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

Re: [PATCH v8 1/9] xen/events: don't allow binding a global virq from any domain


  • To: Julien Grall <julien@xxxxxxx>, xen-devel@xxxxxxxxxxxxxxxxxxxx
  • From: Jürgen Groß <jgross@xxxxxxxx>
  • Date: Tue, 11 Mar 2025 10:51:13 +0100
  • Autocrypt: addr=jgross@xxxxxxxx; keydata= xsBNBFOMcBYBCACgGjqjoGvbEouQZw/ToiBg9W98AlM2QHV+iNHsEs7kxWhKMjrioyspZKOB ycWxw3ie3j9uvg9EOB3aN4xiTv4qbnGiTr3oJhkB1gsb6ToJQZ8uxGq2kaV2KL9650I1SJve dYm8Of8Zd621lSmoKOwlNClALZNew72NjJLEzTalU1OdT7/i1TXkH09XSSI8mEQ/ouNcMvIJ NwQpd369y9bfIhWUiVXEK7MlRgUG6MvIj6Y3Am/BBLUVbDa4+gmzDC9ezlZkTZG2t14zWPvx XP3FAp2pkW0xqG7/377qptDmrk42GlSKN4z76ELnLxussxc7I2hx18NUcbP8+uty4bMxABEB AAHNH0p1ZXJnZW4gR3Jvc3MgPGpncm9zc0BzdXNlLmNvbT7CwHkEEwECACMFAlOMcK8CGwMH CwkIBwMCAQYVCAIJCgsEFgIDAQIeAQIXgAAKCRCw3p3WKL8TL8eZB/9G0juS/kDY9LhEXseh mE9U+iA1VsLhgDqVbsOtZ/S14LRFHczNd/Lqkn7souCSoyWsBs3/wO+OjPvxf7m+Ef+sMtr0 G5lCWEWa9wa0IXx5HRPW/ScL+e4AVUbL7rurYMfwCzco+7TfjhMEOkC+va5gzi1KrErgNRHH kg3PhlnRY0Udyqx++UYkAsN4TQuEhNN32MvN0Np3WlBJOgKcuXpIElmMM5f1BBzJSKBkW0Jc Wy3h2Wy912vHKpPV/Xv7ZwVJ27v7KcuZcErtptDevAljxJtE7aJG6WiBzm+v9EswyWxwMCIO RoVBYuiocc51872tRGywc03xaQydB+9R7BHPzsBNBFOMcBYBCADLMfoA44MwGOB9YT1V4KCy vAfd7E0BTfaAurbG+Olacciz3yd09QOmejFZC6AnoykydyvTFLAWYcSCdISMr88COmmCbJzn sHAogjexXiif6ANUUlHpjxlHCCcELmZUzomNDnEOTxZFeWMTFF9Rf2k2F0Tl4E5kmsNGgtSa aMO0rNZoOEiD/7UfPP3dfh8JCQ1VtUUsQtT1sxos8Eb/HmriJhnaTZ7Hp3jtgTVkV0ybpgFg w6WMaRkrBh17mV0z2ajjmabB7SJxcouSkR0hcpNl4oM74d2/VqoW4BxxxOD1FcNCObCELfIS auZx+XT6s+CE7Qi/c44ibBMR7hyjdzWbABEBAAHCwF8EGAECAAkFAlOMcBYCGwwACgkQsN6d 1ii/Ey9D+Af/WFr3q+bg/8v5tCknCtn92d5lyYTBNt7xgWzDZX8G6/pngzKyWfedArllp0Pn fgIXtMNV+3t8Li1Tg843EXkP7+2+CQ98MB8XvvPLYAfW8nNDV85TyVgWlldNcgdv7nn1Sq8g HwB2BHdIAkYce3hEoDQXt/mKlgEGsLpzJcnLKimtPXQQy9TxUaLBe9PInPd+Ohix0XOlY+Uk QFEx50Ki3rSDl2Zt2tnkNYKUCvTJq7jvOlaPd6d/W0tZqpyy7KVay+K4aMobDsodB3dvEAs6 ScCnh03dDAFgIq5nsB11j3KPKdVoPlfucX2c7kGNH+LUMbzqV6beIENfNexkOfxHfw==
  • Cc: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>, Anthony PERARD <anthony.perard@xxxxxxxxxx>, Michal Orzel <michal.orzel@xxxxxxx>, Jan Beulich <jbeulich@xxxxxxxx>, Roger Pau Monné <roger.pau@xxxxxxxxxx>, Stefano Stabellini <sstabellini@xxxxxxxxxx>
  • Delivery-date: Tue, 11 Mar 2025 09:51:22 +0000
  • List-id: Xen developer discussion <xen-devel.lists.xenproject.org>

On 11.03.25 10:35, Julien Grall wrote:
Hi Juergen,

On 04/02/2025 11:33, Juergen Gross wrote:
Today Xen will happily allow binding a global virq by a domain which
isn't configured to receive it. This won't result in any bad actions,
but the bind will appear to have succeeded with no event ever being
received by that event channel.

Instead of allowing the bind, error out if the domain isn't set to
handle that virq. Note that this check is inside the write_lock() on
purpose, as a future patch will put a related check into
set_global_virq_handler() with the addition of using the same lock.
 > > Signed-off-by: Juergen Gross <jgross@xxxxxxxx>

I see this patch was already committed. But I have a question about the logic.

---
V6:
- new patch
V7:
- move handling domain check inside locked region (Jan Beulich)
- style fix (Jan Beulich)
---
  xen/common/event_channel.c | 21 +++++++++++++++++----
  1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/xen/common/event_channel.c b/xen/common/event_channel.c
index 46281b16ce..cd6f5a1211 100644
--- a/xen/common/event_channel.c
+++ b/xen/common/event_channel.c
@@ -120,6 +120,13 @@ static uint8_t get_xen_consumer(xen_event_channel_notification_t fn)
  /* Get the notification function for a given Xen-bound event channel. */
  #define xen_notification_fn(e) (xen_consumers[(e)->xen_consumer-1])
+static struct domain *__read_mostly global_virq_handlers[NR_VIRQS];
+
+static struct domain *get_global_virq_handler(unsigned int virq)
+{
+    return global_virq_handlers[virq] ?: hardware_domain;
+}
+
  static bool virq_is_global(unsigned int virq)
  {
      switch ( virq )
@@ -469,6 +476,7 @@ int evtchn_bind_virq(evtchn_bind_virq_t *bind, evtchn_port_t port)
      struct domain *d = current->domain;
      int            virq = bind->virq, vcpu = bind->vcpu;
      int            rc = 0;
+    bool           is_global;
      if ( (virq < 0) || (virq >= ARRAY_SIZE(v->virq_to_evtchn)) )
          return -EINVAL;
@@ -478,8 +486,9 @@ int evtchn_bind_virq(evtchn_bind_virq_t *bind, evtchn_port_t port)
      * speculative execution.
      */
      virq = array_index_nospec(virq, ARRAY_SIZE(v->virq_to_evtchn));
+    is_global = virq_is_global(virq);
-    if ( virq_is_global(virq) && (vcpu != 0) )
+    if ( is_global && vcpu != 0 )
          return -EINVAL;
      if ( (v = domain_vcpu(d, vcpu)) == NULL )
@@ -487,6 +496,12 @@ int evtchn_bind_virq(evtchn_bind_virq_t *bind, evtchn_port_t port)
      write_lock(&d->event_lock);
+    if ( is_global && get_global_virq_handler(virq) != d )

What prevent a race between get_global_virq_handler() and set_global_virq_handler()? Also, it is not clear in the implementation of get_global_virq_handler() that it will ever only read global_virq_handlers[virq] once.

set_global_virq_handler() is taking the event_lock of the domain
registered as handler.

So if a domain is registered for handling a virq, d->event_lock is
protecting against the handling domain to be changed. Concurrent
calls of set_global_virq_handler() are handled via taking the
global_virq_handlers_lock spin_lock.


Juergen

Attachment: OpenPGP_0xB0DE9DD628BF132F.asc
Description: OpenPGP public key

Attachment: OpenPGP_signature.asc
Description: OpenPGP digital signature


 


Rackspace

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