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

Re: [Xen-devel] [PATCH 1/2] x86/mm: factor out the code for shattering an l3 PTE


  • To: Hongyan Xia <hongyxia@xxxxxxxxxx>, <xen-devel@xxxxxxxxxxxxxxxxxxxx>
  • From: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
  • Date: Fri, 6 Dec 2019 17:50:13 +0000
  • Authentication-results: esa3.hc3370-68.iphmx.com; dkim=none (message not signed) header.i=none; spf=None smtp.pra=andrew.cooper3@xxxxxxxxxx; spf=Pass smtp.mailfrom=Andrew.Cooper3@xxxxxxxxxx; spf=None smtp.helo=postmaster@xxxxxxxxxxxxxxx
  • Autocrypt: addr=andrew.cooper3@xxxxxxxxxx; prefer-encrypt=mutual; keydata= mQINBFLhNn8BEADVhE+Hb8i0GV6mihnnr/uiQQdPF8kUoFzCOPXkf7jQ5sLYeJa0cQi6Penp VtiFYznTairnVsN5J+ujSTIb+OlMSJUWV4opS7WVNnxHbFTPYZVQ3erv7NKc2iVizCRZ2Kxn srM1oPXWRic8BIAdYOKOloF2300SL/bIpeD+x7h3w9B/qez7nOin5NzkxgFoaUeIal12pXSR Q354FKFoy6Vh96gc4VRqte3jw8mPuJQpfws+Pb+swvSf/i1q1+1I4jsRQQh2m6OTADHIqg2E ofTYAEh7R5HfPx0EXoEDMdRjOeKn8+vvkAwhviWXTHlG3R1QkbE5M/oywnZ83udJmi+lxjJ5 YhQ5IzomvJ16H0Bq+TLyVLO/VRksp1VR9HxCzItLNCS8PdpYYz5TC204ViycobYU65WMpzWe LFAGn8jSS25XIpqv0Y9k87dLbctKKA14Ifw2kq5OIVu2FuX+3i446JOa2vpCI9GcjCzi3oHV e00bzYiHMIl0FICrNJU0Kjho8pdo0m2uxkn6SYEpogAy9pnatUlO+erL4LqFUO7GXSdBRbw5 gNt25XTLdSFuZtMxkY3tq8MFss5QnjhehCVPEpE6y9ZjI4XB8ad1G4oBHVGK5LMsvg22PfMJ ISWFSHoF/B5+lHkCKWkFxZ0gZn33ju5n6/FOdEx4B8cMJt+cWwARAQABtClBbmRyZXcgQ29v cGVyIDxhbmRyZXcuY29vcGVyM0BjaXRyaXguY29tPokCOgQTAQgAJAIbAwULCQgHAwUVCgkI CwUWAgMBAAIeAQIXgAUCWKD95wIZAQAKCRBlw/kGpdefoHbdD/9AIoR3k6fKl+RFiFpyAhvO 59ttDFI7nIAnlYngev2XUR3acFElJATHSDO0ju+hqWqAb8kVijXLops0gOfqt3VPZq9cuHlh IMDquatGLzAadfFx2eQYIYT+FYuMoPZy/aTUazmJIDVxP7L383grjIkn+7tAv+qeDfE+txL4 SAm1UHNvmdfgL2/lcmL3xRh7sub3nJilM93RWX1Pe5LBSDXO45uzCGEdst6uSlzYR/MEr+5Z JQQ32JV64zwvf/aKaagSQSQMYNX9JFgfZ3TKWC1KJQbX5ssoX/5hNLqxMcZV3TN7kU8I3kjK mPec9+1nECOjjJSO/h4P0sBZyIUGfguwzhEeGf4sMCuSEM4xjCnwiBwftR17sr0spYcOpqET ZGcAmyYcNjy6CYadNCnfR40vhhWuCfNCBzWnUW0lFoo12wb0YnzoOLjvfD6OL3JjIUJNOmJy RCsJ5IA/Iz33RhSVRmROu+TztwuThClw63g7+hoyewv7BemKyuU6FTVhjjW+XUWmS/FzknSi dAG+insr0746cTPpSkGl3KAXeWDGJzve7/SBBfyznWCMGaf8E2P1oOdIZRxHgWj0zNr1+ooF /PzgLPiCI4OMUttTlEKChgbUTQ+5o0P080JojqfXwbPAyumbaYcQNiH1/xYbJdOFSiBv9rpt TQTBLzDKXok86LkCDQRS4TZ/ARAAkgqudHsp+hd82UVkvgnlqZjzz2vyrYfz7bkPtXaGb9H4 Rfo7mQsEQavEBdWWjbga6eMnDqtu+FC+qeTGYebToxEyp2lKDSoAsvt8w82tIlP/EbmRbDVn 7bhjBlfRcFjVYw8uVDPptT0TV47vpoCVkTwcyb6OltJrvg/QzV9f07DJswuda1JH3/qvYu0p vjPnYvCq4NsqY2XSdAJ02HrdYPFtNyPEntu1n1KK+gJrstjtw7KsZ4ygXYrsm/oCBiVW/OgU g/XIlGErkrxe4vQvJyVwg6YH653YTX5hLLUEL1NS4TCo47RP+wi6y+TnuAL36UtK/uFyEuPy wwrDVcC4cIFhYSfsO0BumEI65yu7a8aHbGfq2lW251UcoU48Z27ZUUZd2Dr6O/n8poQHbaTd 6bJJSjzGGHZVbRP9UQ3lkmkmc0+XCHmj5WhwNNYjgbbmML7y0fsJT5RgvefAIFfHBg7fTY/i kBEimoUsTEQz+N4hbKwo1hULfVxDJStE4sbPhjbsPCrlXf6W9CxSyQ0qmZ2bXsLQYRj2xqd1 bpA+1o1j2N4/au1R/uSiUFjewJdT/LX1EklKDcQwpk06Af/N7VZtSfEJeRV04unbsKVXWZAk uAJyDDKN99ziC0Wz5kcPyVD1HNf8bgaqGDzrv3TfYjwqayRFcMf7xJaL9xXedMcAEQEAAYkC HwQYAQgACQUCUuE2fwIbDAAKCRBlw/kGpdefoG4XEACD1Qf/er8EA7g23HMxYWd3FXHThrVQ HgiGdk5Yh632vjOm9L4sd/GCEACVQKjsu98e8o3ysitFlznEns5EAAXEbITrgKWXDDUWGYxd pnjj2u+GkVdsOAGk0kxczX6s+VRBhpbBI2PWnOsRJgU2n10PZ3mZD4Xu9kU2IXYmuW+e5KCA vTArRUdCrAtIa1k01sPipPPw6dfxx2e5asy21YOytzxuWFfJTGnVxZZSCyLUO83sh6OZhJkk b9rxL9wPmpN/t2IPaEKoAc0FTQZS36wAMOXkBh24PQ9gaLJvfPKpNzGD8XWR5HHF0NLIJhgg 4ZlEXQ2fVp3XrtocHqhu4UZR4koCijgB8sB7Tb0GCpwK+C4UePdFLfhKyRdSXuvY3AHJd4CP 4JzW0Bzq/WXY3XMOzUTYApGQpnUpdOmuQSfpV9MQO+/jo7r6yPbxT7CwRS5dcQPzUiuHLK9i nvjREdh84qycnx0/6dDroYhp0DFv4udxuAvt1h4wGwTPRQZerSm4xaYegEFusyhbZrI0U9tJ B8WrhBLXDiYlyJT6zOV2yZFuW47VrLsjYnHwn27hmxTC/7tvG3euCklmkn9Sl9IAKFu29RSo d5bD8kMSCYsTqtTfT6W4A3qHGvIDta3ptLYpIAOD2sY3GYq2nf3Bbzx81wZK14JdDDHUX2Rs 6+ahAA==
  • Cc: jgrall@xxxxxxxxxx, Wei Liu <wl@xxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>
  • Delivery-date: Fri, 06 Dec 2019 17:50:37 +0000
  • Ironport-sdr: qBjD/sXvstg+RmW0zfLk63Y9C4Z28aI7jtoNN5PATQA/ZnFDl8RGhjhXb9ECwE7w1ecF7fLBTR VCwced4VONzqJPgElxTs4/AQxqQ5ym8A2F8CuJIG9e7Vj4jOJ3HnjlxxCeTd6C8nD/yZtPeIgp 8usPtz1FSn2jTjlI2Fs4kLY3AFNI841iRs6IiL+gmmWrFJrQQahLepvL1nCUVbghqySprVCgW5 nO4diaI6YqrmSKpuIXX8dpWXxO0t9nVsiAN97B9XoJV7LdyZ7F47xOEAFHEE4gjMjGFKGJ0slx AVM=
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>
  • Openpgp: preference=signencrypt

On 06/12/2019 15:53, Hongyan Xia wrote:
> map_pages_to_xen and modify_xen_mappings are performing almost exactly
> the same operations when shattering an l3 PTE, the only difference
> being whether we want to flush.
>
> Signed-off-by: Hongyan Xia <hongyxia@xxxxxxxxxx>

Just for reviewing purposes, how does this relate to your other posted
series.  Is it independent, a prerequisite, or does it depend on that
series?

> ---
>  xen/arch/x86/mm.c | 85 ++++++++++++++++++++++-------------------------
>  1 file changed, 40 insertions(+), 45 deletions(-)
>
> diff --git a/xen/arch/x86/mm.c b/xen/arch/x86/mm.c
> index 7d4dd80a85..42aaaa1083 100644
> --- a/xen/arch/x86/mm.c
> +++ b/xen/arch/x86/mm.c
> @@ -5151,6 +5151,43 @@ l1_pgentry_t *virt_to_xen_l1e(unsigned long v)
>                           flush_area_local((const void *)v, f) : \
>                           flush_area_all((const void *)v, f))
>  
> +/* Shatter an l3 entry and populate l2. If virt is passed in, also do flush. 
> */
> +static void shatter_l3e(l3_pgentry_t *pl3e, l2_pgentry_t *l2t,
> +        unsigned long virt, bool locking)
> +{
> +    unsigned int i;
> +    l3_pgentry_t ol3e = *pl3e;
> +
> +    for ( i = 0; i < L2_PAGETABLE_ENTRIES; i++ )
> +        l2e_write(l2t + i,
> +                  l2e_from_pfn(l3e_get_pfn(ol3e) +
> +                               (i << PAGETABLE_ORDER),
> +                               l3e_get_flags(ol3e)));

The PTE macros are especially poor for generated asm, and in cases like
this, I'd like to improve things.

In particular, I believe the following code has identical behaviour:

l2_pgentry_t nl2e = l2e_from_intpte(l3e_get_intpte(ol3e));

for ( i = 0; i < L2_PAGETABLE_ENTRIES; i++, nl2e.l2 += PAGETABLE_ORDER )
    l2e_write(l2t + i, nl2e);

(although someone please double check my logic) and rather better asm
generation.  (I also expect there to be some discussion on whether using
n2le.l2 directly is something we'd want to start doing.)

> +
> +    if ( locking )
> +        spin_lock(&map_pgdir_lock);
> +    if ( (l3e_get_flags(ol3e) & _PAGE_PRESENT) &&
> +         (l3e_get_flags(ol3e) & _PAGE_PSE) )

There is a subtle difference between the original code, and the
refactored code, and it depends on the memory barrier from spin_lock().

Previously, it was re-read from memory after the lock, whereas now it is
likely the stale version from before map_pgdir was locked.

Either you can go back to the old version and use *pl3e, or
alternatively use:

    if ( locking )
        spin_lock(&map_pgdir_lock);
    ol3e = ACCESS_ONCE(*pl3e);
    if ( ...

to make it clear that a reread from memory is required.

> +    {
> +        l3e_write_atomic(pl3e, l3e_from_mfn(virt_to_mfn(l2t),

This would probably generate better asm by using the maddr variants so
we don't have a double shift.

> +                                            __PAGE_HYPERVISOR));
> +        l2t = NULL;
> +    }
> +    if ( locking )
> +        spin_unlock(&map_pgdir_lock);
> +    if ( virt )
> +    {
> +        unsigned int flush_flags =
> +            FLUSH_TLB | FLUSH_ORDER(2 * PAGETABLE_ORDER);
> +
> +        if ( (l3e_get_flags(ol3e) & _PAGE_GLOBAL) )
> +                flush_flags |= FLUSH_TLB_GLOBAL;
> +        flush_area(virt, flush_flags);
> +    }
> +    if ( l2t )
> +        free_xen_pagetable(l2t);

This surely needs to NULL out l2t, just so the caller doesn't get any
clever ideas and ends up with a use-after-free?

~Andrew

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel

 


Rackspace

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