May 2017 - tor-commits - lists.torproject.org

[tor/master] take some of the suggestions from cypherpunks on #6892
by arma＠torproject.org 30 May '17

30 May '17

commit c36145899879265e588a4eb76be5abb50e34075f Author: Roger Dingledine <arma(a)torproject.org> Date: Tue May 30 14:17:49 2017 -0400 take some of the suggestions from cypherpunks on #6892 --- doc/torify.1.txt | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/doc/torify.1.txt b/doc/torify.1.txt index 614e3f6..7e49081 100644 --- a/doc/torify.1.txt +++ b/doc/torify.1.txt @@ -17,12 +17,10 @@ SYNOPSIS DESCRIPTION ----------- -**torify** is a simple wrapper that -calls torsocks with a tor-specific +**torify** is a simple wrapper that calls torsocks with a tor-specific configuration file. -It is provided for backward compatibility; probably instead you should -just use torsocks. +It is provided for backward compatibility; instead you should use torsocks. WARNING -------

1 0

[tor/master] Update the torify.1 manpage
by arma＠torproject.org 30 May '17

30 May '17

commit 5860f0a7c8ec6aa9eb542a86b828f69e20251c98 Author: Nick Mathewson <nickm(a)torproject.org> Date: Fri May 26 15:45:39 2017 -0400 Update the torify.1 manpage I went into this to fix 6892 and say "we don't do anything for circuit isolation." But instead I did a fair amount of text-removal to stop implying that torify does anything more than call torsocks. --- changes/torify-manpage | 3 +++ doc/torify.1.txt | 22 +++++++++++----------- 2 files changed, 14 insertions(+), 11 deletions(-) diff --git a/changes/torify-manpage b/changes/torify-manpage new file mode 100644 index 0000000..f8bf56c --- /dev/null +++ b/changes/torify-manpage @@ -0,0 +1,3 @@ + o Documentation: + - Clarify the manpage for the (deprecated) torify script. Closes + ticket 6892. diff --git a/doc/torify.1.txt b/doc/torify.1.txt index 8dca901..614e3f6 100644 --- a/doc/torify.1.txt +++ b/doc/torify.1.txt @@ -17,25 +17,25 @@ SYNOPSIS DESCRIPTION ----------- -**torify** is a simple wrapper that attempts to find the best underlying Tor -wrapper available on a system. It calls torsocks with a tor specific -configuration file. + +**torify** is a simple wrapper that +calls torsocks with a tor-specific +configuration file. -torsocks is an improved wrapper that explicitly rejects UDP, safely resolves DNS -lookups and properly socksifies your TCP connections. + - -Please note that since both method use LD_PRELOAD, torify cannot be applied to -suid binaries. +It is provided for backward compatibility; probably instead you should +just use torsocks. WARNING ------- -When used with torsocks, torify should not leak DNS requests or UDP data. + +When used with torsocks, torify should not leak DNS requests or UDP data. + +torify can leak ICMP data. -Both will leak ICMP data. +torify will not ensure that different requests are processed on +different circuits. SEE ALSO -------- -**tor**(1), **tor-resolve**(1), **torsocks**(1) +**tor**(1), **torsocks**(1) AUTHORS -------

1 0

[tor/master] Merge remote-tracking branch 'argonblue/bug22413'
by nickm＠torproject.org 30 May '17

30 May '17

commit 4a15b3169645a397fc8193f91737604538ac077e Merge: c0559b9 90dd7dc Author: Nick Mathewson <nickm(a)torproject.org> Date: Tue May 30 13:33:49 2017 -0400 Merge remote-tracking branch 'argonblue/bug22413' changes/bug22413 | 4 ++++ configure.ac | 2 +- 2 files changed, 5 insertions(+), 1 deletion(-)

1 0

[tor/master] Check for libzstd >= 1.1
by nickm＠torproject.org 30 May '17

30 May '17

commit 90dd7dc92a7f7f59224bef02e847511bfc6b4489 Author: Taylor Yu <catalyst(a)torproject.org> Date: Tue May 30 13:00:54 2017 -0400 Check for libzstd >= 1.1 The consensus compression code depends on a streaming compression API that is new in libzstd-1.1. Fixes #22413. --- changes/bug22413 | 4 ++++ configure.ac | 2 +- 2 files changed, 5 insertions(+), 1 deletion(-) diff --git a/changes/bug22413 b/changes/bug22413 new file mode 100644 index 0000000..5b522f1 --- /dev/null +++ b/changes/bug22413 @@ -0,0 +1,4 @@ + o Minor bugfixes (directory protocol): + - Check for libzstd >= 1.1 because older versions lack the + necessary streaming API. Fixes bug 22413; bugfix on + 0.3.1.1-alpha. diff --git a/configure.ac b/configure.ac index 59908c7..064644e 100644 --- a/configure.ac +++ b/configure.ac @@ -856,7 +856,7 @@ if test "x$enable_zstd" = "xno"; then have_zstd=no; else PKG_CHECK_MODULES([ZSTD], - [libzstd], + [libzstd >= 1.1], have_zstd=yes, have_zstd=no)

1 0

[tor/master] hs: Correctly validate v3 descriptor encrypted length
by nickm＠torproject.org 30 May '17

30 May '17

commit 5b33d95a3dfe943625d78983bb53be2901a51150 Author: David Goulet <dgoulet(a)torproject.org> Date: Tue May 30 10:27:42 2017 -0400 hs: Correctly validate v3 descriptor encrypted length The encrypted_data_length_is_valid() function wasn't validating correctly the length of the encrypted data of a v3 descriptor. The side effect of this is that an HSDir was rejecting the descriptor and ultimately not storing it. Fixes #22447 Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- changes/bug22447 | 3 +++ src/or/hs_descriptor.c | 23 ++++------------------- src/test/test_hs_descriptor.c | 5 ++--- 3 files changed, 9 insertions(+), 22 deletions(-) diff --git a/changes/bug22447 b/changes/bug22447 new file mode 100644 index 0000000..f5649d6 --- /dev/null +++ b/changes/bug22447 @@ -0,0 +1,3 @@ + o Major bugfixes (hidden service v3): + - HSDir failed to validate the encrypted size of a v3 descriptor and thus + rejecting it. Fixes bug 22447; bugfix on tor-0.3.0.1-alpha. diff --git a/src/or/hs_descriptor.c b/src/or/hs_descriptor.c index f16a2fd..938b7a7 100644 --- a/src/or/hs_descriptor.c +++ b/src/or/hs_descriptor.c @@ -1023,30 +1023,15 @@ cert_parse_and_validate(tor_cert_t **cert_out, const char *data, STATIC int encrypted_data_length_is_valid(size_t len) { - /* Check for the minimum length possible. */ - if (len < HS_DESC_ENCRYPTED_MIN_LEN) { + /* Make sure there is enough data for the salt and the mac. The equality is + * there to ensure that there is at least one byte of encrypted data. */ + if (len <= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN) { log_warn(LD_REND, "Length of descriptor's encrypted data is too small. " "Got %lu but minimum value is %d", - (unsigned long)len, HS_DESC_ENCRYPTED_MIN_LEN); + (unsigned long)len, HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN); goto err; } - /* Encrypted data has the salt and MAC concatenated to it so remove those - * from the validation calculation. */ - len -= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; - - /* Check that it's aligned on the block size of the crypto algorithm. */ - if (len % HS_DESC_PLAINTEXT_PADDING_MULTIPLE) { - log_warn(LD_REND, "Length of descriptor's encrypted data is invalid. " - "Got %lu which is not a multiple of %d.", - (unsigned long) len, HS_DESC_PLAINTEXT_PADDING_MULTIPLE); - goto err; - } - - /* XXX: Check maximum size. Will strongly depends on the maximum intro point - * allowed we decide on and probably if they will all have to use the legacy - * key which is bigger than the ed25519 key. */ - return 1; err: return 0; diff --git a/src/test/test_hs_descriptor.c b/src/test/test_hs_descriptor.c index 02a71aa..97fe191 100644 --- a/src/test/test_hs_descriptor.c +++ b/src/test/test_hs_descriptor.c @@ -587,9 +587,8 @@ test_encrypted_data_len(void *arg) /* No length, error. */ ret = encrypted_data_length_is_valid(0); tt_int_op(ret, OP_EQ, 0); - /* Not a multiple of our encryption algorithm (thus no padding). It's - * suppose to be aligned on HS_DESC_PLAINTEXT_PADDING_MULTIPLE. */ - value = HS_DESC_PLAINTEXT_PADDING_MULTIPLE * 10 - 1; + /* This value is missing data. */ + value = HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; ret = encrypted_data_length_is_valid(value); tt_int_op(ret, OP_EQ, 0); /* Valid value. */

1 0

[tor/release-0.3.0] hs: Correctly validate v3 descriptor encrypted length
by nickm＠torproject.org 30 May '17

30 May '17

commit 5b33d95a3dfe943625d78983bb53be2901a51150 Author: David Goulet <dgoulet(a)torproject.org> Date: Tue May 30 10:27:42 2017 -0400 hs: Correctly validate v3 descriptor encrypted length The encrypted_data_length_is_valid() function wasn't validating correctly the length of the encrypted data of a v3 descriptor. The side effect of this is that an HSDir was rejecting the descriptor and ultimately not storing it. Fixes #22447 Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- changes/bug22447 | 3 +++ src/or/hs_descriptor.c | 23 ++++------------------- src/test/test_hs_descriptor.c | 5 ++--- 3 files changed, 9 insertions(+), 22 deletions(-) diff --git a/changes/bug22447 b/changes/bug22447 new file mode 100644 index 0000000..f5649d6 --- /dev/null +++ b/changes/bug22447 @@ -0,0 +1,3 @@ + o Major bugfixes (hidden service v3): + - HSDir failed to validate the encrypted size of a v3 descriptor and thus + rejecting it. Fixes bug 22447; bugfix on tor-0.3.0.1-alpha. diff --git a/src/or/hs_descriptor.c b/src/or/hs_descriptor.c index f16a2fd..938b7a7 100644 --- a/src/or/hs_descriptor.c +++ b/src/or/hs_descriptor.c @@ -1023,30 +1023,15 @@ cert_parse_and_validate(tor_cert_t **cert_out, const char *data, STATIC int encrypted_data_length_is_valid(size_t len) { - /* Check for the minimum length possible. */ - if (len < HS_DESC_ENCRYPTED_MIN_LEN) { + /* Make sure there is enough data for the salt and the mac. The equality is + * there to ensure that there is at least one byte of encrypted data. */ + if (len <= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN) { log_warn(LD_REND, "Length of descriptor's encrypted data is too small. " "Got %lu but minimum value is %d", - (unsigned long)len, HS_DESC_ENCRYPTED_MIN_LEN); + (unsigned long)len, HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN); goto err; } - /* Encrypted data has the salt and MAC concatenated to it so remove those - * from the validation calculation. */ - len -= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; - - /* Check that it's aligned on the block size of the crypto algorithm. */ - if (len % HS_DESC_PLAINTEXT_PADDING_MULTIPLE) { - log_warn(LD_REND, "Length of descriptor's encrypted data is invalid. " - "Got %lu which is not a multiple of %d.", - (unsigned long) len, HS_DESC_PLAINTEXT_PADDING_MULTIPLE); - goto err; - } - - /* XXX: Check maximum size. Will strongly depends on the maximum intro point - * allowed we decide on and probably if they will all have to use the legacy - * key which is bigger than the ed25519 key. */ - return 1; err: return 0; diff --git a/src/test/test_hs_descriptor.c b/src/test/test_hs_descriptor.c index 02a71aa..97fe191 100644 --- a/src/test/test_hs_descriptor.c +++ b/src/test/test_hs_descriptor.c @@ -587,9 +587,8 @@ test_encrypted_data_len(void *arg) /* No length, error. */ ret = encrypted_data_length_is_valid(0); tt_int_op(ret, OP_EQ, 0); - /* Not a multiple of our encryption algorithm (thus no padding). It's - * suppose to be aligned on HS_DESC_PLAINTEXT_PADDING_MULTIPLE. */ - value = HS_DESC_PLAINTEXT_PADDING_MULTIPLE * 10 - 1; + /* This value is missing data. */ + value = HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; ret = encrypted_data_length_is_valid(value); tt_int_op(ret, OP_EQ, 0); /* Valid value. */

1 0

[tor/master] Merge branch 'maint-0.3.0'
by nickm＠torproject.org 30 May '17

30 May '17

commit c0559b97b506626f6a21f8009475f18522166aa7 Merge: 671c5dc 5b33d95 Author: Nick Mathewson <nickm(a)torproject.org> Date: Tue May 30 13:21:06 2017 -0400 Merge branch 'maint-0.3.0'

1 0

[tor/maint-0.3.0] hs: Correctly validate v3 descriptor encrypted length
by nickm＠torproject.org 30 May '17

30 May '17

commit 5b33d95a3dfe943625d78983bb53be2901a51150 Author: David Goulet <dgoulet(a)torproject.org> Date: Tue May 30 10:27:42 2017 -0400 hs: Correctly validate v3 descriptor encrypted length The encrypted_data_length_is_valid() function wasn't validating correctly the length of the encrypted data of a v3 descriptor. The side effect of this is that an HSDir was rejecting the descriptor and ultimately not storing it. Fixes #22447 Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- changes/bug22447 | 3 +++ src/or/hs_descriptor.c | 23 ++++------------------- src/test/test_hs_descriptor.c | 5 ++--- 3 files changed, 9 insertions(+), 22 deletions(-) diff --git a/changes/bug22447 b/changes/bug22447 new file mode 100644 index 0000000..f5649d6 --- /dev/null +++ b/changes/bug22447 @@ -0,0 +1,3 @@ + o Major bugfixes (hidden service v3): + - HSDir failed to validate the encrypted size of a v3 descriptor and thus + rejecting it. Fixes bug 22447; bugfix on tor-0.3.0.1-alpha. diff --git a/src/or/hs_descriptor.c b/src/or/hs_descriptor.c index f16a2fd..938b7a7 100644 --- a/src/or/hs_descriptor.c +++ b/src/or/hs_descriptor.c @@ -1023,30 +1023,15 @@ cert_parse_and_validate(tor_cert_t **cert_out, const char *data, STATIC int encrypted_data_length_is_valid(size_t len) { - /* Check for the minimum length possible. */ - if (len < HS_DESC_ENCRYPTED_MIN_LEN) { + /* Make sure there is enough data for the salt and the mac. The equality is + * there to ensure that there is at least one byte of encrypted data. */ + if (len <= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN) { log_warn(LD_REND, "Length of descriptor's encrypted data is too small. " "Got %lu but minimum value is %d", - (unsigned long)len, HS_DESC_ENCRYPTED_MIN_LEN); + (unsigned long)len, HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN); goto err; } - /* Encrypted data has the salt and MAC concatenated to it so remove those - * from the validation calculation. */ - len -= HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; - - /* Check that it's aligned on the block size of the crypto algorithm. */ - if (len % HS_DESC_PLAINTEXT_PADDING_MULTIPLE) { - log_warn(LD_REND, "Length of descriptor's encrypted data is invalid. " - "Got %lu which is not a multiple of %d.", - (unsigned long) len, HS_DESC_PLAINTEXT_PADDING_MULTIPLE); - goto err; - } - - /* XXX: Check maximum size. Will strongly depends on the maximum intro point - * allowed we decide on and probably if they will all have to use the legacy - * key which is bigger than the ed25519 key. */ - return 1; err: return 0; diff --git a/src/test/test_hs_descriptor.c b/src/test/test_hs_descriptor.c index 02a71aa..97fe191 100644 --- a/src/test/test_hs_descriptor.c +++ b/src/test/test_hs_descriptor.c @@ -587,9 +587,8 @@ test_encrypted_data_len(void *arg) /* No length, error. */ ret = encrypted_data_length_is_valid(0); tt_int_op(ret, OP_EQ, 0); - /* Not a multiple of our encryption algorithm (thus no padding). It's - * suppose to be aligned on HS_DESC_PLAINTEXT_PADDING_MULTIPLE. */ - value = HS_DESC_PLAINTEXT_PADDING_MULTIPLE * 10 - 1; + /* This value is missing data. */ + value = HS_DESC_ENCRYPTED_SALT_LEN + DIGEST256_LEN; ret = encrypted_data_length_is_valid(value); tt_int_op(ret, OP_EQ, 0); /* Valid value. */

1 0

[tor/release-0.3.0] Merge branch 'maint-0.3.0' into release-0.3.0
by nickm＠torproject.org 30 May '17

30 May '17

commit 2ab1de9f3b9001f3c88cdb9da2e4c4d75c9b5ec6 Merge: 3844579 5b33d95 Author: Nick Mathewson <nickm(a)torproject.org> Date: Tue May 30 13:21:17 2017 -0400 Merge branch 'maint-0.3.0' into release-0.3.0 changes/bug22447 | 3 +++ src/or/hs_descriptor.c | 23 ++++------------------- src/test/test_hs_descriptor.c | 5 ++--- 3 files changed, 9 insertions(+), 22 deletions(-)

1 0

[stem/master] Stacktrace when running individual tests
by atagar＠torproject.org 30 May '17

30 May '17

commit a09396c9ca80040fc15bae8a6ec320549d1e4f55 Author: Damian Johnson <atagar(a)torproject.org> Date: Tue May 30 09:42:12 2017 -0700 Stacktrace when running individual tests When running individual tests we don't trigger static checks. This in turn caused our recent multithreading optimization to throw a stacktrace... Traceback (most recent call last): File "./run_tests.py", line 457, in <module> main() File "./run_tests.py", line 331, in main task.join() File "/home/atagar/Desktop/stem/test/task.py", line 206, in join self.result = self._background_pipe.recv() AttributeError: 'NoneType' object has no attribute 'recv' --- test/task.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/test/task.py b/test/task.py index 1c1b34e..63a9f20 100644 --- a/test/task.py +++ b/test/task.py @@ -202,7 +202,7 @@ class Task(object): self.error = exc def join(self): - if self._is_background_task: + if self._background_process: self.result = self._background_pipe.recv() self._background_process.join()

1 0