subtree updates
meta-openembedded: 0782ea454a..ce0b93fc12:
Alex Kiernan (5):
faad2: Upgrade 2.8.8 -> 2.10.0
onig: Upgrade 6.9.4 -> 6.9.8
jansson: Honour multilib paths
jansson: Backport linker flag fixes
jansson: Default to shared builds
Beniamin Sandu (1):
libnet: update to v1.2 release
Daniel Gomez (4):
gst-instruments: Update 0.2.3 -> 0.3.1+cb8977a
libftdi: Add ftdi-eeprom support
xf86-video-ati: Update 19.1.0 -> 19.1.0+7a6a34af
v4l-utils: Update 1.22.1 -> 1.23.0+fd544473
Gianluigi Spagnuolo (1):
bpftool: add aarch64 to COMPATIBLE_HOST
Hitomi Hasegawa (1):
libsdl: add CVE-2019-14906 to allowlist
Khem Raj (2):
python3-gevent: Avoid building internal version of libev
xterm: Add _GNU_SOURCE via CFLAGS
Lukas Rusak (2):
libwebsockets: add optional support for sd-event loop
libwebsockets: add error check if PACKAGECONFIG contains systemd but DISTRO_FEATURES doesn't
Ming Liu (1):
plymouth: uprev to 22.02.122
William A. Kennington III (2):
gerbera: upgrade 1.9.2 -> 1.11.0
fmt: upgrade 8.1.1 -> 9.1.0
Yi Zhao (4):
freeradius: fix daemon startup warnings
frr: upgrade 8.2.2 -> 8.3.1
libnftnl: upgrade 1.2.2 -> 1.2.3
nftables: upgrade 1.0.4 -> 1.0.5
onkelpit (1):
tio: added tio version 2.0 and 1.47
wangmy (1):
xterm: upgrade 372 -> 373
meta-arm: 52f07a4b0b..0164b4ca7a:
Abdellatif El Khlifi (12):
arm-bsp/u-boot: corstone1000: update initramfs bundle size
arm-bsp/u-boot: corstone1000: upgrade FF-A support
arm-bsp/optee-os: corstone1000: upgrade to v3.18
arm-bsp/optee-spdevkit: corstone1000: drop the support
arm-bsp/corstone1000-initramfs-image: remove obsolete packages
arm-bsp/trusted-services: corstone1000: add secure partitions support
arm-bsp/machine: corstone1000: disable pulling the kernel into the initramfs
arm-bsp/trusted-services: corstone1000: add MHU-driver
arm-bsp/corstone1000-initramfs-image: add TS PSA API tests packages
arm-bsp/linux: corstone1000: use arm-ffa machine feature
arm/secure-partitions: drop use of the recipe
arm/ffa-debugfs: drop use of the kernel module
Adam Johnston (3):
arm-bsp/edk2-firmware: Update edk2/edk2-platforms versions for N1SDP
arm-bsp/edk2-firmware: Add edk2-platforms patches for N1SDP
arm-bsp/trusted-firmware-a: Update TF-A version for N1SDP
Andrei Gherzan (1):
edk2-firmware: Fix configure sed typo
Anton Antonov (1):
Temporary use qemu 7.0.0 for TS CI pipelines
Davidson K (6):
arm-bsp/tc: upgrade version of trusted-firmware-a
arm-bsp/tc: upgrade version of hafnium
arm-bsp/tc: upgrade version of optee
arm-bsp/u-boot: add gnutls-native as dependency
arm-bsp/trusted-firmware-a: add firmware update support for TC
arm-bsp/hafnium: enable Virtual Host Extension for TC
Denys Dmytriyenko (1):
arm-toolchain/gcc,external-arm-toolchain: resolve conflict with gcc headers
Emekcan (8):
arm-bsp/u-boot: Add external system driver to u-boot device tree
arm-bsp/kernel: Add external device driver
arm-bsp/u-boot: Add external system MHUs to u-boot device tree
arm-bsp/kernel: Add rpmsg_arm_mailbox to corstone1000
arm-bsp/test: Adding a test app for external system
arm-bsp/images: Adding external system test to initramfs image
arm-bsp/test: Changing the test app repository
arm-bsp/external-system: Changing the RTX repo
Jiacheng Tang (1):
arm/fvp-base-r-aem: upgrade to version 11.19.14
Joe Slater (1):
arm/packagegroup-ts-tests: fix parse error
Jon Mason (17):
arm-bsp/optee-os: add 3.10 recipe for corstone1000
arm-bsp/optee: rename corstone1000 files
arm/optee-spdevkit: add version to file name
arm/optee-os: add ARMv7 changes to clang patch and update patches
arm/qemuarm-secureboot: remove optee-os version pin
arm/optee: remove old versions
arm/optee-client: move the 3.14 recipe to meta-arm-bsp
arm/hafnium: update to 2.7
arm-bsp/n1sdp: update linux-yocto patches
arm/edk2-firmware: Work around clang issue
arm-bsp/tc: remove hafnium clang patch
layers: convert to langdale compatibility
CI: Remove uniquely zephyr machines
arm-bsp/fvp: move the fvp include file to the include directory
ci: move features only needed by testimage from base
CI: apply a patch so that meta-zephyr is compatible with langdale
Revert "CI: apply a patch so that meta-zephyr is compatible with langdale"
Khem Raj (6):
optee-os: Extend clang pragma fixes to core_mmu_v7.c for 3.18
trusted-services: Pin to use gcc
ffa-debugfs-mod: Exclude from world builds
linux-yocto: Add bbappend for 5.19
hafnium: Add a fix for clang-15 errors
hafnium: Exclude from world builds
Mohamed Omar Asaker (1):
arm-bsp/n1sdp-board-firmware: upgrade to N1SDP-2022.06.22
Peter Hoyes (4):
arm/lib: Specify the FVP environment variables explicitly
arm-bsp/trusted-firmware-m: Make branch names configurable
arm/classes: Migrate TF-M image signing to bbclass
arm-bsp/corstone1000: Refactor image signing to use new bbclass
Ross Burton (3):
gem5/linux-yocto: upgrade to 5.4.205 and fix buildpaths in binaries
Revert "Temporary use qemu 7.0.0 for TS CI pipelines"
runfvp: pass-through environment variables need for GUI applications
Rui Miguel Silva (1):
arm-bsp: trusted-services: fix openamp build
Vishnu Banavath (2):
arm-bsp/ffa-debugfs: update git SHA for v2.1.0
arm-bsp/external-system:corstone1000: build and install external-system
Xueliang Zhong (1):
arm-bsp/n1sdp: upgrade scp-firmware version
Signed-off-by: Patrick Williams <patrick@stwcx.xyz>
Change-Id: I7a07eab9e4aa0bdbdb50602050c3c4caf062acbf
diff --git a/meta-openembedded/meta-networking/recipes-protocols/frr/frr/0001-bgpd-avoid-notify-race-between-io-and-main-pthreads.patch b/meta-openembedded/meta-networking/recipes-protocols/frr/frr/0001-bgpd-avoid-notify-race-between-io-and-main-pthreads.patch
new file mode 100644
index 0000000..40284d6
--- /dev/null
+++ b/meta-openembedded/meta-networking/recipes-protocols/frr/frr/0001-bgpd-avoid-notify-race-between-io-and-main-pthreads.patch
@@ -0,0 +1,151 @@
+From 71ca5b09bc71e8cbe38177cf41e83fe164e52eee Mon Sep 17 00:00:00 2001
+From: Mark Stapp <mstapp@nvidia.com>
+Date: Thu, 8 Sep 2022 16:14:36 -0400
+Subject: [PATCH] bgpd: avoid notify race between io and main pthreads
+
+The "bgp_notify_" apis in bgp_packet.c generate a notification
+to a peer, usually during error handling. The io pthread wants
+to send notifications in a couple of cases during early
+received-packet validation - but the existing api interacts
+with the peer struct itself, and that's not safe.
+
+Add a new api for use by the io pthread, and adjust the main
+notify api so that it can avoid touching the peer struct.
+
+Signed-off-by: Mark Stapp <mstapp@nvidia.com>
+
+CVE: CVE-2022-37035
+
+Upstream-Status: Backport
+[https://github.com/FRRouting/frr/commit/71ca5b09bc71e8cbe38177cf41e83fe164e52eee]
+
+Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
+---
+ bgpd/bgp_io.c | 17 ++++++++---------
+ bgpd/bgp_packet.c | 32 ++++++++++++++++++++++++++++----
+ bgpd/bgp_packet.h | 2 ++
+ 3 files changed, 38 insertions(+), 13 deletions(-)
+
+diff --git a/bgpd/bgp_io.c b/bgpd/bgp_io.c
+index 7af1fae28..f9bb8d518 100644
+--- a/bgpd/bgp_io.c
++++ b/bgpd/bgp_io.c
+@@ -37,7 +37,7 @@
+ #include "bgpd/bgp_debug.h" // for bgp_debug_neighbor_events, bgp_type_str
+ #include "bgpd/bgp_errors.h" // for expanded error reference information
+ #include "bgpd/bgp_fsm.h" // for BGP_EVENT_ADD, bgp_event
+-#include "bgpd/bgp_packet.h" // for bgp_notify_send_with_data, bgp_notify...
++#include "bgpd/bgp_packet.h" // for bgp_notify_io_invalid...
+ #include "bgpd/bgp_trace.h" // for frrtraces
+ #include "bgpd/bgpd.h" // for peer, BGP_MARKER_SIZE, bgp_master, bm
+ /* clang-format on */
+@@ -526,8 +526,8 @@ static bool validate_header(struct peer *peer)
+ return false;
+
+ if (memcmp(m_correct, m_rx, BGP_MARKER_SIZE) != 0) {
+- bgp_notify_send(peer, BGP_NOTIFY_HEADER_ERR,
+- BGP_NOTIFY_HEADER_NOT_SYNC);
++ bgp_notify_io_invalid(peer, BGP_NOTIFY_HEADER_ERR,
++ BGP_NOTIFY_HEADER_NOT_SYNC, NULL, 0);
+ return false;
+ }
+
+@@ -547,9 +547,8 @@ static bool validate_header(struct peer *peer)
+ zlog_debug("%s unknown message type 0x%02x", peer->host,
+ type);
+
+- bgp_notify_send_with_data(peer, BGP_NOTIFY_HEADER_ERR,
+- BGP_NOTIFY_HEADER_BAD_MESTYPE, &type,
+- 1);
++ bgp_notify_io_invalid(peer, BGP_NOTIFY_HEADER_ERR,
++ BGP_NOTIFY_HEADER_BAD_MESTYPE, &type, 1);
+ return false;
+ }
+
+@@ -574,9 +573,9 @@ static bool validate_header(struct peer *peer)
+
+ uint16_t nsize = htons(size);
+
+- bgp_notify_send_with_data(peer, BGP_NOTIFY_HEADER_ERR,
+- BGP_NOTIFY_HEADER_BAD_MESLEN,
+- (unsigned char *)&nsize, 2);
++ bgp_notify_io_invalid(peer, BGP_NOTIFY_HEADER_ERR,
++ BGP_NOTIFY_HEADER_BAD_MESLEN,
++ (unsigned char *)&nsize, 2);
+ return false;
+ }
+
+diff --git a/bgpd/bgp_packet.c b/bgpd/bgp_packet.c
+index 7daac4494..90695219a 100644
+--- a/bgpd/bgp_packet.c
++++ b/bgpd/bgp_packet.c
+@@ -871,8 +871,9 @@ bool bgp_notify_received_hard_reset(struct peer *peer, uint8_t code,
+ * @param data Data portion
+ * @param datalen length of data portion
+ */
+-void bgp_notify_send_with_data(struct peer *peer, uint8_t code,
+- uint8_t sub_code, uint8_t *data, size_t datalen)
++static void bgp_notify_send_internal(struct peer *peer, uint8_t code,
++ uint8_t sub_code, uint8_t *data,
++ size_t datalen, bool use_curr)
+ {
+ struct stream *s;
+ bool hard_reset = bgp_notify_send_hard_reset(peer, code, sub_code);
+@@ -917,8 +918,11 @@ void bgp_notify_send_with_data(struct peer *peer, uint8_t code,
+ * If possible, store last packet for debugging purposes. This check is
+ * in place because we are sometimes called with a doppelganger peer,
+ * who tends to have a plethora of fields nulled out.
++ *
++ * Some callers should not attempt this - the io pthread for example
++ * should not touch internals of the peer struct.
+ */
+- if (peer->curr) {
++ if (use_curr && peer->curr) {
+ size_t packetsize = stream_get_endp(peer->curr);
+ assert(packetsize <= peer->max_packet_size);
+ memcpy(peer->last_reset_cause, peer->curr->data, packetsize);
+@@ -1001,7 +1005,27 @@ void bgp_notify_send_with_data(struct peer *peer, uint8_t code,
+ */
+ void bgp_notify_send(struct peer *peer, uint8_t code, uint8_t sub_code)
+ {
+- bgp_notify_send_with_data(peer, code, sub_code, NULL, 0);
++ bgp_notify_send_internal(peer, code, sub_code, NULL, 0, true);
++}
++
++/*
++ * Enqueue notification; called from the main pthread, peer object access is ok.
++ */
++void bgp_notify_send_with_data(struct peer *peer, uint8_t code,
++ uint8_t sub_code, uint8_t *data, size_t datalen)
++{
++ bgp_notify_send_internal(peer, code, sub_code, data, datalen, true);
++}
++
++/*
++ * For use by the io pthread, queueing a notification but avoiding access to
++ * the peer object.
++ */
++void bgp_notify_io_invalid(struct peer *peer, uint8_t code, uint8_t sub_code,
++ uint8_t *data, size_t datalen)
++{
++ /* Avoid touching the peer object */
++ bgp_notify_send_internal(peer, code, sub_code, data, datalen, false);
+ }
+
+ /*
+diff --git a/bgpd/bgp_packet.h b/bgpd/bgp_packet.h
+index a0eb579db..9f6d772bc 100644
+--- a/bgpd/bgp_packet.h
++++ b/bgpd/bgp_packet.h
+@@ -62,6 +62,8 @@ extern void bgp_open_send(struct peer *);
+ extern void bgp_notify_send(struct peer *, uint8_t, uint8_t);
+ extern void bgp_notify_send_with_data(struct peer *, uint8_t, uint8_t,
+ uint8_t *, size_t);
++void bgp_notify_io_invalid(struct peer *peer, uint8_t code, uint8_t sub_code,
++ uint8_t *data, size_t datalen);
+ extern void bgp_route_refresh_send(struct peer *peer, afi_t afi, safi_t safi,
+ uint8_t orf_type, uint8_t when_to_refresh,
+ int remove, uint8_t subtype);
+--
+2.25.1
+