When navigating to an about: page that doesn't exist (e.g.
"about:somethingthatdoesnotexist"), the docShell will call
nsSecureBrowserUIImpl::OnLocationChange with a request that is null.
Consequently, we can't use that to QueryInterface to a nsISecurityEventSink to
call OnSecurityChange. The previous implementation would use the prior
request's nsISecurityEventSink, which was a bug but luckily this produced the
correct behavior. Since the original docShell the nsSecureBrowserUIImpl was
initialized with is what needs to be notified, we can just QueryInterface that
to an nsISecurityEventSink and call OnSecurityChange directly instead.
Differential Revision: https://phabricator.services.mozilla.com/D6951
--HG--
rename : browser/base/content/test/siteIdentity/browser_tls_handshake_failure.js => browser/base/content/test/siteIdentity/browser_navigation_failures.js
extra : moz-landing-system : lando
This also removes the (afaict, unused) stub implementation from TabParent. The netwerk header
inclusions were necessary because those files included TabParent.h and through it,
nsISecureBrowserUI, but now TabParent.h no longer does that.
Differential Revision: https://phabricator.services.mozilla.com/D6829
--HG--
extra : moz-landing-system : lando
The site identity security indicator machinery treats connections where the TLS
handshake failed as insecure (also referred to as "unknown identity"). Before
bug 1468222, such cases were easily detectable as the SSLStatus field of the
relevant nsITransportSecurityInfo would be null. When we merged nsISSLStatus
into nsITransportSecurityInfo, we didn't take this differentiation into account.
This patch brings back the prior behavior by checking if the securityInfo's
securityState indicates that the handshake failed (i.e. it is
STATE_IS_INSECURE).
Differential Revision: https://phabricator.services.mozilla.com/D6316
--HG--
extra : moz-landing-system : lando
Everything that goes in a PLDHashtable (and its derivatives, like
nsTHashtable) needs to inherit from PLDHashEntryHdr. But through a lack
of enforcement, copy constructors for these derived classes didn't
explicitly invoke the copy constructor for PLDHashEntryHdr (and the
compiler didn't invoke the copy constructor for us). Instead,
PLDHashTable explicitly copied around the bits that the copy constructor
would have.
The current setup has two problems:
1) Derived classes should be using move construction, not copy
construction, since anything that's shuffling hash table keys/entries
around will be using move construction.
2) Derived classes should take responsibility for transferring bits of
superclass state around, and not rely on something else to handle that.
The second point is not a huge problem for PLDHashTable (PLDHashTable
only has to copy PLDHashEntryHdr's bits in a single place), but future
hash table implementations that might move entries around more
aggressively would have to insert compensation code all over the
place. Additionally, if moving entries is implemented via memcpy (which
is quite common), PLDHashTable copying around bits *again* is
inefficient.
Let's fix all these problems in one go, by:
1) Explicitly declaring the set of constructors that PLDHashEntryHdr
implements (and does not implement). In particular, the copy
constructor is deleted, so any derived classes that attempt to make
themselves copyable will be detected at compile time: the compiler
will complain that the superclass type is not copyable.
This change on its own will result in many compiler errors, so...
2) Change any derived classes to implement move constructors instead of
copy constructors. Note that some of these move constructors are,
strictly speaking, unnecessary, since the relevant classes are moved
via memcpy in nsTHashtable and its derivatives.
A previous patch in this bug made the incorrect assumption that we had disabled
the family safety root detection/importing feature by default. In reality, we
enabled it by default in bug 1282871.
In bug 1487258 we moved enterprise root loading to a background thread so as to
not block the main thread. This patch does the same with the family safety
feature.
Differential Revision: https://phabricator.services.mozilla.com/D5484
--HG--
extra : moz-landing-system : lando
I also changed security/certverifier/moz.build a bit while I am here:
* Using '-Xclang' to pass through '-Wall' on clang-cl.
* Now clang-cl will take clang/gcc path because most '-wd****' options have no
effect on clang-cl. '-wd4010' will have an effect, but we already have the
corresponding clang/gcc option ('-Wno-unused-parameter').
--HG--
extra : source : df566a1bd9087cc0bfc03fe19fd9d21bf58f5d9c
nsIAssociatedContentSecurity and nsISecurityInfoProvider are unused as of
bug 832834, so this patch removes them.
Differential Revision: https://phabricator.services.mozilla.com/D5693
--HG--
extra : moz-landing-system : lando
Each instance of CERTCertList creates a PLArena with a chunk size of 2048 bytes,
but only needs space for 3 pointers per certificate in the list. The majority of
the time Gecko uses CERTCertList, we'll store ~3 certificates (although in some
cases we do store a few hundred, such as in tests or the certificate manager).
This is fairly inefficient. This patch starts the process of avoiding using
CERTCertList in Gecko by converting nsNSSCertList (i.e. nsIX509CertList) (as
well as nsNSSCertListEnumerator) to use a more efficient data structure to hold
references to certificates long-term. Future follow-up patches could (and
should) update certificate verification APIs in PSM to avoid CERTCertList as
well.
Depends on D5096
Differential Revision: https://phabricator.services.mozilla.com/D5097
--HG--
extra : moz-landing-system : lando
nsIX509CertList.getRawCertList is only used once and doesn't provide
particularly unique functionality (its one use can easily be re-worked in terms
of other APIs). Removing this API will ease refactoring work to avoid holding
long-lived references to CERTCertList instances in nsNSSCertList.
Differential Revision: https://phabricator.services.mozilla.com/D5096
--HG--
extra : moz-landing-system : lando
Background here is that we are disabling a piece of the downgrade protection in TLS 1.3 and we want to turn it on. We don't know if that is safe, so a pref (and an associated experiment) seems prudent. This is that pref.
Differential Revision: https://phabricator.services.mozilla.com/D4629
--HG--
extra : moz-landing-system : lando
Several source files use DLL_PREFIX/DLL_SUFFIX defines, and they all set
them in moz.build using `DEFINES`. This is problematic for the WSL
build because the quoting gets lost somewhere between bash and cl.exe.
We cannot simply set them globally in moz.configure because their
stringified definitions would conflict with the `set_config` of
DLL_PREFIX/DLL_SUFFIX. Therefore, we globally define
MOZ_DLL_PREFIX/MOZ_DLL_SUFFIX and change all define-related uses of
DLL_PREFIX/DLL_SUFFIX to use their MOZ-equivalents instead.
Move all fields of nsISSLStatus to nsITransportSecurityProvider
Remove nsISSLStatus interface and definition
Update all code and test references to nsISSLStatus
Maintain ability to read in older version of serialized nsISSLStatus. This
is verified with psm_DeserializeCert gtest.
Differential Revision: https://phabricator.services.mozilla.com/D3704
--HG--
extra : moz-landing-system : lando
Move all fields of nsISSLStatus to nsITransportSecurityProvider
Remove nsISSLStatus interface and definition
Update all code and test references to nsISSLStatus
Maintain ability to read in older version of serialized nsISSLStatus. This
is verified with psm_DeserializeCert gtest.
Differential Revision: https://phabricator.services.mozilla.com/D3704
--HG--
extra : moz-landing-system : lando
As of bug 1346297, we don't collect telemetry for the family safety root
feature. At this point, it makes the most sense to disable the entire feature by
default.
Differential Revision: https://phabricator.services.mozilla.com/D4994
--HG--
extra : moz-landing-system : lando
