shadow mode enable?
> Does this code makes sense to the writeable pagetable, and then may
> be one reason that 32bit log-dirty code can work?
>

I don't think we use log-dirty shadow mode - at least when setting up
VMX domains; we only use shadow translate and external modes.  See
vmx_final_setup_guest() in xen/arch/x86/domain.c.  I searched through
the code for 32-bit x86, and I don't recall ever seeing this log-dirty
mode be enabled.  When should we use log-dirty shadow mode ?
Thanks.

Khoa


_______________________________________________
Xen-devel mailing list
[hidden email]
http://lists.xensource.com/xen-devel