package/libupnp18: security bump to version 1.14.0
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
--- package/libupnp18/libupnp18.hash | 6 +++--- package/libupnp18/libupnp18.mk | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-)
package/libupnp18/libupnp18.hash +3 -3
@@ -1,5 +1,5 @@ -# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.8.7/libupnp-1.8.7.tar.bz2.sha1 -sha1 2ea3011180c58b0584f0cb73cc8e685a0a1c4ec8 libupnp-1.8.7.tar.bz2 +# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.14.0/libupnp-1.14.0.tar.bz2.sha1 +sha1 b14cff9ddd7cfe7f0e4bf552387122a31770f51f libupnp-1.14.0.tar.bz2 # Locally computed: -sha256 e38c69b2b67322e67cd53680db9b02c7c1f720a47a3cd626fd89d57d2dca93b8 libupnp-1.8.7.tar.bz2 +sha256 ecb23d4291968c8a7bdd4eb16fc2250dbacc16b354345a13342d67f571d35ceb libupnp-1.14.0.tar.bz2 sha256 c8b99423cad48bb44e2cf52a496361404290865eac259a82da6d1e4331ececb3 COPYING
package/libupnp18/libupnp18.mk +1 -1
@@ -4,7 +4,7 @@ # ################################################################################ -LIBUPNP18_VERSION = 1.8.7 +LIBUPNP18_VERSION = 1.14.0 LIBUPNP18_SOURCE = libupnp-$(LIBUPNP18_VERSION).tar.bz2 LIBUPNP18_SITE = http://downloads.sourceforge.net/project/pupnp/pupnp/libupnp-$(LIBUPNP18_VERSION) LIBUPNP18_CONF_ENV = ac_cv_lib_compat_ftime=no
Signature · 1 line
2.28.0
PATCH 2/3 package/gerbera: security bump to version 1.6.0
- This version is compatible with libupnp 1.14.x which fix CallStranger a.k.a. CVE-2020-12695
- Drop all patches (already in version)
- expat is not needed since version 1.5.0 and
https://github.com/gerbera/gerbera/commit/a4f0cccd6a1f741c55ca69b06cff3a964eebc1f3
- fmt is a mandatory dependency since version 1.5.0 and
https://github.com/gerbera/gerbera/commit/fe81e5fc8898d6e3a53ce30ddaafb8439683f46f
- spdlog is a mandatory dependency since version 1.5.0 and
https://github.com/gerbera/gerbera/commit/615d698fe4dce9d7462022a00c74af1fac7a1003
- pugixml is a mandatory dependency since version 1.5.0 and
https://github.com/gerbera/gerbera/commit/c244006aa04ab2e4c5f3e7003ca727e05440238d
- Set CXX_FILESYSTEM_NO_LINK_NEEDED to ON to avoid a build failure due to check_cxx_source_runs which has been added with
https://github.com/gerbera/gerbera/commit/8ea0fce24ce9b1cf870837c3be984fed50581dfb
- Update indentation in hash file (two spaces)
--- ...build-without-js-taglib-or-atrailers.patch | 51 --------------- ...indLibMagic-cmake-fix-static-linking.patch | 63 ------------------- ...make-FindMatroska-fix-static-linking.patch | 37 ----------- package/gerbera/Config.in | 9 ++- package/gerbera/gerbera.hash | 4 +- package/gerbera/gerbera.mk | 9 ++- 6 files changed, 14 insertions(+), 159 deletions(-) delete mode 100644 package/gerbera/0001-fix-matroska-build-without-js-taglib-or-atrailers.patch delete mode 100644 package/gerbera/0002-cmake-FindLibMagic-cmake-fix-static-linking.patch delete mode 100644 package/gerbera/0003-cmake-FindMatroska-fix-static-linking.patch
package/gerbera/0001-fix-matroska-build-without-js-taglib-or-atrailers.patch deleted +0 -51
@@ -1,51 +0,0 @@ -From 0ac781b0b0deef5c02c32a70ac484f882c3f4dd0 Mon Sep 17 00:00:00 2001 -From: Fabrice Fontaine <[email protected]> -Date: Tue, 24 Dec 2019 18:55:57 +0100 -Subject: [PATCH] fix matroska build without js, taglib or atrailers -MIME-Version: 1.0 -Content-Type: text/plain; charset=UTF-8 -Content-Transfer-Encoding: 8bit - -i2i function is used in matroska_handler.cc but this function is defined -only if defined(HAVE_JS) || defined(HAVE_TAGLIB) || defined(ATRAILERS) -as a result compilation fails if HAVE_MATROSKA is set but HAVE_JS, -HAVE_TAGLIG or ATRAILERS are not. - -Backported from: 0ac781b0b0deef5c02c32a70ac484f882c3f4dd0 - -Signed-off-by: Fabrice Fontaine <[email protected]> -Signed-off-by: J?rg Krause <[email protected]> ---- - src/string_converter.cc | 2 +- - src/string_converter.h | 2 +- - 2 files changed, 2 insertions(+), 2 deletions(-) - -diff --git a/src/string_converter.cc b/src/string_converter.cc -index f669c661..7a3c55d7 100644 ---- a/src/string_converter.cc -+++ b/src/string_converter.cc -@@ -218,7 +218,7 @@ Ref<StringConverter> StringConverter::p2i() - } - #endif - --#if defined(HAVE_JS) || defined(HAVE_TAGLIB) || defined(ATRAILERS) -+#if defined(HAVE_JS) || defined(HAVE_TAGLIB) || defined(ATRAILERS) || defined(HAVE_MATROSKA) - - Ref<StringConverter> StringConverter::i2i() - { -diff --git a/src/string_converter.h b/src/string_converter.h -index 58495430..f75bf833 100644 ---- a/src/string_converter.h -+++ b/src/string_converter.h -@@ -65,7 +65,7 @@ public: - static zmm::Ref<StringConverter> p2i(); - - #endif --#if defined(HAVE_JS) || defined(HAVE_TAGLIB) || defined(ATRAILERS) -+#if defined(HAVE_JS) || defined(HAVE_TAGLIB) || defined(ATRAILERS) || defined(HAVE_MATROSKA) - /// \brief safeguard - internal to internal - needed to catch some - /// scenarious where the user may have forgotten to add proper conversion - /// in the script. --- -2.26.1 -
package/gerbera/0002-cmake-FindLibMagic-cmake-fix-static-linking.patch deleted +0 -63
63 lines not on this page · raw message
package/gerbera/0003-cmake-FindMatroska-fix-static-linking.patch deleted +0 -37
@@ -1,37 +0,0 @@ -From aab2eacbaad10759294f4fd74bbb5ecef3cf3a8d Mon Sep 17 00:00:00 2001 -From: Fabrice Fontaine <[email protected]> -Date: Tue, 24 Dec 2019 22:57:18 +0100 -Subject: [PATCH] cmake/FindMatroska: fix static linking - -Fix static linking with libmatrasoka by adding PC_EBM_LIBRARIES to -EBML_LIBRARIES and PC_MAT_LIBRARIES to MATROSKA_LIBRARIES - -Signed-off-by: Fabrice Fontaine <[email protected]> -[Retrieved from: -https://github.com/gerbera/gerbera/commit/aab2eacbaad10759294f4fd74bbb5ecef3cf3a8d] ---- - cmake/FindMatroska.cmake | 4 ++-- - 1 file changed, 2 insertions(+), 2 deletions(-) - -diff --git a/cmake/FindMatroska.cmake b/cmake/FindMatroska.cmake -index 4b09a5ec..12ca593d 100644 ---- a/cmake/FindMatroska.cmake -+++ b/cmake/FindMatroska.cmake -@@ -23,7 +23,7 @@ FIND_LIBRARY(EBML_LIBRARY ebml - FIND_PACKAGE_HANDLE_STANDARD_ARGS(EBML - REQUIRED_VARS EBML_LIBRARY EBML_INCLUDE_DIR) - if (EBML_FOUND) -- set (EBML_LIBRARIES ${EBML_LIBRARY}) -+ set (EBML_LIBRARIES ${EBML_LIBRARY} ${PC_EBM_LIBRARIES}) - set (EBML_INCLUDE_DIRS ${EBML_INCLUDE_DIR} ) - endif () - MARK_AS_ADVANCED( -@@ -41,7 +41,7 @@ find_library(MATROSKA_LIBRARY matroska - FIND_PACKAGE_HANDLE_STANDARD_ARGS(MATROSKA - REQUIRED_VARS MATROSKA_LIBRARY MATROSKA_INCLUDE_DIR) - if (MATROSKA_FOUND) -- set (MATROSKA_LIBRARIES ${MATROSKA_LIBRARY}) -+ set (MATROSKA_LIBRARIES ${MATROSKA_LIBRARY} ${PC_MAT_LIBRARIES}) - set (MATROSKA_INCLUDE_DIRS ${MATROSKA_INCLUDE_DIR} ) - endif () - MARK_AS_ADVANCED(
package/gerbera/Config.in +6 -3
@@ -5,10 +5,13 @@ config BR2_PACKAGE_GERBERA depends on BR2_TOOLCHAIN_HAS_THREADS depends on BR2_TOOLCHAIN_GCC_AT_LEAST_7 # C++17 optional depends on BR2_INSTALL_LIBSTDCPP + depends on BR2_USE_WCHAR # fmt depends on !BR2_PACKAGE_LIBUPNP # libupnp18 - select BR2_PACKAGE_EXPAT + select BR2_PACKAGE_FMT select BR2_PACKAGE_LIBICONV if !BR2_ENABLE_LOCALE select BR2_PACKAGE_LIBUPNP18 + select BR2_PACKAGE_PUGIXML + select BR2_PACKAGE_SPDLOG select BR2_PACKAGE_SQLITE select BR2_PACKAGE_UTIL_LINUX select BR2_PACKAGE_UTIL_LINUX_LIBUUID
@@ -19,9 +22,9 @@ config BR2_PACKAGE_GERBERA https://gerbera.io -comment "gerbera needs a toolchain w/ C++, threads, gcc >= 7" +comment "gerbera needs a toolchain w/ C++, threads, wchar, gcc >= 7" depends on BR2_USE_MMU depends on BR2_TOOLCHAIN_HAS_ATOMIC depends on !BR2_PACKAGE_LIBUPNP depends on !BR2_INSTALL_LIBSTDCPP || !BR2_TOOLCHAIN_HAS_THREADS || \ - !BR2_TOOLCHAIN_GCC_AT_LEAST_7 + !BR2_USE_WCHAR || !BR2_TOOLCHAIN_GCC_AT_LEAST_7
package/gerbera/gerbera.hash +2 -2
@@ -1,3 +1,3 @@ # Locally computed: -sha256 904a9031c85ac805e4c139f363510226952683d7257acd1dee25ba1e97fd7651 gerbera-1.4.0.tar.gz -sha256 cae4138373be41fd2be75faf41ce7efbcf49fb17d0e05ad1c51cc01ac335b9b6 LICENSE.md +sha256 3a4956ec5fea1101e8daa32d9cfb985db908a49e2ac6137a1b2bf509e2684a6c gerbera-1.6.0.tar.gz +sha256 cae4138373be41fd2be75faf41ce7efbcf49fb17d0e05ad1c51cc01ac335b9b6 LICENSE.md
package/gerbera/gerbera.mk +6 -3
@@ -4,20 +4,23 @@ # ################################################################################ -GERBERA_VERSION = 1.4.0 +GERBERA_VERSION = 1.6.0 GERBERA_SITE = $(call github,gerbera,gerbera,v$(GERBERA_VERSION)) GERBERA_LICENSE = GPL-2.0 GERBERA_LICENSE_FILES = LICENSE.md GERBERA_DEPENDENCIES = \ - expat \ + fmt \ host-pkgconf \ libupnp18 \ + pugixml \ + spdlog \ sqlite \ util-linux \ zlib GERBERA_CONF_OPTS = \ -DWITH_DEBUG=OFF \ - -DWITH_JS=OFF + -DWITH_JS=OFF \ + -DCXX_FILESYSTEM_NO_LINK_NEEDED=ON # Uses __atomic_fetch_add_4 ifeq ($(BR2_TOOLCHAIN_HAS_LIBATOMIC),y)
Signature · 1 line
2.28.0
PATCH 3/3 package/gmrender-resurrect: add libupnp 1.14.x support
In reply to Fabrice FontaineUsing libupnp 1.14.x is needed to fix CallStranger a.k.a. CVE-2020-12695
--- .../0001-Drop-UpnpInit.patch | 174 ++++++++++++++++++ 1 file changed, 174 insertions(+) create mode 100644 package/gmrender-resurrect/0001-Drop-UpnpInit.patch
package/gmrender-resurrect/0001-Drop-UpnpInit.patch new file +174 -0
@@ -0,0 +1,174 @@ +From d3264ebc98a17fae00004dff5f5608bbc07b83e2 Mon Sep 17 00:00:00 2001 +From: Fabrice Fontaine <[email protected]> +Date: Fri, 21 Aug 2020 21:29:00 +0200 +Subject: [PATCH] Drop UpnpInit + +UpnpInit has been dropped from libupnp 1.14.x as it can't be fixed +against CallStranger a.k.a. CVE-2020-12695 so replace it by UpnpInit2 + +Signed-off-by: Fabrice Fontaine <[email protected]> +--- + dist-scripts/centos7/README.md | 2 +- + dist-scripts/debian/gmediarender.1 | 8 ++------ + dist-scripts/fedora/README.md | 2 +- + src/main.c | 13 ++++--------- + src/upnp_device.c | 18 +++++++++--------- + src/upnp_device.h | 2 +- + 6 files changed, 18 insertions(+), 27 deletions(-) + +diff --git a/dist-scripts/centos7/README.md b/dist-scripts/centos7/README.md +index 278d777..ed82fb6 100644 +--- a/dist-scripts/centos7/README.md ++++ b/dist-scripts/centos7/README.md +@@ -45,7 +45,7 @@ Additional configuration is also recommended, sice there's no configuration file + # vi /etc/systemd/system/gmediarender.service.d/customize.conf # or nano, or emacs, or whatever editor you like + [Service] + ExecStart= +- ExecStart=/usr/bin/gmediarender --port=49494 --ip-address=<your_IP_address> -f "DLNA Renderer GMediaRender" ++ ExecStart=/usr/bin/gmediarender --port=49494 --interface-name=<your_interface_name> -f "DLNA Renderer GMediaRender" + + # systemctl daemon-reload + # systemctl start gmediarender.service +diff --git a/dist-scripts/debian/gmediarender.1 b/dist-scripts/debian/gmediarender.1 +index 96123ff..2d663c9 100644 +--- a/dist-scripts/debian/gmediarender.1 ++++ b/dist-scripts/debian/gmediarender.1 +@@ -50,12 +50,8 @@ Usually, it is desirable for the renderer + to show up on controllers under a recognisable and unique name. This is + the option to set that name. + .TP +-.B \-I, \-\-ip\-address \fI\<ip-address\>\fP +-The local IP address the service is running and advertised on. +- +-This can +-only be a single address, and must be explicitly specified (i.e. not +-0.0.0.0). ++.B \-I, \-\-ip\-address \fI\<interface-name\>\fP ++The local interface name the service is running and advertised on. + .TP + .B \-p, \-\-port \fI\<port>\fP + Port to listen to. [49152..65535]. +diff --git a/dist-scripts/fedora/README.md b/dist-scripts/fedora/README.md +index 7b9ea4b..45aa536 100644 +--- a/dist-scripts/fedora/README.md ++++ b/dist-scripts/fedora/README.md +@@ -43,7 +43,7 @@ Additional configuration is also recommended, sice there's no configuration file + # vi /etc/systemd/system/gmediarender.service.d/customize.conf # or nano, or emacs, or whatever editor you like + [Service] + ExecStart= +- ExecStart=/usr/bin/gmediarender --port=49494 --ip-address=<your_IP_address> -f "DLNA Renderer GMediaRender" ++ ExecStart=/usr/bin/gmediarender --port=49494 --interface-name=<your_interface_name> -f "DLNA Renderer GMediaRender" + + # systemctl daemon-reload + # systemctl start gmediarender.service +diff --git a/src/main.c b/src/main.c +index ef720e3..2030c49 100644 +--- a/src/main.c ++++ b/src/main.c +@@ -69,11 +69,7 @@ static gboolean show_transport_scpd = FALSE; + static gboolean show_outputs = FALSE; + static gboolean daemon_mode = FALSE; + +-// IP-address seems strange in libupnp: they actually don't bind to +-// that address, but to INADDR_ANY (miniserver.c in upnp library). +-// Apparently they just use this for the advertisement ? Anyway, 0.0.0.0 would +-// not work. +-static const gchar *ip_address = NULL; ++static const gchar *interface_name = NULL; + static int listen_port = 49494; + + #ifdef GMRENDER_UUID +@@ -92,9 +88,8 @@ static const gchar *mime_filter = NULL; + static GOptionEntry option_entries[] = { + { "version", 0, 0, G_OPTION_ARG_NONE, &show_version, + "Output version information and exit", NULL }, +- { "ip-address", 'I', 0, G_OPTION_ARG_STRING, &ip_address, +- "The local IP address the service is running and advertised " +- "(only one, 0.0.0.0 won't work)", NULL }, ++ { "interface-name", 'I', 0, G_OPTION_ARG_STRING, &interface_name, ++ "The local interface name the service is running and advertised", NULL }, + // The following is not very reliable, as libupnp does not set + // SO_REUSEADDR by default, so it might increment (sending patch). + { "port", 'p', 0, G_OPTION_ARG_INT, &listen_port, +@@ -302,7 +297,7 @@ int main(int argc, char **argv) + listen_port); + return EXIT_FAILURE; + } +- device = upnp_device_init(upnp_renderer, ip_address, listen_port); ++ device = upnp_device_init(upnp_renderer, interface_name, listen_port); + if (device == NULL) { + Log_error("main", "ERROR: Failed to initialize UPnP device"); + return EXIT_FAILURE; +diff --git a/src/upnp_device.c b/src/upnp_device.c +index db65e4f..3151238 100644 +--- a/src/upnp_device.c ++++ b/src/upnp_device.c +@@ -416,13 +416,13 @@ static UPNP_CALLBACK(event_handler, EventType, event, userdata) + + static gboolean initialize_device(struct upnp_device_descriptor *device_def, + struct upnp_device *result_device, +- const char *ip_address, ++ const char *interface_name, + unsigned short port) + { + int rc; + char *buf; + +- rc = UpnpInit(ip_address, port); ++ rc = UpnpInit2(interface_name, port); + /* There have been situations reported in which UPNP had issues + * initializing right after network came up. #129 + */ +@@ -430,13 +430,13 @@ static gboolean initialize_device(struct upnp_device_descriptor *device_def, + static const int kRetryTimeMs = 1000; + while (rc != UPNP_E_SUCCESS && retries_left--) { + usleep(kRetryTimeMs * 1000); +- Log_error("upnp", "UpnpInit(ip=%s, port=%d) Error: %s (%d). Retrying... (%ds)", +- ip_address, port, UpnpGetErrorMessage(rc), rc, retries_left); +- rc = UpnpInit(ip_address, port); ++ Log_error("upnp", "UpnpInit2(interface=%s, port=%d) Error: %s (%d). Retrying... (%ds)", ++ interface_name, port, UpnpGetErrorMessage(rc), rc, retries_left); ++ rc = UpnpInit2(interface_name, port); + } + if (UPNP_E_SUCCESS != rc) { +- Log_error("upnp", "UpnpInit(ip=%s, port=%d) Error: %s (%d). Giving up.", +- ip_address, port, UpnpGetErrorMessage(rc), rc); ++ Log_error("upnp", "UpnpInit2(interface=%s, port=%d) Error: %s (%d). Giving up.", ++ interface_name, port, UpnpGetErrorMessage(rc), rc); + return FALSE; + } + Log_info("upnp", "Registered IP=%s port=%d\n", +@@ -483,7 +483,7 @@ static gboolean initialize_device(struct upnp_device_descriptor *device_def, + } + + struct upnp_device *upnp_device_init(struct upnp_device_descriptor *device_def, +- const char *ip_address, ++ const char *interface_name, + unsigned short port) + { + int rc; +@@ -516,7 +516,7 @@ struct upnp_device *upnp_device_init(struct upnp_device_descriptor *device_def, + webserver_register_buf(srv->scpd_url, buf, "text/xml"); + } + +- if (!initialize_device(device_def, result_device, ip_address, port)) { ++ if (!initialize_device(device_def, result_device, interface_name, port)) { + UpnpFinish(); + free(result_device); + return NULL; +diff --git a/src/upnp_device.h b/src/upnp_device.h +index 3e635e1..8c8e783 100644 +--- a/src/upnp_device.h ++++ b/src/upnp_device.h +@@ -49,7 +49,7 @@ struct upnp_device; + struct action_event; + + struct upnp_device *upnp_device_init(struct upnp_device_descriptor *device_def, +- const char *ip_address, ++ const char *interface_name, + unsigned short port); + + void upnp_device_shutdown(struct upnp_device *device); +-- +2.28.0 +
Signature · 1 line
2.28.0
PATCH 3/3 package/gmrender-resurrect: add libupnp 1.14.x support
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Fabrice" == Fabrice Fontaine <[email protected]> writes:
Using libupnp 1.14.x is needed to fix CallStranger a.k.a. CVE-2020-12695
Signed-off-by Fabrice Fontaine <[email protected]>
It would be good to add a link to the upstream PR (which also explains why you are doing those "unrelated" centos/fedora changes):
https://github.com/hzeller/gmrender-resurrect/pull/214
Signature · 1 line
Bye, Peter Korsgaard
PATCH 3/3 package/gmrender-resurrect: add libupnp 1.14.x support
Le lun. 31 ao?t 2020 ? 22:15, Peter Korsgaard <[email protected]> a ?crit :
Quoted 10 lines
Quoted 5 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
"Fabrice" == Fabrice Fontaine <[email protected]> writes:
Using libupnp 1.14.x is needed to fix CallStranger a.k.a. CVE-2020-12695
Signed-off-by Fabrice Fontaine <[email protected]>It would be good to add a link to the upstream PR (which also explains why you are doing those "unrelated" centos/fedora changes):
Yes, will be in v2.
Quoted 3 lines
--
Bye, Peter Korsgaard
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
In reply to Fabrice FontaineQuoted 1 line · Fabrice Fontaine
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
Again, although this bump indeed fixes those issues, it's a feature version
bump so I'm not sure if it can be called "security bump".
In addition, the libupnp18 package exists because of API incompatibility with
1.6. Are we sure that this problem doesn't repeat itself for 1.14?
Regards,
Arnout
Quoted 33 lines
Signed-off-by Fabrice Fontaine <[email protected]>--- package/libupnp18/libupnp18.hash | 6 +++--- package/libupnp18/libupnp18.mk | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-)package/libupnp18/libupnp18.hash +3
-3@@ -1,5 +1,5 @@ -# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.8.7/libupnp-1.8.7.tar.bz2.sha1 -sha1 2ea3011180c58b0584f0cb73cc8e685a0a1c4ec8 libupnp-1.8.7.tar.bz2 +# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.14.0/libupnp-1.14.0.tar.bz2.sha1 +sha1 b14cff9ddd7cfe7f0e4bf552387122a31770f51f libupnp-1.14.0.tar.bz2 # Locally computed: -sha256 e38c69b2b67322e67cd53680db9b02c7c1f720a47a3cd626fd89d57d2dca93b8 libupnp-1.8.7.tar.bz2 +sha256 ecb23d4291968c8a7bdd4eb16fc2250dbacc16b354345a13342d67f571d35ceb libupnp-1.14.0.tar.bz2 sha256 c8b99423cad48bb44e2cf52a496361404290865eac259a82da6d1e4331ececb3 COPYINGpackage/libupnp18/libupnp18.mk +1
-1@@ -4,7 +4,7 @@ # ################################################################################ -LIBUPNP18_VERSION = 1.8.7 +LIBUPNP18_VERSION = 1.14.0 LIBUPNP18_SOURCE = libupnp-$(LIBUPNP18_VERSION).tar.bz2 LIBUPNP18_SITE = http://downloads.sourceforge.net/project/pupnp/pupnp/libupnp-$(LIBUPNP18_VERSION) LIBUPNP18_CONF_ENV = ac_cv_lib_compat_ftime=no
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
Le dim. 30 ao?t 2020 ? 20:34, Arnout Vandecappelle <[email protected]> a ?crit :
Quoted 11 lines · Fabrice Fontaine
Quoted 1 line
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
Again, although this bump indeed fixes those issues, it's a feature version
bump so I'm not sure if it can be called "security bump".In addition, the libupnp18 package exists because of API incompatibility with
1.6. Are we sure that this problem doesn't repeat itself for 1.14?
There is indeed an API incompatibility between 1.8 and 1.14 related to CallStranger a.k.a. CVE-2020-12695: starting from 1.14, UpnpInit function has been removed as this function can't be fixed against CallStranger because this function takes an IP address and not an interface name.
However, UpnpInit2 is available for more than 10 years and is used by most of the applications (i.e. mpd and vlc) with the exception of gmrender-resurrect (which is patched in this serie).
As soon as libupnp 1.14 is available, I'm planning to update the applications that are still using the legacy libupnp 1.6.x version (i.e. igd2-for-linux and ushare) and drop this unsecure version. I would like to avoid adding a third version of libupnp (i.e. a libupnp114 package) as from a security perspective, all packages should use this version.
Still, I agree that this is not only a "security bump" so I would advise to apply this serie to next and backport it to our LTS branches in a few months.
Quoted 31 lines
Regards,
ArnoutQuoted 27 lines
Signed-off-by Fabrice Fontaine <[email protected]>--- package/libupnp18/libupnp18.hash | 6 +++--- package/libupnp18/libupnp18.mk | 2 +- 2 files changed, 4 insertions(+), 4 deletions(-)package/libupnp18/libupnp18.hash +3
-3@@ -1,5 +1,5 @@ -# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.8.7/libupnp-1.8.7.tar.bz2.sha1 -sha1 2ea3011180c58b0584f0cb73cc8e685a0a1c4ec8 libupnp-1.8.7.tar.bz2 +# From https://sourceforge.net/projects/pupnp/files/pupnp/libupnp-1.14.0/libupnp-1.14.0.tar.bz2.sha1 +sha1 b14cff9ddd7cfe7f0e4bf552387122a31770f51f libupnp-1.14.0.tar.bz2 # Locally computed: -sha256 e38c69b2b67322e67cd53680db9b02c7c1f720a47a3cd626fd89d57d2dca93b8 libupnp-1.8.7.tar.bz2 +sha256 ecb23d4291968c8a7bdd4eb16fc2250dbacc16b354345a13342d67f571d35ceb libupnp-1.14.0.tar.bz2 sha256 c8b99423cad48bb44e2cf52a496361404290865eac259a82da6d1e4331ececb3 COPYINGpackage/libupnp18/libupnp18.mk +0
-0@@ -4,7 +4,7 @@ # ################################################################################
Quoted 6 lines
package/libupnp18/libupnp18.mk line 10
Quoted 6 lines
package/libupnp18/libupnp18.mk +1
-1-LIBUPNP18_VERSION = 1.8.7 +LIBUPNP18_VERSION = 1.14.0 LIBUPNP18_SOURCE = libupnp-$(LIBUPNP18_VERSION).tar.bz2 LIBUPNP18_SITE = http://downloads.sourceforge.net/project/pupnp/pupnp/libupnp-$(LIBUPNP18_VERSION) LIBUPNP18_CONF_ENV = ac_cv_lib_compat_ftime=no
Best Regards,
Fabrice
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
Quoted 14 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Fabrice" == Fabrice Fontaine <[email protected]> writes:
Le dim. 30 ao?t 2020 ? 20:34, Arnout Vandecappelle <[email protected]> a ?crit :
Quoted 11 lines · Fabrice Fontaine
Quoted 1 line
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
Again, although this bump indeed fixes those issues, it's a feature version bump so I'm not sure if it can be called "security bump".
In addition, the libupnp18 package exists because of API incompatibility with 1.6. Are we sure that this problem doesn't repeat itself for 1.14?
Quoted 14 lines
There is indeed an API incompatibility between 1.8 and 1.14 related to CallStranger a.k.a. CVE-2020-12695: starting from 1.14, UpnpInit function has been removed as this function can't be fixed against CallStranger because this function takes an IP address and not an interface name.
However, UpnpInit2 is available for more than 10 years and is used by most of the applications (i.e. mpd and vlc) with the exception of gmrender-resurrect (which is patched in this serie).
As soon as libupnp 1.14 is available, I'm planning to update the applications that are still using the legacy libupnp 1.6.x version (i.e. igd2-for-linux and ushare) and drop this unsecure version. I would like to avoid adding a third version of libupnp (i.e. a libupnp114 package) as from a security perspective, all packages should use this version.
So we would end up with package/libupnp = 1.14.0? Sounds sensible.
Quoted 3 lines
Still, I agree that this is not only a "security bump" so I would advise to apply this serie to next and backport it to our LTS branches in a few months.
Ok. I was still hesitating if I should apply this for 2020.08, but I think you are right - Lets get 2020.08 out the door, merge next and let this stabilize a bit before backporting to 2020.08.x / 2020.02.x.
Signature · 1 line
Bye, Peter Korsgaard
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
Le lun. 31 ao?t 2020 ? 22:14, Peter Korsgaard <[email protected]> a ?crit :
Quoted 30 lines
Quoted 15 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
Quoted 2 lines
"Fabrice" == Fabrice Fontaine <[email protected]> writes:
Le dim. 30 ao?t 2020 ? 20:34, Arnout Vandecappelle <[email protected]> a ?crit :
Quoted 11 lines · Fabrice Fontaine
Quoted 1 line
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
Again, although this bump indeed fixes those issues, it's a feature version bump so I'm not sure if it can be called "security bump".
In addition, the libupnp18 package exists because of API incompatibility with 1.6. Are we sure that this problem doesn't repeat itself for 1.14?
Quoted 14 lines
There is indeed an API incompatibility between 1.8 and 1.14 related to CallStranger a.k.a. CVE-2020-12695: starting from 1.14, UpnpInit function has been removed as this function can't be fixed against CallStranger because this function takes an IP address and not an interface name.
However, UpnpInit2 is available for more than 10 years and is used by most of the applications (i.e. mpd and vlc) with the exception of gmrender-resurrect (which is patched in this serie).
As soon as libupnp 1.14 is available, I'm planning to update the applications that are still using the legacy libupnp 1.6.x version (i.e. igd2-for-linux and ushare) and drop this unsecure version. I would like to avoid adding a third version of libupnp (i.e. a libupnp114 package) as from a security perspective, all packages should use this version.
Quoted 1 line
So we would end up with package/libupnp = 1.14.0? Sounds sensible.
Yes ideally, we should have package/libupnp = 1.14.0. Would it be acceptable/reasonable to bump libupnp from 1.6.x to 1.14.x and remove libupnp18?
If this is acceptable, I'll send a v2 of this serie (with the drop of libupnp18 and the update of ushare/igd2-for-linux).
Quoted 11 lines
Quoted 4 lines
Still, I agree that this is not only a "security bump" so I would advise to apply this serie to next and backport it to our LTS branches in a few months.
Ok. I was still hesitating if I should apply this for 2020.08, but I think you are right - Lets get 2020.08 out the door, merge next and let this stabilize a bit before backporting to 2020.08.x / 2020.02.x.
--
Bye, Peter Korsgaard
Best Regards,
Fabrice
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Fabrice" == Fabrice Fontaine <[email protected]> writes:
Hi,
Quoted 4 lines
Quoted 1 line
So we would end up with package/libupnp = 1.14.0? Sounds sensible.
Yes ideally, we should have package/libupnp = 1.14.0. Would it be acceptable/reasonable to bump libupnp from 1.6.x to 1.14.x and remove libupnp18?
Yes. Both 1.6 and 1.18 are vulnerable, right? So if we want to fix the CVE we need to do so.
Quoted 2 lines
If this is acceptable, I'll send a v2 of this serie (with the drop of libupnp18 and the update of ushare/igd2-for-linux).
Thanks!
Signature · 1 line
Bye, Peter Korsgaard
PATCH 1/3 package/libupnp18: security bump to version 1.14.0
In reply to Arnout VandecappelleQuoted 10 lines
Quoted 1 line
Quoted 1 line
Quoted 1 line
Quoted 1 line
"Arnout" == Arnout Vandecappelle <[email protected]> writes:
Quoted 1 line · Fabrice Fontaine
Fix CallStranger a.k.a. CVE-2020-12695 as well as CVE-2020-13848
Again, although this bump indeed fixes those issues, it's a feature version
bump so I'm not sure if it can be called "security bump".In addition, the libupnp18 package exists because of API incompatibility with
1.6. Are we sure that this problem doesn't repeat itself for 1.14?
And calling it libupnp18 when it is provides 1.14.x is quite confusing :/
Signature · 1 line
Bye, Peter Korsgaard