xen-devel.lists.xenproject.org archive mirror
 help / color / mirror / Atom feed
From: Paul Durrant <paul.durrant@citrix.com>
To: xen-devel@lists.xen.org
Cc: Paul Durrant <paul.durrant@citrix.com>,
	Keir Fraser <keir@xen.org>, Jan Beulich <jbeulich@suse.com>
Subject: [PATCH v7 05/15] x86/hvm: add length to mmio check op
Date: Thu, 9 Jul 2015 14:10:45 +0100	[thread overview]
Message-ID: <1436447455-11524-6-git-send-email-paul.durrant@citrix.com> (raw)
In-Reply-To: <1436447455-11524-1-git-send-email-paul.durrant@citrix.com>

When memory mapped I/O is range checked by internal handlers, the length
of the access should be taken into account.

Signed-off-by: Paul Durrant <paul.durrant@citrix.com>
Cc: Keir Fraser <keir@xen.org>
Cc: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>
---

v7:
- No change

v6:
- Added Andrew's reviewed-by

v5:
- Simplified by leaving mmio_check() implementation alone and
  calling to check last byte if first-byte check passes
---
 xen/arch/x86/hvm/intercept.c |   23 ++++++++++++++++++++---
 xen/include/asm-x86/hvm/io.h |   16 ++++++++++++++++
 2 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/xen/arch/x86/hvm/intercept.c b/xen/arch/x86/hvm/intercept.c
index a44baa1..f9535ea 100644
--- a/xen/arch/x86/hvm/intercept.c
+++ b/xen/arch/x86/hvm/intercept.c
@@ -35,9 +35,20 @@
 static bool_t hvm_mmio_accept(const struct hvm_io_handler *handler,
                               const ioreq_t *p)
 {
+    paddr_t first = hvm_mmio_first_byte(p);
+    paddr_t last = hvm_mmio_last_byte(p);
+
     BUG_ON(handler->type != IOREQ_TYPE_COPY);
 
-    return handler->mmio.ops->check(current, p->addr);
+    if ( !handler->mmio.ops->check(current, first) )
+        return 0;
+
+    /* Make sure the handler will accept the whole access */
+    if ( p->size > 1 &&
+         !handler->mmio.ops->check(current, last) )
+        domain_crash(current->domain);
+
+    return 1;
 }
 
 static int hvm_mmio_read(const struct hvm_io_handler *handler,
@@ -106,7 +117,8 @@ static const struct hvm_io_ops portio_ops = {
 int hvm_process_io_intercept(const struct hvm_io_handler *handler,
                              ioreq_t *p)
 {
-    struct hvm_vcpu_io *vio = &current->arch.hvm_vcpu.hvm_io;
+    struct vcpu *curr = current;
+    struct hvm_vcpu_io *vio = &curr->arch.hvm_vcpu.hvm_io;
     const struct hvm_io_ops *ops = (p->type == IOREQ_TYPE_COPY) ?
                                    &mmio_ops : &portio_ops;
     int rc = X86EMUL_OKAY, i, step = p->df ? -p->size : p->size;
@@ -215,6 +227,9 @@ int hvm_process_io_intercept(const struct hvm_io_handler *handler,
 
     if ( i != 0 )
     {
+        if ( rc == X86EMUL_UNHANDLEABLE )
+            domain_crash(curr->domain);
+
         p->count = i;
         rc = X86EMUL_OKAY;
     }
@@ -331,7 +346,9 @@ bool_t hvm_mmio_internal(paddr_t gpa)
 {
     ioreq_t p = {
         .type = IOREQ_TYPE_COPY,
-        .addr = gpa
+        .addr = gpa,
+        .count = 1,
+        .size = 1,
     };
 
     return hvm_find_io_handler(&p) != NULL;
diff --git a/xen/include/asm-x86/hvm/io.h b/xen/include/asm-x86/hvm/io.h
index 1a37243..a01502a 100644
--- a/xen/include/asm-x86/hvm/io.h
+++ b/xen/include/asm-x86/hvm/io.h
@@ -43,6 +43,22 @@ struct hvm_mmio_ops {
     hvm_mmio_write_t write;
 };
 
+static inline paddr_t hvm_mmio_first_byte(const ioreq_t *p)
+{
+    return p->df ?
+           p->addr - (p->count - 1ul) * p->size :
+           p->addr;
+}
+
+static inline paddr_t hvm_mmio_last_byte(const ioreq_t *p)
+{
+    unsigned long count = p->count;
+
+    return p->df ?
+           p->addr + p->size - 1:
+           p->addr + (count * p->size) - 1;
+}
+
 typedef int (*portio_action_t)(
     int dir, unsigned int port, unsigned int bytes, uint32_t *val);
 
-- 
1.7.10.4

  parent reply	other threads:[~2015-07-09 13:10 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-07-09 13:10 [PATCH v7 00/15] x86/hvm: I/O emulation cleanup and fix Paul Durrant
2015-07-09 13:10 ` [PATCH v7 01/15] x86/hvm: remove multiple open coded 'chunking' loops Paul Durrant
2015-07-09 15:13   ` Jan Beulich
2015-07-09 16:16     ` Paul Durrant
2015-07-09 16:24       ` Jan Beulich
2015-07-09 16:27         ` Paul Durrant
2015-07-09 13:10 ` [PATCH v7 02/15] x86/hvm: change hvm_mmio_read_t and hvm_mmio_write_t length argument Paul Durrant
2015-07-09 13:10 ` [PATCH v7 03/15] x86/hvm: restrict port numbers and uint16_t and sizes to unsigned int Paul Durrant
2015-07-09 15:24   ` Jan Beulich
2015-07-09 16:10     ` Paul Durrant
2015-07-09 16:20       ` Jan Beulich
2015-07-09 16:23         ` Paul Durrant
2015-07-09 16:31           ` Jan Beulich
2015-07-09 13:10 ` [PATCH v7 04/15] x86/hvm: unify internal portio and mmio intercepts Paul Durrant
2015-07-09 13:10 ` Paul Durrant [this message]
2015-07-09 13:10 ` [PATCH v7 06/15] x86/hvm: unify dpci portio intercept with standard portio intercept Paul Durrant
2015-07-09 13:10 ` [PATCH v7 07/15] x86/hvm: unify stdvga mmio intercept with standard mmio intercept Paul Durrant
2015-07-09 15:33   ` Jan Beulich
2015-07-09 16:12     ` Paul Durrant
2015-07-09 16:21       ` Jan Beulich
2015-07-09 16:24         ` Paul Durrant
2015-07-09 13:10 ` [PATCH v7 08/15] x86/hvm: limit reps to avoid the need to handle retry Paul Durrant
2015-07-09 13:10 ` [PATCH v7 09/15] x86/hvm: only call hvm_io_assist() from hvm_wait_for_io() Paul Durrant
2015-07-09 13:10 ` [PATCH v7 10/15] x86/hvm: split I/O completion handling from state model Paul Durrant
2015-07-09 13:10 ` [PATCH v7 11/15] x86/hvm: remove HVMIO_dispatched I/O state Paul Durrant
2015-07-09 13:10 ` [PATCH v7 12/15] x86/hvm: remove hvm_io_state enumeration Paul Durrant
2015-07-09 13:10 ` [PATCH v7 13/15] x86/hvm: use ioreq_t to track in-flight state Paul Durrant
2015-07-09 13:10 ` [PATCH v7 14/15] x86/hvm: always re-emulate I/O from a buffer Paul Durrant
2015-07-09 13:10 ` [PATCH v7 15/15] x86/hvm: track large memory mapped accesses by buffer offset Paul Durrant
2015-07-09 15:46   ` Jan Beulich
2015-07-09 16:05     ` Paul Durrant
2015-07-10  9:27 ` [PATCH v7 00/15] x86/hvm: I/O emulation cleanup and fix | Full Backtrace of domU's X crash caused by SSE2 istruction in attachment Fabio Fantoni
2015-07-10  9:31   ` Paul Durrant
2015-07-10  9:54     ` Fabio Fantoni
2015-07-10 10:09       ` Fabio Fantoni
2015-07-10 10:13         ` Paul Durrant
2015-07-10 10:20         ` Jan Beulich
2015-07-10 10:51           ` Fabio Fantoni
2015-07-10 11:00             ` Jan Beulich
2015-07-09 19:32               ` Zhi Wang
2015-07-10 11:46                 ` Jan Beulich
2015-07-10 11:49               ` Fabio Fantoni

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1436447455-11524-6-git-send-email-paul.durrant@citrix.com \
    --to=paul.durrant@citrix.com \
    --cc=jbeulich@suse.com \
    --cc=keir@xen.org \
    --cc=xen-devel@lists.xen.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).