- tor-commits - lists.torproject.org

[tor/main] fixup! relay: Change DNS timeout label on MetricsPort
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit b37674fec75f14a0be9dbeff458355d91897cbe5 Author: David Goulet <dgoulet(a)torproject.org> Date: Tue Dec 14 16:12:17 2021 -0500 fixup! relay: Change DNS timeout label on MetricsPort Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- src/rust/.cargo/config | 12 ------------ 1 file changed, 12 deletions(-) diff --git a/src/rust/.cargo/config b/src/rust/.cargo/config deleted file mode 100644 index d91e26719d..0000000000 --- a/src/rust/.cargo/config +++ /dev/null @@ -1,12 +0,0 @@ -[source] - - [source.crates-io] - registry = 'https://github.com/rust-lang/crates.io-index' - replace-with = 'vendored-sources' - - [source.vendored-sources] - directory = '' - -[build] - rustflags = [ "-D", "warnings" ] -

1 0

[tor/main] Merge branch 'ticket40527_046_01' into ticket40527_047_01
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit ad6a0ebb119dba09fd1a0d1d6eb2a1362803bce3 Merge: b4c55f3a70 cda7acb35d Author: David Goulet <dgoulet(a)torproject.org> Date: Mon Dec 13 10:48:54 2021 -0500 Merge branch 'ticket40527_046_01' into ticket40527_047_01 changes/ticket40527 | 5 +++ config.rust | 24 +++++++++++++ src/feature/stats/rephist.c | 12 ------- src/rust/.cargo/config | 12 +++++++ src/test/test_stats.c | 82 ++------------------------------------------- 5 files changed, 44 insertions(+), 91 deletions(-) diff --cc config.rust index 0000000000,0000000000..878a231cda new file mode 100644 --- /dev/null +++ b/config.rust @@@ -1,0 -1,0 +1,24 @@@ ++# Used by our cargo build.rs script to get variables from autoconf. ++# ++# The "configure" script will generate "config.rust" from "config.rust.in", ++# and then build.rs will read "config.rust". ++ ++BUILDDIR=/home/dgoulet/Documents/git/tor ++TOR_LDFLAGS_zlib= ++TOR_LDFLAGS_nss=@TOR_LDFLAGS_nss@ ++TOR_LDFLAGS_openssl= ++TOR_LDFLAGS_libevent= ++TOR_ZLIB_LIBS=-lz ++TOR_LIB_MATH=-lm ++TOR_LIBEVENT_LIBS=-levent ++TOR_OPENSSL_LIBS=-lssl -lcrypto ++TOR_LIB_WS32= ++TOR_LIB_GDI= ++TOR_LIB_USERENV= ++CURVE25519_LIBS= ++TOR_SYSTEMD_LIBS=-lsystemd ++TOR_LZMA_LIBS=-llzma ++TOR_ZSTD_LIBS=-lzstd ++LIBS=-lseccomp -lcap ++LDFLAGS= -pie -z relro -z now -rdynamic ++NSS_LIBS= diff --cc src/rust/.cargo/config index 0000000000,0000000000..d91e26719d new file mode 100644 --- /dev/null +++ b/src/rust/.cargo/config @@@ -1,0 -1,0 +1,12 @@@ ++[source] ++ ++ [source.crates-io] ++ registry = 'https://github.com/rust-lang/crates.io-index' ++ replace-with = 'vendored-sources' ++ ++ [source.vendored-sources] ++ directory = '' ++ ++[build] ++ rustflags = [ "-D", "warnings" ] ++

1 0

[tor/main] Merge remote-tracking branch 'tor-gitlab/mr/500' into main
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit 193781e6ef8b1236bfd51eab344abcb0e914e168 Merge: 48d778bc32 eb06d52dae Author: Alexander Færøy <ahf(a)torproject.org> Date: Wed Dec 15 12:46:18 2021 +0000 Merge remote-tracking branch 'tor-gitlab/mr/500' into main changes/ticket40527 | 11 ++++ src/feature/nodelist/networkstatus.c | 1 - src/feature/relay/relay_metrics.c | 2 +- src/feature/stats/rephist.c | 100 ++--------------------------------- src/feature/stats/rephist.h | 2 - src/test/test_stats.c | 82 ++-------------------------- 6 files changed, 18 insertions(+), 180 deletions(-)

1 0

[tor/release-0.4.6] relay: Don't make DNS timeout trigger an overload
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit cda7acb35d40c505dc4d2c3b55d611faab189477 Author: David Goulet <dgoulet(a)torproject.org> Date: Mon Dec 13 10:22:29 2021 -0500 relay: Don't make DNS timeout trigger an overload Tor has configure libevent to attempt up to 3 times a DNS query for a maximum of 5 seconds each. Once that 5 seconds has elapsed, it consider the query "Timed Out" but tor only gets a timeout if all 3 attempts have failed. For example, using Unbound, it has a much higher threshold of timeout. It is well defined in https://www.nlnetlabs.nl/documentation/unbound/info-timeout/ and has some complexity to it. But the gist is that if it times out, it will be much more than 5 seconds. And so the Tor DNS timeouts are more of a "UX issue" rather than a "network issue". For this reason, we are removing this metric from the overload general signal. See https://gitlab.torproject.org/tpo/network-health/team/-/issues/139 for more information. Fixes #40527 Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- changes/ticket40527 | 5 +++ src/feature/stats/rephist.c | 12 ------- src/test/test_stats.c | 82 ++------------------------------------------- 3 files changed, 8 insertions(+), 91 deletions(-) diff --git a/changes/ticket40527 b/changes/ticket40527 new file mode 100644 index 0000000000..631b3d4bb9 --- /dev/null +++ b/changes/ticket40527 @@ -0,0 +1,5 @@ + o Major bugfixes (relay, overload): + - Don't make Tor DNS timeout trigger an overload general state. These + timeouts are different from DNS server timeout. They have to be seen as + timeout related to UX and not because of a network problem. Fixes bug + 40527; bugfix on 0.4.6.1-alpha. diff --git a/src/feature/stats/rephist.c b/src/feature/stats/rephist.c index c3a281a8c2..2bfa14d326 100644 --- a/src/feature/stats/rephist.c +++ b/src/feature/stats/rephist.c @@ -283,18 +283,6 @@ overload_general_dns_assessment(void) return; } - /* Lets see if we can signal a general overload. */ - double fraction = (double) overload_dns_stats.stats_n_error_timeout / - (double) overload_dns_stats.stats_n_request; - if (fraction >= overload_dns_timeout_fraction) { - log_notice(LD_HIST, "General overload -> DNS timeouts (%" PRIu64 ") " - "fraction %.4f%% is above threshold of %.4f%%", - overload_dns_stats.stats_n_error_timeout, - fraction * 100.0, - overload_dns_timeout_fraction * 100.0); - rep_hist_note_overload(OVERLOAD_GENERAL); - } - reset: /* Reset counters for the next period. */ overload_dns_stats.stats_n_error_timeout = 0; diff --git a/src/test/test_stats.c b/src/test/test_stats.c index 3b9a192c75..af035ae54c 100644 --- a/src/test/test_stats.c +++ b/src/test/test_stats.c @@ -721,7 +721,7 @@ test_overload_stats(void *arg) stats_str = rep_hist_get_overload_stats_lines(); tt_assert(!stats_str); - /* Note a DNS overload */ + /* Note a overload */ rep_hist_note_overload(OVERLOAD_GENERAL); /* Move the time forward one hour */ @@ -742,7 +742,7 @@ test_overload_stats(void *arg) /* Now the time should be 2002-01-07 00:00:00 */ - /* Note a DNS overload */ + /* Note a overload */ rep_hist_note_overload(OVERLOAD_GENERAL); stats_str = rep_hist_get_overload_general_line(); @@ -760,7 +760,7 @@ test_overload_stats(void *arg) tt_str_op("overload-fd-exhausted 1 2002-01-07 00:00:00\n", OP_EQ, stats_str); tor_free(stats_str); - /* Move the time forward. Register DNS overload. See that the time changed */ + /* Move the time forward. Register overload. See that the time changed */ current_time += 3600*2; update_approx_time(current_time); @@ -867,81 +867,6 @@ test_overload_stats(void *arg) tor_free(stats_str); } -/** Test the overload stats logic. */ -static void -test_overload_dns_timeout(void *arg) -{ - char *stats_str = NULL; - (void) arg; - - /* Lets simulate a series of timeouts but below our default 1% threshold. */ - - for (int i = 0; i < 1000; i++) { - /* This should trigger 9 timeouts which is just below 1% (10) */ - if (i > 0 && !(i % 100)) { - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - } else { - rep_hist_note_dns_query(0, DNS_ERR_NONE); - } - } - - /* No overload yet. */ - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* Move it 10 minutes in the future and see if we get a general overload. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should NOT trigger the general overload because we are below - * our default of 1%. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* We'll now go above our 1% threshold. */ - for (int i = 0; i < 1000; i++) { - /* This should trigger 10 timeouts which is our threshold of 1% (10) */ - if (!(i % 10)) { - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - } else { - rep_hist_note_dns_query(0, DNS_ERR_NONE); - } - } - - /* Move it 10 minutes in the future and see if we get a general overload. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should trigger the general overload because we are above 1%. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(stats_str); - tor_free(stats_str); - - /* Move 72h in the future, we should NOT get an overload anymore. */ - update_approx_time(approx_time() + (72 * 3600)); - - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* This query should NOT trigger the general overload. */ - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* Move it 10 minutes in the future and see if we get a general overload. We - * have now 100% of requests timing out. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should trigger the general overload with 50% of timeouts. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(stats_str); - tor_free(stats_str); - - done: - tor_free(stats_str); -} - #define ENT(name) \ { #name, test_ ## name , 0, NULL, NULL } #define FORK(name) \ @@ -958,7 +883,6 @@ struct testcase_t stats_tests[] = { FORK(rephist_v3_onions), FORK(load_stats_file), FORK(overload_stats), - FORK(overload_dns_timeout), END_OF_TESTCASES };

1 0

[tor/maint-0.4.6] relay: Don't make DNS timeout trigger an overload
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit cda7acb35d40c505dc4d2c3b55d611faab189477 Author: David Goulet <dgoulet(a)torproject.org> Date: Mon Dec 13 10:22:29 2021 -0500 relay: Don't make DNS timeout trigger an overload Tor has configure libevent to attempt up to 3 times a DNS query for a maximum of 5 seconds each. Once that 5 seconds has elapsed, it consider the query "Timed Out" but tor only gets a timeout if all 3 attempts have failed. For example, using Unbound, it has a much higher threshold of timeout. It is well defined in https://www.nlnetlabs.nl/documentation/unbound/info-timeout/ and has some complexity to it. But the gist is that if it times out, it will be much more than 5 seconds. And so the Tor DNS timeouts are more of a "UX issue" rather than a "network issue". For this reason, we are removing this metric from the overload general signal. See https://gitlab.torproject.org/tpo/network-health/team/-/issues/139 for more information. Fixes #40527 Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- changes/ticket40527 | 5 +++ src/feature/stats/rephist.c | 12 ------- src/test/test_stats.c | 82 ++------------------------------------------- 3 files changed, 8 insertions(+), 91 deletions(-) diff --git a/changes/ticket40527 b/changes/ticket40527 new file mode 100644 index 0000000000..631b3d4bb9 --- /dev/null +++ b/changes/ticket40527 @@ -0,0 +1,5 @@ + o Major bugfixes (relay, overload): + - Don't make Tor DNS timeout trigger an overload general state. These + timeouts are different from DNS server timeout. They have to be seen as + timeout related to UX and not because of a network problem. Fixes bug + 40527; bugfix on 0.4.6.1-alpha. diff --git a/src/feature/stats/rephist.c b/src/feature/stats/rephist.c index c3a281a8c2..2bfa14d326 100644 --- a/src/feature/stats/rephist.c +++ b/src/feature/stats/rephist.c @@ -283,18 +283,6 @@ overload_general_dns_assessment(void) return; } - /* Lets see if we can signal a general overload. */ - double fraction = (double) overload_dns_stats.stats_n_error_timeout / - (double) overload_dns_stats.stats_n_request; - if (fraction >= overload_dns_timeout_fraction) { - log_notice(LD_HIST, "General overload -> DNS timeouts (%" PRIu64 ") " - "fraction %.4f%% is above threshold of %.4f%%", - overload_dns_stats.stats_n_error_timeout, - fraction * 100.0, - overload_dns_timeout_fraction * 100.0); - rep_hist_note_overload(OVERLOAD_GENERAL); - } - reset: /* Reset counters for the next period. */ overload_dns_stats.stats_n_error_timeout = 0; diff --git a/src/test/test_stats.c b/src/test/test_stats.c index 3b9a192c75..af035ae54c 100644 --- a/src/test/test_stats.c +++ b/src/test/test_stats.c @@ -721,7 +721,7 @@ test_overload_stats(void *arg) stats_str = rep_hist_get_overload_stats_lines(); tt_assert(!stats_str); - /* Note a DNS overload */ + /* Note a overload */ rep_hist_note_overload(OVERLOAD_GENERAL); /* Move the time forward one hour */ @@ -742,7 +742,7 @@ test_overload_stats(void *arg) /* Now the time should be 2002-01-07 00:00:00 */ - /* Note a DNS overload */ + /* Note a overload */ rep_hist_note_overload(OVERLOAD_GENERAL); stats_str = rep_hist_get_overload_general_line(); @@ -760,7 +760,7 @@ test_overload_stats(void *arg) tt_str_op("overload-fd-exhausted 1 2002-01-07 00:00:00\n", OP_EQ, stats_str); tor_free(stats_str); - /* Move the time forward. Register DNS overload. See that the time changed */ + /* Move the time forward. Register overload. See that the time changed */ current_time += 3600*2; update_approx_time(current_time); @@ -867,81 +867,6 @@ test_overload_stats(void *arg) tor_free(stats_str); } -/** Test the overload stats logic. */ -static void -test_overload_dns_timeout(void *arg) -{ - char *stats_str = NULL; - (void) arg; - - /* Lets simulate a series of timeouts but below our default 1% threshold. */ - - for (int i = 0; i < 1000; i++) { - /* This should trigger 9 timeouts which is just below 1% (10) */ - if (i > 0 && !(i % 100)) { - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - } else { - rep_hist_note_dns_query(0, DNS_ERR_NONE); - } - } - - /* No overload yet. */ - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* Move it 10 minutes in the future and see if we get a general overload. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should NOT trigger the general overload because we are below - * our default of 1%. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* We'll now go above our 1% threshold. */ - for (int i = 0; i < 1000; i++) { - /* This should trigger 10 timeouts which is our threshold of 1% (10) */ - if (!(i % 10)) { - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - } else { - rep_hist_note_dns_query(0, DNS_ERR_NONE); - } - } - - /* Move it 10 minutes in the future and see if we get a general overload. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should trigger the general overload because we are above 1%. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(stats_str); - tor_free(stats_str); - - /* Move 72h in the future, we should NOT get an overload anymore. */ - update_approx_time(approx_time() + (72 * 3600)); - - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* This query should NOT trigger the general overload. */ - rep_hist_note_dns_query(0, DNS_ERR_TIMEOUT); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(!stats_str); - - /* Move it 10 minutes in the future and see if we get a general overload. We - * have now 100% of requests timing out. */ - update_approx_time(approx_time() + (10 * 60)); - - /* This query should trigger the general overload with 50% of timeouts. */ - rep_hist_note_dns_query(0, DNS_ERR_NONE); - stats_str = rep_hist_get_overload_general_line(); - tt_assert(stats_str); - tor_free(stats_str); - - done: - tor_free(stats_str); -} - #define ENT(name) \ { #name, test_ ## name , 0, NULL, NULL } #define FORK(name) \ @@ -958,7 +883,6 @@ struct testcase_t stats_tests[] = { FORK(rephist_v3_onions), FORK(load_stats_file), FORK(overload_stats), - FORK(overload_dns_timeout), END_OF_TESTCASES };

1 0

[tor/main] fixup! relay: Change DNS timeout label on MetricsPort
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit eb06d52dae9cbb8344f543e1077ca6f0b62a317b Author: David Goulet <dgoulet(a)torproject.org> Date: Tue Dec 14 16:13:00 2021 -0500 fixup! relay: Change DNS timeout label on MetricsPort Signed-off-by: David Goulet <dgoulet(a)torproject.org> --- src/test/test_stats.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/src/test/test_stats.c b/src/test/test_stats.c index a224c21198..c7e0c10845 100644 --- a/src/test/test_stats.c +++ b/src/test/test_stats.c @@ -721,7 +721,7 @@ test_overload_stats(void *arg) stats_str = rep_hist_get_overload_stats_lines(); tt_assert(!stats_str); - /* Note a overload */ + /* Note an overload */ rep_hist_note_overload(OVERLOAD_GENERAL); /* Move the time forward one hour */ @@ -742,7 +742,7 @@ test_overload_stats(void *arg) /* Now the time should be 2002-01-07 00:00:00 */ - /* Note a overload */ + /* Note an overload */ rep_hist_note_overload(OVERLOAD_GENERAL); stats_str = rep_hist_get_overload_general_line();

1 0

[tor/release-0.4.6] Merge branch 'maint-0.4.6' into release-0.4.6
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit a89ebdbc37003f1b0f611618e083a91b2fd36276 Merge: d06bcf7672 cda7acb35d Author: Alexander Færøy <ahf(a)torproject.org> Date: Wed Dec 15 12:43:48 2021 +0000 Merge branch 'maint-0.4.6' into release-0.4.6 changes/ticket40527 | 5 +++ src/feature/stats/rephist.c | 12 ------- src/test/test_stats.c | 82 ++------------------------------------------- 3 files changed, 8 insertions(+), 91 deletions(-)

1 0

[tor/main] Do not count controller-selected paths towards path bias.
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit dd085d42f96e39f9387aa1d7a306100c0d7df305 Author: Nick Mathewson <nickm(a)torproject.org> Date: Mon Nov 15 08:55:47 2021 -0500 Do not count controller-selected paths towards path bias. As a side effect, this fixes a "Bug" warning. Closes #40515. Bugfix on 0.2.4.10-alpha. --- changes/bug40515 | 6 ++++++ src/core/or/origin_circuit_st.h | 6 ++++++ src/feature/client/circpathbias.c | 11 +++++++++++ src/feature/control/control_cmd.c | 2 ++ 4 files changed, 25 insertions(+) diff --git a/changes/bug40515 b/changes/bug40515 new file mode 100644 index 0000000000..d315e28411 --- /dev/null +++ b/changes/bug40515 @@ -0,0 +1,6 @@ + o Minor bugfixes (controller, path bias): + - When a circuit's path is specified, in full or in part, from the + controller API, do not count that circuit towards our path-bias + calculations. (Doing so was incorrect, since we cannot tell whether + the controller is selecting relays randomly.) Resolves a "Bug" + warning. Fixes bug 40515; bugfix on 0.2.4.10-alpha. diff --git a/src/core/or/origin_circuit_st.h b/src/core/or/origin_circuit_st.h index c40e84aed8..ca7c4b1bef 100644 --- a/src/core/or/origin_circuit_st.h +++ b/src/core/or/origin_circuit_st.h @@ -182,6 +182,12 @@ struct origin_circuit_t { */ unsigned first_hop_from_controller : 1; + /** + * If true, this circuit's path has been chosen, in full or in part, + * by the controller API, and it's okay to ignore checks that we'd + * usually do on the path as whole. */ + unsigned int any_hop_from_controller : 1; + /** * Tristate variable to guard against pathbias miscounting * due to circuit purpose transitions changing the decision diff --git a/src/feature/client/circpathbias.c b/src/feature/client/circpathbias.c index 4d27553926..29264135f9 100644 --- a/src/feature/client/circpathbias.c +++ b/src/feature/client/circpathbias.c @@ -363,6 +363,17 @@ pathbias_should_count(origin_circuit_t *circ) return 0; } + /* Ignore circuits where the controller helped choose the path. When + * this happens, we can't be sure whether the path was chosen randomly + * or not. */ + if (circ->any_hop_from_controller) { + /* (In this case, we _don't_ check to see if shouldcount is changing, + * since it's possible that an already-created circuit later gets extended + * by the controller. */ + circ->pathbias_shouldcount = PATHBIAS_SHOULDCOUNT_IGNORED; + return 0; + } + /* Completely ignore one hop circuits */ if (circ->build_state->onehop_tunnel || circ->build_state->desired_path_len == 1) { diff --git a/src/feature/control/control_cmd.c b/src/feature/control/control_cmd.c index b4d7228b51..a1c5e55fe1 100644 --- a/src/feature/control/control_cmd.c +++ b/src/feature/control/control_cmd.c @@ -822,6 +822,8 @@ handle_control_extendcircuit(control_connection_t *conn, circ->first_hop_from_controller = 1; } + circ->any_hop_from_controller = 1; + /* now circ refers to something that is ready to be extended */ first_node = zero_circ; SMARTLIST_FOREACH(nodes, const node_t *, node,

1 0

[tor/main] Merge remote-tracking branch 'tor-gitlab/mr/491' into main
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit 48d778bc3246cd7e459475f1045ad24684ddeb85 Merge: 95b82c4fee dd085d42f9 Author: Alexander Færøy <ahf(a)torproject.org> Date: Wed Dec 15 12:41:00 2021 +0000 Merge remote-tracking branch 'tor-gitlab/mr/491' into main changes/bug40515 | 6 ++++++ src/core/or/origin_circuit_st.h | 6 ++++++ src/feature/client/circpathbias.c | 11 +++++++++++ src/feature/control/control_cmd.c | 2 ++ 4 files changed, 25 insertions(+)

1 0

[tor/main] Limit the number of elements in a consdiff hash line.
by ahf＠torproject.org 15 Dec '21

15 Dec '21

commit 86819229afde13ae8466ee782f4c4bd9ba6f37cd Author: Nick Mathewson <nickm(a)torproject.org> Date: Mon Dec 6 12:35:08 2021 -0500 Limit the number of elements in a consdiff hash line. This avoids performing and then freeing a lot of small mallocs() if the hash line has too many elements. Fixes one case of bug 40472; resolves OSS-Fuzz 38363. Bugfix on 0.3.1.1-alpha when the consdiff parsing code was introduced. --- changes/bug40472 | 6 ++++++ src/feature/dircommon/consdiff.c | 2 +- 2 files changed, 7 insertions(+), 1 deletion(-) diff --git a/changes/bug40472 b/changes/bug40472 new file mode 100644 index 0000000000..d87c1dc2cc --- /dev/null +++ b/changes/bug40472 @@ -0,0 +1,6 @@ + o Minor bugfixes (performance, DoS): + - Fix one case of a not-especially viable denial-of-service attack found + by OSS-Fuzz in our consensus-diff parsing code. This attack causes a + lot small of memory allocations and then immediately frees them: this + is only slow when running with all the sanitizers enabled. Fixes one + case of bug 40472; bugfix on 0.3.1.1-alpha. diff --git a/src/feature/dircommon/consdiff.c b/src/feature/dircommon/consdiff.c index d0f7594ce3..3c38e92dd6 100644 --- a/src/feature/dircommon/consdiff.c +++ b/src/feature/dircommon/consdiff.c @@ -1126,7 +1126,7 @@ consdiff_get_digests(const smartlist_t *diff, { const cdline_t *line2 = smartlist_get(diff, 1); char *h = tor_memdup_nulterm(line2->s, line2->len); - smartlist_split_string(hash_words, h, " ", 0, 0); + smartlist_split_string(hash_words, h, " ", 0, 4); tor_free(h); }

1 0