1134561 - (CVE-2015-0803) Type confusion in HTMLSourceElement::AfterSetAttr

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Nils]

Attachment: testcase

The issue can be triggered while setting a specific attribute (srcset,sizes,media,type) of a <source> element. When iterating over the sibling elements in order to find an HTMLImageElement this function does only check the tag name and not the namespace:

nsresult
HTMLSourceElement::AfterSetAttr(int32_t aNameSpaceID, nsIAtom* aName,
                                const nsAttrValue* aValue, bool aNotify)
{
  // If we are associated with a <picture> with a valid <img>, notify it of
  // responsive parameter changes
  nsINode *parent = nsINode::GetParentNode();
  if (aNameSpaceID == kNameSpaceID_None &&
      (aName == nsGkAtoms::srcset ||
       aName == nsGkAtoms::sizes ||
       aName == nsGkAtoms::media ||
       aName == nsGkAtoms::type) &&
      parent && parent->Tag() == nsGkAtoms::picture) {
    nsString strVal = aValue ? aValue->GetStringValue() : EmptyString();
    // Find all img siblings after this <source> and notify them of the change
    nsCOMPtr<nsINode> sibling = AsContent();
    while ( (sibling = sibling->GetNextSibling()) ) {
      if (sibling->Tag() == nsGkAtoms::img) {
        HTMLImageElement *img = static_cast<HTMLImageElement*>(sibling.get());
        if (aName == nsGkAtoms::srcset) {
          img->PictureSourceSrcsetChanged(AsContent(), strVal, aNotify);
        } else if (aName == nsGkAtoms::sizes) {
          img->PictureSourceSizesChanged(AsContent(), strVal, aNotify);
        } else if (aName == nsGkAtoms::media ||
                   aName == nsGkAtoms::type) {
          img->PictureSourceMediaOrTypeChanged(AsContent(), aNotify);
        }
      }
    }


This results in a cast of an object which isn't an HTMLImageElement to this type. The attached testcase demonstrates this behaviour. ASan output attached as well. ASan detects this as a use-after-free issue.

Comment 1

[Nils]

Attachment: asan.txt

Comment 2

[Robert Longson [:longsonr]]

Attachment: ishtml.txt

Comment 3

[Robert Longson [:longsonr]]

This is going to be sec-critical for the same reason as bug 1134131

Comment 4

[Andrea Marchesini [:baku]]

Attachment: fix1.patch

I agree with smaug about removing Tag() completely, but writing that patch will take a while. In the meantime, this patch fixes this issue.

Comment 5

[Andrea Marchesini [:baku]]

Comment on attachment 8566477 [details] [diff] [review]
fix1.patch

Sorry, I didn't see the previous patch. I guess we submitted the patch more or less at the same time.

Comment 6

[Andrea Marchesini [:baku]]

Comment on attachment 8566465 [details] [diff] [review]
ishtml.txt

Review of attachment 8566465 [details] [diff] [review]:
-----------------------------------------------------------------

Just comparing my patch with yours :)
I let jst the real review of the patch.

::: dom/html/HTMLPictureElement.cpp
@@ +43,5 @@
>  void
>  HTMLPictureElement::RemoveChildAt(uint32_t aIndex, bool aNotify)
>  {
>    // Find all img siblings after this <source> to notify them of its demise
>    nsCOMPtr<nsINode> child = GetChildAt(aIndex);

nsIContent

@@ +46,5 @@
>    // Find all img siblings after this <source> to notify them of its demise
>    nsCOMPtr<nsINode> child = GetChildAt(aIndex);
>    nsCOMPtr<nsIContent> nextSibling;
> +  if (child && child->IsElement() &&
> +      child->AsElement()->IsHTML(nsGkAtoms::source)) {

and remove AsElement/IsElement

::: dom/html/HTMLSourceElement.cpp
@@ +90,3 @@
>      nsString strVal = aValue ? aValue->GetStringValue() : EmptyString();
>      // Find all img siblings after this <source> and notify them of the change
>      nsCOMPtr<nsINode> sibling = AsContent();

nsCOMPtr<nsIContent> sibling = AsContent();

@@ +91,5 @@
>      // Find all img siblings after this <source> and notify them of the change
>      nsCOMPtr<nsINode> sibling = AsContent();
>      while ( (sibling = sibling->GetNextSibling()) ) {
> +      if (sibling->IsElement() &&
> +          sibling->AsElement()->IsHTML(nsGkAtoms::img)) {

and here you can remove IsElement() and AsElement()

@@ +152,4 @@
>      // Find any img siblings after this <source> and notify them
>      nsCOMPtr<nsINode> sibling = AsContent();
>      while ( (sibling = sibling->GetNextSibling()) ) {
> +      if (sibling->IsElement() &&

same here.

Comment 7

[Robert Longson [:longsonr]]

Attachment: ishtml.txt

Address review comments

Comment 8

[Robert Longson [:longsonr]]

Comment on attachment 8566491 [details] [diff] [review]
ishtml.txt

[Security approval request comment]
How easily could an exploit be constructed based on the patch? You'd know you needed to create something using a picture tag in a non-html namespace but not exactly how to do that.
Do comments in the patch, the check-in comment, or tests included in the patch paint a bulls-eye on the security problem? no.

Which older supported branches are affected by this flaw? Firefox 33 onwards preffed off, only preffed on in Firefox 38

If not all supported branches, which bug introduced the flaw? bug 870022, bug 1017875 turned the pref on.

Do you have backports for the affected branches? If not, how different, hard to create, and risky will they be? I think the patch will apply to all branches. 

How likely is this patch to cause regressions; how much testing does it need? This whole area could do with more tests but that seems to be bug 1017878 as far as I can tell.

Comment 9

[Al Billings [:abillings - ex-MoCo]]

Sec-approval for checkin on March 10 or later. We're about to ship so we can't take this for 36.

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9264aec18c62

Andrea asked me to mention that for the most part, this patch was rendered obsolete by bug 1134280 anyway.

Comment 11

[Ryan VanderMeulen [:RyanVM]]

We'll of course need to use the attached patch for the uplifts at least. Robert, can you please request Aurora/Beta approval on this when you get a chance?

Comment 12

[Robert Longson [:longsonr]]

Comment on attachment 8566491 [details] [diff] [review]
ishtml.txt

Approval Request Comment
[Feature/regressing bug #]: bug 870022, bug 1017875 turned the pref on.
[User impact if declined]: exploitable security hole.
[Describe test coverage new/current, TreeHerder]: This whole area could do with more tests but that seems to be bug 1017878 as far as I can tell.
[Risks and why]: Pretty low risk we're just adding extra checks.
[String/UUID change made/needed]: none

Comment 13

[Wes Kocher (:KWierso) (Not reading bugmail; email directly if needed)]

https://hg.mozilla.org/mozilla-central/rev/9264aec18c62

Comment 14

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-aurora/rev/b837e465cdf1
https://hg.mozilla.org/releases/mozilla-beta/rev/97c57043b3fc